Datasets:

sordonia

/

redpajama-sample_from_valid_all

like 1

high_school_physics

Modern camera lenses require several glass elements in order to properly project an image onto a planar surface. This makes for a heavy and expensive lens, particularly when a large aperture is desired. The complexity of lenses could be significantly reduced by projecting onto a curved surface instead. A CCD manufactured as a sector of a sphere allows a lens to create a sharp image without having to account for spherical aberration. Obviously this presents some manufacturing challenges, the most difficult of which is forming the silicon. Once a wafer is obtained however, the rest of the process isn't too difficult. Photolithography also benefits from the design, as manufacturing equipment needn't account for spherical aberrations either. Note link is a PDF. Indeed I do. Here I researched spherical geometry to ensure sector was the proper term, and I carelessly ended up eliminating an entire dimension. Agreed, a good idea. It may be technically difficult, though - the wafers on which the CCDs are built are sliced from a cylindrical ingot and then polished to present a single crystal plane on the surface. Not sure how you'd do this as a spherical surface. Either carve out a curved surface from a wafer several milimetres thick, or get a flat wafer and use molecular beam epitaxy to build up the surface? If you carve it, the silicon crystal plane will be all over the place. Adding a 4th dimension projection would be one obvious solution. For watching it means that the viewer has to maintain a fixed distance from the spherical-section screen. Do the optics maintain their relative simplicity if you want to +/- zoom ? Would the surface be concave or convex? For some reason I originally thought convex, but I'm guessing it would actually be concave, matching the function of the eye. Also, given a high enough resolution, could this not be done in software? In essence, by “offsetting” each pixel based on the difference between where it is in space and where it would be if the sensor were curved. You could do that, but the granularity of the photo will be slightly higher at the edges than the middle simply due to angle of incidence. I had this idea a few years ago -- simply from the fact that the retina doesn't suffer from blue/red shifts at the edges -- and got the blunt reply of "too expensive to manufacture". Even if you could manufacture it, the curvature would have to be matched to a fixed focal length. This isn't an enormous problem; cameras like the Fuji x100 have shown there's still a demand for fixed lens cameras. Another problem though is that it doesn't actually solve, for instance, chromatic aberration. The edge-of-frame effects are gone (transverse chrom abs) but the loss of sharpness and blooming would remain (longitudinal chrom abs). There's an excellent website I found <link> which explains lens aberrations really well. Thoroughly worth reading the whole site if you have an interest in optics. //could this not be done in software?// That is a very good point. //the granularity of the photo will be slightly higher at the edges// That is also true. However, one could have different pixel sizes at the edges...no, hang on, you can't tesselate squares of sizes that increase smoothly in both directions. Maybe the software solution is simplest, with resampling to give a constant granularity. On the other hand, is it really so difficult to make lenses that correct for the planarity? The lens already has multiple elements to provide focus and achromaticity - how many extra lens elements are actually needed to correct for planarity? //could this not be done in software?// No. If the edges of the frame are out of focus for whatever reason, no amount of software will recover that. However, the latest generation of digital SLRs match the profile of the lens and do correct what they can (distortion/edge of frame colour shifts) in-body. But would the edges be out of focus or just distorted? "Distortion" with regards to photographic lenses means straight lines being represented as curved on the image. Chromatic and spherical aberration are a loss of sharpness because some of the light is out of focus. Curving the image plane does not stop this, it simply makes the effect uniform across the frame. Multi-element lens designs actually reduce these effects. The one thing a curved image plane would prevent would be astigmatism. But I really do recommend the toothwalker site for better explanations. One thing I should also mention - there are cameras that actually use single element lenses these days: disposable film cameras. Many of these do actually curve the film plane to try and improve the image, of course practical constraints mean it can only be curved in one plane. Assuming space-time is curved, as allowed by GR theory, just make the camera really really big. It's seems somewhat ironic complaining about "aberration" on here.. You only have to approximate a sphere, though, so within any individual pixel the crystal plane can be flawless. Imagine starting with a perfectly flat wafer and then somehow building up each pixel by the correct number of crystal lattice units, so that the pixels are individually flat but the overall shape of the surface is spherical. Making the interconnects work could be tricky. Actually, I've recently had cause to deal with silicon wafers which are thin enough to be very flexible, so perhaps this isn't as unfeasible as I'd first thought. Flexible probably doesn't have the precision required, and it almost definitely can't do a bowl shape with the required accuracy. In order to get the precision for an optical application, you're almost definitely either going to have to somehow force the crystal to grow in shape, or be able to grind it to shape. That I like. Is there a "catalytic" surface that silicon will condense on, that can also be removed/dissolved? Could you entice the entire CCD sensor structure to "grow" (self-assembling)? Maybe use custom-engineered enzymes? Ain't gonna work. You need a single crystal plane. well, check the <link>. They're growing single crystal turbine blades, so it's not a massive intellectual leap to suppose they can make the whole 'plane in one crystal. I think the furry dice hanging from the rear view mirror might prove technically challenging, however. More seriously, collimated beams are a bit of a pain to make, but after that, optics gets very easy. It's tough to get everything lined up in x & y, adding in another dimension with tolerances down in the individual nanometres is not to be encouraged. You would mount the flexible wafer on a solid curved support. You could at least get curvature in one dimension. If the curved sensor substrate were not only flexible, but rubbery, it could be sucked from behind, to produce not only a very accurately shaped concavity, but also a variable one. If the lens elements were also able to variably bulge and space apart, it could be possible to build a remarkably simple varifocal (a bit like a zoom in effect, but simpler because it doesn’t track focus like a zoom lens does*). * I have a Kiron 30-80mm varifocal lens. It looks like a zoom, but optically is much simpler (although it’s marvellously over-engineered and heavy) with the downside that as you “zoom” in or out (push-pull the single grip barrel) it drops out of focus and you have to refocus (rotate the barrel). You’d be surprised how utterly irritating that is to actually use, each shot, yet the heft and feel almost make up for it.

{'timestamp': '2019-04-21T00:43:15Z', 'url': 'http://www.halfbakery.com/idea/Curved_20CCD', 'language': 'en', 'source': 'c4'}

high_school_physics

\section{\@startsection{section}{1}{\z@}% {15pt \@plus 3\p@ \@minus 3\p@}% {4\p@}% \sectionfont\raggedright\hst[13pt]}} \renewcommand\subsection{\@startsection{subsection}{2}{\z@}% {10pt \@plus 3\p@ \@minus 2\p@}% {.1\p@}% \ssectionfont\raggedright }} \renewcommand\subsubsection{\@startsection{subsubsection}{3}{\z@}% {10pt \@plus 1\p@ \@minus .3\p@}% {.1\p@}% \sssectionfont\raggedright}} \makeatother \let\WriteBookmarks\relax \def1{1} \def.001{.001} \shorttitle{NectarCAM detectors performances} \shortauthors{Tsiahina et al.} \graphicspath{{Images/}} \title [mode = title]{Measurement of performance of the NectarCAM photodetectors} \author[1]{A. Tsiahina}[type=editor, auid=000,bioid=1, prefix=, role=, orcid= ] \cormark[1] \fnmark[1] \ead{[email protected]} \author[1] {P. Jean} \author[1] {J.-F. Olive} \author[1] {J. Kn\"odlseder} \author[1] {C. Marty} \author[1] {T. Ravel} \author[1] {C. Jarnot} \author[2] {B. Biasuzzi} \author[3] {J. Bolmont} \author[4] {F. Brun} \author[3] {S. Caroff} \author[5] {E. Delagnes} \author[6] {S. Fegan} \author[6] {G. Fontaine} \author[7] {D. Gascon} \author[4] {J.-F. Glicenstein} \author[8] {D. Hoffmann} \author[3] {S. Karkar} \author[3] {J.-P. Lenain} \author[7] {J. Paredes} \author[9] {P.-O. Petrucci} \author[10] {J. Prast} \author[7] {M. Rib\'o} \author[11] {S. Rivoire} \author[7] {A. Sanuy} \author[2] {P. Sharma} \author[2] {T. Suomijarvi} \author[12] {L.A. Tejedor} \author[3] {F. Toussenel} \author{for NectarCAM} \author{CTA} \address[1] {Institut de Recherche en Astrophysique et Plan\'etologie, CNRS-INSU, Universit\'e Paul Sabatier, 9 avenue Colonel Roche, BP 44346, 31028 Toulouse Cedex 4, France} \address[2] {Laboratoire de Physique des 2 infinis, Irene Joliot-Curie,IN2P3/CNRS, Universit\'e Paris-Saclay, Universit\'e de Paris, 15 rue Georges Clemenceau, 91406 Orsay, Cedex, France} \address[3] {Sorbonne Universit\'e, Universit\'e Paris Diderot, Sorbonne Paris Cit\'e, CNRS/IN2P3, Laboratoire de Physique Nucl\'eaire et de Hautes Energies, LPNHE, 4 Place Jussieu, F-75005 Paris, France} \address[4] {IRFU, CEA, Universit\'e Paris-Saclay, B\^at 141, 91191 Gif-sur-Yvette, France} \address[5] {IRFU/DEDIP, CEA, Universit\'e Paris-Saclay, B\^at 141, 91191 Gif-sur-Yvette, France} \address[6] {Laboratoire Leprince-Ringuet, \'Ecole Polytechnique (UMR 7638, CNRS/IN2P3, Institut Polytechnique de Paris), 91128 Palaiseau, France} \address[7] {Departament de F\'isica Qu\`antica i Astrof\'isica, Institut de Ci\`encies del Cosmos, Universitat de Barcelona, IEEC-UB, Mart\'i i Franqu\`es, 1, 08028, Barcelona, Spain} \address[8] {Aix Marseille Univ, CNRS/IN2P3, CPPM, 163 Avenue de Luminy, 13288 Marseille cedex 09, France} \address[9] {Univ. Grenoble Alpes, CNRS, IPAG, 414 rue de la Piscine, Domaine Universitaire, 38041 Grenoble Cedex 9, France} \address[10] {LAPP, Univ. Grenoble Alpes, Univ. Savoie Mont Blanc, CNRS-IN2P3, 9 Chemin de Bellevue - BP 110, 74941 Annecy Cedex, France} \address[11] {Laboratoire Univers et Particules de Montpellier, Universit\'e de Montpellier, CNRS/IN2P3, CC 72, Place Eug\`ene Bataillon, F-34095 Montpellier Cedex 5, France} \address[12] {EMFTEL department and IPARCOS, Universidad Complutense de Madrid, 28040 Madrid, Spain} \begin{abstract}[S U M M A R Y] \noindent NectarCAM is a camera for the medium-sized telescopes of the Cherenkov Telescope Array (CTA), which covers the energy range of 100 GeV to 30~TeV. The camera is equipped with 265 focal plane modules (FPMs). Each FPM comprises 7 pixels, each consisting of a photo-multiplier tube, a preamplifier, an independently controlled power supply, and a common control system. We developed a dedicated test bench to validate and qualify the industrial FPM production and to measure the performance of each FPM in a dark room before its integration in the camera. We report the measured performance of 61~FPM prototypes obtained with our experimental setup. We demonstrate that the gains of the photo multiplier tubes are stable and that pulse widths, transit time spreads, afterpulse rates and charge resolutions are within the specifications for NectarCAM. \end{abstract} \begin{keywords} Gamma-ray astronomy \sep Cherenkov telescope \sep Photo Multiplier Tube \end{keywords} \maketitle \section{Introduction} The Cherenkov Telescope Array (CTA) Observatory will provide a leap forward in sensitivity in the 0.02-300~TeV gamma-ray domain compared to the existing imaging atmospheric Cherenkov telescopes (IACT) such as H.E.S.S., VERITAS and MAGIC. The CTA Observatory will improve our understanding of particle acceleration processes in the Universe through the observation of very-high-energy gamma-ray sources (such as active galactic nuclei, supernova remnants, pulsar wind nebulae, see \cite{CTAScience:2017}). It will also allow us to probe the interstellar and intergalactic media and to explore physics beyond the standard model by searching for signatures of dark matter or effects of quantum gravity. The CTA Observatory will be composed of Cherenkov telescopes of three sizes. These telescopes will be installed at the Instituto de Astrof\'{i}sica de Canarias (IAC) site on La Palma (Canary Islands, Spain) and at the European Southern Observatory (ESO) site at Paranal (Chile), with up to eight large-sized telescopes (LST; 4 at La Palma, and 4 at Paranal), 40 medium-sized telescopes (MST; 15 at La Palma and 25 at Paranal), and 70 Small-Sized Telescopes (SST; all at Paranal). CTA is about to complete its pre-construction phase, and will soon enter the construction phase. A prototype LST was constructed at the La Palma site and is currently under commissioning. Completion of the array is foreseen in 2025, and the science operations are expected to last over 30 years. CTA will use the IACT technique to image atmospheric air showers produced by gamma and cosmic rays when interacting with the Earth's atmosphere. For general reference about the CTA project and the IACT techniques, see \cite{ExpAstron:2011}. Two types of cameras are currently developed for the MSTs: NectarCAM \cite{NectarCAM:2015} and FlashCam \cite{FlashCAM2012, FlashCAM2017}. Both cameras are optimized for the detection of gamma-ray showers in the 0.1 - 30 TeV energy range with a field of view of 8$^\circ$ and a pixel size of 0.18$^\circ$. NectarCAM cameras are designed, developed, tested and commissioned by a consortium of 16 European institutes under French leadership. Among the participating institutes, the Institut de Recherche en Astrophysique et Plan\'etologie (IRAP) laboratory is responsible for the development and provision of the focal plane instrumentation composed of photomultiplier tubes (PMTs) and their associated electronics, providing the power, the signal amplification and the control. The focal plane is divided into 265 independent modules, each equipped with seven PMTs and the electronics to drive them. In mid-2018, a batch of 61 prototype modules were produced. These have been extensively tested with a dedicated test bench developed and built at IRAP. The aim of these tests was to validate the functioning of the modules produced by an industrial partner, verify that their performance meets the NectarCAM specifications, and to perform an initial calibration before their integration into the NectarCAM camera at Institut de Recherche sur les lois Fondamentales de l'Univers (IRFU), Saclay. The tests also allowed us to evaluate the test bench itself for later use by an industrial partner, where it will serve for validation of mass-produced modules in the CTA construction phase. In section 2, we describe the design of a NectarCAM focal plane module (FPM). The design of our test bench, which allows functional tests and performance measurements on seven FPMs in parallel, is described in the next section. In section 4, we present the methods and the results of the most important performance tests obtained with the 61 focal plane module prototypes, namely: the stability of PMT gains, the charge resolution, the single photo-electron pulse shape (width and transit-time spread) and the afterpulse rate. The measured performance is compared with previous measurements and discussed in the context of the CTA requirements \cite{MSTReq:LevelA, MSTReq:LevelB, HintonMSTReq14}. \section{The Focal Plane Module of NectarCAM}\label{Text:FPM_Nect} NectarCAM has a total of 1855 pixels, sensitive to individual photons, organised in 265 individual FPMs (see \circled{A} in Figure \ref{FIG:FPM_overview}). Each FPM contains seven detector units (DU, \circled{E}), connected to an interface board (IB, \circled{F}, \circled{G}) equipped with a micro-controller. Each DU is composed of a PMT (\circled{C}) soldered to a High-Voltage Pre-Amplifier board (HVPA, \circled{D}) and enclosed in a grounded metal tube. The detectors are seven-dynode PMTs (R12992-100 from Hamamatsu \cite{HamamatsuPMTDataSheet}) specially designed for CTA in order to optimize the performance of the CTA imaging cameras (see \cite{Toyama15}). These PMTs have a good quantum efficiency (up to $40$\%), a high photo-electron collection efficiency, a short pulse width, a short transit time spread and low after-pulse probability. Finally, for astrophysical observations, Winston cones (WC, \circled{B}) are placed in front of each DU as a light guide in order to maximize the light collection at the photocathode. Since the aim of our tests is to characterise the FPMs, we do not use the WCs in our test bench; they are tested independently at Institut de Plan\'etologie et d'Astrophysique de Grenoble (IPAG) laboratory (see \cite{Hen13} and \cite{Hen17}), and the full NectarCAM detector chain is subsequently tested at IRFU. \begin{figure}[H] \centering \includegraphics[width=7cm]{FPM_desc_v3} \caption{ Overview of a Focal Plane Module (FPM). \protect\circled{A} is the FPM equipped with 7 Winston Cones (WC), \protect\circled{B} is a WC, \protect\circled{C} a PMT from Hamamatsu, \protect\circled{D} is a HVPA board, \protect\circled{E} is a Detector Unit (DU), \protect\circled{F} is the Interface Board (IB) and \protect\circled{G} is a PMT soldered to a HVPA board and connected to the IB.} \label{FIG:FPM_overview} \end{figure} The HVPA boards that are soldered to every PMT provide high voltage to the dynodes and the photocathode through a Cockcroft-Walton (CW) type electronic circuit. The CW provides a high voltage of up to 1550~V. It has been designed so that the high voltage stability (offset and ripple) is smaller than 1~V to ensure that the PMT gain fluctuations will not affect the charge resolution. In addition to providing high voltage to the PMT, the HVPA also amplifies the anode signal through a dedicated PACTA (pre-amplifier for CTA) ASIC developed by the University of Barcelona (see \cite{Sanuy2012}). In order to accommodate a dynamic range from ~1 up to ~3000 photo-electrons, PACTA splits the anode signal into two lines: a high-gain (HG) line with typical trans-impedance of $1200~\Omega$ with a maximum deviation of $\pm 60~\Omega$ and a low-gain (LG) line with typical trans-impedance of $79~\Omega$ with a maximum deviation of ${+3}/{-5}~\Omega$. The mean and standard deviation of the HG/LG ratio have been measured to be 15 and 2, respectively. Each HVPA is connected to the IB that sets the seven desired high voltages with a closed-loop control system. The IB also routes the HG and LG signals to the front-end electronic board (FEB) on $100~\Omega$ impedance lines. Furthermore, the IB digitalizes the load currents (total current of the HVPA) and the CW current consumption. The latter is a measure of the anode current and, therefore, of the photocathode illumination. The CW current is compared via an analog circuit on the IB to a pre-defined, tunable threshold level in order to protect the PMTs in case of high levels of illumination. Such high levels may arise, for example, if a very bright star passes through the field of view, or if the camera is accidentally exposed to light. If the CW current exceeds the defined threshold, the IB will automatically shut down the PMT high voltage in less than $\sim$~100~ms. The IB communicates with the FEB via a SPI protocol. The FEB remotely controls the IB in order to get/set slow control data such as PMTs voltages, ramp speed voltage (high voltage increment per second), status, HVPA current consumption, CW current consumption and CW threshold currents. Each FPM is connected to a FEB provided by Laboratoire de Physique Nucl\'eaire et de hautes \'Energies (LPNHE). The FEB functionalities include the triggering system of the camera \footnote{The FEB generates the trigger signals that will be transferred to the digital trigger board in order to elaborate the camera trigger.}, final amplification of the signal through an ACTA (amplifier for CTA) ASIC (see \cite{Sanuy2012b}), analog to digital conversion of amplified anode signals, data packet assembly and transmission to the computer that controls the module. The readout system of the FEB is based on the NECTAr chip, which performs the sampling, storage of the analog signal and digitization of two differential lines with a 12-bit Analog to Digital Converter (ADC) and a sampling rate from 0.5 to 2~GHz (see \cite{Naumann2012}). For each triggered event, the FEB delivers 14 waveforms which are the digitized anode currents of each HG/LG lines sampled with a nominal period of 1~ns (i.e. a sampling rate of 1~GHz). The FEBs were developped by IRFU, LPNHE, Laboratoire Univers et Particules de Montpellier (LUPM), Laboratoire de Physique Subatomique et de Cosmologie (LPSC) and Institut de Ci\`encies del Cosmos, University of Barcelona (ICCUB). \section{The FPM test bench}\label{Text:FPM_TB} The FPM test bench is composed of a light generation and a readout system which are connected to the FPM under test (see Figure \ref{FIG:FPM_TB_diag}). Both systems are controlled by a central computer. The light generation system is composed of a LED (called nano LED hereafter) that produces nanosecond duration UV flashes ($\lambda = 400 {~\rm nm}$) and a set of LEDs emitting continuous white light to simulate the night sky background (NSB) noise. The nanosecond flashes are produced by a LED (HERO HUVL400-510) that is driven by a LED pulser system designed by LPNHE. The light intensity is adjustable by a 8-bit DAC (with a non-linear scale); the measured light pulse intensity resolution is $\sim$~4\%. We add a controlled motorized neutral-density filter in front of the nano LED to have an extended dynamic range in pulse intensity, in particular to produce flashes at the level of a single photo-electron (ph.e.) per PMT. Without the filter, the maximum flash intensity (obtained with the LED pulser tuned to its maximum value) produces $\sim$ 5000 ph.e. in the PMT photocathode. With the filter, the minimum flash intensity (obtained with the LED pulser tuned to its minimum value) produces $\sim$ 0.03 ph.e. The attenuation factor of the filter is 12.5 \footnote{The attenuation factor was measured with an optical spectrometer and verified with PMTs using photon pulses.}. The flash LED pulser provides two analog triggering signals: one to drive the LED and the other one to trigger the readout system. The delay between the two analog signals is set remotely in order to fine tune the signal position in the readout window relative to the corresponding trigger signal. Eight white LEDs, behind a diffuser frame (see \circled{I} in Figure \ref{FIG:FPM_TB_pic}) and powered by a controlled DC power supply, simulate the NSB with rates ranging from 0.003 to 15~ph.e./ns produced in PMT photocathodes. The NSB level is monitored by a Vega light meter (Vega model 7Z01560 from Ophir Photonics) equipped with a NIST photodiode sensor installed in the plane of the FPM photocathodes. The Vega light meter is controlled by the computer via an RS232 port. We use seven NectarCAM FEBs as the readout system of our test bench, allowing us to test seven FPMs (49 PMT DUs) simultaneously. Each FEB amplifies, samples and digitizes the anode signals of FPMs (see section \ref{Text:FPM_Nect}) and transmits the resulting waveforms to the computer. The backends of the FEBs are connected to NectarCAM digital trigger board backplanes (DTB) that link the FEBs to the computer via ethernet cables and a hub (see \cite{Schwanke15}). The backplanes also transmit the external trigger signal produced by the flash LED pulser to the FEBs. The seven DTBs, produced by Deutsches Elektronen-Synchrotron (DESY), Zeuthen, convert the analog triggering signal to digital triggering signals which are delivered to the FEBs. The trigger frequency theoretically ranges up to 156~kHz although we are limited by the communication protocols to 500~Hz for measurements with 7~FPMs. The computer communicates with the FEB via UDP and TCP/IP protocols. It receives the digitized pulse currents of each PMT and the slow control data (see section \ref{Text:FPM_Nect}). The computer sends control commands to the FEB to manage waveform acquisitions and to tune the high voltage of each PMT. It also controls the flash LED via a TCP/IP protocol, the filter wheel via an Arduino and the DC power that supplies the eight white LEDs via USB connection. Our FPM test bench is installed in a dark room ($1.8\times 3.6$~${\rm m^2}$) with a controlled average air temperature of 18~$^{\circ}$C with an uncertainty of 2~${^\circ}$C (see Figure \ref{FIG:FPM_TB_pic}). The distance between the light generation system and the photocathodes plane is 103~cm. The computer that controls all the elements of the test bench with Python 2.7 scripts is installed outside the dark room. \begin{figure}[H] \centering \includegraphics[width=15cm]{Images/BTFPM1_diagram} \caption{The FPM test bench diagram.} \label{FIG:FPM_TB_diag} \end{figure} \begin{figure}[H] \centering \begin{subfigure}[t]{.4\textwidth} \centering \includegraphics[height=5.5cm]{Images/FPM_TB.jpg} \caption{}\label{fig:fig_a} \end{subfigure} \begin{subfigure}[t]{.4\textwidth} \centering \includegraphics[height=5.5cm]{Images/NSB_LEDs.jpg} \caption{}\label{fig:fig_b} \end{subfigure} \caption{Picture of the FPM test bench. In (\subref{fig:fig_a}) \protect\circled{H} is the dark room at IRAP laboratory, the frame \protect\circled{I} is the light generation system and \protect\circled{J} is the module rack. In (\subref{fig:fig_b}) is the other side of the frame \protect\circled{I} with the 8 white NSB LEDs fully powered and at the center the flash LED covered by the motorized ND filter.} \label{FIG:FPM_TB_pic} \end{figure} \section{Measurements}\label{Text:Meas} In order to validate the proper functioning of the FPMs after their assembly and prior to their integration into NectarCAM we perform a series of measurements on each of them using the IRAP test bench. These measurements verify that their performance meets a series of NectarCAM targets concerning, for example, the PMT gain, the stability of the high-voltage supply, the signal-to-noise ratio, and the afterpulsing rate. The measurements also yield a series of ``start-of-life'' calibrations for each DU, which can then be used as a reference throughout its lifetime; these include the initial gain versus voltage curve, afterpulsing fraction, single photo-electron spectrum, and impulse response. In this section, we report the measurement methods and the results of the tests used to verify the NectarCAM detector properties with our experimental setup. These results are compared with those presented in previous studies and with the NectarCAM requirements. \subsection{PMT gain measurements, tuning and stability with NSB rate}\label{Text:Gain} The PMT gain is the multiplication factor of one photo-electron charge. CTA plans to operate the PMTs with a nominal gain of 40000 in order to guarantee a detector lifetime that is compliant with an observatory lifetime of 30 years while achieving a sufficient signal-to-noise ratio that still allows the detection of individual photo-electrons. CTA requires the systematic uncertainty in the determination of the gain to be less than 3\% (see \cite{Brown2015}). A first series of tests therefore ensures that a gain of 40000 can be achieved for each PMT within the allowable voltage range for its CW supply, and that the gain is sufficiently stable over the range of night-sky background photon rates expected in operation. Several methods are used to measure the gain of PMTs for Cherenkov telescopes (e.g. see \cite{Bellamy94}, \cite{Hanna08}, \cite{FeganReport}, \cite{Takahashi18}). The method adopted in the present study is the single photo-electron fitting which is widely used in high-energy astrophysics instrumentation. Its main advantage is that, regardless of the flash LED pulser metrological characteristics (i.e. thermal stability, flash fluctuations), the number of photo-electrons can be determined in the fit. During operations, with NectarCAMs mounted on MST structures, an integrated single photo-electron calibration system will measure the gains with the same method (see \cite{Biasuzzi20}). The single photo-electron fitting method is based on fitting the position in ADC counts of the single photo electron peak in charge spectra accumulated at low luminosity of incident light. The charges are obtained by integrating the anode currents in the time window where the pulses are expected. Since the anode currents are digitized (see section \ref{Text:FPM_Nect}), the charges are estimated with a simple sum of the samples in the given time window (see Figure \ref{FIG:real_spe_plot}). The charge unit is measured in ADC counts. It contains the baseline, also called pedestal (which fluctuates due to electronic noise) and the possible photo-electron induced pulse contributions. The latter contribution to the charge is called the net charge; it is estimated by subtracting the baseline value from the measured charge. For illustration, a typical charge spectrum is shown in Figure \ref{FIG:myspefitter_plot}. The first peak, situated around 3350 ADC counts, is produced by events without any photo-electron; its mean value represents a programmable "pedestal" offset, while its width is determined by the noise level of the PMT and its acquisition chain. The second peak, at around 3420 ADC counts, corresponds to the case that a single photo electron was produced in the photo cathode due to the incidence of a photon. This second peak is called the Single Photo-Electron (SPE) peak. A third peak, corresponding to two photo-electrons, expected at 3490 ADC counts, is not seen in Figure \ref{FIG:myspefitter_plot} because the source luminosity is very low. \begin{figure}[H] \centering \includegraphics[width=7cm]{Images/SPE_real_sample_for_paper_Plot_style_73_2021-02-26_10-20-45.jpg} \caption{Example of a digitized single photo-electron anode current pulse measured with the FPM test bench.} \label{FIG:real_spe_plot} \end{figure} The measurement, with the FPM test bench, is performed in a low-intensity regime of the pulsed nano-LED, such that one photo-electron is produced in the photocathode for every five light flashes, on average. The number of collected events is 50000 and the charges are estimated by integrating the anode signal in a 14~ns time window (see section \ref{Text:PulseShape}). Before starting the measurement, FPMs are warmed up for about one hour to ensure stability of the electronics and PMT responses. The gain is derived by fitting the charge spectrum to the analytical model adopted by \cite{Biasuzzi20} and \cite{Caroff19}, using a downhill simplex method (Nelder-Mead). The analytical model is the same as the one used by \cite{Biasuzzi20} and \cite{Caroff19}, for the NectarCAM project. It consists of a Gaussian, parametrized by its centroid and its width ($\sigma_{ped}$), to account for the pedestal peak and a single photo-electron distribution convolved with a Poisson law to account for the distribution of the number of photo-electrons produced per light pulse. The single photo-electron distribution is based on a Gaussian centered to the SPE peak and a half Gaussian\footnote{The normalized distribution of a half Gaussian centered on the origin is given by ${f(y)={\frac {\sqrt {2}}{\sigma {\sqrt {\pi }}}}\exp \left(-{\frac {y^{2}}{2\sigma ^{2}}}\right)\quad y\geq 0}$ and $f(y) = 0$ when $y < 0$. $\sigma$ is the standard deviation of the normal distribution.\\} to account for the low-charge component. This model is suggested by simulations of photomultiplier response to single photo-electron (\cite{Anthony18}). The half Gaussian could be attributed to events with a lower electron multiplication at the first dynodes, such as when the photo-electron is produced at the first dynode instead of at the photocathode or when the photo-electron misses the first dynode or when electrons are backscattered from the dynode (see \cite{Bellamy94}, \cite{Takahashi18} and \cite{KL12}). The fraction of the single photo-electron events in the half Gaussian is called the low charge fraction. In our analyses, the low-charge fraction is fixed to a value of 0.17 per single photo-electron. This value is an average derived from measurements at high gain with few PMTs (see the conclusion for further discussions of the low-charge component fraction). The width of the low charge component distribution is related to the gain and SPE resolution using the relationship described in \cite{Caroff19}. The single photo-electron distribution is convolved with a normal distribution with a width of $\sigma_{ped}$ to account for the charge dispersion due to electronic noise fluctuations. A sample spectrum and its best-fit parameters are shown in Figure \ref{FIG:myspefitter_plot}. Among the fitted parameters of the model, we obtain the mean single photo-electron charge which is the mean value of the best fitting single photo-electron distribution relative to the mean pedestal\footnote{A charge value relative to the mean pedestal is called a net charge.}. Considering the camera measurement chain (signal pre-amplification by the PACTA, line impedances and amplification by the ACTA), a single photo-electron amplified with a gain of 40000 corresponds to a mean net charge of $\bar{Q}_{SPE} \equiv$ 58 ADC counts. The other parameters provided by the fit are the pedestal characteristics (mean and standard deviation $\sigma_{ped}$), the single photo-electron rms and the intensity (mean number of photo-electrons per light pulse). The total number of free parameters is 7. \begin{figure}[H] \centering \includegraphics[width=8cm]{Images/SPE_for_paper_data_RUN3_FPM_v2_PROTO_Drawer_6_ch0_Plot_style_44_2019-09-23_18-19-59_v2} \caption{Example of a charge spectrum recorded with 50000 events. The blue curve shows the measured spectrum, the red curve shows the best fitting model. The best fitting parameters are included as a legend.} \label{FIG:myspefitter_plot} \end{figure} A PMT gain of 40000 (nominal gain) is obtained by tuning the high voltage such that the mean single photo-electron charge is 58 ADC. In such a condition, the high voltage is called the nominal high voltage. In principle this nominal high voltage could be found iteratively by taking SPE spectra, fitting the parameters, and adjusting the voltage using some ad-hoc gain vs voltage function. However, the large number of required events (the spectrum is made with 50000 events to get a best fit of the mean single photo-electron charge with a statistical uncertainty of $\sim$~0.5\%) and the computing time of the model fit (around 1 minute per spectrum in our i7 8 core computer) prevent us from using this method for a large number of FPMs due to the long acquisition and analysis durations. The high voltage tuning method has to be fast because, during the mass production phase, FPMs have to be calibrated quickly along the production flow to reduce the operation costs in industry. Instead, we implemented an iterative two-step method that uses the mean charge produced by intense light pulses (measurement in a high-intensity regime) as a tracer of the gain variation as a function of high voltage. The intensity of the flash LED pulser is tuned such that the mean number of photo-electrons is about 60 ph.e. per event. With such an intensity, the signal from the HG line is strong enough while staying far from saturation (at $\sim$ 200 ph.e.). The mean charge is measured with 4000 events by integrating the anode pulses in a 16~ns time window. In this condition the accuracy of the resulting mean net charge is better than 0.5\%. We estimate the gain assuming that the mean net charge is the sum of individual photo-electron charges, i.e. the gain is the measured mean net charge divided by the known number of photo-electrons. In a first step, the number of photo-electrons produced in each PMT photocathode is derived from three first acquisitions performed at an initial high voltage: (i) a single photo-electron spectrum measured in the low-intensity regime to estimate the initial PMT gain, (ii) a mean charge measurement in the high-intensity regime that produces about 60 ph.e. per light pulse and (iii) a measurement without light pulse to estimate the electronic pedestal with 4000 events. The mean net charge $\bar{Q}_{init}$ is obtained by subtracting from the mean charge the mean electronic pedestal value and the number of photo-electrons is the mean net charge divided by the initial gain $G_{init}$. With those values, we derive $\bar{Q}_{nom}$, the mean net charge expected when the gain is 40000 : $\bar{Q}_{nom} = \bar{Q}_{init} \times 40000 / G_{init}$. The second step is an iterative process that tunes the high voltage to the nominal value $HV_{nom}$ until the measured mean net charge reaches the $\bar{Q}_{nom}$ value. This step consists of consecutive high-intensity and pedestal acquisitions (as the previous cases (ii) and (iii)) at several high voltages. The change in high voltage between each iteration is based on the observation that the gain varies as a power-law of the high voltage (e.g. see Figure 7 of \cite{Toyama15} and Figure 4 of \cite{Mir17}). The iterations are stopped once the difference between the measured mean net charge and the $\bar{Q}_{nom}$ target value is less than 0.5\%. This accuracy limit is determined by our precision limit of the high-voltage measurement on the IB that was designed to have a setting and measurement accuracy of better than 1~V. Finally the last step consists of measuring the photo-electron spectrum with 50000 events in the low-intensity regime in order to fit the final gain with its uncertainty as well as other parameters which characterize the performance of the PMT (mean and rms values of the electronic noise and single photo-electron rms). This method is faster than the interpolation method in the SPE regime. We need only a few thousand events to get a reliable mean net charge value. However its drawback is that we have to know the mean number of photo-electrons with an initial single photo-electron fit so the error of the initial gain measurement propagates through the other measurements. The nominal voltages are all between 920~V and 1150~V, well within the range that can be generated by the CW supply (see section \ref{Text:FPM_Nect}), leaving considerable room to increase the voltages during the lifetime of the PMTs to compensate for gain decrease with time, or should CTA wish to take data with a higher gain value. The nominal high voltage determination method has been developed and extensively tested on 7 FPM prototypes. For a batch of 49 PMTs, the total process duration is about 50~min. Nominal high voltages are reached in less than 5 iterations (see Figure \ref{FIG:Iterations}). Figure \ref{FIG:Gain_hist} shows the histogram of the resulting nominal gains. The mean value of the gain is 40012 and its standard deviation is 355. The standard deviation is $\sim$~0.9\% of the mean. This quantifies the accuracy of the iterative method used to determine the nominal high voltage. \begin{figure}[H] \centering \includegraphics[width=8cm]{HV_nominal_tracking_7FPMs_49DUs_Plot_style_19_2021-02-26_10-32-48} \caption{Nominal high voltage tuning iterations for 7 FPMs (49 PMTs). The dashed red line shows the goal value of 40000. The gain estimations are done with the methods described in section \ref{Text:Gain}.} \label{FIG:Iterations} \end{figure} The nominal high voltage values of the 427 PMTs measured with our test bench are in agreement with those provided by Hamamatsu (see Figure \ref{FIG:Nom_vs_Hama}), the mean of the differences is 0.5\%, and the standard deviation of the differences is 1\%. At their nominal gain, the excess noise factor\footnote{This factor is ENF = $\sqrt{1+\sigma_{SPE}^2}$ , see equation 1 of \cite{Mir17}, with $\sigma_{SPE}$ the best fitting single photo-electron rms (i.e. SPE RMS in Figure \ref{FIG:myspefitter_plot}). It describes the effect of the charge dispersion caused by the dynode amplification system.} (ENF) of the 427 PMTs derived from the best fitting photo-electron spectrum models, has a mean value of 1.106 and a standard deviation of 0.005, with minimal and maximal values of 1.096 and 1.131, respectively. This is in good agreement with the previous measurements (ENF ranging from 1.05 to 1.12) made by \cite{Mir17} with the same kind of 7-dynodes PMTs. The mean and the standard deviation of the RMS of the electronics noise pedestal obtained with the best fitting parameters of the photo-electron spectrum model are 20.5\% and 0.6\% of the single photo-electron peak, respectively. The measurements made in the high-intensity regime offer the opportunity to estimate the PMT gain with the photo-statistics (or F-factor) method (see \cite{FeganReport}, \cite{Bencheikh92}, \cite{Biller95}, \cite{Mir_int_20}) for comparison. This method uses the Poissonian-law based relation between the mean number of photo-electrons and the variance of the measured charges. The gains obtained with the photo-statistics method are similar to the ones obtained with the single photo-electron fitting method; the mean ratio of the photo-statistics gains to the single photo-electron gains is 0.981 with a standard deviation of 0.025. \begin{figure}[H] \centering \includegraphics[width=8cm]{SPE_nominal_distribution_with_powerlaw_method_Plot_style_24_2019-09-23_16-34-25} \caption{Distribution of the gain at nominal high voltage for 7 FPMs (49 PMTs). The red line shows the mean gain value of 40012. The standard deviation of the distribution is 355.} \label{FIG:Gain_hist} \end{figure} \begin{figure}[H] \centering \includegraphics[width=8cm]{Nominal_vs_Hamamatsu_voltages_Plot_style_0_2021-03-02_15-22-07} \caption{The nominal high voltages for 427 PMTs from our test bench vs nominal high voltages from Hamamatsu data. The green line shows the one-to-one relation between the two voltages.} \label{FIG:Nom_vs_Hama} \end{figure} It has been observed that PMT gain increases with increasing photo-cathode current, which in an IACT is largely determined by the night-sky background (NSB) level, due to the impedances of the last dynodes (e.g. see \cite{Zhong89}). Indeed, at high photo-electron rate, the current in the last stages of the voltage divider circuit is so large that it enhances the potential difference between the last dynodes, leading to an enhancement of the global PMT gain. That effect has been minimised by design, using low-impedance resistors in the last stages of the voltage divider. For NectarCAM we require that the gain variation with NSB rate should be below $2\%$ up to NSB rates of $\sim$~1~GHz, which corresponds to the maximum NSB rate expected during observations at CTA sites. For the whole batch of 61 FPMs, we did multiple pedestal and high-intensity regime (with 60 ph.e. pulses) acquisitions with NSB rates ranging from $\nu_{nsb}= $ 4~MHz to 15~GHz. For each PMT and each NSB rate, we estimated the gain with the measurement of the mean net charge using the method described above. We estimate the NSB rate from the variation of the pedestal width with illumination. Such a variation is expected due to the single photo-electron pulses induced by NSB photons, which produce fluctuations of the electronic baseline. The NSB rate is derived from pedestal acquisitions with the formula indicated in Eq. \ref{eq:nsb_estimation} where $\sigma_{ped}(\nu_{nsb})$ is the standard deviation of measured charges with a NSB rate $\nu_{nsb}$, $\sigma_{ped}(\nu_{nsb}=0)$ is the standard deviation of measured charges without NSB (i.e. the rms of electronic noise), $G$ is the gain of the PMT, $T$ is the duration of the integration window (in our case we choose $T = 32$~ns). The subtraction of 2~ns from the duration $T$ accounts for time window edge effects since, with a width of $\sim$2.3~ns FWHM (see section \ref{Text:PulseShape}), the anode pulses produced by NSB induced photo-electrons, that start within $\leq$ 2~ns of the end of the time window are not fully integrated. \begin{equation}\label{eq:nsb_estimation} \nu_{nsb} = \frac{\sigma_{ped}^2(\nu_{nsb}) - \sigma_{ped}^2(\nu_{nsb}=0)}{G^2*{\rm ENF}^2*(T-2)} \end{equation} \noindent This formula is based on the fact that the variance of the baseline ($\sigma_{ped}^2$) increases with the intensity of random background light (see \cite{Bencheikh92}). Its validity has been verified at low NSB rate by comparing the rates obtained with Eq. \ref{eq:nsb_estimation} with the photo-electron counting rates; the differences between the two rates are $<$~10\% for NSB rates ranging from 2~MHz to 200~MHz. We compute the relative gain with a reference at $\nu_{nsb}=$~0.3~GHz which is representative of the NSB rate expected during observations with a MST telescope pointing towards the Galactic plane. Figure \ref{FIG:NSB_3D_map} shows the distribution of the relative gain variation versus NSB rate of the 427 PMTs. As expected, the gains increase at high NSB rates. However, below 1~GHz most of the relative gain variations are below 2\%, which is compliant with the NectarCAM requirements. \begin{figure}[H] \centering \includegraphics[width=8cm]{NSB_3D_map_log_freqs_Plot_style_12bis_2020-01-16_15-22-11} \caption{Distribution of the relative gain variation of the 61 FPMs (427 PMTs). The sum of entries in each vertical column is 100\%. The orange line shows the limit on PMT gain variation of 2\%, below a NSB rate of 1~GHz. \label{FIG:NSB_3D_map} \end{figure} \subsection{Charge resolution}\label{Text:ChargeRes} The relative charge resolution is the relative dispersion of the measurements of the number of photo-electrons detected with a PMT when it is illuminated with light pulses of known intensity. The charge resolution is an important characteristic for CTA since it directly influences the final energy resolution for gamma-ray observations. Definitive measurements of the NectarCAM charge resolution will be done at the integration site at IRFU, once the full system is assembled, as this requires a realistic in-camera temperature profile and RFI (Radio-Frequency Interference) environment, which can only be achieved when the full system is assembled and operated in realistic conditions. In this section, we describe the method we developed to evaluate the charge resolution of the FPMs with our test bench. The charge resolution is measured in the presence of simulated NSB light in order to reproduce some of the conditions expected during observations. The charge resolution tests verify that the signal-to-noise ratio (SNR) is within specification over almost four orders of magnitude of pulse amplitude. This SNR test serves as a "system test" for the FPM, verifying the integrity of the signal chain, and that there are no unexpected sources of noise present (pickup from the CW supply, ringing in the pre-amplifier, etc). In the production phase, failure of these SNR tests would require the manufacturer to diagnose and repair the faulty DU or FPM boards. The charge resolution is measured with PMT high voltages set to their nominal values (i.e. a gain of 40000 -- see section \ref{Text:Gain}) and with the NSB simulation system tuned to produce a rate of $\sim$~0.125~ph.e./ns (i.e. a NSB rate of $\sim$~125~MHz\footnote{Actually, the NSB rates measured with our PMTs had a mean of 124~MHz with a standard deviation of 15~MHz depending on their position on the FPM test bench.}), as specified by CTA (\cite{HintonMSTReq14}). During a data taking run, the PMTs are illuminated with 4000 light pulses at a given intensity emitted by the nano LED. Several runs are performed with light pulse intensities ranging from $\sim$~1 to $\sim$~5000 ph.e. per light pulse. An additional acquisition is made without light pulses to estimate the pedestal characteristics and to control the NSB rate. For each set of 4000 events, we extract the charges by integrating the anode pulses in a 16~ns time window and calculate the mean net charge value $\bar{Q}$ and the standard deviation of the distribution $\sigma_Q$. The charge resolution is then given by $r = \nicefrac{\sigma_Q}{\bar{Q}}$. At nominal PMT gain, the anode signal of the HG line is near saturation when $\bar{Q} \gtrsim$~160 ph.e. Consequently, for larger pulses, the charge is measured with the anode signal from the LG line. In the overlap range when the anode signal of the LG line is significant enough and the one of the HG line does not saturate (from $\bar{Q}$ = 70 to $\bar{Q}$ = 150 ph.e.), we measured an average HG/LG mean net charge ratio of $\langle r_{hg/lg} \rangle$ = 13.58 with a standard deviation of 0.29 for all the data of the 427 PMTs. As expected, this mean net charge ratio is in quite good agreement with the ratio of the PACTA gains (see section \ref{Text:FPM_Nect}). Figure \ref{FIG:Charge_res_3D_map} shows the distribution of the charge resolutions measured with the 427 PMTs as a function of the mean net charge of the pulse $\bar{N}_{pe}$ expressed in photo-electron unit; i.e. $\bar{N}_{pe} = \bar{Q} / \bar{Q}_{SPE}$ for charges measured with the HG line and $\bar{N}_{pe} = \bar{Q} / \bar{Q}_{SPE} \times r_{hg/lg}$ for charges measured with the LG line. The dashed curve is the theoretical lower limit corresponding to the Poissonian statistics ($\nicefrac{1}{\sqrt{\bar{N}_{pe}}}$), broadened by the ENF of the PMT. The blue curve shows the limit on charge resolution for acceptance of the DU/FPM for use in NectarCAM for a NSB rate of 125 MHz, while the green curve shows the goal for a NSB rate of 125 MHz. The measured charge resolutions for all of the 427 DUs are within the acceptable bounds. As expected, for high mean net charges (above 100 ph.e.), the measured charge resolution asymptotically tends toward the Poissonian limit broadened by the ENF because, in that case, the main source of fluctuations is the number of photo-electrons, well above other sources such as the electronic noise or NSB. This is not the case for low mean net charges whose fluctuations are due to a combination of Poissonian, electronic noise and NSB fluctuations. The drop of the charge resolution for charges $\gtrsim$~4000~ph.e. (see Figure \ref{FIG:Charge_res_3D_map}) is not an improvement of the charge resolution. It originates from an underestimation of charges due to the saturation of the anode signal in the LG line. Indeed, the electronics chain (PACTA, ACTA and ADC) was designed to process anode current pulses up to $\sim$ 3000 ph.e. when the PMT gain is 40000 (see section \ref{Text:FPM_Nect}). The presented charge resolution does not take into account the possible systematic effects that could deteriorate the overall charge resolution when the PMTs are installed in the camera, such as e.g. the variations of the pedestal levels and PMT gains with temperature, which lead to inaccurate estimation of charges. The impact of those systematic effects on the charge resolution at the camera level will be presented elsewhere, in a dedicated publication. \begin{figure}[H] \centering \includegraphics[width=8cm]{Images/Charge_Resolution_3D_map_Plot_style_13_2020-08-27_15-10-57.jpg} \caption{Charge resolution distribution measured with a NSB rate of $\sim$~125~MHz for 61 FPMs (427 PMTs). The sum of entries in each vertical column is 100\%} \label{FIG:Charge_res_3D_map} \end{figure} \subsection{Pulse shape}\label{Text:PulseShape} The first level of trigger (so called level 0) of the NectarCAM electronics is set to the level of $\sim$~5 ph.e. amplitude of the anode signal. The timing performance of this trigger is directly related to the impulse response function of the system for a single photo-electron. Due to the limited bandwidth of the detection chain, the impulse response function is Gaussian-like with a finite rise time and a delay (transit time) after the photo-electron generation. The transit time is affected by fluctuations that are determined by the point of illumination on the photocathode and the propagation of the electron cascade within the PMT. In this section, we present measurements of the single photo-electron pulse shape as well as its timing properties. We measured the single photo-electron pulse shapes of the 427 PMTs with the data acquired for the single photo-electron measurements at nominal gain (see section \ref{Text:Gain}). In order to avoid events from pedestal or low-charge components and events produced by two photo-electrons, we selected pulses with net charges $Q$ of $\pm$~2 ADC counts around the single photo-electron peak $\bar{Q}_{peak}$ of the photo-electron spectrum: i.e. $Q \in [\bar{Q}_{peak} - 2, \bar{Q}_{peak} + 2]$ with $\bar{Q}_{peak} \simeq1.14 \cdot\bar{Q}_{SPE} \simeq 66$ ADC counts. To get a better timing resolution in order to determine the mean pulse shape from the acquisition, we re-binned each pulse from the sampled resolution of 1~ns to a resolution of 0.1~ns by quadratically interpolating the measured values. Before the quadratic interpolation, the arrival time of each pulse is estimated by fitting each measured pulse with an analytical model composed of three Gaussian functions, where the first Gaussian accommodates the main pulse, and the two other Gaussians model oscillations that occur in the trailing wing of the pulse. The position of the first Gaussian is used to estimate the arrival time of the pulse within the sampling window. The arrival time is the time delay with respect to the trigger signal from the Flash LED pulser. The interpolated pulses, with their sampled resolution of 0.1~ns, are shifted by their respective time delays. The resulting shifted pulses are stacked together and then averaged. Figure \ref{FIG:Stacked_pulses} shows the resulting averaged single photo-electron pulses of the 427 PMTs. The shape of mean pulses is in agreement\footnote{The delays of the second and third pulse peaks, as well as the ratio of the amplitude of the main pulse peak to the amplitude of the second pulse peak are similar.} with that measured directly at the output of the PACTA with an oscilloscope (internal NectarCAM note \cite{Sanuy_private20}). The pulse is characterized by a first impulse with an amplitude of about 23 ADC counts, followed by damped oscillations with a frequency of about 280 MHz. The total duration of the pulse is about 12~ns when the second oscillation is accounted for. Figure \ref{FIG:FWHM_pulses} presents the distribution of the single-photo-electron pulse widths (FWHM) of the 427 PMTs measured with the extracted pulse shapes of Figure \ref{FIG:Stacked_pulses}. The average width of the primary pulse is 2.29~ns FWHM with a standard deviation of 0.05~ns. This value is smaller than the mean value of $\simeq$~3~ns measured by \cite{Toyama15} and the value of 2.86~ns with a standard deviation of 0.05~ns provided by Hamamatsu which can be explained by the fact that we average the pulses after charge selection and time delay correction. The LST team measured a single photoelectron pulse width of 2.6~ns FWHM by minimizing the pulse broadening due to the dispersion in time delay with a mask which has small pinholes in front of the photocathodes of 8-dynode PMTs (see \cite{LSTmeeting2020}). Without selection and time delay correction we measure an average pulse width of 3.09~ns with a standard deviation of 0.10~ns, in agreement with \cite{Toyama15} and the values measured by Hamamatsu. The dispersion in time delays is attributed to differences in the electron path lengths within the PMT for different incident photon locations on the photocathode. Electrons generated at the center of the photocathode have different trajectories to the first dynode than electrons generated at the outer parts. The time delay dispersion is characterized by the transit time spread (TTS) which is the FWHM of the time delay distribution. Figure \ref{FIG:TTS_distribution} shows the distribution of the TTS values measured with the 427 PMTs. We measure TTS values with a mean of 1.76~ns and a standard deviation of 0.28~ns, in agreement with (1.81 $\pm$ 0.12)~ns measured by \cite{LSTmeeting2020} and with the Hamamatsu technical notes ($<$~2~ns) but not with the initial requested specifications for the PMTs of CTA ($<$~1.5~ns FWHM, see \cite{Toyama15}). \begin{figure}[H] \centering \includegraphics[width=8cm]{SPEs_stacking_61FPMs_427DUs_Plot_style_27_2021-03-10_09-48-56} \caption{Mean single photo-electron pulses of the 427 PMTs measured with a gain of 40000. One curve is plotted for each PMT.} \label{FIG:Stacked_pulses} \end{figure} \begin{figure}[H] \centering \includegraphics[width=8cm]{SPE_pulses_fwhms_Plot_style_29_2021-03-10_10-34-18} \caption{Distribution of the single photo-electron pulse width of the 427 PMTs at a gain of 40000. The average and standard deviation of the distribution are 2.29 ns and 0.055 ns, respectively.} \label{FIG:FWHM_pulses} \end{figure} \begin{figure}[H] \centering \includegraphics[width=8cm]{TTS_with_spe_all_61FPMs_Plot_style_27b_2021-03-10_10-25-55} \caption{Distribution of the TTS values measured with the 427 PMTs at a gain of 40000. The average and standard deviation of the distribution are 1.76 ns (red line) and 0.28 ns, respectively.} \label{FIG:TTS_distribution} \end{figure} \subsection{Afterpulses}\label{Text:AfterPulses} Afterpulses are produced by residual ions within the PMT that can release up to $\sim$~50 ph.e. when they interact with the photocathode or with a dynode. They degrade the performance of the Cherenkov camera in several ways. During observations, these spurious pulses mimic the signal of Cherenkov photons and may increase the electronics dead-time of a camera due to an increased trigger and read-out rate. If their rate is too high the trigger threshold level has to be increased to reduce the dead-time, resulting in an increase of the energy threshold of gamma rays; afterpulses have a direct impact on the energy threshold of a Cherenkov camera (see \cite{Mir97}). Moreover, afterpulses affect the separation of events due to gamma from the (background) events due to hadrons in the shower image analyses. The PMTs were designed by Hamamatsu, in collaboration with CTA, to have an afterpulse probability of $<$~2~$\times$ 10$^{-4}$ when the pulse threshold is set to 4 ph.e. (see \cite{Toyama15}). Here we verify that the PMTs furnished to NectarCAM meet this specification. With the FPM test bench, the afterpulse rate was measured by illuminating a FPM with a continuous rate of simulated NSB photons. The simulated NSB photons produce afterpulses which might occur by chance in the acquisition windows. The distribution of the net charges measured in the acquisition windows allows highlighting the contribution of afterpulses since the accelerated residual ions produce on average more photo-electrons per interaction than the NSB photons. The NSB rate was set to a value that allows for a clear detection of individual photon pulses from the NSB and the characterization of their waveforms, while producing enough afterpulses. The acquisition system was triggered at a frequency of 800 Hz. We used PMT gains of 100000, which is the same setting that is also used by Hamamatsu to determine the afterpulse probabilities. Adoption of such a gain improves the pulse detection and the accuracy of the amplitude and net charge measurements. The number of afterpulses is obtained by counting the number of pulses with amplitude and net charge larger than a given number of photo-electrons. To estimate the afterpulse probability, the resulting number counts are divided by the total number of photo-electrons which is derived from the measured NSB rate. Figure \ref{FIG:Afterpulse_probability} presents the event probability as a function of the threshold net charge in units of photo-electrons, obtained from 1.5 $\times$ 10$^6$ waveforms with a duration of 48~ns each \footnote{48 ns is the maximum duration of an acquisition with our test bench.} and a mean NSB rate of $\sim$~20 MHz. The shape of the distribution shows the contributions of NSB pulses and afterpulses in the event counts. The probability distribution is dominated by NSB pulses in the low charge range (up to $\sim$ 3.5 ph.e.) and by afterpulses in the high charge range. The afterpulse probability ranges from 0.6 $\times$ 10$^{-4}$ to 0.9 $\times$ 10$^{-4}$ for charges $>$ 4 ph.e., which is in agreement with the PMT specifications and previous measurements (see \cite{Toyama15}, \cite{Mir17} and \cite{Mir16}). \begin{figure}[H] \centering \includegraphics[width=8cm]{Images/Afterpulses_paper_Plot_style_67_2020-08-27_16-21-50.jpg} \caption{Afterpulse probabilities for the 7 PMTs of a focal plane module prototype measured at a gain of 100k for comparison with Hamamatsu measurements. The dashed line shows the Hamamatsu specifications. The probability distribution is dominated by NSB photons below $\sim$ 3.5 ph.e.} \label{FIG:Afterpulse_probability} \end{figure} \section[Conclusions]{Conclusions} We developed a test bench to measure the performance of NectarCAM focal plane instrumentation. The test bench is fully functional and the measurements, performed with 61~FPMs (427 PMTs), show that the performance of the focal plane modules is in agreement with our expectations, and with measurements done previously on individual PMTs. Our tests validate the design of the FPMs for use with NectarCAM and CTA. For each PMT we find high-voltage values producing a nominal gain of 40000 that are similar to those determined by Hamamatsu. We measured excess-noise factors with a mean of 1.106 with a standard deviation of 0.005. At nominal gain, the dark pedestal RMS values of $\sim$~20\% of the photo-electron peak are compliant with NectarCAM requirements. The gains are stable within $\pm$~2$\%$ when PMTs are subject to NSB rates ranging from a few MHz up to $\sim$~1~GHz. The charge resolutions measured from 1 to $\sim$~3000~ph.e., with a NSB rate of 125~MHz, are within specification. The mean single photo-electron pulse widths and transit time spreads are 3.09~ns with a standard deviation of 0.10~ns and 1.76~ns with a standard deviation of 0.28~ns, respectively, consistent with the values measured by \cite{HamamatsuPMTDataSheet}, \cite{Toyama15}, \cite{Mir17} and \cite{Mir16} . After correction of the time delay with respect to the trigger, we obtained a mean single photo-electron pulse intrinsic width of 2.29~ns with a standard deviation of 0.05~ns. The afterpulse rates measured with 7~PMTs are $< 2\times10^{-4}$ for charges $>$ 4 photo-electrons, in accordance with the specification of Hamamatsu. We calculated the autocorrelation functions of single photoelectron pulses measured with 427 PMTs of the Focal Plane Modules for NectarCAM. After removing the contribution of noise fluctuations, the width of the auto-correlated single photoelectron pulses are of $\sim$~3.2 ns in average. The bandwidths of the full acquisition chain, calculated with the Fast Fourier Transform of auto-correlated pulses, are about $\sim$~77 MHz. The average bandwidth of PMTs alone is $\sim$~82 MHz, after correction of the electronic chain (PACTA, ACTA and Nectar chip) response. The slight difference between those values shows that the electronic chain does not reduce significantly the bandwidth of PMTs; the electronic chain response is adapted to the response of PMTs. The similarity between the average auto-correlated single photoelectron pulses and the auto-correlation of the extracted mean single photoelectron pulses suggests that the method to extract the single photoelectron pulse does not twist too much the shape of single photoelectron pulse. After validation with our test bench, the 61~FPMs were installed in a NectarCAM prototype camera which was mounted on a MST telescope prototype (see \cite{Garczarczyk18}) at Adlershof (Germany) during May--June 2019 to perform measurements and tests under realistic operation conditions and to record images of air showers (e.g. see Figure~4 of \cite{Glicenstein19}). The results of that observation campaign will be presented elsewhere (\cite{NCam20}). We upgraded our test bench to allow the characterization of the 265 FPMs that are produced by industry to furbish the NectarCAM Qualification Model (QM), which fully reflects the design of the final model. The 61 FPMs are spare parts of the QM. One of the improvements consists of replacing the wooden shelf on which FPMs are installed up to now (see Figure \ref{FIG:FPM_TB_pic}) by a robust mechanical structure. That structure is also designed to enhance the ergonomy for easy and reliable plugging and unplugging of FPMs by the test operator. With regards to measurements and analyses, we have identified a need to improve the method for estimating the PMT gains. So far we estimate the gains by assuming a low-charge component fraction of 0.17 per single photo-electron for each PMT but our analysis suggests that this fraction may vary significantly among PMTs. This may result in systematic uncertainties of about 1.3$\%$ on the gain (see \cite{Caroff19}). In order to reduce these uncertainties, we will implement a new method in our test bench, where we will estimate the low-charge component from single photo-electron measurements at high gain. During the mass production of cameras for CTA, the FPMs of NectarCAM will be produced in industry including a characterization of the modules by the manufacturer using a dedicated test bench before delivery. The fully automated test bench will perform a series of tests per batch of 7 FPMs. That test bench will comprise the same elements as the FPM test bench described in this paper but it will be mounted on a mechanical structure consisting of a dark box with a hinged lid so that it can be used within a production chain during daylight. Moreover, the measurements that will be performed in industry will be optimized to reduce time and cost. In that way, fully qualified and characterized FPMs for the NectarCAM can be delivered by an industrial partner, avoiding the need for costly performance checks at the lab and guaranteeing the delivery of homogeneous, high-quality hardware for NectarCAM on CTA. \section*{Acknowledgement} The authors are grateful for the support of the R\'egion Occitanie and Microtec. The authors want to thank R. Mirzoyan and D. Nakajima for useful discussions. This work was conducted in the context of the NectarCAM Project of CTA. The CTA Consortium gratefully acknowledge financial support from the following agencies and organizations listed on this webpage : https://www.cta-observatory.org/consortium\_acknowledgments/ . \medskip \noindent This paper has gone through an internal review by the CTA Consortium. \printcredits \footnotesize

{'timestamp': '2021-10-13T02:23:05', 'yymm': '2110', 'arxiv_id': '2110.06030', 'language': 'en', 'url': 'https://arxiv.org/abs/2110.06030'}

high_school_physics

This unusual clock has a transverse movement and stands dramatically as an unsupported sphere on any flat surface. Touch it, and it wobbles gently to and fro, gradually coming to rest in the same upright position. The spherical cap is the second hand and the two segments show minutes and hours in numerical format. This contemporary clock is battery operated and is an impressive 15cm in diameter (6 inches).

{'timestamp': '2019-04-24T10:58:41Z', 'url': 'http://www.crowstone.co.uk/lp/halley_silver_large_spherical_table_clock/', 'language': 'en', 'source': 'c4'}

high_school_physics

Sign In to Financial Professional Site Institutional Mutual Funds Variable Insurance Products Defined Contribution (401k) Workplace 529 Program Portfolio Matrix Review Fund Commentary Sales Charges & Breakpoints Fund Comparison Social Security Optimizer Developing Your Practice Franklin Templeton Academy RIA Resources Volatility Resources for Your Clients Here's a helpful set of tools you can use to talk with clients now. Franklin New York Tax-Free Income Fund - A1 May 31 2019 Dividend Policy and Distribution Date Changes Effective 5/1/19 - Read More On 5/1/19, this fund began declaring dividends daily, but continues to pay out monthly. It changed its distribution policy from fixed pay, where dividend rates are set based on projections, to variable pay, where income is distributed as it’s earned. The fund also changed its pay date to the end of each month. It made a smaller supplemental distribution in late April and paid its first variable pay distribution in late May. Please see the prospectus supplement for more information. Mar 01 2019 Updated Pricing Schedule for Class A/A1 Long-Term Fixed Income Funds - Read More Effective 3/1/19, the maximum initial sales charge for purchases of less than $100,000 of Franklin Templeton’s long-term fixed income funds (Class A and A1 shares) was lowered to 3.75%. Please see the prospectus supplement for more information. If you have any questions, please contact your financial advisor or call Shareholder Services. Sep 10 2018 Changes to Class A Shares - Read More Class A shares (NASDAQ: FNYTX) were renamed Class A1 and closed to new investors. Existing Class A1 shareholders may continue to add money to their accounts. A new Class A share (NASDAQ: FNYQX) launched with a different expense structure, in which new shareholders can invest. The fund also changed certain sales charges and commissions paid to dealers on Class A and A1 shares. Please see the prospectus supplement for more information. If you have additional questions, please contact your financial advisor or call Franklin Templeton. FNYTX Franklin New York Tax-Free Income Fund Share Class A1 New investors must invest in new Class A shares. This A1 share class is available only for existing investors as of 9/7/18. Distributions & Pricing Dividend Distributions Further Information Ex-Dividend Fund: This type of fund declares periodic dividends, which are paid to shareholders who own the fund as of a certain date - known as the "record date". After the record date, the fund is said to be trading "ex-dividend"(literally, without the dividend). An investor who buys shares during the interval between the record date and the payment of the dividend does not receive the dividend. A fund that's in the ex-dividend period is marked with an "x" in newspaper listings. Daily Accrual Fund: This type of fund typically declares income dividends each day that its net asset value is calculated. These dividends remain in the fund until the designated reinvestment or payment date and are distributed to the shareholders, generally once a month, as additional shares or cash, respectively. This fund is a Daily Accrual fund Rates and Yields Further Information 30-day Standardized Yield (for non-money market funds) is the fund's net income per share (dividends minus accrued expenses) for the 30-day period ending on the last day of the month, divided by the maximum offering price on that day (annualized). Taxable Equivalent Yield is the yield that must be received from a taxable investment to deliver the same after-tax return as a tax-free investment. 7-day Current Yield reflects the interest income per share a money market fund earned on its investments for the last 7 days (annualized). Does not include reinvested dividends. 7-day Effective Yield is calculated like the 7-day current yield, but it includes reinvested dividends. Read important information about results and other investment disclosures Taxable Equivalent Distribution Rate As of 01/17/2020 (updated daily) 30-Day Standardized yield As of 12/31/2019 (updated monthly) Taxable Equivalent Yield As of 12/31/2019 (updated monthly) Distribution Rate and Breakpoints As of 01/17/2020 (updated daily) At Net Asset Value TWELVE-MONTH-ROLLING DISTRIBUTIONS PER SHARE As of 01/18/2020 (updated daily) View historical data Distribution amounts may be estimates Further Information Estimated distributions can change depending on current market conditions and number of shares outstanding. Changes can occur up to the record date. Estimates don't include ordinary income distributions the funds may be making. A fund's capital gain distributions per share will be the same for each share class. All dates and distributions are subject to board approval. Amount ($) Reinvestment Date Reinvestment Price ($) Total Distributions Per Share 0.026066 Daily Daily Dec 31 Dec 31 11.17 Tax-Exempt Income 0.023605 Daily Daily Nov 29 Nov 29 11.17 0.025400 Daily Daily Oct 31 Oct 31 11.18 0.026630 Daily Daily Sep 30 Sep 30 11.22 0.025653 Daily Daily Aug 30 Aug 30 11.33 0.028117 Daily Daily Jul 31 Jul 31 11.15 0.024679 Daily Daily Jun 28 Jun 28 11.10 0.027716 Daily Daily May 31 May 31 11.10 0.009800 Daily Daily Apr 30 Apr 26 10.98 0.028300 Daily Daily Apr 3 Apr 1 10.92 0.028300 Daily Daily Mar 5 Mar 1 10.77 0.028300 Daily Daily Feb 5 Feb 1 10.75 Pricing at NAVAs of 01/17/2020 [no-data-msg] Daily Fund Prices and Breakpoints [further-information] Net Asset Value (NAV) is the value of one share of the fund on a given day. NAV change is the change from the last quoted price. The Public Offering Price (POP) is the purchase price for each share of the fund on a given day. It includes the maximum initial sales charge. As of 01/17/2020 View Historical Data [price-info-text] Share Prices ($) NAV (Net Asset Value) NAV Change POP (Public Offering Price) Pop and Sales Charge at Breakpoints 3.75% $11.71 52-week Range ($) NAV Highon 09/04/2019 NAV Lowon 01/23/2019 All investments involve risks, including possible loss of principal. Because municipal bonds are sensitive to interest rate movements, the fund's yield and share price will fluctuate with market conditions. Bond prices generally move in the opposite direction of interest rates. Thus, as prices of bonds in the fund adjust to a rise in interest rates, the fund's share price may decline. Because the fund invests principally in a single state, it is subject to greater risk of adverse economic and regulatory changes in that state than a geographically diversified fund. Puerto Rico municipal bonds have been impacted by recent adverse economic and market changes, which may cause the fund's share price to decline. Changes in the credit rating of a bond, or in the credit rating or financial strength of a bond's issuer, insurer or guarantor, may affect the bond's value. The fund may invest a significant part of its assets in municipal securities that finance similar types of projects, such as utilities, hospitals, higher education and transportation. A change that affects one project would likely affect all similar projects, thereby increasing market risk. CFA® and Chartered Financial Analyst® are trademarks owned by CFA Institute. For investors subject to the alternative minimum tax, a small portion of fund dividends may be taxable. Distributions of capital gains are generally taxable. To avoid the imposition of 28% backup withholding on all fund distributions and redemption proceeds, U.S. investors must be properly certified on Form W-9 and non-U.S. investors on Form W-8BEN. Indexes are unmanaged, and one cannot invest directly in an index. Index returns do not reflect any fees, expenses or sales charges. Your clients should carefully consider a fund's investment goals, risks, charges and expenses before investing. Download a prospectus, which contains this and other information. Your clients should read the prospectus carefully before they invest or send money. Effective 9/10/18 Class A shares closed to new investors and were renamed Class A1 shares. This share class is only available for shareholders who purchased prior to to the close of market on 9/7/18. All investments involve risks, including possible loss of principal. Because municipal bonds are sensitive to interest rate movements, the fund's yield and share price will fluctuate with market conditions. Bond prices generally move in the opposite direction of interest rates. Thus, as prices of bonds in the fund adjust to a rise in interest rates, the fund's share price may decline. Changes in the credit rating of a bond, or in the credit rating or financial strength of a bond's issuer, insurer or guarantor, may affect the bond's value. The fund may invest a significant part of its assets in municipal securities that finance similar types of projects, such as utilities, hospitals, higher education and transportation. A change that affects one project would likely affect all similar projects, thereby increasing market risk. The fund's 30-day standardized yield is calculated over a trailing 30-day period using the yield to maturity on bonds and/or the dividends accrued on stocks. It may not equal the fund's actual income distribution rate, which reflects the fund's past dividends paid to shareholders. Percentage of the fund's returns explained by movements in the Barclays Municipal Bond Index. 100 equals perfect correlation to the index. Based on the 3-year period ended as of the date of the calculation. A measure of the fund's volatility relative to the market, as represented by the Barclays Municipal Bond Index. A beta greater than 1.00 indicates volatility greater than the market. Based on the 3-year period ended as of the date of the calculation. ©2019 Morningstar, Inc. All rights reserved. The information contained herein: (1) is proprietary to Morningstar and/or its content providers; (2) may not be copied or distributed and (3) is not warranted to be accurate, complete or timely. Neither Morningstar nor its content providers are responsible for any damages or losses arising from any use of this information. Taxable Equivalent Yield: The Taxable Equivalent Yield assumes the maximum regular federal income tax rate, maximum New York State and New York City personal income tax rates, and the Medicare tax in effect on June 2017. For investors subject to the alternative minimum tax, a small portion of fund dividends may be taxable. Distributions of capital gains may be taxable. For net asset value (NAV) purchases, a contingent deferred sales charge (CDSC) may apply to shares redeemed within 18 months. The CDSC for equity, alternative and multi-asset funds is 1% and the CDSC for long-, intermediate- and limited-term fixed income funds is currently 0.75%. Effective on or about March 10, 2020, the CDSC will be 1% for all funds. Effective 6/1/2019, the fund's benchmark is the Bloomberg Barclays New York Municipal Bond Index. The Linked Bloomberg Barclays New York Muni Bond/Muni Bond Index reflects performance of the Bloomberg Barclays Municipal Bond Index from the fund's inception through 5/31/2019 and the Bloomberg Barclays New York Municipal Bond Index thereafter.

{"pred_label": "__label__wiki", "pred_label_prob": 0.5158865451812744, "wiki_prob": 0.5158865451812744, "source": "cc/2020-05/en_middle_0076.json.gz/line209191"}

high_school_physics

For a finite real reflection group $W$ with Coxeter element $\gamma$ we give a case-free proof that the closed interval, $[I, \gamma]$, forms a lattice in the partial order on $W$ induced by reflection length. Key to this is the construction of an isomorphic lattice of spherical simplicial complexes. We also prove that the greatest element in this latter lattice embeds in the type W simplicial generalised associahedron, and use this fact to give a new proof that the geometric realisation of this associahedron is a sphere.

{'timestamp': '2019-04-26T08:18:22Z', 'url': 'https://arrow.dit.ie/scschmatart/150/', 'language': 'en', 'source': 'c4'}

high_school_physics

\section{Introduction} In general, like-pairing correlations, such as proton-proton ($pp$) and neutron-neutron ($nn$) pairing usually adopted in understanding the nuclear superfluidity, have isovector spin-singlet ($T=1, J=0$) mode and manifest themselves as nuclear odd-even mass staggering. They contribute to keeping spherical properties due to their $J =$ even couplings against the deformation treated in the mean field. On the other hand, neutron-proton ($np$) pairing correlations in the residual interaction are expected to play meaningful roles in $N \simeq Z$ nuclear structure, and relevant nuclear electro-magnetic (EM) and weak transitions because protons and neutrons in these nuclei occupy the same (or nearby) orbitals leading to the maximum spatial overlap. The $np$ pairing correlations have two different modes, viz. isoscalar (IS) spin-triplet ($T=0, J=1$) and isovector (IV) spin-singlet ($T=1, J=0$) \cite {Chen-Gos, Wolter, Goodman,simkovic03}. The IV $np$ pairing part, whose spin-singlet gives rise to anti-aligned pair, can be investigated by the study of $T=0$ and $T=1$ excited states in even-even and odd-odd nuclei. But the IS $np$ pairing part has a spin-aligned pair structure. For example, deuteron ground state as well as $np$-scattering phase shift analyses indicate strong attractive $np$ pairing features due to the strong $^{3}S_{1}$ tensor force. Inside nuclei, they are believed to contribute mainly to the mean field. But, even after converted to a mean field, it is quite natural to conjecture that there should still remain such strong attractive interactions due to the $np$ pairing in the residual interaction, although direct data implying the IS $np$ pairing inside nuclei are still controversial even in the $N \simeq Z$ nuclear structure study Recently, with the advent of modern radioactive beam facilities producing the $N \simeq Z$ nuclei, competition or coexistence of these two IV and IS $np$ pairing in the residual interaction for $N \simeq Z$ nuclei is emerging as an interesting topic in the nuclear structure. Detailed reports about the recent progress regarding the $np$ pairing correlations in nuclear structure can be found at Refs. \cite{Sweden16,Brend16}. In particular, the IS spin-triplet part in the $np$ pairing has been argued to be very elusive compared to the IV spin-singlet part stemming from the like- and unlike-pairing correlations. Moreover, the deformation may also affect the IS and IV pairing correlations or vice versa because the IS $np$ pairing has $J=$ odd coupling inducing (non-spherical) deformation contrary to the IV $np$ mode keeping $J=$ even coupling. Importance of the $np$ pairing has been discussed two decades ago in our early reports of the single- and double-$\beta$ decay transitions \cite {Ch93,pan} within a spherical quasi-particle random phase approximation (QRPA) framework with a realistic two-body interaction given by the Brueckner $G$-matrix based on the CD Bonn potential. But these works did not explicitly include the deformation, and the IS $np$ pairing was taken into account effectively by a renormalization of the IV $np$ pairing strength. Recent study \cite{Gamb15} regarding the relation of the deformation and the $np$ pairing correlations addressed extensively, by a model combining shell model techniques and mean-field, that the large coexistence of the IV and IS may be found and the deformation can induce the quenching of the IV $np$ pairing. Furthermore, recent experimental data for the M1 spin transition reported strong IV quenching in the $N =Z$ $sd$-shell nuclei \cite{Matsu15}, whose nuclei are thought to be largely deformed. It means that the IV quenching giving rise to IS condensation may become of an important ingredient for understanding the nuclear deformation in those nuclei. In Refs. \cite{Ha17-2,Ha18}, we demonstrated that such IS condensation is really feasible in those $sd$- and $pf$-shell nuclei, in particular, for large deformation region. Similar discussions were carried out by other authors \cite{Bai14,Nils04}. But the deformation was not explicitly taken into account. Ref. \cite{Bai14} argued that the IS $np$ pairing may affect the low-lying Gamow-Teller (GT) state near to the isobaric analogue resonance (IAR) for $pf$-shell $N=Z+2$ nuclei by studying the GT data \cite{Fujita14}. Ref. \cite{Nils04} has performed a self-consistent $pn$-QRPA based on a relativistic HFB approach, which clearly demonstrated the importance of the $np$ pairing, specifically, IV $np$ pairing, for proper understanding the GT peaks. But, very recent calculation of the GT strength for Ni isotopes by $pn$-QRPA \cite{Sadiye18} is based only on the like-pairing correlations in the spherical QRPA. Main interests in this work are how to interrelate the $np$ pairing in the residual interaction with the deformation considered in the mean field on the Gamow-Teller (GT) transition because the IS pairing may compete with the deformation features due to its coupling to non-spherical $J =$ odd states. Most studies have focused on the $N = Z$ nuclei because the $np$ pairing is expected to be larger rather than other $N \neq Z$ nuclei. However, as shown in a recent work \cite{Bertch11}, the nuclear structure of the $N \simeq Z$ nuclei may also be more or less affected by the $np$ pairing correlations. In our recent papers \cite{Ha15,Ha15-1}, we developed a deformed QRPA (DQRPA) approach by explicitly including the deformation \cite{Ha15}, in which all effects by the deformation and the like-pairing ($nn$ and $pp$) correlations are consistently treated at the Deformed BCS (DBCS) and RPA stages. But the $np$ pairing correlations were taken into account only at the DBCS with a schematic interaction and a $G$-matrix type interaction \cite{Ha15-1}. Very recently both effects are simultaneously considered and applied to the GT strength study of $^{24,26}$Mg at Ref. \cite{Ha16} and other $N=Z$ $sd$-shell nuclei at Ref. \cite{Ha17}. We argued that the $np$ pairing correlations as well as the deformation may affect the $sd$-shell nuclear structure and their GT response functions. Along these preceding papers, here, we extend our applications to the GT strength study of $pf$-shell nuclei in order to understand the interplay of the pairing correlations and the deformation in the medium heavy nuclei. In this work, we also investigate how such IS condensation and deformation affect the GT strength distribution in $pf$-shell nuclei, because roles of the IS and IV pairings in the deformed mean field and their effects on the GT strength distributions are still remained to be intensively discussed in those medium heavy nuclei. Our results are presented as follows. In section II, a brief pedagogical explanation of the formalism is introduced. Numerical GT results for $^{56,58}$N i and $^{62,64}$Ni are discussed in Sec. III. A summary and conclusions are done in Sec. IV. \section{Theoretical Formalism} The $np$ pairing correlations change the conventional quasi-particle concept, {\it i.e.}, quasi-neutron (quasi-proton) composed by neutron (proton) particle and its hole state, to quasi-particle 1 and 2 which may mix properties of the quasi-proton and quasi-neutron. Here we explain briefly the formalism, DBCS and DQRPA, to be applied for GT transition strength distributions of some Ni isotopes, in which we include all types pairings as well as the deformation. We start from the following DBCS transformation between creation and annihilation operators for real (bare) and quasi-particle states \cite{Ha15} \begin{equation} \label{eq:HFB_1} \left( \begin{array}{c} a_{1}^{\dagger} \\ a_{2}^{\dagger} \\ a_{\bar{1}} \\ a_{\bar{2}} \end{array}\right)_{\alpha} = \left( \begin{array}{cccc} u_{1p} & u_{1n} & v_{1p} & v_{1n} \\ u_{2p} & u_{2n} & v_{2p} & v_{2n} \\ -v_{1p} & -v_{1n} & u_{1p} & u_{1n} \\ -v_{2p} & -v_{2n} & u_{2p} & u_{2n} \end{array}\right)_{\alpha} \left( \begin{array}{c} c_{p}^{\dagger} \\ c_{n}^{\dagger} \\ c_{\bar{p}} \\ c_{\bar{n}} \end{array}\right)_{\alpha} ~. \end{equation} Hereafter Greek letter denotes a single-particle state (SPS) of neutron and proton with a projection $\Omega $ of a total angular momentum on a nuclear symmetry axis. The projection $\Omega$ is treated as the only good quantum number in the deformed basis. We assume the time reversal symmetry in the transformation coefficient and do not allow mixing of different SPSs ($\alpha$ and $\beta$) to a quasi-particle in the deformed state. But, in a spherical state representation of Eq. (\ref{eq:HFB_1}), the quasi-particle states would be mixed with different particle states in the spherical state because each deformed state (basis) is expanded by a linear combination of the spherical state (basis) (see Fig. 1 at Ref. \cite{Ha15}). In this respect, the DBCS is another representation of the deformed HFB transformation in the spherical basis. If we discard the $np$ pairing, Eq. (\ref{eq:HFB_1}) is reduced to the conventional BCS transformation in a deformed basis. The other merit is that, by expanding all deformed wave functions constructed from the deformed harmonic oscillator basis into the spherical basis, we may exploit the Wigner-Eckardt theorem for more complicated physical operators in the deformed states. Finally, using the transformation of Eq. (\ref{eq:HFB_1}), the following DBCS equation was obtained \begin{equation} \label{eq:hfbeq} \left( \begin{array}{cccc} \epsilon_{p}-\lambda_{p} & 0 & \Delta_{p {\bar p}} & \Delta_{p {\bar n}} \\ 0 & \epsilon_{n}-\lambda_{n} & \Delta_{n {\bar p}} & \Delta_{n {\bar n}} \\ \Delta_{p {\bar p}} & \Delta_{p {\bar n}} & -\epsilon_{p} + \lambda_{p} & 0 \\ \Delta_{n {\bar p}} & \Delta_{n {\bar n}} & 0 & -\epsilon_{n} + \lambda_{n} \end{array}\right)_{\alpha} \left( \begin{array}{c} u_{\alpha'' p} \\ u_{\alpha'' n} \\ v_{\alpha'' p} \\ v_{\alpha'' n} \end{array}\right)_{\alpha} = E_{\alpha \alpha''} \left( \begin{array}{c} u_{\alpha'' p} \\ u_{\alpha'' n} \\ v_{\alpha'' p} \\ v_{\alpha'' n} \end{array}\right)_{\alpha}, \end{equation} where $E_{\alpha \alpha''}$ is the energy of a quasi-particle denoted as $\alpha''$(=1 or 2) in the $\alpha$ state. The pairing potentials in the DBCS Eq. (\ref{eq:hfbeq}) were calculated in the deformed basis by using the $G$-matrix obtained from the realistic Bonn CD potential for nucleon-nucleon (N-N) interaction as follows \begin{equation} \label{eq:gap} \Delta_{{p \bar{p}_\alpha}} = \Delta_{\alpha p \bar{\alpha}p} = - \sum_{\gamma} \Big[ \sum_{J, a, c } g_{\textrm{pp}} F_{\alpha a \bar{\alpha} a}^{J0} F_{\gamma c \bar{\gamma} c}^{J0} G(aacc,J,T=1)\Big] (u_{1p_{\gamma}}^* v_{1p_{\gamma}} + u_{2p_{\gamma}}^* v_{2p_{\gamma}}) ~, \end{equation} \begin{eqnarray} \label{eq:gap_pn} \Delta_{{p \bar{n}_\alpha}} = \Delta_{\alpha p \bar{\alpha}n} = &-& \sum_{\gamma} \Bigg[ \Big[\sum_{J, a, c} g_{\textrm{np}} F_{\alpha a \bar{\alpha} a}^{J0} F_{\gamma c \bar{\gamma} c}^{J0} G(aacc,J,T=1)\Big] Re(u_{1n_{\gamma}}^* v_{1p_{\gamma}} + u_{2n_{\gamma}}^* v_{2p_{\gamma}}) \\ \nonumber &+& \Big[ \sum_{J, a, c} g_{\textrm{np}} F_{\alpha a \bar{\alpha} a}^{J0} F_{\gamma c \bar{\gamma} c}^{J0} iG(aacc,J,T=0)\Big] Im (u_{1n_{\gamma}}^* v_{1p_{\gamma}} + u_{2n_{\gamma}}^* v_{2p_{\gamma}}) \Bigg]~, \end{eqnarray} where $F_{ \alpha a {\bar \alpha a }}^{JK}=B_{a}^{\alpha}~B_{a}^{\alpha} ~{(-1)^{j_{a} -\Omega_{\alpha}}}~C^{JK}_{j_{a} \Omega_{\alpha} j_{a}-\Omega_{\alpha}}(K=\Omega_{\alpha} - \Omega_{\alpha})$ was introduced to describe the $G$-matrix in the deformed basis with the expansion coefficient $B_{\alpha}$ calculated as \cite{Ha15} \begin{equation} \label{eq:sps_ex} B_{a}^{\alpha} = \sum_{N n_z \Sigma} C_{l \Lambda { 1 \over 2} \Sigma}^{j \Omega_{\alpha}} A_{N n_z \Lambda}^{N_0 l}~b_{N n_z \Sigma} ~,~A_{N n_z \Lambda}^{N_0 l n_r} =<N_0 l \Lambda|N n_z \Lambda >~. \end{equation} Here $K$ is a projection number of the total angular momentum $J$ onto the $z$ axis and selected as $K=0$ at the DBCS stage because we considered pairings of the quasi-particles at $\alpha$ and ${\bar\alpha}$ states. $G(aacc ~J T)$ represents a two-body (pairwise) scattering matrix in the spherical basis, where all possible scattering of the nucleon pairs above Fermi surface were taken into account. In the present work, we have included all possible $J$ values with the $K =0$ projection. $\Delta_{\alpha n \bar{\alpha}n}$ is the similar to Eq. (\ref{eq:gap}) where $n$ was replaced by $p$. In order to renormalize the $G$-matrix due to the finite Hilbert space, strength parameters, $g_{\textrm{pp}}$, $g_{\textrm{nn}}$, and $g_{\textrm{np}}$ were multiplied with the $G$-matrix \cite{Ch93} by adjusting the pairing potentials, $\Delta_{p {\bar p}}$ and $\Delta_{n {\bar n}}$, in Eq. (\ref{eq:gap}) to the empirical pairing gaps, $\Delta_{p}^{emp}$ and $\Delta_{n}^{emp}$, which were evaluated by a symmetric five mass term formula for the neighboring nuclei \cite{Ha15-1} using empirical masses. The theoretical $np$ pairing gaps were calculated as \cite{Ch93,Bend00} \begin{equation} \delta_{np}^{th.} = - [ ( H_{gs}^{12} + E_1 + E_2 ) - ( H_{gs}^{np} + E_p + E_n)]~. \end{equation} Here $H_{gs}^{12} (H_{gs}^{np}) $ is a total DBCS ground state energy with (without) $np$ pairing and $ E_1 + E_2 (E_p + E_n)$ is a sum of the lowest two quasi-particle energies with (without) $np$ pairing potential $\Delta_{n{\bar p}}$ in Eq. (\ref{eq:hfbeq}). For the mean field energy $\epsilon_{p(n)}$ in Eq. (\ref{eq:hfbeq}) we exploited a deformed Woods-Saxon potential \cite{cwi} with the universal parameter set. By taking the same approach as used in the QRPA equation in Ref. \cite{Suhonen}, our Deformed QRPA (DQRPA) equation including the $np$ pairing correlations was summarized as follows, \begin{eqnarray}\label{qrpaeq} &&\left( \begin{array}{cccccccc} ~ A_{\alpha \beta \gamma \delta}^{1111}(K) &~ A_{\alpha \beta \gamma \delta}^{1122}(K) & ~ A_{\alpha \beta \gamma \delta}^{1112}(K) &~ A_{\alpha \beta \gamma \delta}^{1121}(K) & ~ B_{\alpha \beta \gamma \delta}^{1111}(K) &~ B_{\alpha \beta \gamma \delta}^{1122}(K) & ~ B_{\alpha \beta \gamma \delta}^{1112}(K) &~ B_{\alpha \beta \gamma \delta}^{1121}(K) \\ ~ A_{\alpha \beta \gamma \delta}^{2211}(K) &~ A_{\alpha \beta \gamma \delta}^{2222}(K) & ~ A_{\alpha \beta \gamma \delta}^{2212}(K) &~ A_{\alpha \beta \gamma \delta}^{2221}(K) & ~B_{\alpha \beta \gamma \delta}^{2211}(K) & ~B_{\alpha \beta \gamma \delta}^{2222}(K) & ~B_{\alpha \beta \gamma \delta}^{2212}(K) & ~B_{\alpha \beta \gamma \delta}^{2221}(K)\\ ~A_{\alpha \beta \gamma \delta}^{1211}(K) & ~A_{\alpha \beta \gamma \delta}^{1222}(K) & ~A_{\alpha \beta \gamma \delta}^{1212}(K) & ~A_{\alpha \beta \gamma \delta}^{1221}(K) & ~B_{\alpha \beta \gamma \delta}^{1211}(K) & ~B_{\alpha \beta \gamma \delta}^{1222}(K) & ~B_{\alpha \beta \gamma \delta}^{1212}(K) & ~B_{\alpha \beta \gamma \delta}^{1221}(K)\\ ~A_{\alpha \beta \gamma \delta}^{2111}(K) & ~A_{\alpha \beta \gamma \delta}^{2122}(K) & ~A_{\alpha \beta \gamma \delta}^{2112}(K) & ~A_{\alpha \beta \gamma \delta}^{2121}(K) & ~B_{\alpha \beta \gamma \delta}^{2111}(K) & ~B_{\alpha \beta \gamma \delta}^{2122}(K) & ~B_{\alpha \beta \gamma \delta}^{2112}(K) & ~B_{\alpha \beta \gamma \delta}^{2121}(K)\\ & & & & & & & \\ - B_{\alpha \beta \gamma \delta}^{1111}(K) & -B_{\alpha \beta \gamma \delta}^{1122}(K) & -B_{\alpha \beta \gamma \delta}^{1112}(K) & -B_{\alpha \beta \gamma \delta}^{1121}(K) & - A_{\alpha \beta \gamma \delta}^{1111}(K) & -A_{\alpha \beta \gamma \delta}^{1122}(K) & -A_{\alpha \beta \gamma \delta}^{1112}(K) & -A_{\alpha \beta \gamma \delta}^{1121}(K)\\ - B_{\alpha \beta \gamma \delta}^{2211}(K) & -B_{\alpha \beta \gamma \delta}^{2222}(K) & -B_{\alpha \beta \gamma \delta}^{2212}(K) & -B_{\alpha \beta \gamma \delta}^{2221}(K) & - A_{\alpha \beta \gamma \delta}^{2211}(K) & -A_{\alpha \beta \gamma \delta}^{2222}(K) & -A_{\alpha \beta \gamma \delta}^{2212}(K) & -A_{\alpha \beta \gamma \delta}^{2221}(K)\\ - B_{\alpha \beta \gamma \delta}^{1211}(K) & -B_{\alpha \beta \gamma \delta}^{1222}(K) & -B_{\alpha \beta \gamma \delta}^{1212}(K) & -B_{\alpha \beta \gamma \delta}^{1221}(K) & - A_{\alpha \beta \gamma \delta}^{1211}(K) & -A_{\alpha \beta \gamma \delta}^{1222}(K) & -A_{\alpha \beta \gamma \delta}^{1212}(K) & -A_{\alpha \beta \gamma \delta}^{1221}(K) \\ - B_{\alpha \beta \gamma \delta}^{2111}(K) & -B_{\alpha \beta \gamma \delta}^{2122}(K) & -B_{\alpha \beta \gamma \delta}^{2112}(K) & -B_{\alpha \beta \gamma \delta}^{2121}(K) & - A_{\alpha \beta \gamma \delta}^{2111}(K) & -A_{\alpha \beta \gamma \delta}^{2122}(K) & -A_{\alpha \beta \gamma \delta}^{2112}(K) & -A_{\alpha \beta \gamma \delta}^{2121}(K) \\ \end{array} \right)\\ \nonumber ~~&& \times \left( \begin{array}{c} {\tilde X}_{(\gamma 1 \delta 1)K}^{m} \\ {\tilde X}_{(\gamma 2 \delta 2)K}^{m} \\ {\tilde X}_{(\gamma 1 \delta 2)K}^{m} \\ {\tilde X}_{(\gamma 2 \delta 1)K}^{m} \\ \\ {\tilde Y}_{(\gamma 1 \delta 1)K}^{m} \\ {\tilde Y}_{(\gamma 2 \delta 2)K}^{m} \\ {\tilde Y}_{(\gamma 1 \delta 2)K}^{m}\\{\tilde Y}_{(\gamma 2 \delta 1)K}^{m} \end{array} \right) = \hbar {\Omega}_K^{m} \left ( \begin{array}{c} {\tilde X}_{(\alpha 1 \beta 1)K}^{m} \\{\tilde X}_{(\alpha 2 \beta 2)K}^{m} \\ {\tilde X}_{(\alpha 1 \beta 2)K}^{m} \\ {\tilde X}_{(\alpha 2 \beta 1)K}^{m}\\ \\ {\tilde Y}_{(\alpha 1 \beta 1)K}^{m} \\ {\tilde Y}_{(\alpha 2 \beta 2)K}^{m} \\ {\tilde Y}_{(\alpha 1 \beta 2)K}^{m} \\ {\tilde Y}_{(\alpha 2 \beta 1)K}^{m} \end{array} \right) ~, \end{eqnarray} where the amplitudes ${\tilde X}^m_{(\alpha \alpha'' \beta \beta'')K }$ and ${\tilde Y}^m_{(\alpha \alpha'' \beta \beta'')K}$ in Eq. (\ref{qrpaeq}) stand for forward and backward going amplitudes from state ${ \alpha \alpha'' }$ to ${\beta \beta''}$ state \cite{Ch93}. Our DQRPA equation is very general because we include the deformation as well as all kinds of pairing correlations still remained in the mean field. If we switch off the $np$ pairing, all off-diagonal terms in the A and B matrices in Eq. (\ref{qrpaeq}) disappear with the replacement of 1 and 2 into $p$ and $n$. Then the DQRPA equation is decoupled into pp + nn + pn + np DQRPA equations \cite{saleh}. For charge conserving (or neutral current) reactions, pp + nn DQRPA can describe the M1 spin or EM transitions on the same nuclear species, while np + pn DQRPA describes the GT(+/-) transitions in the charge exchange (or charged current) reactions. Here it should be pointed out that np DQRPA is different from pn DQRPA because of the deformation. The A and B matrices in Eq. (\ref{qrpaeq}) are given by \begin{eqnarray} \label{eq:mat_A} A_{\alpha \beta \gamma \delta}^{\alpha'' \beta'' \gamma'' \delta''}(K) = && (E_{\alpha \alpha''} + E_{\beta \beta''}) \delta_{\alpha \gamma} \delta_{\alpha'' \gamma''} \delta_{\beta \delta} \delta_{\beta'' \delta''} - \sigma_{\alpha \alpha'' \beta \beta''}\sigma_{\gamma \gamma'' \delta \delta''}\\ \nonumber &\times& \sum_{\alpha' \beta' \gamma' \delta'} [-g_{pp} (u_{\alpha \alpha''\alpha'} u_{\beta \beta''\beta'} u_{\gamma \gamma''\gamma'} u_{\delta \delta''\delta'} +v_{\alpha \alpha''\alpha'} v_{\beta \beta''\beta'} v_{\gamma \gamma''\gamma'} v_{\delta \delta''\delta'} ) ~V_{\alpha \alpha' \beta \beta',~\gamma \gamma' \delta \delta'} \\ \nonumber &-& g_{ph} (u_{\alpha \alpha''\alpha'} v_{\beta \beta''\beta'}u_{\gamma \gamma''\gamma'} v_{\delta \delta''\delta'} +v_{\alpha \alpha''\alpha'} u_{\beta \beta''\beta'}v_{\gamma \gamma''\gamma'} u_{\delta \delta''\delta'}) ~V_{\alpha \alpha' \delta \delta',~\gamma \gamma' \beta \beta'} \\ \nonumber &-& g_{ph} (u_{\alpha \alpha''\alpha'} v_{\beta \beta''\beta'}v_{\gamma \gamma''\gamma'} u_{\delta \delta''\delta'} +v_{\alpha \alpha''\alpha'} u_{\beta \beta''\beta'}u_{\gamma \gamma''\gamma'} v_{\delta \delta''\delta'}) ~V_{\alpha \alpha' \gamma \gamma',~\delta \delta' \beta \beta' }], \end{eqnarray} \begin{eqnarray} \label{eq:mat_B} B_{\alpha \beta \gamma \delta}^{\alpha'' \beta'' \gamma'' \delta''}(K) = &-& \sigma_{\alpha \alpha'' \beta \beta''} \sigma_{\gamma \gamma'' \delta \delta''} \\ \nonumber &\times& \sum_{\alpha' \beta' \gamma' \delta'} [g_{pp} (u_{\alpha \alpha''\alpha'} u_{\beta \beta''\beta'}v_{\gamma \gamma''\gamma'} v_{\delta \delta''\delta'} +v_{\alpha \alpha''\alpha'} v_{{\bar\beta} \beta''\beta'}u_{\gamma \gamma''\gamma'} u_{{\bar\delta} \delta''\delta'} ) ~ V_{\alpha \alpha' \beta \beta',~\gamma \gamma' \delta \delta'}\\ \nonumber &- & g_{ph} (u_{\alpha \alpha''\alpha'} v_{\beta \beta''\beta'}v_{\gamma \gamma''\gamma'} u_{\delta \delta''\delta'} +v_{\alpha \alpha''\alpha'} u_{\beta \beta''\beta'}u_{\gamma \gamma''\gamma'} v_{\delta \delta''\delta'}) ~ V_{\alpha \alpha' \delta \delta',~\gamma \gamma' \beta \beta'} \\ \nonumber &- & g_{ph} (u_{\alpha \alpha''\alpha'} v_{\beta \beta''\beta'}u_{\gamma \gamma''\gamma'} v_{\delta \delta''\delta'} +v_{\alpha \alpha''\alpha'} u_{\beta \beta''\beta'}v_{\gamma \gamma''\gamma'} u_{\delta \delta''\delta'}) ~ V_{\alpha \alpha' \gamma \gamma',~\delta \delta' \beta \beta'}], \end{eqnarray} where $u$ and $v$ coefficients are determined from DBCS Eq. (\ref{eq:hfbeq}). The two body interactions $V_{\alpha \beta,~\gamma \delta}$ and $V_{\alpha \delta,~\gamma \beta}$ are particle-particle and particle-hole matrix elements of the residual $N-N$ interaction $V$, respectively, in the deformed state, which are detailed at Ref. \cite{Ha15-1}. The GT transition amplitudes from the ground state of an initial (parent) nucleus to the excited state of a daughter nucleus, {\it i.e.} the one phonon state $K^+$ in a final nucleus, are written as \begin{eqnarray} \label{eq:phonon} &&< K^+, m | {\hat {GT}}_{K }^- | ~QRPA > \\ \nonumber &&= \sum_{\alpha \alpha''\rho_{\alpha} \beta \beta''\rho_{\beta}}{\cal N}_{\alpha \alpha''\rho_{\alpha} \beta \beta''\rho_{\beta} } < \alpha \alpha''p \rho_{\alpha}| \sigma_K | \beta \beta''n \rho_{\beta}> [ u_{\alpha \alpha'' p} v_{\beta \beta'' n} X_{(\alpha \alpha''\beta \beta'')K}^{m} + v_{\alpha \alpha'' p} u_{\beta \beta'' n} Y_{(\alpha \alpha'' \beta \beta'')K}^{m}], \\ \nonumber &&< K^+, m | {\hat {GT}}_{K }^+ | ~QRPA > \\ \nonumber &&= \sum_{\alpha \alpha'' \rho_{\alpha} \beta \beta''\rho_{\beta}}{\cal N}_{\alpha \alpha'' \beta \beta'' } < \alpha \alpha''p \rho_{\alpha}| \sigma_K | \beta \beta''n \rho_{\beta}> [ u_{\alpha \alpha'' p} v_{\beta \beta'' n} Y_{(\alpha \alpha'' \beta \beta'')K}^{m} + v_{\alpha \alpha'' p} u_{\beta \beta'' n} X_{(\alpha \alpha'' \beta \beta'')K}^{m} ]~, \end{eqnarray} where $|~QRPA >$ denotes the correlated QRPA ground state in the intrinsic frame and the nomalization factor is given as $ {\cal N}_{\alpha \alpha'' \beta \beta''} (J) = \sqrt{ 1 - \delta_{\alpha \beta} \delta_{\alpha'' \beta''} (-1)^{J + T} }/ (1 + \delta_{\alpha \beta} \delta_{\alpha'' \beta''}). $ The forward and backward amplitudes, $X^m_{(\alpha \alpha'' \beta \beta'')K}$ and $Y^m_{(\alpha \alpha'' \beta \beta'')K}$, are related to $\tilde{X^m}_{(\alpha \alpha'' \beta \beta'')K}=\sqrt2 \sigma_{\alpha \alpha'' \beta \beta''} X^m_{(\alpha \alpha'' \beta \beta'')K}$ and $\tilde{Y^m}_{(\alpha \alpha'' \beta \beta'')K}=\sqrt2 \sigma_{\alpha \alpha'' \beta \beta''} Y^m_{(\alpha \alpha'' \beta \beta'')K}$ in Eq. (\ref{qrpaeq}), where $\sigma_{\alpha \alpha'' \beta \beta''}$ = 1 if $\alpha = \beta$ and $\alpha''$ = $\beta''$, otherwise $\sigma_{\alpha \alpha'' \beta \beta'' }$ = $\sqrt 2$ \cite{Ch93,Ha17}. The particle model space for all Ni isotopes was exploited up to $N = 5 \hbar \omega$ for a deformed basis and up to $N = 10 \hbar \omega$ for a spherical basis. In our previous papers \cite{Ha15,Ha15-1,Ha2013}, SPSs obtained from the deformed Woods-Saxon potential were shown to be sensitive to the deformation parameter $\beta_2$, which causes the shell evolution by the deformation. In this work, we allow some small variation within 20 \% from theoretical $\beta_2$ values in Table I to test the deformation dependence. \section{Numerical results for the GT transition strength distribution for $^{56,58}$Ni and $^{62,64}$Ni} \begin{table} \caption[bb]{ Deformation parameter $\beta_2^{E2}$ from the experimental E2 transition data \cite{Ram01} and theoretical calculations, $\beta_2^{RMF}$ and $\beta_2^{FRDM}$, by Relativistic Mean Field (RMF) \cite{Lala99} and FRDM model \cite{Mol16} for some Ni isotopes. Empirical pairing gaps evaluated from the five-point mass formula \cite{Ch93} are also shown. Last column denotes Ikeda sum rule for GT transition \cite{Ha15} as a percentage ratio to $3(N-Z)$. \\} \setlength{\tabcolsep}{2.0 mm} \begin{tabular}{cccccccc}\hline Nucleus & $|\beta_2^{E2}|$ & $\beta_2^{RMF}$&$\beta_2^{FRDM}$ & $\Delta_p^{\textrm{emp}}$ & $\Delta_n^{\textrm{emp}}$ & $\delta_{np}^{\textrm{emp}}$ & {GT($\%$)} \\ \hline \hline $^{56}$Ni & 0.173 &0. & 0. & 2.077 & 2.150 & 1.107 & 99.6 \\ $^{58}$Ni & 0.182 &-0.001 & 0. & 1.669 & 1.349 & 0.233 & 98.6 \\ $^{62}$Ni & 0.197 & 0.093 &0.107 & 1.747 & 1.639 & 0.465 & 99.1 \\ $^{64}$Ni & 0.179 &-0.091 &-0.094&1.747 & 1.602 & 0.454 & 99.2 \\ \hline \end{tabular} \label{tab:beta2} \end{table} For direct comprehension of feasible effects due to the $np$ pairing and the deformation in the GT transition, first, we took two Ni isotopes, $^{56,58}$Ni, which are known as the nuclei affected easily by the $np$ pairing correlations \cite{Saga16,Bai13}. $^{56}$Ni is thought to be almost spherical because of its double magic number. But, in this work, we allowed small deformation, from spherical ($\beta_2$ = 0.0) to small deformation ($\beta_2$ = 0.15), in order to study the correlations of the deformation and the $np$ pairing. We note that if we take a $\alpha$-cluster model for $^{56}$Ni, the ground state may be deformed \cite{Darai11}. Second, we calculated the GT strength distribution for $^{62,64}$Ni, which have more excess neutrons with finite deformation. Moreover, they have stronger $np$ pairing gaps, almost twice, rather than $^{58}$Ni as shown in Table I, where we show the empirical pairing gaps, the deformation parameter $\beta_{2}$ deduced from E2 transition data and theoretical estimations for Ni isotopes. Also we show Ikeda sum rule results for GT strength distribution. \begin{figure} \includegraphics[width=0.45\linewidth]{56_gtm_wonp} \includegraphics[width=0.45\linewidth]{56_gtm_wnp} \caption{(Color online) Gamow-Teller (GT) transition strength distributions B(GT(--)) of $^{56}$Ni. Experimental data by $^{56}$Ni(p,n) in panel (a) are from Ref. \cite{Sasano11}. Results of (b) - (d) in the left (right) hand side are without (with) the $np$ pairing. Our results are presented by the excitation energy from parent nucleus. } \label{fig1} \end{figure} Recent calculations by the QRPA with particle-vibration coupling (PVC) based on the Skyrme interaction \cite{Niu14} addressed that the PVC contribution may spread or redistribute the GT strength for the double magic nucleus and also other Ni isotopes. But, if we recollect that the PVC contribution originates from the particle or hole propagation inside nuclei, these contributions can be taken into account by the Brueckner $G$-matrix in the Brueckner HF (BHF) approach, which already includes such nucleon-nucleon interaction inside nuclei through the Bethe-Goldstone equation. In the following, we discuss our numerical results for the GT strength distributions for $^{56,58}$Ni and $^{62,64}$Ni. Figure \ref{fig1} presents GT strength distributions for $^{56}$Ni(p,n) reaction by our DQRPA. Left (right) panels are results without (with) the $np$ pairing correlations. In the left panel, the more deformation scatters the distribution to the bit higher energy regions because of the repulsive particle-hole ($p-h$) interaction. But, the two peaks peculiar to this GT distribution data were not reproduced well only by the deformation. Namely, the 2nd high energy peak does not appear enough to explain the data. In the right panel, we showed the $np$ pairing effects, which push the distribution to the higher energy region even without the deformation, if one compares the results in the left panel to the right panel with the same deformation. Contrary to the $p-h$ repulsive force by the deformation, the $np$ pairing is mainly attractive, by which the Fermi energy difference of protons and neutrons, $\Delta_f = \epsilon_f^p - \epsilon_f^n$, is reduced by its attractive interaction and consequently gives rise to high-lying GT transitions between deeply bound neutron and proton single particle states \cite{Ha17}. As a result, the two peaks and their magnitudes appear explicitly by the $np$ pairing. \begin{figure} \includegraphics[width=0.6\linewidth]{56_sum_gtm} \caption{(Color online) Running sums for the GT (--) strength distributions in Fig.\ref{fig1} (b)-(d) for $^{56}$Ni. } \label{fig2} \end{figure} This feature becomes significant in the running sum in Fig. \ref{fig2}, if one notes differences of the solid lines (with the $np$) and the dashed lines (without the $np$). Therefore, the deformation just scatters the strength distributions to a higher energy region by the repulsive $p-h$ interaction, but the $np$ pairing shifts the distribution to a higher energy region in a more concentrated form owing to the attractive interaction. \begin{figure} \includegraphics[width=0.65\linewidth]{56_occu_a} \includegraphics[width=0.65\linewidth]{56_occu_b} \caption{(Color online) Occupation probabilities of neutrons and protons in $^{56}$Ni of the single particle state energy (SPSE) given by Nilsson basis \cite{Nilsson}. Left (right) panels are with (without) neutron-proton pairing for $\beta_2=0.02$ ((a) and (b)) and $0.15$ ((c) and (d)), respectively. } \label{fig3} \end{figure} \begin{figure} \includegraphics[width=0.45\linewidth]{56_gtm_is} \includegraphics[width=0.45\linewidth]{56_sum_gtm_is} \caption{(Color online) Isoscalar (IS) $np$ pairing effects on the GT (--) strength distribution (left) and its running sum (right) for $^{56}$Ni with $\beta_2$ = 0.05. Panels in the left side show results without the IS $np$ pairing (a) ($T=0*0$), with the normal IS $np$ pairing (b) ($T =0*1$) and with the enhanced IS $np$ pairing (c) ($T =0*3$), respectively. Panel (e) is their running sums.} \label{fig4} \end{figure} Figure \ref{fig3} shows change of the Fermi surface by the deformation and the $np$ pairing correlations. The larger deformation makes the more smearing as shown Fig.\ref{fig3} (c) and (d). But the $np$ pairing gives rise to more effective smearing, which leads to the deeper Fermi energies, compared to the deformation effect, if we compare results in Fig.\ref{fig3} (a) and (c) to those in Fig.\ref{fig3} (b) and (d). Recently, the IV quenching was reported at the $M1$ spin strength distribution of $N =Z$ $sd$-shell nuclei \cite{Matsu15}. Because this quenching may mean the condensation of the IS pairing in the $np$ pairing, we tested the IS pairing effects on the GT distribution \cite{Ha17-2,Ha18}. Figure \ref{fig4} reveals the effects of the enhanced IS condensation with a small deformation $\beta_2 =$ 0.05. The left panels explicitly show the shift of the GT strength distribution to a higher energy region with the increase of IS pairing, {\it i.e.} the case without IS (a), normal IS (c) used in the results of Fig. \ref{fig1} and the enhanced IS (d), where we retain the IV pairing to ensure the IS effect. We took the enhanced IS pairing factor as 3.0 as argued in our previous paper \cite{Ha17-2,Ha18}. Because the IS pairing causes more attractive interaction, as shown in $^3 S_1$ state in the $np$ interaction, rather than the IV coupling, the shift of the GT strength distributions by the $np$ pairing is mainly attributed to the IS coupling condensation. This trend is also found in the results of Fig. \ref{fig1}. The IS and IV $np$ effects also manifest themselves in the GT running sum of Fig.\ref{fig4} (e). Not only the IV effect but also the IS effect are shown to be salient on the GT strength in this $N=Z$ nucleus. \begin{figure} \includegraphics[width=0.45\linewidth]{58_gtm_wonp} \includegraphics[width=0.45\linewidth]{58_gtm_wnp} \caption{(Color online) GT strength distributions B(GT(--)) for $^{58}$Ni. Experimental data by $^3$He beam in panel (a) are from Ref. \cite{Fujita07}. Left (right) panel (b)-(d) are the results without (with) the $np$ pairing correlations, respectively.}\label{fig5} \end{figure} In order to study the $np$ pairing effects in $N >Z$ nucleus, we present the GT results for $^{58}$Ni in Fig. \ref{fig5}. In Ref. \cite{Bai14}, extensive discussions of the GT states of $pf$-shell $N = Z+2$ nuclei have been done for the experimental data \cite{Fujita14}. They addressed that the IS $np$ pairing could be a driving force to create low-lying GT states for those nuclei because the IS $np$ pairing induces only transitions with same $j$ value near to the Isobaric Analogue Resonance (IAR) owing to the isospin operator. If we look into detail those results in Fig.\ref{fig5}, similar effects can be found at low-lying GT states around 12 MeV region, whose distributions become spilled out to a low-lying energy region by the IS $np$ pairing contribution, but the dependence on the deformation is larger than that by the $np$ pairing and its strength by the IS $np$ pairing is small compared to the main GT peak, as explained in Ref. \cite{Bai14}. If we compare to the results $^{56}$Ni in Fig.\ref{fig1}, the $np$ pairing does not show drastic effects as shown in the results without (left) and with (right) the $np$ pairing. It means that the $np$ pairing effects become the smaller with the increase of $N-Z$ number. However, this trend is shown to digress for $^{62,64}$Ni, as shown later on. Figure \ref{fig6} presents results of the GT(+) states and their experimental data \cite{Cole06}, where the deformation effect is more significant than the $np$ pairing effect. \begin{figure} \includegraphics[width=0.45\linewidth]{58_gtp_wonp} \includegraphics[width=0.45\linewidth]{58_gtp_wnp} \caption{(Color online) The same as in Fig. \ref{fig5}, but for B(GT(+)) of $^{58}$Ni. Experimental data by {\it t} beam in panel (a) are from Ref. \cite{Cole06}. } \label{fig6} \end{figure} \begin{figure} \includegraphics[width=0.45\linewidth]{58_sum_gtm} \includegraphics[width=0.45\linewidth]{58_sum_gtp} \caption{(Color online) Running sums for the GT (--) and GT (+) strength distributions in Figs. \ref{fig5} and \ref{fig6} (b)-(d) for $^{58}$Ni, respectively. Here we used the general quenching factor ${(0.74)}^2$ for theoretical calculations. GT(--) data are taken from the isospin decomposition with $(p,p')$ scattering data \cite{Fujita07}. But GT(+) data are normalized to the lowest B(GT) by the calibration from $\beta$-decay \cite{Cole06}. Therefore, these data do not satisfy the Ikeda sum rule (ISR) for GT strengths. }\label{fig7} \end{figure} For $^{58}$Ni, the deformation effect turned out to be more important rather than the $np$ pairing correlations, as confirmed in the running sum in Fig. \ref{fig7}. Differences between colored solid and dot-dashed (dotted and dashed) curves by the $np$ pairing correlations are smaller than those by the deformation compared to the results in Fig. \ref{fig2} for $^{56}$Ni case. However, the $np$ pairing correlations turned out to play a role of explaining properly the GT running sum as shown in Fig. \ref{fig7}. \begin{figure} \includegraphics[width=0.45\linewidth]{62_gtm} \includegraphics[width=0.45\linewidth]{64_gtm} \caption{(Color online) Same as Fig. \ref{fig5}, but for $^{62}$Ni (left) and $^{64}$Ni (right). Experimental data for $^{64}$Ni by $^3$He beam in the right panel are from Ref. \cite{64data}. Results are shown with and without $np$ pairing correlations, respectively.} \label{fig8} \end{figure} Figure \ref{fig8} provides the GT strength distributions for $^{62,64}$Ni, which shows stronger $np$ pairing effects rather than those of $^{58}$Ni in Fig.\ref{fig5}. That is, the $np$ pairing separates explicitly the GT distribution into a low-lying and a high-lying part adjacent to GTGR position. If we recollect that the $np$ pairings of $^{62,64}$Ni are almost twice of that of $^{58}$Ni as shown in Table I, this separation is quite natural. Moreover, the larger deformation of those nuclei scatters those distributions rather than those of $^{58}$Ni. If we note the shift of the GTGR position to the reasonable region around 11 MeV, which is consistent with the data in Ref. \cite{Fujita11}, the $np$ pairing effect is indispensable even for these $pf$-shell $N \neq Z $ nuclei, $^{62,64}$Ni. More experimental data for these nuclei would be helpful for more definite conclusion for the $np$ pairing correlations effect on the GT transition strength distributions. Recently, many papers argued the tensor force effect on the GT strength \cite{Bai14,Alex06,Urata17,Bernard16}. The tensor force effect in Ref. \cite{Bai14}, which is based on the zero range pairing force usually adopted in the Skyrme-Hartree-Fock (SHF) approach, shifts the high-lying GT state downward a few MeV owing to the its attractive force property. However, this trend may become changed by the deformation because of the following facts. The angular momentum $J$ is not a good quantum number in Nilsson basis, as a result it can be split by the deformation. It means that several angular momentum components are mixed in a deformed SPS, which makes direct understanding of a role of the tensor force in a deformed mean field fuzzy. Also the deformation changes the level density around the Fermi level, which leads to some increase or decrease of the pairing correlations, depending on the Fermi surface as argued in Refs. \cite{Alex06,Urata17}. For example, recent calculation in Ref. \cite{Bernard16} showed that the tensor force could be attractive, but sometimes it could be repulsive along with the SPS evolution with the deformation. The tensor force in our DQRPA approach is explicitly taken into account on the residual interaction by the G-matrix calculated from the Bonn potential which explicitly includes the tensor force in the potential. In the mean field, the tensor force is implicitly included because of the phenomenological Woods-Saxon potential globally adjusted to some nuclear data. But, in order to discuss the separated tensor force effect on the GT strength distribution, the present approach needs more detailed analysis of the tensor force effect on the mean field as done in Ref. \cite{Alex06}, in which a tensor type potential is derived for reproducing the G-matrix calculation. But the approach is beyond the present calculation and would be a future project. \section{Summary and conclusion} In summary, in order to discuss the effects of the IS and IV $np$ pairing correlations and the deformation, we calculated the GT strength distribution of $^{56,58}$Ni and $^{62,64}$Ni, which are $N=Z$ and $N-Z=$2 ,6, and 8 nucleus, respectively. The $np$ pairing effects turned out to be able to properly explain the GT strength although the deformation was shown to be also the other important property. In particular, the IS condensation part played a meaningful role to explaining the GT strength distribution of $N=Z$ nucleus, $^{56}$Ni, whose GT strength distribution was shifted to a bit higher energy region by the reduction of the Fermi energy difference of proton and neutron due to the attractive $np$ pairing. For $^{58}$Ni, the deformation was more influential rather than the $np$ pairing. But for $^{62,64}$Ni, the situation is reversed because the $np$ pairing correlations are stronger than $^{58}$Ni. Namely, the $np$ pairing divides the GT strength distribution into a low and high energy region. Therefore, the deformation treated by a mean field approach can be balanced by the spherical property due to the IV $np$ pairing coming from the unlike $np$ pairing as well as the like-pairing correlations. But the IS spin-triplet $np$ part, which contributes more or less to the deformation property due to its coupling to odd $J$ states, may give rise to more microscopic deformation features which cannot be included in the deformed mean field approach, and push the GT states to a bit high energy region. But all of the present results are based on a phenomenological Woods-Saxon potential. More self-consistent approaches are desirable for further definite conclusions on the IS $np$ pairing correlations. Finally, the GT strength distribution as well as the M1 spin transition strength are shown to be useful for investigating the IS and IV pairing properties. \section*{Acknowledgement} This work was supported by the National Research Foundation of Korea (Grant Nos. NRF-2015R1D1A4A01020477, NRF-2015K2A9A1A06046598, NRF-2017R1E1A1A01074023). \newpage \section*{References}

{'timestamp': '2018-11-13T02:17:01', 'yymm': '1811', 'arxiv_id': '1811.04589', 'language': 'en', 'url': 'https://arxiv.org/abs/1811.04589'}

high_school_physics

Other Operating Charges During the three months ended June 29, 2018, the Company recorded other operating charges of $225 million. These charges primarily consisted of $111 million related to the Company's productivity and reinvestment program. In addition, other operating charges included $60 million of CCR asset impairments and $34 million related to costs incurred to refranchise certain of our North America bottling operations. Costs related to refranchising include, among other items, internal and external costs for individuals directly working on the refranchising efforts, severance, and costs associated with the implementation of information technology systems to facilitate consistent data standards and availability throughout our North America bottling system. Other operating charges also included $22 million related to tax litigation expense. Refer to Note 8 for additional information related to the tax litigation. Refer to Note 12 for additional information on the Company's productivity and reinvestment program. Refer to Note 15 for information on how the Company determined the asset impairment charges. Refer to Note 16 for the impact these charges had on our operating segments. During the six months ended June 29, 2018, the Company recorded other operating charges of $761 million. These charges primarily consisted of $450 million of CCR asset impairments and $206 million related to the Company's productivity and reinvestment program. In addition, other operating charges included $79 million related to costs incurred to refranchise certain of our North America bottling operations. Costs related to refranchising include, among other items, internal and external costs for individuals directly working on the refranchising efforts, severance, and costs associated with the implementation of information technology systems to facilitate consistent data standards and availability throughout our North America bottling system. Other operating charges also included $27 million related to tax litigation expense. Refer to Note 8 for additional information related to the tax litigation. Refer to Note 12 for additional information on the Company's productivity and reinvestment program. Refer to Note 15 for information on how the Company determined the asset impairment charges. Refer to Note 16 for the impact these charges had on our operating segments. During the three months ended June 30, 2017, the Company recorded other operating charges of $826 million. These charges primarily consisted of $653 million of CCR asset impairments and $87 million related to the Company's productivity and reinvestment program. In addition, other operating charges included $47 million related to costs incurred to refranchise certain of our bottling operations. Costs related to refranchising include, among other items, internal and external costs for individuals directly working on the refranchising efforts, severance and costs associated with the implementation of information technology systems to facilitate consistent data standards and availability throughout our North America bottling system. Other operating charges also included $14 million related to the impairment of a Venezuelan intangible asset and $19 million related to tax litigation expense. Refer to Note 1 for additional information about the Venezuelan intangible asset and Note 15 for information on how the Company determined the asset impairment charges. Refer to Note 8 for additional information related to the tax litigation. Refer to Note 12 for additional information on the Company's productivity and reinvestment program. Refer to Note 16 for the impact these charges had on our operating segments. During the six months ended ended June 30, 2017, the Company recorded other operating charges of $1,116 million. These charges primarily consisted of $737 million of CCR asset impairments and $226 million related to the Company's productivity and reinvestment program. In addition, other operating charges included $86 million related to costs incurred to refranchise certain of our bottling operations, $34 million related to impairments of Venezuelan intangible assets and $25 million related to tax litigation expense. Refer to Note 1 for additional information about the Venezuelan intangible assets and Note 15 for information on how the Company determined the asset impairment charges. Refer to Note 8 for additional information related to the tax litigation. Refer to Note 12 for additional information on the Company's productivity and reinvestment program. Refer to Note 16 for the impact these charges had on our operating segments. Other Nonoperating Items During the three and six months ended June 30, 2017, the Company recorded a net charge of $38 million related to the extinguishment of long-term debt. Equity Income (Loss) — Net During the three and six months ended June 29, 2018, the Company recorded net charges of $33 million and $84 million, respectively. During the three and six months ended June 30, 2017, the Company recorded a net gain of $37 million and a net charge of $21 million, respectively. These amounts represent the Company's proportionate share of significant operating and nonoperating items recorded by certain of our equity method investees. Refer to Note 16 for the impact these items had on our operating segments. During the three months ended June 29, 2018, the Company recorded charges of $86 million related to pension settlements and net charges of $102 million due to the refranchising of certain bottling territories in North America. The Company also recorded an other-than-temporary impairment charge of $52 million related to one of our equity method investees. These charges were partially offset by a net gain of $36 million related to the refranchising of our Latin American bottling operations and a net gain of $36 million related to realized and unrealized gains and losses on equity securities and trading debt securities as well as realized gains and losses on available-for-sale debt securities. Refer to Note 1 and Note 4 for additional information on equity and debt securities. Refer to Note 2 for additional information on refranchising activities. Refer to Note 16 for the impact these items had on our operating segments. During the six months ended June 29, 2018, the Company recorded charges of $86 million related to pension settlements and net charges of $104 million due to the refranchising of certain bottling territories in North America. The Company also recorded an other-than-temporary impairment charge of $52 million related to one of our equity method investees and a net loss of $49 million related to realized and unrealized gains and losses on equity securities and trading debt securities as well as realized gains and losses on available-for-sale debt securities. Additionally, the Company recognized a net loss of $33 million primarily related to the reversal of the cumulative translation adjustments resulting from the substantial liquidation of the Company's former Russian juice operations and charges of $21 million primarily related to payments made to convert the bottling agreements for certain North America bottling partners' territories to a single form of CBA with additional requirements. These charges were partially offset by a net gain of $36 million related to the refranchising of our Latin American bottling operations. Refer to Note 1 and Note 4 for additional information on equity and debt securities. Refer to Note 2 for additional information on refranchising activities and North America conversion payments. Refer to Note 16 for the impact these items had on our operating segments. During the three months ended June 30, 2017, the Company recognized a gain of $445 million related to the integration of Coca‑Cola West Co., Ltd. ("CCW") and Coca-Cola East Japan Co., Ltd. ("CCEJ") to establish Coca-Cola Bottlers Japan Inc., now known as Coca-Cola Bottlers Japan Holdings Inc. ("CCBJHI"). In exchange for our previously existing equity interests in CCW and CCEJ, we received an approximate 17 percent equity interest in CCBJHI with a fair market value of $1,112 million as of April 1, 2017. The Company accounts for its 17 percent interest in CCBJHI as an equity method investment based on our equity ownership percentage, our representation on CCBJHI's Board of Directors, material intercompany transactions and other governance rights. The Company also recognized a $25 million gain as a result of Coca‑Cola FEMSA, an equity method investee, issuing additional shares of its stock during the period at a per share amount greater than the carrying value of the Company's per share investment and a gain of $9 million related to refranchising a substantial portion of our China bottling operations. These gains were partially offset by a net charge of $214 million due to the refranchising of certain bottling territories in North America and charges of $109 million primarily related to payments made to convert the bottling agreements for certain North America bottling partners' territories to a single form of CBA with additional requirements. The Company also incurred a charge of $26 million related to our former German bottling operations. Refer to Note 2 for additional information on refranchising activities and the conversion payments. Refer to Note 16 for the impact these items had on our operating segments. During the six months ended June 30, 2017, the Company recognized net charges of $711 million due to the refranchising of certain bottling territories in North America and charges of $215 million primarily related to payments made to convert the bottling agreements for certain North America bottling partners' territories to a single form of CBA with additional requirements. The Company also incurred a charge of $26 million related to our former German bottling operations and net charges of $15 million resulting from special termination benefits and curtailment credits related to refranchising certain of our North America bottling operations. These charges were partially offset by a gain of $445 million related to the integration of CCW and CCEJ to establish CCBJHI. In exchange for our previously existing equity interests in CCW and CCEJ, we received an approximate 17 percent equity interest in CCBJHI. The Company also recognized a $25 million gain as a result of Coca‑Cola FEMSA, an equity method investee, issuing additional shares of its stock during the period at a per share amount greater than the carrying value of the Company's per share investment and a gain of $9 million related to refranchising a substantial portion of our China bottling operations. Refer to Note 2 for additional information on refranchising activities and the conversion payments. Refer to Note 13 for additional information on the special termination benefit charges and curtailment credits. Refer to Note 16 for the impact these items had on our operating segments. Name: ko_SignificantOperatingAndNonoperatingItems The entire disclosure for an event or transaction that is unusual in nature or infrequent in occurrence, or both. -SubTopic 20 -URI http://asc.fasb.org/subtopic&trid=2122313 Name: us-gaap_UnusualOrInfrequentItemsDisclosureTextBlock

{"pred_label": "__label__cc", "pred_label_prob": 0.7350722551345825, "wiki_prob": 0.2649277448654175, "source": "cc/2021-04/en_head_0001.json.gz/line938751"}

high_school_physics

← Additionally, nitric oxide (Simply no), endothelin-1 (ET-1), vasopressin, and vascular endothelial growth factor (VEGF) are also been shown to be impaired mainly because regulators of kidney vascular tone in people with diabetes (58, 65, 72C74) 1991;354:82 → From co-localization with GFP in GAD67-GFP slices, it appears that interneurons contribute to this small populace of neurons that maintain dendritic morphology following NMDA (Figure 2C) Posted on January 13, 2022 by Reginald Arnold From co-localization with GFP in GAD67-GFP slices, it appears that interneurons contribute to this small populace of neurons that maintain dendritic morphology following NMDA (Figure 2C). timecourse of these experiments (90 min post-NMDA washout). Formation of regularly-spaced, spherical dendritic varicosities (dendritic beading) was rarely observed, except when NMDA was applied in Ca2+-free ACSF. Under these conditions, beading appeared predominant in interneurons, as assessed from experiments with GAD67-GFP (neo) mice. Ca2+-removal was associated with significantly better preservation of dendritic structure (MAP2) following NMDA exposure, and other ionic fluxes (sensitive to Gd3+ and spermine) may contribute to residual damage occurring in Ca2+-free conditions. These results suggest that irregularly shaped dendritic swelling is usually a Ca2+-dependent degenerative event that may be quite different from Ca2+-impartial dendritic beading, and can be a predominant type of injury in CA1 pyramidal neurons in slice. ischemia (Hori and Carpenter, 1994; Matesic and Lin, 1994) and excitatory amino acid agonist exposure (Olney et al., 1979; Sloviter and Dempster, 1985) and (Park et al., 1996; Hasbani et al., 1998; Al-Noori and Swann, 2000; Ikegaya et al., 2001; Oliva et al., 2002). While many detailed studies of mechanisms of dendritic beading have used neuronal cultures or slice culture preparations, less is known about the progression of dendritic damage in mature slice preparations. Regular, spherical dendritic beading occurs in mature neurons exposed to cold, and is rapidly reversible (within ~5 min) with this non-lethal stimulus (Kirov Dot1L-IN-1 et al., 2004). In contrast, other studies report irreversible dendritic beading in mature CA1 pyramidal neurons following oxygen-glucose deprivation (OGD) or glutamate agonist exposure, and it was concluded that this beading underlies irrecoverable changes Dot1L-IN-1 in slice light transmittance (Andrew et al., 1999; Jarvis et al., 1999; Obeidat et al., 2000). However, in those studies reporting irreversible beading, at least a number of the specific varicosities demonstrated made an appearance formed irregularly, with an increase of heterogeneous distribution of sizes and dendrite constrictions than typically continues to be referred to for dendritic beading in neuronal ethnicities. We analyzed reactions to NMDA in severe hippocampal pieces lately, to investigate systems underlying lack of MAP2 from dendrites (Hoskison and Shuttleworth, 2006). In that scholarly study, most preparations had been examined at past due timepoints after a short NMDA publicity, and it had been difficult to judge dendritic structure because of extensive MAP2 reduction. However, in a single experiment, following NMDA exposure immediately, significant irregular bloating was noticed, with little proof regularly-spaced dendritic beads. The purpose of the present function was to examine the features of dendritic structural adjustments in CA1 dendrites third , excitotoxic excitement, with particular focus on the development of effects, as well as the reliance on extracellular Ca2+ focus. Utilizing a stimulus (30M NMDA, 10 min) matched up to that creating reversible dendritic beading in isolated ethnicities (Recreation area et al., 1996; Faddis et al., 1997; Hasbani et al., 1998) and hippocampal cut ethnicities (Ikegaya et al., 2001), we’ve examined the hypothesis Dot1L-IN-1 a predominant kind of dendritic damage in mature CA1 neurons can be Ca2+-reliant and continual, with structural features quite dissimilar from structured dendritic beading. The long lasting nature of pervasive abnormal swellings that people describe here shows that this sort of dendritic injury may donate to continual neuronal damage and dysfunction connected with excitotoxic injury in these slices. On the other hand, problems under Ca2+-free of charge conditions created some beading, but this specific damage was limited to a subpopulation of dendrites in CA1 that seemed to consist of procedures of interneurons. EXPERIMENTAL Methods Hippocampal slice planning and stimulation Tests had been carried out relative to the Country wide Institute of Wellness recommendations for the humane treatment of lab pets, and the process for these methods was reviewed yearly from the Institutional Pet Care and Make use of Committee in the College or university of New Mexico College of Medicine. All attempts were designed to minimize pet struggling and the real amount of pets utilized. Acute hippocampal pieces (250m) had been ready from adult mice (4C6 weeks old) and managed in the same way as useful for electrophysiological research, as referred to previously (Hoskison and Shuttleworth, 2006). Quickly, mice had been deeply anesthetized with an assortment of ketamine and xylazine (85mg/ml and 15mg/ml, respectively; 200l s.c.) and decapitated. Brains had been removed and put into ice-cold cutting remedy and coronal areas had been cut utilizing CLTC a Vibratome (Complex Items International, Dot1L-IN-1 St Louis MO). Pieces from each pet had been divided between treatment and control organizations, and transferred into room-temperature ACSF initially. After warming to 37C and keeping for one hour, ACSF again was changed, and slices had been maintained at space temp for 1.5C3 hours before you begin experiments. FVB/N mice (Harlan) had been used for some tests. Because this stress is proven to. This entry was posted in Estrogen Receptors. Bookmark the permalink.

{"pred_label": "__label__cc", "pred_label_prob": 0.5956661105155945, "wiki_prob": 0.4043338894844055, "source": "cc/2022-05/en_middle_0092.json.gz/line1442196"}

high_school_physics

Trio accused of orchestrating international drug sales scheme Three people accused of being part of an international drug trafficking ring are in custody on federal charges alleging they used the darknet to obtain drugs and offer narcotics for sale. The defendants are also accused of shipping pound quantities of methamphetamine to buyers in the Philippines, New Zealand, Poland and other foreign countries. An eight-count grand jury indictment returned on March 22 charges the three defendants with conspiracy to distribute methamphetamine and substantive drug trafficking offenses. The defendants are Michael Alan Goldberg, 34, who allegedly directed the conspiracy, even after being taken into federal custody in June 2018; Donnica Mae Rabulan, 31, Goldberg’s wife; and James Caleb Kueker, 41, a resident of the Hollywood Hills. Rabulan, who in addition to the conspiracy charge faces four counts of distributing methamphetamine, was arraigned on April 1 in United States District Court. Kueker, who also is charged with one count of possession with the intent to distribute ketamine, as well as with one count of possession of a firearm in furtherance of a drug trafficking crime, was also arraigned on April 1. Rabulan and Kueker each entered not guilty pleas and were ordered to stand trial on June 4. Goldberg, who also faces one count of intentionally distributing methamphetamine, is scheduled to be arraigned on April 8. The eight-count indictment followed a criminal complaint filed on March 1 that outlined how the three defendants used the darknet to purchase narcotics with cryptocurrency such as Bitcoin, and then made 59 shipments of methamphetamine and other illegal drugs to international buyers. During the course of the alleged conspiracy, Goldberg was arrested on an unrelated fraud case and was sentenced to 42 months in federal prison. In May 2018, a shipping company security manager in Van Nuys contacted federal law enforcement about seven parcels destined for the Philippines that were suspected of containing controlled substances, according to an affidavit in support of the complaint. The seven packages cumulatively contained nearly 21 kilograms of methamphetamine, and had contact information with telephone numbers linked to Goldberg, the affidavit states. The indictment alleges that Rabulan shipped methamphetamine to international buyers, while Kueker collected tens of thousands of dollars in illicit proceeds from the sales on Goldberg’s behalf. Search warrants executed on March 11 led to the discovery of suspected drug labs at Kueker’s homes in the Hollywood Hills and Valley Village, as well as a loaded AK-47-style rifle in the bedroom of Kueker’s Hollywood Hills residence, court papers state. Federal agents also recovered more than $150,000 in suspected drug proceeds in a safe deposit box registered to Kueker, according to court documents. Three residents of the house were arrested and were charged in separate indictments alleging possession with intent to distribute a controlled substance. If convicted on all counts, each defendant faces a statutory maximum sentence of life in federal prison. The conspiracy count carries a mandatory minimum sentence of 10 years in prison. The count alleging firearm possession in furtherance of a drug trafficking crime carries a five-year mandatory minimum prison sentence. Man charged for collision that killed scooter rider Suspect faces tax evasion charges Defendants charged for scheme targeting seniors December 10, 2020 Another defendant with links to Huizar pleads guilty September 3, 2020 Authorities fight sexual harassment in housing June 4, 2020 Former Beverly Hills resident faces federal immigration charges January 9, 2020 WeHo doctor charged for scheme involving sales of growth hormone July 3, 2019 Suspect faces tax evasion charges April 18, 2019 Three guilty in international money laundering scheme January 3, 2019 WeHo resident charged in new federal fraud case November 15, 2018 Parking lot operator sentenced to federal prison for fraud August 30, 2018

{"pred_label": "__label__wiki", "pred_label_prob": 0.7341157793998718, "wiki_prob": 0.7341157793998718, "source": "cc/2021-04/en_middle_0007.json.gz/line1107335"}

high_school_physics

\section{Introduction} \label{sec:Intro} The fractional quantum Hall effect (FQHE) at filling factor $\nu = 1/3$ in the lowest Landau level (LLL) was the first to be experimentally observed~\cite{Tsui82} and subsequently understood, as a result of Laughlin's trial wave function~\cite{Laughlin83}. The FQHE has since proven to be an incredibly rich platform for exploring the physics of strongly correlated electron systems. After the initial observation, many additional fractions were observed in the LLL at $\nu = n/(2pn\pm1)$ with $n$ and $p$ positive integers. These fractions are understood as an integer quantum Hall effect (IQHE) of composite fermions, where a composite fermion (CF) is an emergent particle consisting of an electron bound to an even number of vortices~\cite{Jain89, Jain07, Halperin20}. In contrast, the FQHE in the second LL (SLL) of GaAs is less well understood. Interestingly, even the physical origin of the $\nu=7/3$ FQHE, which corresponds to 1/3 filled SLL, has not been conclusively established. Exact diagonalization studies~\cite{Ambrumenil88, Peterson08, Peterson08b, Balram13b, Kusmierz18, Balram20} have convincingly shown that the actual state at $\nu=7/3$ for a zero width system is an incompressible FQHE state. However, the overlap of the exact ground state with the Laughlin state is not large, typically less than $60\%$ for systems accessible to exact diagonalization studies~\cite{Ambrumenil88, Peterson08, Peterson08b, Balram13b, Kusmierz18, Balram20}. [In contrast, the overlap of the Coulomb ground state at $\nu=1/3$ in the LLL with the Laughlin state is greater than $98\%$ for up to $N=15$ electrons~\cite{Balram20}.] Furthermore, the excitations of the 7/3 FQHE in exact diagonalization studies are qualitatively different from those at 1/3, and exact diagonalization studies also do not show a clearly identifiable branch of low-energy excitations, called the magnetoroton or the CF-exciton mode, as observed in the LLL~\cite{Girvin85, Girvin86, Dev92, Scarola00, Jolicoeur17}. As a result, the precise nature of the state at $\nu=7/3$ has remained a topic of debate~\cite{Read99, Balram13b, Johri14, Zaletel15, Peterson15, Jeong16, Balram19a, Balram20}. In a recent work, Balram \emph{et al}.\cite{Balram19a} have proposed, inspired by the parton paradigm for the FQHE~\cite{Jain89b}, that the $\nu=7/3$ FQHE is a $\mathbb{Z}_n$ topological superconductor, wherein bound states of $n$ composite bosons~\cite{Zhang89} undergo Bose-Einstein condensation. This generalizes the Zhang-Hansson-Kivelson theory of the 1/3 Laughlin state as a Bose-Einstein condensate of composite bosons~\cite{Zhang89}, with $\mathbb{Z}_1$ corresponding to the Laughlin wave function. While the different $\mathbb{Z}_n$ states share many topological quantum numbers, a key distinction between them is that the elementary quasiparticle has a charge of $-e/(3n)$, where $-e$ is the charge of the electron. Variational calculations in Ref.~\cite{Balram19a} suggest that the best candidate is the $\mathbb{Z}_3$ state, which has lower energy than the Laughlin state in the thermodynamic limit, and also a higher overlap with the exact SLL Coulomb ground state for systems where such a calculation is possible. The experimental observations at $\nu = 7/3$ are, however, largely consistent with the Laughlin state. In particular, shot noise~\cite{Dolev08, Venkatachalam11,Dolev11} and scanning single-electron transistor~\cite{Venkatachalam11} experiments at $\nu = 7/3$ have measured quasiparticles of charge $-e/3$. This raises the question: Why are experimental measurements consistent with the Laughlin state while theory suggests that better variational states exist? This question has motivated the present study. There can be several reasons for the discrepancy between theory and experiment. The theoretical calculations mentioned above do not include the effects of finite width, Landau level mixing, screening, and disorder, which can affect the variational comparisons. We consider in this article the competition between the different $\mathbb{Z}_n$ states as a function of the quantum well width. Our primary result is the prediction of a phase transition from the $\mathbb{Z}_4$ state at small widths into the Laughlin ($\mathbb{Z}_1$) state when the quantum well width exceeds approximately 1.5 magnetic lengths. We also predict a similar phase transition at $\nu=1/3$ in the zeroth Landau level of bilayer graphene as a function of the magnetic field. \section{$\mathbb{Z}_n$ parton wave function at $\nu=1/3$} \label{sec:wftns} The parton theory generalizes the Jain CF states~\cite{Jain07} to a larger class of candidate wave functions~\cite{Jain89b}. In the parton theory, one considers fractionalizing electrons into a set of fictitious particles called partons. The partons are fractionally charged, have the same density as electrons, and have filling factor $\nu_\alpha$, where $\alpha$ labels the parton species. An incompressible state is achieved when each parton species is in an IQHE state, i.e. $\nu_\alpha = n_\alpha$, with $n_\alpha$ an integer. (More generally, we can place the partons in any known incompressible states.) The partons are of course unphysical and must be combined back into physical electrons, which is equivalent to setting the parton coordinates $z_j^\alpha$ equal to the parent electron coordinates $z_j$, i.e. $z_j^\alpha = z_j$ for all $\alpha$. (The quantity $z_j = x_j - iy_j$ is the complex coordinate of the $j$th electron.) The resulting wave functions, labeled ``$n_1n_2n_3...$," are given by \begin{equation} \Psi^{n_1n_2n_3...}_\nu = \mathcal{P}_{\rm LLL} \prod_{n_\alpha} \Phi_{n_\alpha}(\{z_j\}), \end{equation} where $\Phi_n$ is the Slater determinant wave function for the state with $n$ filled Landau levels, and $\mathcal{P}_{\rm LLL}$ denotes projection into the LLL, as appropriate in the high field limit. The partons can also experience magnetic fields anti-parallel to the field experienced by electrons; these correspond to negative filling factors, which we denote as $\bar{n}$, with $\Phi_{\bar{n}}=\Phi_{-n}=\Phi_n^*$. To ensure that each parton species has the same density as the electron density, the charge of each parton species is given by $e_\alpha = -\nu e / \nu_\alpha$. The relation $\sum_\alpha e_\alpha=-e$ implies that the electron filling factor is given by $\nu = [\sum_\alpha \nu_\alpha^{-1}]^{-1}$. The Laughlin wave function at $\nu=1/3$ can be interpreted as the $111$ parton state. The Jain $n/(2pn+1)$ states appear as the $n11...$ states and the Jain $n/(2pn-1)$ states as $\bar{n}11...$; these correspond to the wave function $\Psi_{n/(2pn\pm1)} = \mathcal{P}_{\rm LLL} \Phi_{\pm n} \Phi_1^{2p}$. Many other parton states have recently been shown to be plausible for SLL and other FQHE~\cite{Wu17, Balram18, Bandyopadhyay18, Balram18a, Faugno19, Balram19, Kim19, Balram19a, Balram20, Faugno20a, Balram20a}. These states often have exotic properties, such as non-Abelian anyonic excitations~\cite{Wen91}. For $\nu = 1/3$, Balram \emph{et al}. proposed the $\mathbb{Z}_n$ parton states described by the wave function \begin{equation} \Psi_{1/3}^{\mathbb{Z}_n} = \mathcal{P}_{\rm LLL}\Phi_n\Phi_{\bar{n}}\Phi_1^3 \sim \Psi_{n/(2n+1)}\Psi_{n/(2n-1)} \Phi_1^{-1}, \end{equation} where in the last step we redefine the wave function as $[\mathcal{P}_{\rm LLL}\Phi_n \Phi_1^2] [\mathcal{P}_{\rm LLL}\Phi_{\bar{n}} \Phi_1^2]\Phi_1^{-1}$. (This grouping is chosen to facilitate the use of Jain-Kamilla projection method in our numerical simulations \cite{Jain07,Jain97, Jain97b, Davenport12, Moller05, Balram15a}. It is accepted and has been shown for many cases that the topological properties of the state do not depend on the details of the projection method \cite{Balram16b}.) Because the factor $\Phi_n\Phi_{\bar{n}}$ is real, all the $\mathbb{Z}_n$ states occur at the same ``shift"~\cite{Wen92} $\mathcal{S}=3$ in the spherical geometry. The physical interpretation of the wave function as a superconductor of composite bosons arises from the fact that $\Phi_n\Phi_{\bar{n}}$ represents a $\mathbb{Z}_n$ superconductor of electrons~\cite{Barkeshli13,Balram19a}, and the factor $\Phi_1^3$ attaches three vortices to each electron to convert it into a composite boson. The elementary excitation corresponds to an excitation in the factor $\Phi_n$ or $\Phi_{\bar{n}}$ and has a charge of magnitude $e/(3n)$. \section{Finite width phase diagram for $\nu=7/3$ in GaAs} We will use the spherical geometry~\cite{Haldane83} in our calculations, which considers $N$ electrons on the surface of a sphere subjected to a total flux $2Q\phi_0$, with $\phi_0=hc/e$ and $2Q$ is a positive integer. The radius of the sphere is $R=\sqrt{Q}\ell$, where $\ell=\sqrt{\hbar c/eB}$ is the magnetic length. The state with $n$ filled Landau levels can only be constructed for particle number $N$ divisible by $n$ and $N\geq n^2$. The same is true of the $\mathbb{Z}_n$ state. We approximate the confining potential as an infinite square well of width $w$ that results in a transverse wave function given by a sine function. The problem of electrons in the SLL interacting with the Coulomb interaction is equivalent to that of electrons in the LLL interacting with an effective interaction; the effective interaction for this system is given in Ref.~\cite{Toke08}. We consider well widths up to 5 magnetic lengths. (For convenience, we use the disk pseudopotentials for our calculations in the spherical geometry; this should not cause any corrections because we will perform the calculation for large systems and take the thermodynamic limit.) \begin{figure}[tbhp] \includegraphics[width=0.49\textwidth]{SLLTL} \caption{Thermodynamic extrapolations for the energy differences $\Delta E=E(\mathbb{Z}_n)-E_{\rm Laughlin}$ for various $\mathbb{Z}_n$ states at $\nu=7/3$. The results are for zero quantum well width. The $\mathbb{Z}_4$ state has the lowest energy in the thermodynamic limit.} \label{fig: TLs} \end{figure} \begin{figure}[tbhp] \includegraphics[width=0.49\textwidth]{SLLFW} \caption{Thermodynamic energies of various $\mathbb{Z}_n$ states at $\nu=7/3$ as a function of the quantum well width $w$. All energies are quoted relative to the Laughlin state. The $x$-axis is the quantum well width in units of magnetic length. A phase transition from the $\mathbb{Z}_4$ to the Laughlin state is seen to occur at $w\sim1.5 \ell$.} \label{fig: FWSLL} \end{figure} At zero width, the $\mathbb{Z}_4$ state has the lowest energy in the thermodynamic limit, as seen in Fig.~\ref{fig: TLs}. Here and below, all energies are quoted in units of $e^2/(\epsilon \ell)$ where $\epsilon$ is the dielectric constant of the material. (We note that the $\mathbb{Z}_4$ and $\mathbb{Z}_5$ states were not studied in Ref.~\cite{Balram19a}. Also, we cannot definitively rule out the $\mathbb{Z}_3$ state within the numerical accuracy of our calculations.) We have similarly determined the thermodynamic energies for quantum wells of various widths [see Appendix~\ref{sec: ThermoLim}]. In Fig.~\ref{fig: FWSLL}, we show how the energies of several $\mathbb{Z}_n$ states, measured relative to the energy of the Laughlin state, evolve as we increase the well-width. For a $w<1.5\ell$, the $\mathbb{Z}_4$ state has the lowest variational energy in the thermodynamic limit. For $w>1.5\ell$, the Laughlin state is preferred in our calculation, suggesting that the $\mathbb{Z}_4$ state should only be observed in samples with sufficiently low quantum well widths and/or low density. We add here that because of the numerical uncertainty in the thermodynamic energy differences and our simple model for the finite width, the critical value of $1.5\ell$ should be taken only as a first estimate. \section{$\mathbb{Z}_n$ state in bilayer graphene} We next ask if similar physics can appear elsewhere. We expect to find a transition in the zeroth Landau level of bilayer graphene (BLG) as a function of the magnetic field. The zeroth Landau level of BLG is exactly equivalent to the LLL of GaAs when the magnetic field is infinite and continuously interpolates to the SLL of GaAs as the magnetic field is decreased. As such, we expect that a $\mathbb{Z}_n$ state is stabilized below a critical field and the Laughlin state is favored above the critical field. The Coulomb interaction between electrons can be parameterized using Haldane pseudopotentials $V_m$~\cite{Haldane83}, which is the energy of two electrons in a relative angular momentum state $m$ in the disk geometry. The pseudopotentials in the zeroth LL of bilayer graphene are given by \begin{equation} V^{0-{\rm BLG}}_m(\theta) = \int_{0}^{\infty} dq~F^{0-{\rm BLG}}(\theta,q)e^{-q^2}L_m(q^2).\label{BLGVm} \end{equation} where the Fourier transformed form factor is \cite{Apalkov11} \begin{equation} F^{0-{\rm BLG}}(\theta,q) = \Bigg[\sin^{2}(\theta)L_{1}\Big(\frac{q^2}{2}\Big)+\cos^{2}(\theta)L_{0}\Big(\frac{q^2}{2}\Big) \Bigg]^2. \label{eq:ZerothLL_BLG_form_factor} \end{equation} Here we have set the magnetic length to unity, $L_r(x)$ is the $r$th order Laguerre polynomial and $\theta$ is a parameter that varies between $0$ and $\pi/2$ to control the relative proportion of the $n=0$ and $n=1$ LLs in the two-component wave function. At $\theta=0$ the form factor is that of the LLL of GaAs while for $\theta = \pi/2$, it is exactly the form factor for the SLL in GaAs. At the mid-way point, $\theta =\pi/4$, the form factor is that of the $n=1$ LL in monolayer graphene~\cite{Balram15c}. The value of $\theta$ is related to the magnetic field by $\tan^2{(\theta)}\propto \ell /(\hbar v_F)\propto 1/\sqrt{B}$, where $v_F$ is the Fermi velocity. For very large magnetic fields, we anticipate the physics of the LLL of GaAs. As the magnetic field is lowered, we first expect to see the physics of monolayer graphene appear (which has been shown to be well described by the composite fermion theory~\cite{Balram15c}) that eventually gives way to states exhibiting the physics of the SLL of GaAs at very small magnetic fields. \begin{figure}[tbhp] \includegraphics[width=0.4\textwidth]{BiGrNoSpinEdit} \caption{Energies of $\mathbb{Z}_n$ states in bilayer graphene, measured relative to the energy of the Laughlin state, as a function of the tangent of the mixing angle $\theta$. All energies represent thermodynamic limits. The $\mathbb{Z}_4$ state is seen to be favored for $\theta \gtrapprox 1.45$. The top axis shows $\hbar v_F/\ell$ in units of meV (see text for relation between $\theta$ and $\hbar v_F/\ell$). Energies are shown only in the vicinity of the transition. } \label{fig:BiGr} \end{figure} We construct an effective interaction as shown in Appendix~\ref{sec: effInt}, and obtain the thermodynamic energies of various candidate states as a function of $\theta$. The angle $\theta$ is related to measurable quantities through $\tan{\theta}= t \ell/(\sqrt{2}\hbar v_F)$ where $t$ is the hopping integral and $v_F$ is the Fermi velocity~\cite{Apalkov11}. Taking $t\sim 350\text{meV}$, as obtained from DFT calculations at zero magnetic field~\cite{Jung14}, we obtain $\hbar v_F/\ell = 350\text{meV}/(\sqrt{2}\tan{\theta})$. The top axis in Fig.~\ref{fig:BiGr} shows $\hbar v_F/\ell$ in units of meV. We find that the transition from the $\mathbb{Z}_{4}$ to the Laughlin state occurs approximately at $\hbar v_F/\ell\sim30$ meV. For graphene, with a typical Fermi velocity of $10^6$ $m/s$, this corresponds to a magnetic field strength of $B\approx 1.4$T. \section{Landau level mixing and spin} It is natural to ask if $\mathbb{Z}_n$ parton states can be relevant for $\nu=1/3$ in the LLL. We have performed extensive calculations as a function of quantum well width and density, also including LL mixing. We use a self-consistent LDA calculation to determine the transverse electron density at zero magnetic field at several electron densities and quantum well widths. We further include LL mixing through the so-called fixed phase diffusion Monte Carlo method~\cite{Ortiz93, Zhang16, Zhao18}. For all parameters we have considered, the Laughlin state remains the lowest energy state. The detailed results are given in Appendix~\ref{sec: ThermoLim}. Recent experiments have mapped out the spin polarization of the SLL in GaAs quantum wells~\cite{Yoo19}. They observe an anomalous spin depolarization between fillings factors 1/5 and 1/3. The Laughlin state is fully spin polarized, but $\mathbb{Z}_n$ states with $n>1$ allow for the possibility of spin-unpolarized or spin-partially polarized states. The generalization is analogous to that for Jain CF states to spin-singlet or partially spin-polarized states~\cite{Wu93, Park98, Jain07, Balram15a, Balram17}. Specifically, the $\mathbb{Z}_n$ state can be generalized to include spin as \begin{eqnarray} \Psi_{1/m}^{\mathbb{Z}_n(n,0;\bar{n}_\uparrow,\bar{n}_\downarrow)} &=& \mathcal{P}_{\rm LLL}\Phi_n\Phi_{\bar{n}_\uparrow,\bar{n}_\downarrow}\Phi_1^m \nonumber \\ &\sim & \Psi_{n/(2n+1)} \Phi_1^{m-4} \mathcal{P}_{\rm LLL} \Phi^*_{n_\uparrow,n_\downarrow} \Phi_1^2 \label{Znspin} \end{eqnarray} where $n =n_\uparrow + n_\downarrow$, and $\Phi_{n_\uparrow,n_\downarrow}$ represents the state with $n_\uparrow$ spin-up and $n_\downarrow$ spin-down filled Landau levels. These wave functions can be shown to satisfy the Fock cyclic conditions~\cite{Jain07}. In the above wave function, we have made the $\bar{n}$-parton spinful. An analogous wave function $\Psi_{1/m}^{\mathbb{Z}_n(n_\uparrow,n_\downarrow;\bar{n},0)}$ can be written where the $n$-parton is endowed with spin. Which configuration is preferred depends on the interaction. Our detailed calculations, shown in Appendix~\ref{sec: ThermoLim}, demonstrate that the fully spin-polarized states have better variational energies for all interactions considered in this article. \section{Discussion} \label{sec: discussion} Our work was motivated by an apparent discrepancy between theory and experiment for the FQHE at $\nu=7/3$: while theory finds the $\mathbb{Z}_4$ parton state to have lower energy than the Laughlin state, experiments are consistent with the latter. We find that when we take into account finite width corrections, there is a transition from the $\mathbb{Z}_4$ state into the Laughlin state at width $\sim 1.5 \ell$. All experimental observations of the 7/3 state appear to be for larger widths and thus fall in the region where the Laughlin state is favored. (Large mobilities, necessary for an observation of the 7/3 state, are typically obtained for relatively wide quantum wells because that minimizes the effect of interface roughening. One may alternatively go to low densities. The 7/3 state has been observed at very low densities~\cite{Pan14, Samkharadze17}, but even there, with 7/3 state occurring at $B\approx 0.9$T, the width of 65 nm translates approximately into 2.5 $\ell$.) It may be possible to decrease both the quantum well width and the density to get into the regime where the $\mathbb{Z}_4$ state is predicted. If a phase transition is observed (for example, by gap closing and reopening as a function of the density), it would provide evidence in favor of an unconventional $\mathbb{Z}_n$ state at small widths, and also of the role of large width in stabilizing the Laughlin state at 7/3. We note that a previous exact diagonalization calculation has also shown that the magnetoroton branch, absent at zero width, appears by the time the quantum well width is three magnetic lengths~\cite{Jolicoeur17}. Another exact diagonalization study has shown that finite width stabilizes the 7/3 Laughlin state~\cite{Balram20}. An additional experimental quantity, namely the chiral central charge, can be measured in thermal Hall conductance measurements and can sometimes distinguish between states with different topological content~\cite{Kane97, Cappelli02}. For the $\mathbb{Z}_n$ states, the chiral central charge is independent of the value of $n$ and so the thermal Hall conductance is predicted to be the same for all of these states. The measured value of the thermal Hall conductance at $7/3$ is consistent with all of these states \cite{Banerjee17b}. The Hall viscosity of all $\mathbb{Z}_n$ states is identical because they all have the same shift. These states are, however, not topologically equivalent as they have different topological entanglement entropies~\cite{Balram19a}. The clearest experimental signature distinguishing the states will be the charge of the fundamental quasiparticles. The Laughlin state has charge $-e/3$ quasiparticles while the $\mathbb{Z}_n$ state has charge $-e/(3n)$ quasiparticles. These quasiparticles can, in principle, be detected through scanning electron transistor experiments~\cite{Venkatachalam11}. The situation for shot noise experiments is more subtle. As Balram \emph{et al}. argued~\cite{Balram19a}, the $-e/(3n)$ quasiparticles are gapped at the edge and only the $-e/3$ quasiparticles can be excited at arbitrarily low temperatures. It may be possible, however, that the $-e/(3n)$ quasiparticles become relevant in shot noise experiments at somewhat elevated temperatures (or voltage bias). We note that the so-called anti-Read-Rezayi $4$-cluster (aRR$4$) state~\cite{Read99} also provides a plausible candidate wave function for the $7/3$ FQHE~\cite{Peterson15}. The energy of the aRR$4$ state is equal to the energy of the Laughlin state within numerical uncertainty~\cite{Peterson15}, in contrast to our $\mathbb{Z}_4$ state which has lower energy than Laughlin's. Furthermore, the aRR$4$ state has overlaps of 0.77 and 0.59 with the exact ground state for 10~\cite{Kusmierz18} and 12 particles, whereas the $\mathbb{Z}_2$ state has an overlap of 0.87 for 10 particles and $\mathbb{Z}_3$ has an overlap of 0.93 for 9 particles~\cite{Balram19a}. (The $\mathbb{Z}_4$ state requires a minimum of 16 particles, for which we cannot obtain overlaps.) Finally, assuming equilibration of all edge modes, the thermal Hall measurements at 7/3 are inconsistent with the chiral central charge of the aRR$4$ state~\cite{Banerjee17b}. \begin{acknowledgments} The work at Penn State was supported by the U. S. Department of Energy, Office of Basic Energy Sciences, under Grant no. DE-SC0005042. Some portions of this research were conducted with Advanced CyberInfrastructure computational resources provided by The Institute for CyberScience at The Pennsylvania State University. Some of the numerical calculations reported in this work were carried out on the Nandadevi supercomputer, which is maintained and supported by the Institute of Mathematical Science’s High Performance Computing center. One of us (Th.J.) acknowledges CEA-DRF for providing CPU time on the supercomputer COBALT at GENCI-CCRT. \end{acknowledgments}

{'timestamp': '2020-11-25T02:25:14', 'yymm': '2011', 'arxiv_id': '2011.12195', 'language': 'en', 'url': 'https://arxiv.org/abs/2011.12195'}

high_school_physics

Hubble maps 3-D structure of ejected material around erupting star by NASA's Goddard Space Flight Center Hubble's Wide Field Camera 3 imaged the double-star system T Pyxidis, or T Pyx, over a four-month period. T Pyx is a recurrent nova, erupting every 12 to 50 years. T Pyx's latest outburst was in April 2011. The star is the white blob in the middle of each image. Credit: NASA, ESA, A. Crotts, J. Sokoloski, and H. Uthas (Columbia University), and S. Lawrence (Hofstra University) (Phys.org) —A flash of light from a stellar outburst has provided a rare look at the 3-D structure of material ejected by an erupting nova. Astronomers used NASA's Hubble Space Telescope to observe the light emitted by the close double-star system T Pyxidis, or T Pyx, a recurring nova, during its latest outburst in April 2011. A nova erupts when a white dwarf, the burned-out core of a sun-like star, has siphoned off enough hydrogen from a companion star to trigger a thermonuclear runaway. As hydrogen builds up on the surface of the white dwarf, it becomes hotter and denser until it detonates like a colossal hydrogen bomb, leading to a 10,000-fold increase in brightness in a little more than one day. Nova explosions are extremely powerful, equal to a blast of one million billion tons of dynamite. T Pyx erupts every 12 to 50 years. Contrary to some predictions, the astronomers were surprised to find the ejecta from earlier outbursts stayed in the vicinity of the star and formed a disk of debris around the nova. The discovery suggests material continues expanding outward along the system's orbital plane, but it does not escape the system. "We fully expected this to be a spherical shell," says Arlin Crotts of Columbia University, a member of the research team. "This observation shows it is a disk, and it is populated with fast-moving ejecta from previous outbursts." Team member Stephen Lawrence of Hofstra University in Hempstead, N.Y., will present the results Tuesday at the American Astronomical Society meeting in Indianapolis. Team member Jennifer Sokoloski, also of Columbia University and co-investigator on the project, suggests these data indicate the companion star plays an important role in shaping how material is ejected, presumably along the system's orbital plane, creating the pancake-shaped disk. The disk is tilted about 30 degrees from face-on toward Earth. Using Hubble's Wide Field Camera 3, the team took advantage of the blast of light emitted by the erupting nova to trace the light's path as it lit up the disk and material from previous ejecta. The disk is so vast, about a light-year across, that the nova's light cannot illuminate all of the material at once. Instead, the light sweeps across the material, sequentially illuminating parts of the disk, a phenomenon called a light echo. The light reveals which parts of the disk are nearer to Earth and which sections are farther away. By tracing the light, the team assembled a 3-D map of the structure around the nova. "We've all seen how light from fireworks shells during the grand finale will light up the smoke and soot from shells earlier in the show," Lawrence said. "In an analogous way, we're using light from T Pyx's latest outburst and its propagation at the speed of light to dissect its fireworks displays from decades past." Although astronomers have witnessed light propagating through material surrounding other novae, this is the first time the immediate environment around an erupting star has been studied in three dimensions. Astronomers have studied light echoes from other novae, but those phenomena illuminated interstellar material around the stars instead of material ejected from them. The team also used the light echo to refine estimates of the nova's distance from Earth. The new distance is 15,600 light-years from Earth. Previous estimates were between 6,500 and 16,000 light-years. T Pyx is located in the southern constellation Pyxis, or the Mariner's Compass. The team is continuing to analyze the Hubble data to develop an outflow model. T Pyx has a history of outbursts. Besides the 2011 event, other previous known eruptions were seen in 1890, 1902, 1920, 1944, and 1966. Astronomers call erupting stars novae, Latin for "new," because they abruptly appear in the sky. A nova quickly begins to fade in several days or weeks as the hydrogen is exhausted and blown into space. The team also includes Helena Uthas of Columbia University. The team's results will appear online June 5 and will be published in the June 20, 2013, issue of the Astrophysical Journal Letters. Sokoloski is the paper's lead author. Accurate distance measurement resolves major astronomical mystery Journal information: Astrophysical Journal Letters Provided by NASA's Goddard Space Flight Center Citation: Hubble maps 3-D structure of ejected material around erupting star (2013, June 4) retrieved 18 July 2019 from https://phys.org/news/2013-06-hubble-d-ejected-material-erupting.html New table-top gravitational-wave detector proposed Light echoes from V838 Mon Spitzer and Hubble discover strobe-like flashes in suspected binary protostar Hubble reveals the ring nebula's true shape Massive white dwarf in our galaxy may go supernova Astronomers discover a new black hole in our galaxy cantdrive85 Their theory doesn't match the observation, for just about any other theory that qualifies as a falsification. Not so in astrophysics, GR is the Teflon theory. Fleetfoot If the star is accreting from a disc, it would be expected to have a very high spin so it seems logical that the material would remain concentrated around the equator. The "centrifugal" effect would also tend to concentrate the ejecta in that plane so perhaps this isn't so surprising. GSwift7 Contrary to some predictions Their theory doesn't match the observation No, you are mis-interpreting what they said. The differences between predictions here are for the most part, just a matter of what values you pick for certain unknown variables, such as the mass and density of the material, as well as the spatial distribution of the material. The fact that the ejecta is thrown out in a disk, in stead of a sphere, makes it more compact and therefore holds itself together better, like a rubber band. You should like this idea, the disk is full of plasma and it's a circle, so you can pretend that there's a huge magnetic field and electric current in the disk if you want. OH OH even better, the material is confined to the disk by the double layer sandwiching the whole system, with positive. Yeah, that's the ticket. "Contrary to some predictions, the astronomers were surprised to find the ejecta from earlier outbursts stayed in the vicinity of the star and formed a disk of debris around the nova." In 1994-95 Hubble imagery of T Pyx revealed a shell of several thousand knots surrounding the star: http://en.wikiped..._big.jpg A *reasonable* assumption was that this shell represented a spherical distribution of knots ejected by T Pyx in previous eruptions. The 2011 outburst of the star allowed astronomers to gain actual information on the 3D distribution of the knots and through the study of light echoes of the eruption found "the more-or-less circular appearance of the known remnant is due not to spherical symmetry, but to a 2D disk-like structure." A preprint of the new study is available here:

{"pred_label": "__label__wiki", "pred_label_prob": 0.8587141036987305, "wiki_prob": 0.8587141036987305, "source": "cc/2019-30/en_head_0019.json.gz/line1723890"}

high_school_physics

Invest for Good with IQ Dual Impact ETFs Our suite of ESG ETFs seek to impact your portfolio's potential and the world. Longevity & Growth Tax Efficient Investing Combating Inflation Generating Income Managing Volatility IndexIQ ETFs Variable Portfolios Gain From Our Insights Our investment managers offer unparalleled domain expertise and diversity of thought, generating deeper insights alongside strong conviction to deliver better outcomes. Access to our extensive library of product and market insights and materials. Product Prospectuses and Reports Our Portfolio Managers Global Boutiques Click here for more complete Morningstar ratings information. MainStay WMC Value Fund seeks long-term appreciation of capital. A fundamental value strategy Bottom-up, fundamental research-based approach seeking a compelling combination of value, quality and capital return characteristics. Wellington's boutique approach Investment team has autonomy to set the portfolio's philosophy and process, while leveraging the resources of a large, global firm. Focus on quality at a discount Seeks financially sound, temporarily out-of-favor companies that provide above-average total return potential at below average valuation. Class A & INV: 5.5% maximum initial sales charge; a 1% CDSC may be imposed on certain redemptions made within 18 months of the date of purchase on shares that were purchased without an initial sales charge. Class B: CDSC up to 5% if redeemed within six years. Class C: 1% CDSC if redeemed within one year. Class I: No initial sales charge or CDSC. Returns represent past performance which is no guarantee of future results. Current performance may be lower or higher. Investment return and principal value will fluctuate, and shares, when redeemed, may be worth more or less than their original cost. No initial sales charge applies on investments of $1 million or more (and certain other qualified purchases). However, a contingent deferred sales charge of 1.00% may be imposed on certain redemptions made within 18 months of the date of purchase on shares that were purchased without an initial sales charge. Expenses stated are as of the fund's most recent prospectus. Effective April 26, 2021, the Fund replaced both of its subadvisors and modified its principal investment strategies. The past performance prior to that date reflects the Fund’s prior subadvisors and principal investment strategies. Performance reflects a contractual fee waiver and/or expense limitation agreement for Class I shares in effect through 2/28/23, without which total returns may have been lower. This agreement renews automatically for one-year terms unless written notice is provided before the start of the next term or upon approval of the Board. Rotate your phone for full Morningstar Tools Experience Does not reflect the effect of any sales charge, which would reduce performance shown. Performance for other sales charges will differ based on differences in sales charges and expense structure. View all Holdings Top Sectors (%) Distribution & Yields Distributions may be comprised of ordinary income, net capital gains, and/or a return of capital (ROC) of your investment in the fund. Because the distribution rate and the 12-month rate may include a ROC, they should not be confused with yield or income. Please refer to the most recent Section 19 Notice, if applicable, for additional information regarding the composition of distributions. Final determination of a distribution’s tax character will be made on Form 1099 DIV sent to shareholders each January. Distribution Rate: The distribution rate measures the percentage return in the form of dividends. It is calculated daily by annualizing the most recent dividend distribution and dividing by the daily share price (NAV or POP). If the Fund did not make a distribution as of the latest scheduled distribution date, "N/A" will be displayed. 12-month Rate: The 12-month rate measures the percentage return in the form of dividends. It is calculated monthly by taking the sum of the trailing 12-month dividend payments divided by the last month's ending share price (NAV or POP) plus any capital gains distributed over previous 12 months. If the Fund did not make any distributions over the previous 12 months, "N/A" will be displayed. The 30 Day SEC Yield is calculated by dividing the net investment income per share for the first 30 days of the month by the offering price per share at the end of that period. The yield reflects the dividends and interest earned during the period, after the deduction of the Fund's expenses. Yield reflects a fee waiver and/or expense limitation agreement without which the 30 Day SEC Yield would have been lower. Dividend distributions are the distribution of a dividend to mutual fund shareholders as of a certain date. The following Funds declare daily dividends: MainStay MacKay California Tax Free Opportunities, MainStay Floating Rate, MainStay MacKay High Yield Municipal Bond, MainStay MacKay Infrastructure Bond, MainStay Money Market, MainStay MacKay New York Tax Free Opportunities, MainStay MacKay Tax Advantaged ShortTerm Bond and MainStay MacKay Tax Free Bond. Click here for MainStay Funds sales charge and breakpoint information. Fees & Expenses A global asset manager with expertise and solutions across equity, fixed income, multi-asset and alternatives. One of the world’s largest independent investment management firms, offering comprehensive investment capabilities built on rigorous, proprietary research spanning nearly all segments of the global capital markets, including equity, fixed income, multi-asset and alternative strategies. Adam Illfelder Please enter one email address. Sender Name * Sender Email * Your email address is used only to identify you to the recipient and in case of transmission errors. Neither your address nor the recipient's address will be sold or used for any purpose other than transmission of this message. MainStay WMC Value Fund Class A: MAPAX | Class B: MAPBX | Class C: MMPCX | Class I: MUBFX | Class INV: MSMIX | Class R1: MAPRX | Class R2: MPRRX | Class R3: MMAPX | Class R6: MMPDX BEFORE YOU INVEST Before considering an investment in the Fund, you should understand that you could lose money. Growth-oriented common stocks and other equity type securities (such as preferred stocks, convertible preferred stocks and convertible bonds) may involve larger price swings and greater potential for loss than other types of investments. The principal risk of investing in value funds is that the price of the security may not approach its anticipated value. Investing in mid-cap stocks may carry more risk than investing in stocks of larger, more well-established companies. Issuers of convertible securities may not be as financially strong as those issuing securities with higher credit ratings and are more vulnerable to economic changes. Foreign securities are subject to interest rate, currency exchange rate, economic, and political risks. These risks may be greater for emerging markets. Environmental, Social and Governance (ESG) managers may take into consideration factors beyond traditional financial information to select securities, which could result in relative investment performance deviating from other strategies or broad market benchmarks, depending on whether such sectors or investments are in or out of favor in the market. Further, ESG strategies may rely on certain values-based criteria to eliminate exposures found in similar strategies or broad market benchmarks, which could also result in relative investment performance deviating. Russell 1000® Value Index measures the performance of the large-cap value segment of the U.S. equity universe. It includes those Russell 1000® Index companies with lower price-to-book ratios and lower expected growth values. An investment cannot be made directly into an index. Standard Deviation measures how widely dispersed a fund's returns have been over a specified period of time. A high standard deviation indicates that the range is wide, implying greater potential for volatility. Alpha measures a fund's risk-adjusted performance and is expressed as an annualized percentage. Beta is a measure of historical volatility relative to an appropriate index (benchmark) based on its investment objective. A beta greater than 1.00 indicates volatility greater than the benchmark's. R-Squared measures the percentage of a fund's movements that result from movements in the index. Sharpe Ratio shown is calculated for the past 36-month period by dividing annualized excess returns by annualized standard deviation. Annual Turnover Rate is as of the most recent annual shareholder report. The Morningstar Rating™ for funds, or "star rating", is calculated for managed products (including mutual funds, variable annuity and variable life subaccounts, exchange-traded funds, closed-end funds, and separate accounts) with at least a three-year history. Exchange-traded funds and open-ended mutual funds are considered a single population for comparative purposes. It is calculated based on a Morningstar Risk-Adjusted Return measure that accounts for variation in a managed product's monthly excess performance, placing more emphasis on downward variations and rewarding consistent performance (this does not include the effects of sales charges, loads, and redemption fees). The top 10% of products in each product category receive 5stars, the next 22.5% receive 4 stars, the next 35% receive 3 stars, the next 22.5% receive 2 stars, and the bottom 10% receive 1 star. The Overall Morningstar Rating for a managed product is derived from a weighted average of the performance figures associated with its three-, five-, and 10-year (if applicable) Morningstar Rating metrics. The weights are: 100% three-year rating for 36-59 months of total returns, 60% five-year rating/40% three-year rating for 60-119 months of total returns, and 50% 10-year rating/30% five-year rating/20% three-year rating for 120 or more months of total returns. While the 10-year overall star rating formula seems to give the most weight to the 10-year period, the most recent three-year period actually has the greatest impact because it is included in all three rating periods. GLOBAL BOUTIQUES GLOBAL BOUTIQUES Candriam GoldPoint Partners IndexIQ Kartesia MacKay Shields Madison Capital Funding NYL Investors PA Capital Tristan Capital Partners Sign up to get the latest on ideas and insights that matter most to you. Consider the Funds' investment objectives, risks, charges, and expenses carefully before investing. The prospectus, or summary prospectus, and the statement of additional information include this and other relevant information about the Funds and are available by calling (888) 474-7725 for IndexIQ ETFs and 800-624-6782 for MainStay Funds®. Read the prospectus carefully before investing. The strategies discussed are strictly for illustrative and educational purposes and are not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. There is no guarantee that any strategies discussed will be effective. The MainStay Funds® are managed by New York Life Investment Management LLC and distributed through NYLIFE Distributors LLC, 30 Hudson Street, Jersey City, NJ 07302, a wholly owned subsidiary of New York Life Insurance Company. NYLIFE Distributors LLC is a Member of FINRA/SIPC. IndexIQ® is an indirect wholly owned subsidiary of New York Life Investment Management Holdings LLC and serves as the advisor to the IndexIQ ETFs. ALPS Distributors, Inc. (ALPS) is the principal underwriter of the ETFs. NYLIFE Distributors LLC is a distributor of the ETFs. NYLIFE Distributors LLC is located at 30 Hudson Street, Jersey City, NJ 07302. ALPS Distributors, Inc. is not affiliated with NYLIFE Distributors LLC. NYLIFE Distributors LLC is a Member FINRA/SIPC. “New York Life Investments” is both a service mark, and the common trade name, of certain investment advisors affiliated with New York Life Insurance Company. The mutual funds may be offered and sold only to persons in the United States. The products and services of New York Life Investments' boutiques are not available in all jurisdictions or regions where such provision would be contrary to local laws or regulations. AUM: As of September 30, 2021. Assets under management includes the assets of the investment advisors affiliated with New York Life Insurance Company. AUM beginning in 2012 excludes Assets under Administration. AUM is reported in USD. AUM that are not denominated in USD are converted using spot rates as of the reporting date. ©2021 New York Life Investment Management LLC. All rights reserved. You are now leaving the New York Life Investments Web Site. You assume total responsibility and risk for your use of the site(s) to which you are linking. The information being provided is strictly as a courtesy. When you link to any of the web sites provided herewith, you are leaving this site. Neither the Company, New York Life nor any of its subsidiaries, make any representations as to the completeness or accuracy of information that is provided at these sites. Nor are they liable for any direct or indirect technical or systems issues or any consequences arising out of your access to or your use of third-party technologies, sites, information and programs made available through this site. Retirement Specialist New York Life Investments uses cookies to personalize and improve your site experience and provide you with information about products and services tailored to your interests. When selecting ‘Accept and Continue’ you accept all categories of cookies. For more information about the types of cookies we use and how you can disable certain cookies, please visit our “Online Privacy Policy” which can be found under “Privacy and Other Policies". Please be aware that certain types of cookies are necessary to browse our website and therefore cannot be disabled. These necessary cookies do not collect any personal information about you. Please select your role so we can personalize the site to meet your needs Almost There ! Please enter your email to download:

{"pred_label": "__label__cc", "pred_label_prob": 0.5949418544769287, "wiki_prob": 0.4050581455230713, "source": "cc/2022-05/en_middle_0074.json.gz/line114240"}

high_school_physics

District of Massachusetts Outreach & Initiatives Drug/Violent Crime Violent Extremism Prevention Victim Witness Program Victim Witness Resources Daily Court Calendars Springfield Court Calendar Worcester Court Calendar Search Case Information Civil and Criminal Debt Repayment Guide to Federal Court and Legal Terms Business/Vendors Assistant U.S. Attorneys Administrative & Professional Staff U.S. Attorneys » District of Massachusetts » News Two Men Arrested for Conspiracy to Distribute Fentanyl BOSTON – Two men were arrested on Wednesday, Dec. 11, 2019, and charged with conspiracy to distribute fentanyl. Steven Perez, of Lowell, and Anthony Holloway, of Methuen, were arrested and charged with conspiracy to distribute and possession with intent to distribute fentanyl. Following an initial appearance in federal court in Boston, Perez and Holloway were detained pending a detention and probable cause hearing set for Dec. 19, 2019. According to the charging documents, over the course of several months, law enforcement investigated Perez and Holloway’s fentanyl distribution operation. The men allegedly used an elderly housing complex run by the Methuen Housing Authority to package drugs and conduct drug sales. The charge of conspiracy to distribute and to possess with intent to distribute fentanyl carries a sentence of up to 20 years in prison, at least three years of supervised release and a fine of $1 million. Sentences are imposed by a federal district court judge based upon the U.S. Sentencing Guidelines and other statutory factors. United States Attorney Andrew Lelling; Joseph R. Bonavolonta, Special Agent in Charge of the Federal Bureau of Investigation’s Boston Field Office; and Methuen Police Chief Joseph Solomon made the announcement. Assistant U.S. Attorney Philip C. Cheng of Lelling’s Narcotics and Money Laundering Unit is prosecuting the case. The details contained in the charging documents are allegations. The defendants are presumed innocent unless and until proven guilty beyond a reasonable doubt in a court of law. USAO - Massachusetts Making sure victims of federal crimes are treated with compassion, fairness and respect. Information on the Boston Marathon bombing cases

{"pred_label": "__label__wiki", "pred_label_prob": 0.5475289821624756, "wiki_prob": 0.5475289821624756, "source": "cc/2020-05/en_head_0038.json.gz/line679173"}

high_school_physics

The mosaics accommodate significantly more charge per unit area than previously estimated for contact electrification, but the overall/“net” charge on an electrified surface remains relatively small due to the “compensation” between the (+) and the (–) regions. sounds much like an analog of feromagnetic materials, that are not magnetized. they have small domains inside that are magnetized, but they are distributed at random and they cancel each other out. When an external magnetic field is applied, the domains alighned in that direction begin to grow at the expense of the other, and the "domain walls" shift appropriately.

{'timestamp': '2019-04-22T06:12:46Z', 'url': 'https://www.overunityresearch.com/index.php?topic=1529.0', 'language': 'en', 'source': 'c4'}

high_school_physics

GlossaryRodgers Reidy2022-02-24T10:17:24+00:00 Home » Glossary Who is qualified to be appointed as Receiver or Receiver and Manager? Pursuant to Companies Act 2016 Section 372, “Any person who is an approved liquidator referred to in section 433 shall be qualified to be appointed as receiver or receiver and manager.”. Section 385 state that an approved liquidator may be appointed as receiver or receiver and manager if an application is made to the court in the event that the company is wound up by court order. How does one appoint a Receiver or Receiver and Manager? Section 374 of the Companies Act 2016 states that the power to appoint could be made under: a. Any instrument that confers on a debenture holder or charge holder; b. Any instrument that creates a charge in respect of property and undertaking of a company confers on the charge holder; or c. By the Court. What are the differences between Creditors’ Voluntary Winding-Up and Members’ Voluntary Winding-Up? CREDITORS’ VOLUNTARY WINDING-UP The creditors’ voluntary winding-up is initiated by the directors and approved by shareholders and creditors of the company. An Extraordinary General Meeting (“EGM”) of shareholders and a creditors’ meeting be convened to resolve that the company be wound up within five (5) weeks from the lodgement of Declaration of Inability to Continue Business. The provisional liquidator is appointed at the board of directors’ meeting for the purpose of overseeing the affairs of the company thereafter until the holding of the EGM and the creditors’ meeting. Where there is a difference between the company’s and the creditors’ nominee(s), the creditors’ choice of liquidator(s) will prevail over that of the Board of Directors’ and members’ nomination. MEMBERS’ VOLUNTARY WINDING-UP Members’ voluntary winding-up is initiated by the company’s directors and approved by the shareholders due to business environment, dormant status of the company and/or corporate restructuring involving a group of companies. Prior to initiating a members’ voluntary liquidation, it is imperative that the requirements of Section 443 of the Companies Act 2016 are complied with whereby the directors of the company are required to confirm the solvency position of the company and is able to pay all its debt in full within twelve (12) months from the commencement of liquidation. What is the power of the Receiver or Receiver and Manager? The 6th Schedule [Section 383] of the Companies Act 2016 outlined the power of the Receiver or Receiver and Manager. What are the powers of a Liquidator? The power of an appointed liquidator in a company winding-up by court is outlined in the Companies Act 2016 under Schedule 12 (Secs. 472 & 486) and Schedule 11 for voluntary winding-up. What are the fees involved? The professional fees chargeable for court liquidation is guided under Table C of Companies (Winding-up) Rules 1972: A. Where the Official receiver acts as interim Liquidator only: i. RM1.00 for each member, creditor or debtor ii. On the value of the company’s assets as estimated in the statement of affairs a. 1% on the first RM100,000 or fraction thereof; b. ½% above RM100,000. B. Where the Official Receiver is liquidator: ii. Upon the total net assets realized: a. 6% on the first RM50,000 or fraction thereof; b. 5% on the first RM50,000 or fraction thereof; c. 4% above RM100,000. C. Upon the amount available for distribution to Creditors or contributories and each time a distribution is made: i. 4% on the first RM50,000; ii. 3% on the next RM50,000; iii. 2% above RM100,000. The professional fees structure for court liquidation, creditors’ voluntary winding-up, members’ voluntary winding-up and receivership can be based on the estimates of the time expected to incurred and resources to facilitate the execution of the role appointed and complexity of the engagement. The fees structure will be revised on a regular basis to commensurate with the level of work involved. In additional to this, there will be ancillary expenses like legal fees, insurance premium, security personnel, utilities and rates, taxes like corporate tax and goods and services tax, consultants professional fees for example tax, audit and property surveyor, and any other expenses incurred that will facilitate the appointed officer to carry out the duties and responsibilities. Organise a meeting today.

{"pred_label": "__label__cc", "pred_label_prob": 0.5287399888038635, "wiki_prob": 0.4712600111961365, "source": "cc/2023-06/en_head_0060.json.gz/line249663"}

high_school_physics

SensorView Demo Posted on October 17, 2019 November 25, 2019 by Derek Wallace Overhead power lines are used to transmit and distribute electrical power across long distances, forming a power or national grid. Steel tower structures, pylons or wooden poles suspend electricity-carrying conductors across a range of operating voltages. Business and residential consumers are normally supplied by high, medium and low voltage lines that are categorized depending on their destinations. High voltage transmission lines are considered to be from 69KV up to around 100KV for bulk transmission to large consumers. Medium voltage sub-transmission lines are between 1KV and 69KV for distribution in urban and rural areas. Low voltage distribution lines are below 1KV and likely destined for residential or small business consumers. In the USA, overhead power lines are more common due to the life-cycle cost of those underground, being 2-4 times the cost of overhead. Underground cables are also more likely subject to ground movement. The North American bulk electric system consists of more than 360,000 miles of transmission lines, including approximately 180,000 miles of high voltage lines. Transmission systems are generally administered by a Regional Transmission Organization (RTO) or an Independent System Operator (ISO) which operates the region’s electricity grid and administers the region’s wholesale electricity markets. Over 3,200 Utility companies generate, transmit and distribute electricity for sale to over 145 million customers. Overhead power lines and their supporting infrastructure may fail over time due to ageing, natural phenomena or accidental contact by vehicles and humans, resulting in power outages and potential wildfires in dry vegetation regions if cables remain live. Conductors can sag to breach safe height limits due to fallen branches, broken insulators and failed fixtures. In hot weather cable expansion and heavy electrical loading causes sagging, along with ice formation in cold weather. Conductors can fall to ground level, especially if tensioning or supporting equipment completely fails or if a vehicle collides with a tower or pole. Rotted or poles damaged over time can also lead to the same result. Entanglement of vehicles with overhead lines is a major problem, especially on farms. Incidents of combine harvesters, cranes, tipping trailers, Hiab loader cranes and mobile elevated work platforms contacting power lines have been widely reported. The UK’s CAA and distribution network operators recommend a 50m exclusion zone for drone operation near overhead lines. Other associated problems include anglers getting fishing lines entangled, kites and metallic helium balloons coming into contact and trees growing too close or through distribution power lines. In the USA, power outages due to weather are the most common type of disruptive events. These events include hurricanes, tropical storms, tornadoes, snow and ice storms and flooding. In the UK, the Health & Safety Executive (HSE) publish guidance notes for the construction, agriculture and tree care industries when working near overhead power lines. This includes planning and preparation, eliminating danger, controlling access and work. It also gives information as to the procedures to follow if a worker or vehicle does come into contact with a conductor. In the period 2012 to 2016, over 3000 haulage and transport vehicles were reported to have come into contact with overhead power lines, resulting in 59 people receiving injuries. In the last five years, there have been more than 1,140 near misses involving electricity on farmland. Eight people have been killed in the last two years. In the USA, the Electrical Safety Foundation International (ESFI) is a non-profit organization dedicated exclusively to promoting electrical safety at home and in the workplace. It states that 36% of all electrically-related workplace fatalities were caused by overhead power lines between 2011 and 2017. Reported Incidents July 2012 (UK) – A jogger running on a footpath came into contact with an 11KV power cable. Sections of a porcelain insulator had disintegrated on a wooden pole which supported the cable, causing it to sag. August 2012 (USA) – Two employees in a boom truck were installing a billboard sign when a piece of metal conduit used to secure the sign came into contact with an energized power line. Both employees suffered electrical burns, one of them was later pronounced dead in hospital. 2017 (UK) – An 18-year-old was electrocuted when he stepped out of his tractor cab after his tipping trailer came into contact with an overhead line. October 2018 (USA) – Two employees were operating a concrete pump truck when the boom and hose touched an overhead power line carrying 13KV. Both employees were electrocuted. February 2019 (UK) – A member of the public in the Western Isles moved fallen high voltage overhead power lines across a main road. Fortunately, the cable was not live. July 2019 (USA) – An employee was operating a Skytrax telehandler traveling to a staging area to pick up roofing materials. The boom was in the elevated position and made contact with an energized 69KV power line. The employee was electrocuted. August 2019 (UK) – A 12-year-old boy was killed after he was believed to be swinging a cable on a railway bridge when it came into contact with a power line. September 2019 (UK) – Western Power Distribution (WPD) issued a warning after responding to three incidents in which hot air balloons came into contact with overhead lines in the previous 18 months. CNIguard’s OverLine is an Overhead Line Monitoring System (OHLMS) that monitors the integrity of overhead conductors and poles used to distribute low and medium voltage electricity. The scalable solution addresses public and worker safety, alerting electricity distribution network operators to immediate infrastructure incidents or failures to improve performance and response to events. Find out more about OverLine here or contact us to request a full product demonstration. IIoT Networks Explained Investment from Energy Innovation Capital CNIguard collaborates with Senet and Semtech CNIguard welcomes new Chairman Introducing the Sentir S2.1 manhole monitoring system New IoT system to prevent manhole cover explosions Collaboration is preventing manhole explosions and fires © 2021 CNIguard | UK +44-207-100-5678 | USA +1-212-764-0100 Website by Avid Creative | Privacy Policy | Cookie Policy For a demo of the SensorView dashboard, please complete the form below. GasMarshal InDetect SensorView SensorBridge SensorCore SensorLink

{"pred_label": "__label__cc", "pred_label_prob": 0.6886797547340393, "wiki_prob": 0.3113202452659607, "source": "cc/2021-04/en_middle_0022.json.gz/line552489"}

high_school_physics

Latest Seminar Topics for Electrical Engineering Students Best Seminar topics on the latest innovative technology for electrical and electronics engineering students. 0.0.1 Best Seminar topics on the latest innovative technology for electrical and electronics engineering students. 1 Vortex .Bladeless Wind Generator 1.0.1 THE FUTURE OF WIND TURBINES? NO BLADES 2 Rain Power: Harvesting Energy from the Sky 2.0.1 Superconductivity: Present and Future Applications 2.0.2 Future energy solutions 2.0.3 Slotless Ac Induction Motor Vortex .Bladeless Wind Generator THE FUTURE OF WIND TURBINES? NO BLADES Day by day energy demand is rising, to meet that demand in the future, civilization will be forced to research and develop alternative energy sources. In order to survive the energy crisis, many engineers in the energy industry are inventing new ways to generate energy from nonconventional energy sources. Their idea is the Vortex, a bladeless wind turbine that looks like a giant rolled joint shooting into the sky. The Vortex has the same goals as conventional wind turbines: To turn breezes into kinetic energy that can be used as electricity. A more efficient, cost-effective, and environmentally friendly way to produce energy. Rain Power: Harvesting Energy from the Sky Trying to find alternative sources of energy has proven to be an extraordinary feat, allowing us to use everything from sun to the motion of the ocean. But there is still one plentiful source of renewable energy which has so far remained pretty much untapped: rain. Getting energy from falling water droplets might seem like an obvious, ‘why didn’t I think of that’ idea, but so far no-one has really exploited this plentiful (albeit somewhat unreliable) energy source. Now a team from CEA/Leti-Minatec has created a system that is capable of recovering kinetic energy from the impact of falling raindrops. To capture the raindrops’ mechanical energy, the scientists used a PVDF (polyvinylidene fluoride) polymer, a piezoelectric material that converts mechanical energy into electrical energy. When a raindrop impacts the 25-micrometer-thick PVDF, the polymer starts to vibrate. Electrodes embedded in the PVDF are used to recover the electrical charges generated by the vibrations. Superconductivity: Present and Future Applications Superconductivity is a phenomenon of exactly zero electrical resistance and rejection of magnetic fields occurring in certain materials when cooled below a critical temperature. Superconductor-based products are extremely environmentally friendly compared to their conventional counterparts. They generate no greenhouse gases and are cooled by non-flammable liquid nitrogen (nitrogen comprises 80% of our atmosphere) as opposed to conventional oil coolants that are both flammable and toxic. They are also typically at least 50% smaller and lighter than equivalent conventional units which translate into economic incentives. These benefits have given rise to the ongoing development of many new applications in the following sectors: Superconductivity Application in Power System If we have a conductor with no loss, we can make more efficient and reliable electrical infrastructure. Superconductor, which is zero resistance, is one of the promising solutions to make innovation on electric grids. Superconductors enable lots of applications to support our aging and heavily loaded electric power infrastructure – for example, in generators, transformers, high tension underground cables, synchronous condensers and fault current limiters. Other promising applications in power system are Superconducting Synchronous Condenser (DSC :SuperVar) and Superconducting motor. SuperVar is a good solution as reactive power compensator which can be applied to increase power transmission capability on the voltage stability limited system. The high power density and electrical efficiency of superconductor wire result in highly compact, powerful devices and systems that are more reliable, efficient, and environmentally. Future energy solutions Here’s a list of some of the solutions, that might help to power our sustainable energy future. Underwater Wind Turbine Biomass Briquettes Micro nuclear reactor Enhanced Geothermal Systems Photosynthetic Solar Cell Green Hydrocarbons Magenn Air Rotor System (Mars Slotless Ac Induction Motor The present invention is a rotating induction motor that is capable of providing higher peak torque than a conventional design, which achieves the shortcomings of the prior art by in regard to iron saturation by a slotless design; removing the iron slot provides more space for the conductor. The motor comprises a stator and a concentric rotor, separated from the stator by an air gap. The rotor has rotor bars and rotor windings. The stator is slot-less and comprises surface mounted conductors separated from each other by suitable insulation. An advantage of this design is that the motor does not exhibit typical behavior at high currents; there is no saturation effect A game-changing solar cell that cheaply and efficiently converts atmospheric carbon dioxide directly into usable hydrocarbon fuel, using only sunlight for energy.Unlike conventional solar cells, which convert sunlight into electricity that must be stored in heavy batteries, the new device essentially does the work of plants, converting atmospheric carbon dioxide into fuel, solving two crucial problems at once. A solar farm of such “artificial leaves” could remove significant amounts of carbon from the atmosphere and produce energy-dense fuel efficiently.“The new solar cell is not photovoltaic — it’s photosynthetic,” Green Hydrocarbons (Biofuels) Renewable hydrocarbon biofuels are fuels produced from biomass through a biological and thermochemical processes. These biofuels are similar to petroleum gasoline, diesel, or jet fuel in chemical makeup and are therefore considered alternative fuels. Biofuels can be used in vehicles without engine modifications and can utilize existing petroleum distribution systems. Magenn Air Rotor System (Mars) The Magenn Air Rotor System (MARS) is the next generation of wind turbines with cost and performance advantages over existing systems. MARS is a lighter-than-air tethered wind turbine that rotates about a horizontal axis in response to the wind, generating electrical energy. This electrical energy is transferred down the tether for consumption, or to a set of batteries or the power grid. Helium sustains the Magenn Air Rotor System, which ascends to an altitude as selected by the operator for the best winds. Its rotation also generates the “Magnus” effect. This aerodynamic phenomenon provides additional lift, keeps the MARS device stabilized, positions MARS within a very controlled and restricted location, and finally, causes MARS to pull up overhead to maximize altitude rather than drift downwind on its tether. It’s become mandatory rather than option to go for the renewable source of energy today in the whole world. For the same requirements we need advance options for future, hence MARS proves its excellence to use for better future. Amorphous Core Transformer An amorphous (non-crystalline) form of steel which has very high electrical resistance and low coercivity compared to previously used transformer steels, properties which reduce electrical losses and allow a given transformer to operate at higher ratings.Image result “The alloys of boron, silicon, phosphorus, and other glass formers with magnetic metals (iron, cobalt, nickel) have high magnetic susceptibility, with low coercivity and high electrical resistance. Usually the conductivity of a metallic glass is of the same low order of magnitude as of a molten metal just above the melting point. The high resistance leads to low losses by eddy currents when subjected to alternating magnetic fields, a property useful for e.g. transformer magnetic cores. Their low coercivity also contributes to low loss.” Intelligent Transformer Substations for Future-Proof Power Distribution Today‘s transformer substations, originally designed for a merely unidirectional energy flow and equipped with conventional transformers, are no longer capable of coping with the effects of volatile power sources. The consequences are more and more frequent supply breakdowns in the classical distribution grid, with ever increasing downtimes. In order to reduce such downtimes notably and to limit the associated blackout costs, quick adjustments to the changed load conditions must be possible. Benefits of intelligent transformer substations Monitoring and assurance of power quality Controlling of overload situations Minimization of loss of power grid revenue by notably reduced interruption times Optimisationof grid expansion Object monitoring of the transformer substation Tomorrow’s digital power grids How will a future electricity grid manage the demands of induction cooking, charging electric cars and roof-installed solar panels? The answer is Smart Grids, which involves digitisation of the electricity grid. The aim of a new research center called CINELDI (Centre for Intelligent Electricity Distribution – to empower the future Smart Grid) is to develop systems as part of tomorrow’s adaptable, robust and intelligent energy system. “Smart Grids provide output and energy efficiency, and make it easier to exploit renewable energy sources. They can also help towards removing the need to expand existing grid capacity – something which would be unavoidable if Smart Grids hadn’t entered the stage”, says SINTEF’s Gerd Kjølle, who will be heading the CINELDI center. Major demands are placed on the electricity grid when we cook our meals on induction hobs and charge up our electric cars at the same time. But Smart Grids enable us, for example, to give the grid operator permission to disconnect consumption linked to water heating, thus avoiding the need to expand grid capacity. http://www.sintef.no/en/projects/cineldi/ Smart Transformer for Smart Grid Smart transformers are transformers which integrates a general transformer’s features with monitoring software and communication technology among others. Power transformers are one of the most important and complex component of the electricity generation and transmission process. An owner of a power transformer has to face certain challenges such as high maintenance cost for extending the life of a transformer among others. Failure of a power transformer can lead to dire consequences such as revenue loss, life loss and legal consequences among others.Such challenges have led to the advent of smart transformers. These new-age transformers are designed to allow the maintenance of the power load independently. They can regulate the voltage constantly, while maintaining contact with the smart grid infrastructure, and deliver the optimal amount of power as and when required. They are programmed to allow remote monitoring and can act on any power fluctuations instantly. Provision for remote monitoring of critical components of the transformer, through in-built sensors, increases the reliability of continuous power distribution and reduces chances of transformer failure. Faults can be recognized remotely on a real-time basis and actions can be taken faster. The smart transformers are also more environment-friendly as they reduce energy consumption and thus reduce the emission of greenhouse gasses. They are also equipped to protect electrical equipment by safeguarding them from voltage fluctuations leading to increasing the longevity of that equipment. Bendable Lithium-ion Battery A new bendable lithium-ion battery that can flex and twist could power wearable devices and one day is used to develop a flexible smartphone, according to Panasonic, which is developing the new battery.Although it’s still in the early stages of development, the battery already has been tested to withstand twists, bends and other deformations while maintaining its ability to hold a charge, according to Panasonic. Triboelectric Nanogenerators (TENG) The TENGs rely on the triboelectric effect, by which certain materials become electrically charged when rubbed against another type of material. When the materials are in contact, electrons flow from one to the other, but when the materials are separated, the one receiving electrons will hold a charge. If these two materials are then connected by a circuit, a small current will flow to equalize the charges. By continuously repeating the process, an alternating electrical current can be produced to generate power. Research on Intelligent Transport Systems (ITS) covers a wide field. ITS comprise combinations of communication, computer and control technology developed and applied in transport to improve system performance, transport safety, efficiency, productivity, service, environment, energy and mobility. ITS can be applied to transport infrastructure, as well as to vehicles, such as cars, trucks and trains. These systems can be used in both passenger and freight transport to improve service quality and transport management. Solar nanotech-powered clothing A scientist who has developed filaments that harvest and store the sun’s energy — and can be woven into textiles. The breakthrough would essentially turn jackets and other clothing into wearable, solar-powered batteries that never need to be plugged in. A high-energy-density sugar biobattery based on a synthetic enzymatic pathway The rapidly growing demand for powering portable electronic devices is driving the development of better batteries with features such as enhanced energy-storage densities, high levels of safety, fast rechargeability, biodegradability and small environmental footprints1. The rechargeable lithium-ion battery is often the system of choice because it offers a high energy density, has a flexible and light-weight design and has a longer lifespan than comparable battery technologies. The energy-storage density of a typical lithium-ion battery is ~0.54 MJ kg−1 (that is, 150 Wh kg−1). The widespread use of metal-catalysed batteries also raises many concerns, primarily related to safety, toxic metal pollution and the availability of costly, limited, irreplaceable or rare metal resources. Enzymatic fuel cells (EFCs) are emerging electrobiochemical devices that directly convert chemical energy from a variety of fuels into electricity using low-cost biocatalyst enzymes. Inspired by living cells that can utilize complex organic compounds, for example, starch and glycogen) as stored energy sources, sugar-powered EFCs represent the next generation of biodegradable, highly safe biobatteries. Compared with microbial fuel cells, EFCs usually generate much higher power density in terms of mW cm−2. This feature highlights their great potential for powering a variety of portable electronic devices in the near future Pramil2018-10-29T09:47:37+05:30Seminar Topic| About the Author: Pramil How to turn your million dollar idea into a reality

{"pred_label": "__label__cc", "pred_label_prob": 0.5320702195167542, "wiki_prob": 0.46792978048324585, "source": "cc/2019-30/en_middle_0090.json.gz/line915468"}

high_school_physics

\section{Introduction} The blue compact dwarf emission-line galaxy I\,Zw\,18 is famous for being the most metal poor galaxy known so far. Its oxygen abundance is about 2\% the solar value, as first shown by Searle and Sargent (1972), and then confirmed by many studies (e.g. Lequeux et al. 1979, French 1980, Kinman \& Davidson 1981, Pagel et al. 1992, Legrand et al. 1997, Izotov et al. 1997b, V\'{\i}lchez \& Iglesias-P\'{a}ramo 1998). Because of this, I\,Zw\,18 has played an essential role in the determination of the primordial helium mass fraction. Also, due to its extreme properties, I\,Zw\,18 has been a choice target for studies of star formation history in blue compact galaxies (Dufour \& Hester 1990, Hunter \& Thronson 1995, Dufour et al. 1996, De Mello et al. 1998, Aloisi et al. 1999), of the elemental enrichment in dwarf galaxies (Kunth \& Sargent 1986, Kunth et al. 1995) and of the interplay between star formation and the interstellar medium (Martin 1996, van Zee et al. 1998). An important clue is the distribution of the oxygen abundance inside the H~{\sc ii}\ regions (Skillman \& Kennicutt 1993, V\'{\i}lchez \& Iglesias-P\'{a}ramo 1998) and in the neutral gas (Kunth et al. 1994, Pettini \& Lipman 1995, van Zee et al. 1998). Another clue is the carbon and nitrogen abundance (Garnett et al. 1997, V\'{\i}lchez \& Iglesias-P\'{a}ramo 1998, Izotov \& Thuan 1999). On the whole, there is a general consent about an intense and recent burst of star formation in I\,Zw\,18 - which provides the ionizing photons - following previous star formation episodes. How exactly has the gas been enriched with metals during the course of the evolution of I\,Zw\,18 remains to be better understood. Much of our understanding (or speculations) on the chemical evolution of I\,Zw\,18 (and other galaxies in general) relies on the confidence placed in the chemical abundances derived from the lines emitted in the H~{\sc ii}\ regions. These are generally obtained using standard, empirical, methods which have been worked out years ago, and rely on the theory of line emission in photoionized gases. Photoionization models are, most of the time, used merely as a guide to evaluate the temperature of the low excitation regions once the characteristic temperature of the high excitation zones has been obtained through the [O~{\sc iii}] $\lambda$4363/5007\ ratio. They also serve to provide formulae for estimating the correction factors for the unseen ionic species of a given element. Direct fitting of the observed emission line spectrum by tailored photoionization models provides more accurate abundances only if all the relevant line ratios are perfectly reproduced by the model (which is rarely the case in model fitting history) and if the model reproducing all the observational constraints is unique. One virtue of model fitting, though, is that it permits to check whether the assumptions used in abundance determinations are correct for a given object. For example, there is the long standing debate whether so-called ``electron temperature fluctuations'' (see e.g. Mathis 1995, Peimbert 1996, Stasi\'{n}ska 1998) are present in H~{\sc ii}\ regions to a sufficient level so as to significantly affect elemental abundance determinations. If a photoionization model is not able to reproduce all the temperature sensitive line ratios, the energy balance is not well understood, and one may question the validity of abundance determinations. Also, photoionization models are a potential tool (see e.g. Esteban et al. 1993, Garc\'{\i}a-Vargas 1996, Stasi\'nska \& Schaer\-er 1997, Crowther et al.\ 1999) to uncover the spectral distribution of the ionizing radiation field, thus providing information on the ionizing stars, their evolutionary status and the structure of their atmospheres. These two points are a strong motivation for a photoionization model analysis of I\,Zw\,18. There have already been a few such attempts in the past (Dufour et al. 1988, Campbell 1990, Stevenson et al. 1993). None of those models were, however, able to reproduce the He~{\sc ii} $\lambda$4686\ line, known to exist in I\,Zw\,18 since the work of French (1980). The reason is that, in those models, the spectral distribution of the ionizing radiation was that of a single star whose radiation field was interpolated from a grid of plane-parallel, LTE model atmospheres for massive stars. Recently, Wolf-Rayet stars have been identified in I\,Zw\,18 through the characteristic bump they produce at 4650~\AA\ (Izotov et al. 1997a, Legrand et al. 1997). Spherically expanding non-LTE model atmospheres for hot Wolf-Rayet stars with sufficiently low wind densities (Schmutz et al. 1992) do predict an output of radiation above the He~{\sc ii}\ ionization edge, which might, at least qualitatively, provide a natural explanation for the narrow He~{\sc ii} $\lambda$4686\ line observed in I\,Zw\,18. Schaerer (1996) has, for the first time, synthesized the broad (stellar) and narrow (nebular) He~{\sc ii}\ features in young starbursts using the Geneva stellar evolution tracks and appropriate stellar model atmospheres. He then extended his computations to the metallicity of I\,Zw\,18 (De Mello et al. 1998). In this paper, we use the emergent radiation field from the synthetic starburst model presented in De Mello et al. (1998) to construct photoionization models of I\,Zw\,18. One of the objectives is to see whether this more realistic ionizing radiation field permits, at the same time, to solve the electron temperature problem encountered in previous studies. Former photoionization models predicted too low a [O~{\sc iii}] $\lambda$4363/5007\ ratio, unless specific geometries were adopted (Dufour et al. 1988, Campbell 1990), which later turned out to be incompatible with Hubble Space Telescope ({\em HST}) images. The synthetic starburst model we use is based on spherically expanding non-LTE stellar atmosphere models for main sequence stars (Schaerer \& de Koter 1997) and for Wolf-Rayet stars (Schmutz et al.\ 1992). These models have a greater heating power than the LTE model atmospheres of same effective temperature (see Fig.\ 3; also Schaerer \& de Koter) The progression of the paper is as follows. In Section 2, we discuss in more detail the photoionization models proposed previously for I\,Zw\,18 and show in what respect they are not consistent with recent observations. In Section 3, we present our own model fitting methodology, including a description of the computational tools. In Section 4, we describe the models we have built for I\,Zw\,18, and discuss the effects of the assumptions involved in the computations. Our main results are summarized in Section 5. \section{Previous photoionization models of I\,Zw\,18} The first attempt to produce a photoionization model for I\,Zw\,18 is that of Dufour et al. (1988). Their observational constraints were provided by spectra obtained in an aperture of 2.5\arcsec $\times$ 6\arcsec of the NW region combined with IUE observations yielding essentially the C~{\sc iii}] $\lambda$1909\ line. Using Shields's photoionization code NEBULA, they modelled the object as an ionization bounded sphere of constant density $n$ = 100~cm$^{-3}$\ and adjustable volume filling factor $\epsilon$ so as to reproduce the observed [O~{\sc iii}] $\lambda$5007/[O~{\sc ii}] $\lambda$3727\ ratio. The ionizing radiation was provided by a central source of radiation, represented by the LTE model atmospheres of Hummer \& Mihalas (1970), modified to take into account the low metallicity of I\,Zw\,18. Discarding the He~{\sc ii}\ problem, they obtained a model that was reasonably successful except that it had an O$^{++}$\ temperature, T(O$^{++}$), marginally smaller than observed (17200~K) compared to the value of 18100 (+1100, -1000)~K derived directly from their observed [O~{\sc iii}] $\lambda$4363/5007\ = (2.79 $\pm$ 0.35) $\times$ 10$^{-2}$ (the errors quoted being 2$\sigma$) \footnote{A summary of various measurements of [O~{\sc iii}] $\lambda$4363/5007\ and corresponding electron temperatures is shown in Fig.\ 2.}. This model was obtained for an effective temperature of 45000~K. These authors showed that, because of the dominant role played by Ly$\alpha$ cooling in I\,Zw\,18, it was impossible, for the adopted geometry, to produce a model with noticeably higher T(O$^{++}$), by varying the free parameters at hand. Even increasing the effective temperature did not raise T(O$^{++}$) appreciably, because then the ionization parameter had to be lowered in order to maintain [O~{\sc iii}] $\lambda$5007/[O~{\sc ii}] $\lambda$3727\ at its observed level, and this resulted in a greater H$^{0}$\ abundance, thus enhancing Ly$\alpha$ excitation. Dufour et al. then proposed a composite model, in which the [O~{\sc iii}]\ line would be mainly produced around high temperature stars (T$_{eff}$\ $>$ 38000~K) and the [O~{\sc ii}]\ line would be mainly emitted around stars of lower T$_{eff}$\ ($<$ 37000~K). Alternatively, one could have, around a star of T$_{eff}$\ $<$ 45000~K, a high ionization component emitting most of the [O~{\sc iii}]\ and a low ionization component emitting most of the [O~{\sc ii}]. Since then, the {\em HST} images (Hunter \& Thronson 1995, Meurer et al. 1995, Dufour et al. 1996, De Mello et al. 1998) have revealed that the NW region appears like a shell of ionized gas about 5\arcsec\ in diameter, encircling a dense star cluster. Thus the geometries proposed by Dufour et al. (1988), although quite reasonable a priori, do not seem to apply to the case under study. Campbell (1990), using Ferland's photoionization code CLOUDY, constructed models to fit the spectral observations of Lequeux et al. (1979) obtained through a slit of 3\arcsec.8 $\times$ 12\arcsec.4. These observations were giving a [O~{\sc iii}] $\lambda$4363/5007\ ratio of (3.75 $\pm$ 0.35) $\times$ 10$^{-2}$. With a constant density, ionization bounded spherical model and a LTE Kurucz stellar atmosphere with metallicity 1/10 solar, in which the adjustable parameters were O/H, T$_{eff}$, $n$ and $\epsilon$, Campbell obtained a best fit model that had [O~{\sc iii}] $\lambda$4363/5007\ = 3.07~10$^{-2}$, i.e. much lower than the value she aimed at reproducing. She then proposed a density gradient model, in which the inner regions had a density $n$ $>$~10$^{5}$~cm$^{-3}$, so as to induce collisional deexcitation of [O~{\sc iii}] $\lambda$5007. Applying standard abundance derivation techniques to this model yields an oxygen abundance higher by 70\% than the input value. This led Campbell to conclude that I~Zw18 was not as oxygen poor as previously thought. The density gradient model of Campbell (1990) can be checked directly using the density sensitive [Ar~{\sc iv}] $\lambda$4741/4713\ ratio. The only observations giving this line ratio are those of Legrand et al. (1997), and they indicate a density in the Ar$^{+++}$\ region lower than 100~cm$^{-3}$ . Direct images with the {\em HST} do not support Campbell's density gradient model either, since, as stated above, the appearance of the H~{\sc ii}\ region is that of a shell surrounding the excitation stars. Stevenson et al. (1993), using a more recent version of CLOUDY, constructed a spherical, ionization bounded constant density photoionization model to fit their own data. They used as an input an extrapolation of the Kurucz LTE model atmospheres. Their modelling procedure was very similar to that of Campbell (1990) for her constant density model. Their best fit model had O/H = 1.90~10$^{-5}$ and returned [O~{\sc iii}] $\lambda$4363/5007\ = 2.79~10$^{-2}$, to be compared to their observed value of (3.21 $\pm$ 0.42) $\times$ 10$^{-2}$. What complicates the discussion of the three studies above is that they use different codes with probably different atomic data, and they aim at fitting different sets of observations. Nevertheless, it is clear that all those models have difficulties in reproducing the high [O~{\sc iii}] $\lambda$4363/5007\ observed. They have other weak points, as noted by their authors. For example, Dufour et al. (1988) and Stevenson et al. (1993) comment on the unsatisfactory fitting of the sulfur lines. However, the atomic data concerning sulfur are far less well established than those concerning oxygen, therefore the discrepancies are not necessarily meaningful. Besides, it is not surprising that, with a simple density structure, one does not reproduce perfectly at the same time the [O~{\sc iii}]/[O~{\sc ii}]\ and [S~{\sc iii}]/[S~{\sc ii}]\ ratios. The most important defect shared by the three models just discussed is that they predict no He~{\sc ii} $\lambda$4686\ emission. This is simply due to the fact that they used an inadequate input stellar radiation field. With the presently available stellar population synthesis models for the exciting stars of giant H~{\sc ii}\ regions which make use of more realistic model atmospheres (Schaerer \& Vacca 1998), and especially models that are relevant for the Wolf-Rayet stages of massive stars, it is interesting to reanalyze the problem. Using simple photon counting arguments, De Mello at al. (1998) have already shown that a starburst with a Salpeter initial mass function and an upper mass limit of 150~M$_{\odot}$\ could reproduce the equivalent width of the Wolf-Rayet features and of the narrow He~{\sc ii} $\lambda$4686\ emission line in I\,Zw\,18. It is therefore interesting, using the emergent radiation field from such a synthetic stellar population, to see whether one can better reproduce the [O~{\sc iii}] $\lambda$4363/5007\ ratio observed in I\,Zw\,18, with a model that is more compatible with the density structure constrained by the {\em HST} images. \section{Our model fitting methodology } \subsection{Computational tools and input parameters} As in the previous studies, we concentrate on the so-called NW component, seen in the top of Fig.\ 1, which shows the WFPC2 H$\alpha$\ image of the {\em HST} (cf.\ Fig.\ 1 of De Mello et al.\ 1998). Throughout the paper, we adopt a distance to I~Zw~18 of 10 Mpc, assuming $H_{o}$ = 75 km~s$^{-1}$~Mpc$^{-1}$, as in many studies (Hunter \& Thronson 1995, Martin 1996, van Zee et al. 1998) \footnote{Izotov et al. (1999) have submitted a paper suggesting a distance of 20 Mpc to I Zw 18. Should this be the case, the conclusions of our paper that are linked to the ionization structure and the temperature of the nebula would hardly be changed. The total mass of the ionized gas would be larger, roughly by a factor 2$^{3}$.}. \subsubsection{The stellar population} We use the same model for the stellar population as described in De Mello et al. (1998). It is provided by a evolutionary population synthesis code using stellar tracks computed with the Geneva code at the appropriate metallicity (1/50~Z$_{\odot}$). The stellar atmospheres used are spherically expanding non-LTE models for WR stars (Schmutz et al.\ 1992) and O stars ({\em CoStar} models at $Z=0.004$, Schaerer \& de Koter 1997), and Kurucz models at [Fe/H]=-1.5 for the remainder. More details can be found in De Mello et al. (1998) and Schaerer \& Vacca (1998). We assume an instantaneous burst of star formation, with an upper mass limit of 150~M$_{\odot}$\ and a lower mass limit of 0.8 ~M$_{\odot}$. Since all observational quantities considered here depend only on the properties of massive stars, the choice for $M_{\rm low}$ has no influence for the results of this paper. It merely serves as an absolute normalisation. The total initial mass of the stars is adjusted in such a way that, at a distance of 10~Mpc, the flux at 3327~\AA\ is equal to 1.7~10$^{-15}$ ~erg~s$^{-1}$~cm$^{-2}$~\AA$^{-1}$, the value measured in the flux calibrated WFPC2 F336W image of De Mello et al.\ (1998) within a circle of 2.5\arcsec\ radius centered on the NW region (see Fig.\ 1). This flux is dominated by the latest generation of stars in I\,Zw\,18, so that our normalization is hardly sensitive to the previous star formation history in the NW region of I\,Zw\,18. It yields a total stellar mass of 8.7~10$^{4}$~M$_{\odot}$, at a distance of 10~Mpc. Actually, most of the flux comes from a region much smaller in size, and our photoionization modelling is made with the ionizing cluster located at the center of the nebula and assuming that its spatial extension is negligible. We consider that the observed ultraviolet flux is only negligibly affected by extinction \footnote{A direct fitting of the ultraviolet stellar continuum by population synthesis models, of which we became aware after the paper had been submitted, yields C(\ifmmode {\rm H}\beta \else H$\beta$\fi) $<$ 0.06 (Mas-Hesse \& Kunth 1999).}. For an extinction C(\ifmmode {\rm H}\beta \else H$\beta$\fi) of 0.04, such as estimated by Izotov \& Thuan (1998), the corrected flux would be only about 10\% larger, which is insignificant in our problem. Other observers give values of C(\ifmmode {\rm H}\beta \else H$\beta$\fi) ranging between 0. and 0.2. If C(\ifmmode {\rm H}\beta \else H$\beta$\fi) were as large as 0.20, as estimated by V\'{\i}lchez \& Iglesias-P\'{a}ramo (1998) and some other observers, the true stellar flux would be a factor two higher. However, all the determinations of C(\ifmmode {\rm H}\beta \else H$\beta$\fi), except the one by Izotov \& Thuan (1998), do not take into account the underlying stellar absorption at \ifmmode {\rm H}\beta \else H$\beta$\fi\, and therefore overestimate the reddening. A further cause of overestimation of C(\ifmmode {\rm H}\beta \else H$\beta$\fi), which applies also to the work of Izotov \& Thuan (1998), is that the intrinsic H$\alpha$/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio is assumed to be the case B recombination value, while collisional excitation of H$\alpha$\ is not negligible in the case of I\,Zw\,18 as noted by Davidson \& Kinman (1985). We will come back to this below. \subsubsection{The nebula} The photoionization computations are performed with the code PHOTO using the atomic data listed in Stasi\'{n}ska \& Leitherer (1996). The code assumes spherical geometry, with a central ionizing source. The diffuse radiation is treated assuming that all the photons are emitted outwards in a solid angle of 2$\pi$, and the transfer of the resonant photons of hydrogen and helium is computed with the same outward only approximation, but multiplying the photo-absorption cross-section by an appropriate factor to account for the increased path length due to scattering (Adams 1975). The nebular abundances used in the computations are those we derived from the spectra of Izotov and Thuan (1998) for the NW component of I\,Zw\,18, with the same atomic data as used in the photoionization code. For helium, however, we adopted the abundance derived by Izotov \& Thuan (1998) for the SE component, as stellar absorption contaminates the neutral helium lines in the NW component. The nominal value of the temperature derived from [O~{\sc iii}] $\lambda$4363/5007\ is 19800~K. This value was used to compute the ionic abundances of all the ions except O$^{+}$, N$^{+}$\ and S$^{+}$, for which a value of 15000~K was adopted (this is the typical value returned by our photoionization models for I\,Zw\,18). The electron density deduced from [S~{\sc ii}] $\lambda$6731/6717\ is 140~cm$^{-3}$, and this density was adopted in the computation of the ionic abundances of all species. The ionization correction factors to compute the total element abundances were those of Kingsburgh \& Barlow (1994), which are based on photoionization models of planetary nebulae and are also suitable for H~{\sc ii}\ regions. They give slightly smaller oxygen abundances (by a few \%) than the traditional ionization correction factors which assume that the O$^{+++}$\ region is coextensive with the He$^{++}$\ (we did not iterate on the ionization correction factors after our photoionization model analysis since this would have not changed any of the conclusions drawn in this paper). The carbon abundance used in the computations follows from the C/O ratio derived by Garnett et al. (1997) from {\em HST} observations of I\,Zw\,18. The abundances of the elements not constrained by the observations (Mg, Si) and (Cl, Fe) have been fixed to 10$^{-7}$ and 10$^{-8}$ respectively. Table 1 presents the abundance set used in all the computations presented in the paper. As already noted by previous authors, at the metallicity of I\,Zw\,18, the heavy elements (i.e. all the elements except hydrogen and helium) play a secondary role in the thermal balance. Their role in the absorption of ionizing photons is completely negligible. Any change of abundances, even that of helium, compatible with the observed intensities of the strong lines, will result in a very small change in the electron temperature, and we have checked that the effect they will induce in the predicted spectra are small compared to the effects discussed below. \begin{table} \caption{Input abundances relative to hydrogen for the models of I\,Zw\,18 (by number).} \begin{flushleft} \begin{tabular}{rr} \hline He & 7.60 10$^{-2}$\\ C & 3.03 10$^{-6}$ \\ N & 3.89 10$^{-7}$ \\ O & 1.32 10$^{-5}$ \\ Ne & 2.28 10$^{-6}$ \\ S & 3.72 10$^{-7}$ \\ Ar & 9.13 10$^{-8}$ \\ \hline \end{tabular} \end{flushleft} \end{table} We do not include dust in the computations. While it is known that, in general, dust mixed with the ionized gas may absorb some of the ionizing photons, and contribute to the energy balance of the gas by photoelectric heating and collisional cooling (e.g. Baldwin et al. 1991, Borkowski \& Harrington 1991, Shields \& Kennicutt 1995), the expected effect in I\,Zw\,18 is negligible, since the dust-to-gas ratio is believed to be small at such metallicities (cf.\ Lisenfeld \& Ferrara 1998). The case of I\,Zw\,18 is thus very interesting for photoionization modelling, since due to the very low metallicity of this object, the number of unconstrained relevant parameters is minimal. \subsection{Fitting the observational constraints} In judging the value of our photoionization models, we do not follow the common procedure of producing a table of intensities relative to \ifmmode {\rm H}\beta \else H$\beta$\fi\ to be compared to the observations. A good photoionization model is not only one which reproduces the observed line ratios within the uncertainties. It must also satisfy other criteria, like being compatible with what is known from the distribution of the ionized gas, and what is known of the ionizing stars themselves. On the other hand, many line ratios are not at all indicative of the quality of a photoionization model. For example, obviously, two lines arising from the same atomic level like [O~{\sc iii}] $\lambda$5007\ and [O~{\sc iii}] $\lambda$4959 have intensity ratios that depend only on the respective transition probabilities. In H~{\sc ii}\ regions, the ratio of hydrogen Balmer lines (if case B applies) is little dependent on the physical conditions in the ionized gas, and this is why it can be used to determine the reddening. The ratios of the intensities of neutral helium lines do depend somewhat on the electron density distribution and on selective absorption by dust of pseudo-resonant photons, (Clegg \& Harrington 1989, Kingdon \& Ferland 1995), and these are introduced in photoionization models. In the case of the NW component of I\,Zw\,18, the observed neutral helium lines are affected by absorption from stars or interstellar sodium (Izotov \& Thuan 1998, V\'{\i}lchez \& Iglesias-P\'{a}ramo 1998), and cannot be easily used as constraints for photoionization models. Generally speaking, once line ratios indicative of the electron temperature (like [O~{\sc iii}] $\lambda$4363/5007, [N~{\sc ii}] $\lambda$5755/6584), of the electron density (like [S~{\sc ii}] $\lambda$6731/6717, [Ar~{\sc iv}] $\lambda$4741/4713) and of the global ionization structure (like [O~{\sc iii}] $\lambda$5007/[O~{\sc ii}] $\lambda$3727\ or [S~{\sc iii}] $\lambda$9532/[S~{\sc ii}] $\lambda$6725) have been fitted, the ratios of all the strong lines with respect to \ifmmode {\rm H}\beta \else H$\beta$\fi\ are necessarily reproduced by a photoionization model whose input abundances were obtained from the observations. The only condition is that the atomic data to derive the abundances and to compute the models should be the same. Problems may arise only if the empirical ionization factors are different from the ones given by the model, or if there is insufficient information on the distribution of the electron temperature or density inside the nebula (in the case of I\,Zw\,18 no direct information is available on the temperature in the low ionization zone, but we adopted a value inspired by the models). Therefore, intensity ratios such as [O~{\sc iii}] $\lambda$5007/\ifmmode {\rm H}\beta \else H$\beta$\fi, [Ne~{\sc iii}] $\lambda$3869/\ifmmode {\rm H}\beta \else H$\beta$\fi, [N~{\sc ii}] $\lambda$6584/\ifmmode {\rm H}\beta \else H$\beta$\fi, [Ar~{\sc iii}] $\lambda$7135/\ifmmode {\rm H}\beta \else H$\beta$\fi\ or C~{\sc iii}] $\lambda$1909/\ifmmode {\rm H}\beta \else H$\beta$\fi\ are not a measure of the quality of the photoionization model. To judge whether a photoionization model is acceptable, one must work with outputs that are significantly affected by the physical processes on which the photoionization model is based, i.e. the transfer of the ionizing radiation, the processes determining the ionization equilibrium of the various atomic species and the thermal balance of the gas. Table 2 lists the quantities that can be used in the case of I\,Zw\,18, given the observational information we have on the object. The value of the \ifmmode {\rm H}\beta \else H$\beta$\fi\ flux is derived from the H$\alpha$\ flux measured in a circle of radius $\theta$=2.5\arcsec\ (shown in Fig. 1), assuming C(\ifmmode {\rm H}\beta \else H$\beta$\fi) = 0. The line ratios He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi, [O~{\sc iii}] $\lambda$4363/5007, [S~{\sc ii}] $\lambda$6731/6717, [O~{\sc iii}] $\lambda$5007/[O~{\sc ii}] $\lambda$3727, [S~{\sc iii}] $\lambda$6312/[S~{\sc ii}] $\lambda$6725\ and [O~{\sc i}] $\lambda$6300/\ifmmode {\rm H}\beta \else H$\beta$\fi\ are the values observed by Izotov \& Thuan (1998) in the rectangular aperture whose approximate position is shown in Fig. 1. It is important to define in advance the tolerance we accept for the difference between our model predictions and the observations. This must take into account both the uncertainty in the observational data, the fact that the spectra were taken through an aperture not encompassing the whole nebula, and the fact that the nebula does not have a perfect, spherical symmetric structure. This latter aspect is, of course, difficult to quantify, and the numbers given in Column 3 of Table 2 are to be regarded rather as guidelines. In Column 4, we indicate which is the dominant factor determining the adopted tolerance : the signal-to-noise, or the geometry. For example, such ratios as He~{\sc ii}/\ifmmode {\rm H}\beta \else H$\beta$\fi, [O~{\sc iii}]/[O~{\sc ii}], [S~{\sc iii}]/[S~{\sc ii}]\ or [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ are obviously more dependent on geometrical effects than [O~{\sc iii}] $\lambda$4363/5007. Note that, even for that ratio, the tolerance given in Table 2 is larger than the uncertainty quoted by Izotov \& Thuan (1998). The reason is that the many observations of the NW component of I\,Zw\,18 made over the years, with different telescopes, detectors, and apertures, yield distinct values for this ratio, as shown in Fig. 2. In view of this figure, a tolerance of 10\% with respect with the [O~{\sc iii}] $\lambda$4363/5007\ ratio measured by Izotov \& Thuan (1998) seems reasonable. The status of the [S~{\sc ii}] $\lambda$6731/6717\ ratio is somewhat different. It indicates the average electron density in the zone emitting [S~{\sc ii}]. This is very close to an input parameter, since photoionization models are built with a given density structure. However, because the density deduced from [S~{\sc ii}] $\lambda$6731/6717\ is not the total hydrogen density but the electron density in the region emitting [S~{\sc ii}], and because the density is not necessarily uniform, it is important to check that the model returns an [S~{\sc ii}] $\lambda$6731/6717\ value that is compatible with the observations. For the total \ifmmode {\rm H}\beta \else H$\beta$\fi\ flux, we accept models giving F(\ifmmode {\rm H}\beta \else H$\beta$\fi) larger than the observed value, on account of the fact that the coverage factor of the ionizing source by the nebula may be smaller than one. \setcounter{table}{1} \begin{table*}[htb] \caption{Observables that a photoionization model of I\,Zw\,18 should fit (references in Sect.\ 3.2). } \begin{flushleft} \begin{tabular}{llllll} \hline Quantity & Value & Tolerance & Major source of uncertainty & Symbol in Figs.\ 4--6 \\ \noalign{\smallskip} \hline \noalign{\smallskip} F(\ifmmode {\rm H}\beta \else H$\beta$\fi) [erg~cm$^{-2}$~s$^{-1}$] & 4.0~10$^{-14}$ & + 0.5 dex & geometry (see text) & circle\\ angular radius $\theta$ [arc sec ] & 2.5 & $\pm$ 0.08 dex & geometry & cross \\ He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi & 0.034 & $\pm$ 0.2 dex & geometry (see text) & square \\ [O~{\sc iii}] $\lambda$4363/5007\ & 3.28~10$^{-2}$ & $\pm$ 0.04 dex & S/N & open triangle\\ [S~{\sc ii}] $\lambda$6731/6717\ & 1.3 & $\pm$ 0.04 dex & S/N & diamond \\ [O~{\sc iii}] $\lambda$5007/[O~{\sc ii}] $\lambda$3727\ & 6.82 & $\pm$ 0.1 dex & geometry & filled triangle \\ [O~{\sc i}] $\lambda$6300/\ifmmode {\rm H}\beta \else H$\beta$\fi\ & 0.007 & $\pm$ 0.3 dex & geometry & plus \\ [S~{\sc iii}] $\lambda$6312/[S~{\sc ii}] $\lambda$6725\ & 0.173 & $\pm$ 0.2 dex & geometry & asterisk \\ \noalign{\smallskip} \hline \end{tabular} \end{flushleft} \end{table*} Thus, in the following, we compute photoionization models with the input parameters as defined in Section 3.1, and see how they compare with the constraints specified in Table 2. We will not examine the effects of varying the elemental abundances, since, as mentioned above, they are negligible in our problem. Uncertainties in the measured stellar flux have only a small impact on our models, and are therefore not discussed here. Similarly, we discard the effects of an error in the distance $d$ to I~Zw~18. These are not crucial on the ionization structure of a model designed to fit the observed flux at 3327\AA, since the mean ionization parameter varies roughly like $d^{2/3}$. What we mainly want to see is whether, with our present knowledge, we can satisfactorily explain the observed properties of I\,Zw\,18. As will be seen, the gas density distribution plays an important role. \section{Climbing the ladder of sophistication} \subsection{The ionizing radiation field } Before turning to proper photoionization modelling, it is worthwhile examining the gross properties of the ionizing radiation field of the synthetic stellar population model we are using, and compare it to single star model atmospheres. Two quantities are particularly relevant. One is $Q({\rm{He^{+}}})$/$Q({\rm{H^{0}}})$, the ratio of the number of photons above 54.4 and 13.6~eV emitted by the ionizing source. This ratio allows one to estimate the He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi\ that would be observed in a surrounding nebula, by using simple conservation arguments leading to the formula: He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi\ = 2.14 $Q({\rm{He^{+}}})$/$Q({\rm{H^{0}}})$\ (taking the case B recombination coefficients given in Osterbrock 1989). As is known, this expression is valid if the nebula is ionization bounded and the observations pertain to the whole volume. It is less commonly realized that it also assumes the average temperature in the He$^{++}$\ region to be the same as in the entire H~{\sc ii}\ region. This is may be far from true, as will be shown below, so a correction should account for that. Another assumption is that all the photons above 54.4~eV are absorbed by He$^{+}$\ ions. This is not what happens in objects with a low ionization parameter. There, the residual neutral hydrogen particles are sufficiently numerous to compete with He$^{+}$. In such a case, the expression above gives an upper limit to the nebular He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio. In spite of these difficulties, $Q({\rm{He^{+}}})$/$Q({\rm{H^{0}}})$\ remains a useful quantity to estimate the intensity of the nebular He~{\sc ii} $\lambda$4686\ line. Fig. 3a shows the variation of $Q({\rm{He^{+}}})$/$Q({\rm{H^{0}}})$\ as a function of starburst age for the synthetic model population we are considering. As already stated in De Mello et al. (1998), the strength of the He~{\sc ii}\ nebular line in I\,Zw\,18 indicates a starburst between 2.9 and 3.2~Myr. Another important ratio is $Q({\rm{He^{0}}})$/$Q({\rm{H^{0}}})$, sometimes referred to as the hardness parameter of the ionizing radiation field. It provides a qualitative measure of the heating power of the stars. We have represented this quantity in Fig. 3c. We see that, as the starburst ages, its heating power gradually declines, and shows only a very mild bump at ages around 3~Myr, where the Wolf-Rayet stars are present. As we will show below this modest increase of the heating power is not sufficient to explain the high electron temperature observed in I\,Zw\,18. For comparison, we show in Figs.\ 3b and d respectively, the values of $Q({\rm{He^{+}}})$/$Q({\rm{H^{0}}})$\ and $Q({\rm{He^{0}}})$/$Q({\rm{H^{0}}})$\ as a function of the stellar effective temperature for the LTE model atmospheres of Kurucz (1991), the {\em CoStar} model atmospheres corresponding to main sequence stars (Schaerer \& de Koter 1997) and for the model atmospheres for Wolf-Rayet stars of Schmutz et al.\ (1992). The {\em CoStar} models show an increased He$^{+}$\ ionizing flux compared to Kurucz models which have a negligible flux even for very low metallicities ([Fe/H]$=-1.5$). The reasons for this difference have been discussed in Schaerer \& de Koter (1997). In addition to the $T_{\rm eff}$\ dependence, $Q({\rm{He^{+}}})$\ from WR models depend strongly on the wind density. He$^{+}$\ ionizing photons are only predicted by models with sufficiently thin winds (cf.\ Schmutz et al.\ 1992). Figure 3d shows the increase of the hardness of the radiation field, at a given $T_{\rm eff}$, between the spherically expanding non-LTE models for O and WR stars and the traditional Kurucz models (see discussion in Schaerer \& de Koter 1997). This provides a greater heating power which, as will be shown later, is however still insufficient to explain the observations. \subsection{I\,Zw\,18 as a uniform sphere} We start with the easiest and most commonly used geometry in photoionization modelling: a sphere uniformly filled with gas at constant density, occupying a fraction $\epsilon$ of the whole nebular volume. The free parameters of the models are then only the age of the starburst, the gas density and the filling factor. Each model is computed starting from the center, and the computations are stopped either when the [O~{\sc iii}]/[O~{\sc ii}]\ ratio has reached the observed value given in Table 2, or when the gas becomes neutral. In other words, we examine also models that are not ionization bounded, in contrast to previous studies. Figure 4 shows our diagnostic diagram for a series of models having a density $n$= 100~cm$^{-3}$\ and a filling factor $\epsilon$=0.01. The left panel shows the computed values of log F(\ifmmode {\rm H}\beta \else H$\beta$\fi) + 15 (open circles), log He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi\ + 2 (squares), angular radius $\theta$ (crosses), [O~{\sc iii}] $\lambda$4363/5007\ $\times$ 100 (open triangles), [S~{\sc ii}] $\lambda$6731/6717\ (diamonds), log ([O~{\sc iii}]/[O~{\sc ii}]) (black triangles), log ([S~{\sc iii}]/[S~{\sc ii}]) (asterisks) and log [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ +3 (plus) as a function of the starburst age. The black circles correspond to the value of log F(\ifmmode {\rm H}\beta \else H$\beta$\fi) + 15 that the nebula would have if it were ionization bounded. Thus, by comparing the positions of an open circle and a black circle, at a given abscissa, one can immediately see whether the model is density bounded and how much diffuse H$\alpha$\ or \ifmmode {\rm H}\beta \else H$\beta$\fi\ emission is expected to be emitted outside the main body of the nebula. In the right panel, the observed values are represented on the same vertical scale and with the same symbols as the model predictions. The tolerances listed in Table 2 are represented as vertical error bars (the horizontal displacement of the symbols has no particular meaning). We readily see that the age of the starburst is important only for the He~{\sc ii} $\lambda$4686\ line, the other quantities varying very little for ages 2.7--3.4 Myr. Therefore, for the following runs of models, we adopt an age of 3.1~Myr. In principle, one can always adjust the age for the model to reproduce the observed He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio exactly. Figure 5 shows the same sort of diagnostic diagram as Fig. 4 for a series of models with $n$ = 100~cm$^{-3}$\ and varying filling factor. For a filling factor around 0.1 or larger, with the adopted electron density, the model is ionization bounded, and its [O~{\sc iii}]/[O~{\sc ii}]\ is larger than observed. For filling factors smaller than that, the gas distribution is more extended, so that the general ionization level drops. The observed [O~{\sc iii}]/[O~{\sc ii}]\ can then only be reproduced for a density bounded model. In such a case, the \ifmmode {\rm H}\beta \else H$\beta$\fi\ radiation actually produced by the nebula is smaller than if all the ionizing photons were absorbed in the nebula. A filling factor of 0.002 -- 0.05 gives values of [O~{\sc iii}]/[O~{\sc ii}], F(\ifmmode {\rm H}\beta \else H$\beta$\fi) and $\theta$ in agreement with the observations. But such models give [O~{\sc iii}] $\lambda$4363/5007\ too small compared with the observations, and [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ below the observed value by nearly two orders of magnitude. It is interesting, though, to understand the qualitative behavior of these line ratios as $\epsilon$ decreases. [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ decreases because the model becomes more and more density bounded in order to reproduce the observed [O~{\sc iii}]/[O~{\sc ii}]\, and levels off at $\epsilon$ = 0.1, because the ionization parameter of the model is then so small that the [O~{\sc i}]\ is gradually emitted by residual neutral oxygen in the main body of the nebula and not in the outskirts. [O~{\sc iii}] $\lambda$4363/5007\ decreases as $\epsilon$ decreases, because of the increasing proportion of L$\alpha$ cooling as the ionization parameter drops. One can build other series of models with different values of $n$ that are still compatible with the observed [S~{\sc ii}] $\lambda$6731/6717. Qualitatively, their behavior is the same and no acceptable solution is found. Interestingly, Fig. 5 shows that models with $\epsilon$ $\ge$ 0.1 have [O~{\sc iii}] $\lambda$4363/5007\ marginally compatible with the observations ( [O~{\sc iii}] $\lambda$4363/5007\ = 3.03 10$^{-2}$ for $\epsilon$ = 0.1), but such models have too large [O~{\sc iii}]/[O~{\sc ii}] ($>$ 10 compared of the observed value 6.8) and too small angular radius ($<$ 1.6\arcsec\ instead of the observed value 2.5\arcsec). Note, by the way, that such models, being optically thick, return a rather large [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi, actually close to the observed value, and a [S~{\sc iii}]/[S~{\sc ii}]\ compatible with the observations. However, we do not use [S~{\sc iii}]/[S~{\sc ii}]\ as a primary criterion to judge the validity of a model, since experience with photoionization modelling of planetary nebulae shows that it is difficult to reproduce at the same time the sulfur and the oxygen ionization structure of a given object, and, in principle, one expects the atomic data for oxygen to be more reliable than those for sulfur. The strongest argument against models with $\epsilon > 0.1$ is their angular size, which is definitely too small compared with the observations. This remains true even when considering a reasonable error on the distance since, with the condition that we impose on the flux at 3327~\AA\ to be preserved, the angular radius of a model goes roughly like $d^{1/3}$. This illustrates the importance of taking into account other parameters in addition to line ratios to accept or reject a photoionization model. \subsection{I~ Zw~18 as a spherical shell } The series of models presented above had mainly a pedagogical value, but they are obviously incompatible with the observed morphology in H$\alpha$. The next step is to consider a model consisting of a hollow, spherical shell of constant density, similar to the one constructed by Garc\'{\i}a-Vargas et al. (1997) for NGC\,7714 for example. In such a case, there is an additional free parameter, $R_{in}$, the radius of the inner boundary of the shell. It is fixed, more or less, by the appearance of the H$\alpha$\ image. Figure 6 shows a diagnostic diagram for a series of models with $R_{in}$ = 2.25 10$^{20}$~cm (corresponding to an angular radius of 1.5\arcsec), $n$ = 100~cm$^{-3}$\, and varying $\epsilon$. The qualitative behavior is similar to that seen for the uniform sphere models presented in Fig. 1, but the [O~{\sc iii}] $\lambda$4363/5007\ ratio is now even lower (it never exceeds 2.5~10$^{-2}$ in this series). This is because of the enhanced role of L$\alpha$ cooling, which is strong in all the parts of the nebula, while for the full sphere model, in the zone close to the star, the ionization parameter is very high and consequently the population of neutral hydrogen very small. Apart from the [O~{\sc iii}] $\lambda$4363/5007\ problem, models with $\epsilon$ = 0.002 -- 0.05 are satisfactory as concerns the main diagnostics ([O~{\sc iii}]/[O~{\sc ii}], F(\ifmmode {\rm H}\beta \else H$\beta$\fi) and $\theta$). The models become progressively density bounded towards smaller values of $\epsilon$ $<$ 0.02, meaning that there is a leakage of ionizing photons. From Fig. 6, one sees that these photons are enough to produce an H$\alpha$\ emission in an extended diffuse envelope that is at least comparable in strength to the total emission from the dense shell. This is in agreement with Dufour \& Hester's (1990) ground-based observation of extended H$\alpha$\ emission surrounding the main body of star formation. \subsection{Other geometries} Closer inspection of the {\em HST} H$\alpha$\ image shows that the gas emission is incompatible with a spherical bubble. This is illustrated in Fig.\ 7, where the observed cumulative surface brightness profile within radius $r$ (dashed line) is compared to the expected profiles for constant density spherical shells of various inner radii (solid lines). The theoretical profiles are obtained assuming that the temperature is uniform in the gas, but taking into account a reasonable temperature gradient in the model hardly changes the picture. Clearly, the observed profile is not compatible with a spherical density distribution. The column density of emitting matter in the central zone of the image is too small. One must either have an incomplete shell with some matter stripped off from the poles, or even a more extreme morphology like an diffuse axisymmetric body with a dense ringlike equator seen face on. Such geometries are actually common among planetary nebulae (Corradi \& Schwartz 1995) and nebulae surrounding luminous blue variables (Nota et al. 1995), being probably the result of the interaction of an aspherical stellar wind from the central stars (Mellema 1995, Frank et al. 1998) and are also suggested to exist in superbubbles and supergiant shells and to give rise to the blow-out phenomenon (Mac Low et al. 1989, Wang \& Helfand 1991, Tenorio-Tagle et al. 1997, Oey \& Smedley 1998, Martin 1998). Does the consideration of such a geometry help in solving the [O~{\sc iii}] $\lambda$4363/5007\ problem? In the case of a spherical shell with some matter lacking at the poles, our computations overestimate the role of the diffuse ionizing radiation, which is supposed to come from a complete shell. Since the heating power of the diffuse ionizing radiation is smaller than that of the stellar radiation, one may be underestimating the electron temperature. As a test, we have run a model switching off completely the diffuse radiation, so as to maximize the electron temperature. The effect was to increase the O$^{++}$\ temperature by only 200~K. In the case of a ring with some diffuse matter seen in projection inside the ring, the gas lying close to the stars would be at a higher electron temperature than the matter of the ring, and one expects that an adequate combination of the parameters describing the gas distribution interior to and inside of the ring might reproduce the [O~{\sc iii}] $\lambda$4363/5007\ measured in the aperture shown in Fig. 1. However, we notice that the region of high [O~{\sc iii}] $\lambda$4363/5007\ is much larger than that. It extends over almost 20\arcsec\ ( V\'{\i}lchez \& Iglesias-P\'{a}ramo 1998) and [O~{\sc iii}] $\lambda$4363/5007\ is neither correlated with [O~{\sc iii}]/[O~{\sc ii}]\ nor with the H$\alpha$\ surface brightness. Therefore, one cannot explain the high [O~{\sc iii}] $\lambda$4363\ by the emitting matter being close to the ionizing stars. In passing we note that the observations of V\'{\i}lchez \& Iglesias-P\'{a}ramo (1998) show the nebular He~{\sc ii} $\lambda$4686\ emission to be extended as well (although not as much as [O~{\sc iii}] $\lambda$4363). If this emission is due to photoionization by the central star cluster, as modelled in this paper, this means that the H$\alpha$\ ring is porous, since He~{\sc ii} $\lambda$4686\ emission necessarily comes from a region separated from the stars by only a small amount of intervening matter. In summary, we must conclude that, if we take into account all the available information on the structure of I\,Zw\,18, we are not able to explain the high observed [O~{\sc iii}] $\lambda$4363/5007\ ratio with our photoionization models. In our best models, [O~{\sc iii}] $\lambda$4363/5007\ is below the nominal value of Izotov \& Thuan (1998) by 25 - 35\%. \subsection{Back to the model assumptions} Can our lack of success be attributed to an improper description of the stellar radiation field? After all, we know little about the validity of stellar model atmospheres in the Lyman continuum (see discussion in Schaerer 1998). Direct measurements of the EUV flux of early B stars revealed an important EUV excess (up to $\sim$ 1.5 dex) with respect to plane parallel model atmospheres (Cassinelli et al.\ 1995, 1996), whose origin has been discussed by Najarro et al.\ (1996), Schaerer \& de Koter (1997) and Aufdenberg et al.\ (1998). For O stars a similar excess in the {\em total} Lyman continuum output is, however, excluded from considerations of their bolometric luminosity and measurements of H~{\sc ii}\ region luminosities (Oey \& Kennicutt 1997, Schaerer 1998). The hardness of the radiation field, which is crucial for the heating of the nebula, is more difficult to test. Some constraints can be obtained by comparing the line emission of nebulae surrounding hot stars with the results of photoionization models (Esteban et al.\ 1993, Pe\~{n}a et al.\ 1998, Crowther et al.\ 1999), but this is a difficult task, considering that the nebular emission depends also on its geometry. Although the hardness predicted by the {\em CoStar} O stars models permits to build grids of photoionization models that seem to explain the observations of Galactic and LMC H~{\sc ii}\ regions (Stasi\'nska \& Schaerer 1997), the constraints are not sufficient to prove or disprove the models. To check the effect of a harder radiation field, we have run a series of models where the radiation field above 24.6~eV was arbitrarily multiplied by a factor 3 (raising the value of $Q({\rm{He^{0}}})$/$Q({\rm{H^{0}}})$\ from 0.33 to 0.59, corresponding to $T_{\rm eff}$\ from $\ga$ 40000 K to $\sim$ 100\,000~K or a blackbody of the same temperature (cf.\ Fig.\ 3d). This drastic hardening of the radiation field resulted in an increase of [O~{\sc iii}] $\lambda$4363/5007\ of only 10\%. It is only by assuming a blackbody radiation of 300\,000~K (which has $Q({\rm{He^{0}}})$/$Q({\rm{H^{0}}})$\ = 0.9) that one approaches the observed [O~{\sc iii}] $\lambda$4363/5007. A model similar to those presented in Fig. 6 but with such a radiation field gives [O~{\sc iii}] $\lambda$4363/5007\ =3.15 10$^{-2}$. But is has a He~{\sc ii} $\lambda$4686/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio of 0.53, which is completely ruled out by the observations. Of course, a blackbody is probably not the best representation for the spectral energy distribution of the radiation emitted by a very hot body, but in order to explain the emission line spectrum of I~Zw~18 by stars, one has to assume an ionizing radiation strongly enhanced at energies between 20 -- 54 eV, but not above 54.4 eV, compared to the model of De Mello et al.\ (1998). If this is realistic cannot be said at the present time. We have also checked the effect of heating by additional X-rays that would be emitted due to the interaction of the stellar winds with ambient matter (see Martin 1996 for such X-ray observations), by simply adding a bremsstrahlung spectrum at T = 10$^{6}$ or T = 10$^{7}$ K to the radiation from the starburst model. As expected, the effect on [O~{\sc iii}] $\lambda$4363/5007\ was negligible, since the X-rays are mostly absorbed in the low ionization regions (they do raise the temperature in the O$^{+}$\ zone to T$_{e}$ $\simeq$ 16\,000~K). As already commented by previous authors, changing the elemental abundances does not improve the situation. Actually, even by putting all the elements heavier than helium to nearly zero abundance, [O~{\sc iii}] $\lambda$4363/5007\ is raised by only 7\%. Varying the helium abundance in reasonable limits does not change the problem. The neglect of dust is not expected to be responsible for the too low [O~{\sc iii}] $\lambda$4363/5007\ we find in our models for I~Zw~18. Gas heating by photoemission from grains can contribute by as much as 20\% to the electron thermal balance when the dust-to-gas ratio is similar to that found in the Orion nebula. But, as discussed by Baldwin et al. (1991), it is effective close to the ionizing source where dust provides most of the opacity. Besides, the proportion of dust in I\,Zw\,18 is expected to be small, given the low metallicity. Extrapolating the relation found by Lisenfeld \& Ferrara (1998) between the dust-to-gas ratio and the oxygen abundances in dwarf galaxies to the metallicity of I\,Zw\,18 yields a dust-to-gass mass ratio 2 to 3 orders of magnitudes smaller than in the solar vicinity. It remains to examine the ingredients of the photoionization code. We first stress that the [O~{\sc iii}] $\lambda$4363/5007\ problem in I\,Zw\,18 has been encountered by various authors using different codes, even if the observational material and the points of emphasis in the discussion were not the same in all the studies. As a further test, we compared the same model for I~Zw~18 run by CLOUDY 90 and by PHOTO and the difference in the predicted value of [O~{\sc iii}] $\lambda$4363/5007\ was only 5\%. One must therefore incriminate something that is wrongly treated in the same way in many codes. One possibility that comes to mind is the diffuse ionizing radiation, which is treated by some kind of outward only approximation in all codes used to model I~Zw~18. However, we do not expect that an accurate treatment of the diffuse ionizing radiation would solve the problem. Indeed, comparison of codes that treat more accurately the ionizing radiation with those using an outward only approximation shows only a negligible difference in the electron temperature (Ferland et al.\ 1996). Besides, as we have shown, even quenching the diffuse ionizing radiation does not solve the problem. Finally, one can also question the atomic data. The most relevant ones here are those governing the emissivities of the observed [O~{\sc iii}]\ lines and the H~{\sc i}\ collision strengths. The magnitude of the discrepancy we wish to solve would require modifications of the collision strengths or transition probabilities for [O~{\sc iii}]\ of about 25\%. This is much larger than the expected uncertainties and the differences between the results of different computations for this ion (see discussion in Lennon and Burke 1994 and Galavis et al. 1997). Concerning L$\alpha$ excitation, dividing the collision strength by a factor 2 (which is far above any conceivable uncertainty, see e.g. Aggarwal et al. 1991) modifies [O~{\sc iii}] $\lambda$4363/5007\ only by 2\% because L$\alpha$ acts like a thermostat. We are therefore left with the conclusion that the [O~{\sc iii}] $\lambda$4363/5007\ ratio cannot be explained in the framework of photoionization models alone. \subsection{Condensations and filaments} Another failure of our photoionization models is that they predict too low [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ compared to observations. This is a general feature of photoionization models and it is often taken as one of the reasons to invoke the presence of shocks. However, it is well known that the presence of small condensations of filaments enhances the [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio, by reducing the local ionization parameter. Another possibility is to have an intervening filament located at a large distance from the ionizing source, whose projected surface on the aperture of the observations would be small. In order to see under what conditions such models can quantitatively account for the observed [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio, we have mimicked such a situation by computing a series of ionization bounded photoionization models for filaments of different densities located at various distances from the exciting stars. For simplicity, we assumed that there is no intervening matter between the filaments and the ionizing source. The models were actually computed for complete shells. The radiation coming from a filament can be simply obtained by multiplying the flux computed in the model by the covering factor $f$ by which the filament is covering the source. \begin{table*} \caption{Line intensities relative to \ifmmode {\rm H}\beta \else H$\beta$\fi\ for models of filaments. } \begin{flushleft} \begin{tabular}{l|l|lll|ll} \hline \noalign{\smallskip} & Main body & \multicolumn{3}{l|}{Filaments of various $n$} & \multicolumn{2}{l}{Filaments at various distances} \\ \hline \noalign{\smallskip} $n$ [cm$^{-3}$] & 10$^2$ $^a$ & 10$^4$ & 10$^5$ & 10$^6$ & 10$^2$ & 10$^2$ \\ $\theta_{\rm in}$ [\arcsec] & & 1.5 & 1.5 & 1.5 & 20. & 100. \\ \noalign{\smallskip} \hline \noalign{\smallskip} F(\ifmmode {\rm H}\beta \else H$\beta$\fi)$^b$ & 2.02e-13 & 3.33e-13 & 3.32e-13 & 3.35e-13 & 3.41e-13 & 3.47e-13 \\ [O~{\sc i}] $\lambda$6300\ & 1.04e-4 & 5.28e-2 & 1.57e-1 & 3.83e-1 & 6.90e-2 & 2.81e-1 \\ [O~{\sc ii}] $\lambda$3727\ & 2.15e-1 & 2.41e-1 & 5.39e-2 & 1.19e-2 & 4.71e-1 & 2.85e-1 \\ [O~{\sc iii}] $\lambda$4363\ & 3.46e-2 & 3.63e-5 & 1.12e-6 & 1.04e-7 & 1.20e-5 & 1.74e-7 \\ [O~{\sc iii}] $\lambda$5007\ & 1.42 & 3.76e-3 & 1.41e-4 & 1.25e-5 & 1.41e-3 & 3.08e-5 \\ [S~{\sc ii}]\ $\lambda$6717 & 5.58e-4 & 5.20e-2 & 5.23e-2 & 5.27e-2 & 1.87e-1 & 5.70e-1 \\ [S~{\sc ii}]\ $\lambda$6731. & 4.19e-3 & 8.04e-2 & 9.16e-2 & 1.03e-1 & 1.35e-1 & 3.96e-1 \\ \noalign{\smallskip} \hline \noalign{\smallskip} \noalign{\small $^a$ The model for the main body is density bounded (see text)} \noalign{\small $^b$ The total \ifmmode {\rm H}\beta \else H$\beta$\fi\ flux is given in erg s$^{-1}$ for a covering factor $f$=1. } \end{tabular} \end{flushleft} \end{table*} Table 3 presents the ratios with respect to \ifmmode {\rm H}\beta \else H$\beta$\fi\ of the [O~{\sc i}] $\lambda$6300, [O~{\sc ii}] $\lambda$3727, [O~{\sc iii}] $\lambda$5007, [S~{\sc ii}] $\lambda$6717 and [S~{\sc ii}] $\lambda$6731 for these models. The first column of the table corresponds to a photoionization model for the main body of the nebula with $R_{in}$ = 2.25 10$^{20}$ cm (corresponding to $\theta_{\rm in}=$ 1.5\arcsec), $n$=100~cm$^{-3}$, and $\epsilon$=0.01, and density bounded so as to obtain the observed [O~{\sc iii}]/[O~{\sc ii}]\ (this is one of the models of Fig. 6). It can be readily seen that, for an intervening filament of density 10$^{2}$~cm$^{-3}$\ located at a distance of 500~pc from the star cluster, or for condensations of density 10$^{6}$~cm$^{-3}$, one can reproduce the observed [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ line ratio without strongly affecting the remaining lines, not even the density sensitive [S~{\sc ii}] $\lambda$6731/6717\ ratio, if one assumes a covering factor $f$ of about 0.1. This explanation may appear somewhat speculative. However, one must be aware that the morphology of the ionized gas in I\,Zw\,18 shows filaments in the main shell as well as further out and this has been amply discussed in the literature (Hunter \& Thronson 1995, Dufour et al. 1995, Martin 1996). Our point is that, even in the framework of pure photoionization models, if one accounts for a density structure suggested directly by the observations, the strength of the [O~{\sc i}]\ line can be easily understood. We note that the condensations or filaments that produce [O~{\sc i}]\ are optically thick, and therefore their neutral counterpart should be seen in H~{\sc i}. Unfortunately, the available H~{\sc i}\ maps of I\,Zw\,18 (van Zee et al. 1998) do not have sufficient spatial resolution to reveal such filaments. But these authors show that the peak emission in H~{\sc i}\ and H$\alpha$\ coincide (in their paper, the peak emission in H$\alpha$\ actually refers to the whole NW component), and find that the entire optical system, including the diffuse emission, is embedded in a diffuse, irregular and clumpy neutral cloud. In such a situation, it is very likely that some clumps or filaments, situated in front of the main nebula, and having a small covering factor, produce the observed [O~{\sc i}]\ line. By using photoionization models to explain the observed [O~{\sc i}]/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio, one can deduce the presence of density fluctuations, even if those are not seen directly. Such density fluctuations could then be, more directly, inferred from density sensitive ratios of [Fe~{\sc ii}] lines, such as seen and analyzed in the Orion nebula by Bautista \& Pradhan (1995). \subsection{A few properties of I\,Zw\,18 deduced from models} Although we have not built a completely satisfactory photoionization model reproducing all the relevant observed properties of I\,Zw\,18, we are not too far from it. We describe below some properties of the best models that may be of interest for further empirical studies of this object. Tables 4 and 5 present the mean ionic fractions and mean ionic temperatures (as defined in Stasi\'{n}ska 1990) for the different elements considered, in the case of the best fit models with a uniform sphere and a spherical shell respectively, with $n$=100~cm$^{-3}$, and $\epsilon$=0.01. It can be seen that, while both geometries yield similar ionic fractions for the main ionization stages, the relative populations of the highly charged trace ions are very different. In the case of the uniform sphere, the proportion of O$^{+++}$, for example, is twice as large as in the shell model. Also, the characteristic temperature of ions with high ionization potential are much higher in the case of the filled sphere, for reasons commented on earlier. As a result, the OIV]1394/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio is 9.7~10$^{-3}$ in the first case and 8.6~10$^{-4}$ in the second. This line is too weak to be measured in I\,Zw\,18, of course, but it is useful to keep this example in mind for the study or more metal rich objects. It is interesting to point out that, in the case of the uniform sphere, the total flux in the He~{\sc ii} $\lambda$4686\ line is smaller than in the case of the shell model (3.1~10$^{-15}$ vs. 3.3~10$^{-15}$ erg~cm$^{-2}$~s$^{-1}$) despite the fact that the ionic fractions of He$^{++}$\ are similar. This is because the He$^{++}$\ region is at a much higher temperature (25\,000~K versus 18\,000~K). Tables 4 and 5 can be used for estimating the ionization correction factors for I\,Zw\,18. Caution should, however, be applied regarding especially the ionization structure predicted for elements of the third row of Mendeleev's table. Experience in photoionization modelling of planetary nebulae, where the observational constraints are larger, shows that the ionization structure of these elements is rarely satisfactorily reproduced with simple models (Howard et al. 1997, Pe\~na et al., 1998). The total ionized mass in the NW component is relatively well determined, since we know the total number of ionizing photons, the radius of the emitting region and have an estimate of the mean ionization parameter through the observed [O~{\sc iii}]/[O~{\sc ii}]. Indeed, at a given $Q({\rm{H^{0}}})$, for a constant density sphere with a filling factor $\epsilon$, $n^{2} \epsilon$ is proportional to the cube of the radius, while $n \epsilon^{2}$ is proportional to the cube of the ionization parameter. Of course, we have just made the point that the NW component of I\,Zw\,18 is not a sphere. Nevertheless, we have an order of magnitude estimate, which is of 3.~10$^{5}$~M$_{\odot}$\ at $d$ = 10~Mpc (this estimate varies like $d^{3}$). Finally, it is important to stress, as already mentioned above, that H$\alpha$\ is partly excited by collisions. In all our models for the main body of the nebula, H$\alpha$/\ifmmode {\rm H}\beta \else H$\beta$\fi\ lies between 2.95 and 3, while the case B recombination value is 2.7. This means that the reddening of I\,Zw\,18 is smaller than the value obtained using the case B recombination value at the temperature derived from [O~{\sc iii}] $\lambda$4363/5007. If we take the observations of Izotov \& Thuan (1998), who also correct for underlying stellar absorption, we obtain C(\ifmmode {\rm H}\beta \else H$\beta$\fi)=0. \section{Summary and conclusions} We have built photoionization models for the NW component of I~Zw~18 using the radiation field from a starburst population synthesis at appropriate metallicity (De Mello et al. 1998) that is consistent with the Wolf-Rayet signatures seen in the spectra of I~Zw~18. The aim was to see whether, with a nebular density structure compatible with recent {\em HST} images, it was possible to explain the high [O~{\sc iii}] $\lambda$4363/5007\ ratio seen in this object, commonly interpreted as indicative of electron temperature $\simeq$ 20\,000K. For our photoionization analysis we have focused on properties which are relevant and crucial model predictions. For the observational constraints we have not only used line ratios, but also other properties such as the integrated stellar flux at 3327~\AA\ and the observed angular radius of the emitting region as seen by the {\em HST}. Care has also been taken to include tolerances on model properties which may be affected by deviations from various simple geometries. We have found that [O~{\sc iii}] $\lambda$4363/5007\ cannot be explained by pure photoionization models, which yield too low an electron temperature. We have considered the effects due to departure from spherical symmetry indicated by the {\em HST} images. Indeed these show that the NW component of I~Zw~18 is neither a uniform sphere nor a spherical shell, but rather a bipolar structure with a dense equatorial ring seen pole on. We have discussed the consequences that an inaccurate description of the stellar ionizing radiation field might have on [O~{\sc iii}] $\lambda$4363/5007\, as well as additional photoionization by X-rays. Finally, we have considered possible errors in the atomic data. All these trials were far from solving the electron temperature problem, raising the [O~{\sc iii}] $\lambda$4363/5007\ ratio by only a few percent while the discrepancy with the observations is on the 30\% level. Such a discrepancy means that we are missing a heating source whose power may be of the same magnitude as that of the stellar ionizing photons. It is also possible that the unknown energy source is not so powerful, but acts in such a way that small quantities of gas are emitting at very high temperatures, thus boosting the [O~{\sc iii}] $\lambda$4363\ line. Shocks are of course one of the options (Peimbert et al. 1991, Martin 1996, 1997), as well as conductive heating at the interface of an X-ray plasma with optically visible gas (Maciejewski et al. 1996). Such ideas need to be examined quantitatively, and applied to the case of I~Zw~18, which we shall attempt in a future work. What are the consequences of our failure in understanding the energy budget on the abundance determinations in I~Zw~18? It depends on how the electron temperature is distributed in the O$^{++}$ zone. As emphasized by Peimbert (1967, 1996) over the years (see also Mathis et al. 1998), the existence of even small zones at very high temperatures will boost the lines with high excitation threshold like [O~{\sc iii}] $\lambda$4363, so that the temperature derived from [O~{\sc iii}] $\lambda$4363/5007\ will overestimate the average temperature of the regions emitting [O~{\sc iii}] $\lambda$5007. Consequently, the true O/H ratio will be larger than the one derived by the standard methods. The C/O ratio, on the other hand, will be smaller than derived empirically, because the ultraviolet C~{\sc iii}] $\lambda$1909\ line will be extremely enhanced in the high temperature regions. Such a possibility was invoked by Garnett et al. (1997) to explain the high C/O found in I~Zw~18 compared to other metal-poor irregular galaxies. Presently, however, too little is known both theoretically and observationally to estimate quantitatively this effect, and it is not excluded that the abundances derived so far may be correct within 30\%. Obviously, high spatial resolution mapping of the [O~{\sc iii}] $\lambda$4363/5007\ ratio in I~Zw~18 would be valuable to track the origin of the high [O~{\sc iii}] $\lambda$4363\ seen in spectra integrated over a surface of about 10~\arcsec $^{2}$. Beside demonstrating the existence of a heating problem in I~Zw~18, our photoionization model analysis led to several other results. The intensity of the nebular He~{\sc ii} $\lambda$4686\ line can be reproduced with a detailed photoionization model having as an input the stellar radiation field that is consistent the observed Wolf-Rayet features in I~Zw~18. This confirms the results of De Mello et al. (1998) based on simple Case B recombination theory. By fitting the observed [O~{\sc iii}]/[O~{\sc ii}]\ ratio and the angular size of the NW component with a model where the stellar radiation flux was adjusted to the observed value, we were able to show that the H~{\sc ii}\ region is not ionization bounded and about half of the ionizing photons are leaking out of it, sufficient to explain the extended diffuse H$\alpha$\ emission observed in I~Zw~18. While the [O~{\sc i}]\ emission is not reproduced in simple models, it can easily be accounted for by condensations or by intervening filaments on the line of sight. There is no need to invoke shocks to excite the [O~{\sc i}]\ line, although shocks are probably involved in the creation of the filaments, as suggested by Dopita (1997) in the context of planetary nebulae. The intrinsic H$\alpha$/\ifmmode {\rm H}\beta \else H$\beta$\fi\ ratio is significantly affected by collisional excitation: our photoionization models give a value of 3.0, to be compared to the case B recombination value of 2.75 used in most observational papers. Consequently, the reddening is smaller than usually estimated, with C(\ifmmode {\rm H}\beta \else H$\beta$\fi) practically equal to 0. Our models can be used to give ionization correction factors appropriate for I~Zw~18 for more accurate abundance determinations. However, the largest uncertainty in the abundances of C, N, O and Ne ultimately lies in the unsolved temperature problem. It would be, of course, of great interest to find out whether other galaxies share with I~Zw~18 this [O~{\sc iii}] $\lambda$4363/5007\ problem. There are at least two other cases which would deserve a more thorough analysis. One is the starburst galaxy NGC 7714, for which published photoionization models (Garc\'{\i}a-Vargas et al. 1997) also give [O~{\sc iii}] $\lambda$4363/5007\ smaller than observed. However, it still needs to be demonstrated that this problem remains when modifying the model assumptions (e.g. the gas density distribution, possible heating of the gas by photolelectric effect on dust particles etc...). In the case of NGC 2363, the photoionization models of Luridiana et al. (1999) that were built using the oxygen abundances derived directly from the observations yielded a [O~{\sc iii}] $\lambda$4363/5007\ ratio marginally compatible with the observations. These authors further argued that, due to the presence of large spatial temperature fluctuations, the true gas metallicity in this object is higher than derived by empirical methods. In such a case, the [O~{\sc iii}] $\lambda$4363/5007\ ratio becomes even more discrepant. It might well be that additional heating sources exist in giant H~{\sc ii}\ regions, giving rise to such large temperature variations and and enhancing the [O~{\sc iii}] $\lambda$4363\ emission. As mentioned above, such a scenario needs to be worked out quantitatively. Further detailed observational and theoretical studies of individual objects would be helpful, since we have shown that with insufficient observational constraints, high [O~{\sc iii}] $\lambda$4363/5007\ may actually be produced by photoionization models. The effort is worthwhile, since it would have implications both on our understanding of the energetics of starburst galaxies and on our confidence in abundance derivations. \begin{acknowledgements} This project was partly supported by the ``GdR Galaxies''. DS acknowledges a grant from the Swiss National Foundation of Scientific Research. We thank Duilia De Mello for providing the {\em HST} images and Jean-Fran\c{c}ois Le Borgne for help with IRAF. During the course of this work, we benefited from conversations with Jose V\'{\i}lchez, Rosa Gonz\'{a}lez-Delgado, Enrique P\'{e}rez, Yurij Izotov, Trinh Xuan Thuan. Thanks are due to Valentina Luridiana, Crystal Martin and Claus Leitherer for reading the manuscript. \end{acknowledgements}

{'timestamp': '1999-09-13T14:03:26', 'yymm': '9909', 'arxiv_id': 'astro-ph/9909203', 'language': 'en', 'url': 'https://arxiv.org/abs/astro-ph/9909203'}

high_school_physics

Cosmological ruler could help us get the measure of dark energy Space 17 March 2017 What’s driving them apart? NASA/HST By Matt Reynolds Measuring cosmological distances is tricky at the best of times, but it is even harder if your ruler isn’t up to scratch. A new cosmological standard ruler could make the difference – and help astrophysicists figure out why the universe is expanding at an accelerating rate. Ever since we realised that the universe’s expansion is accelerating – a discovery that won Adam Riess and his colleagues the 2011 Nobel prize in physics – we have been trying to work out why. The phenomenon is usually put down to the existence of dark energy, a mysterious force that appears to be pushing the universe’s matter further and further apart. Accurate measurements of how the distance between clusters of galaxies has changed over time could tell us whether the effect of dark energy on the universe is increasing. But our current way of measuring this has a problem: it relies on assumptions that are difficult to check. Currently, astronomers measure distances by taking advantage of the fact that the distribution of galaxies throughout the universe fluctuates in predictable ways, a relic of sound waves that echoed through the early universe. Infant echoes In its infancy, the universe was a hot soup of matter that was distributed in a mostly uniform fashion, with a few dense spots of dark matter. About 30,000 years after the big bang, gravity made normal matter collapse around those dense spots, but pressure from photons caused it to rebound outward again. This cosmic bounce created acoustic waves, called baryon acoustic oscillations (BAO), that expanded outwards in a spherical shape and carried normal matter with them. These waves kept expanding until the cooling of the universe halted their progress, freezing each of them at roughly the same point in their expansion. The resulting dense patches of matter at the edge of the waves and their points of origin were more likely to form galaxies, creating a lasting imprint of the acoustic waves that rang out in the early universe. “All our galaxies are close to the surface of these spherical shells or their centre,” says Glenn Starkman at Case Western Reserve University in Ohio. Because each of these spherical shells is the same size, astrophysicists use the distance between a central galaxy cluster and galaxy clusters at the edge of the wave – about 500 million light years – as a kind of cosmological ruler. Blurry boundaries But it isn’t perfect: as the universe aged, this neat structuring of galaxies blurred as gravity and magnetic fields pulled galaxies in different directions. That means that to use the technique mathematical models are needed to make a series of assumptions about how galaxies have shifted over time. Some BAO distances could be spot on, but others are off by several per cent, and there is no way to tell which ones are which. Now, Starkman and his colleagues have come up with a ruler that sidesteps the need for these assumptions. Instead of following the galaxies’ motions, the team measure distance relative to an unmoving mathematical midpoint in the BAOs called the linear point. Starkman says the technique is up to four times more accurate than existing methods. Being able to measure very large cosmological distances is critical to understanding the nature of dark energy, says Will Percival at the University of Portsmouth in the UK. But while using the linear point appears to reduce the number of errors in estimating cosmological distances, he says, it’s not yet clear whether using it will yield more precise measurements than existing methods. For more precise measurements, we will have to wait for the results of future galaxy surveys such as the European Space Agency’s Euclid mission, which will observe areas of the universe that aren’t polluted by light from the Milky Way. The data from these surveys will be at least 10 times more precise than existing surveys, Percival says, and will provide a good opportunity to test whether the proposed cosmological ruler is better than current measurements. Reference: arxiv.org/abs/1703.01275 Magazine issue 3118 , published 25 March 2017 On front line of climate change as Maldives fights rising seas Vision saved by first induced pluripotent stem cell treatment

{"pred_label": "__label__wiki", "pred_label_prob": 0.880020797252655, "wiki_prob": 0.880020797252655, "source": "cc/2019-30/en_head_0007.json.gz/line1056372"}

high_school_physics

Q: Can I use windows icons in my .Net application I want to use Icons that appear on certain windows and dialog boxes in the Windows OS such is the Information bubble icon Here in my application. My application will be distributed free of charge under a closed source software license. So can I use the windows icons in my closed source free application? Note: My application is a .net application and is developed in Visual Studio.

{'language': 'en', 'url': 'https://superuser.com/questions/1586031', 'timestamp': '2023-03-29', 'source': 'stackexchange', 'question_score': '1'}

high_school_physics

\chapter{Assembly, Installation and Engineering Issues} \label{ch:AssemblyAndInstallation} \input{AssemblyAndInstallation/chapter12.1.tex} \input{AssemblyAndInstallation/chapter12.2.tex} \input{AssemblyAndInstallation/chapter12.3.tex} \input{AssemblyAndInstallation/chapter12.4.tex} \input{AssemblyAndInstallation/chapter12.5.tex} \section{Introduction} After the completion of civil construction, through systems integration, assembly, installation and commissioning, all subsystems will be formed as a whole JUNO detector. In this process, we need to establish various engineering regulations and standards, and to coordinate subsystems' assembly, installation, testing and commissioning, especially their onsite work. \subsection{Main Tasks} Main tasks of the integration work include: \begin{itemize} \item To prepare plans and progress reports of each phase; \item To establish a project technology review system; \item To standardize the executive technology management system; \item To have strictly executive on-site work management system; \item To develop and specify security management system on-site; \item To prepare common tools and equipment for each system, and to guarantees project progress; \item To coordinate the installation progress of each system according to the on-site situation. \end{itemize} \subsection{Contents} The work contents mainly include: \begin{itemize} \item To summarize the design, and review progress of each subsystem; \item To organize preparation work for installation in the experiment region; \item To inspect and certify Surface Buildings, underground Tunnels, and Experiment Hall with relevant utilities; \item To coordinate technology interfaces between key parts; \item To coordinate the procedure of assembly and installation both on surface and underground; \end{itemize} \section{ Design Standards, Guidelines and Reviews} \subsection{Introduction} There will be a document to outline the mechanical design standards or guidelines that will be applied to the design work. It also describes the review process of engineering design that will be implemented to ensure that experimental equipment meets all requirements for performance and safety. The following is a brief summary of the guidance that all mechanical engineers/designers should follow in the process. For reasons ranging from local safety requirements to common sense practices, the following information should be understood and implemented in the design process. \subsection{Institutional Standards} When specific institutional design standards or guidelines exist, they should be followed. The guidelines outlined are not meant to replace but instead to supplement institutional guidelines. The majority of equipment and components built for the JUNO Experiment will be engineered and designed at home institutions, procured or fabricated at commercial vendors, then eventually delivered, assembled, installed, tested and operated at the JUNO Experimental facilities, Jiangmen, China. The funding agencies in other countries as well as Jiangmen, China will have some guidelines in this area, as would your home institutions. Where more than one set of guidelines exist, use whichever are the more stringent or conservative approaches. \subsection{Design Loads} The scope of engineering analysis should take handling, transportation, assembly and operational loads into account as well as thermal expansion considerations caused by potential temperature fluctuations. For basic stability and a sensible practice in the design of experimental components an appropriate amount of horizontal load (0.10~g) should be applied. In addition, seismic requirements for experimental equipment is based on a National Standard of People's Republic of China Code of Seismic Design of Buildings GB 50011-2010. At the location of the JUNO Experiment the seismic fortification intensity is grade 7 with a basic seismic acceleration of 0.10~g horizontal applied in evaluating structural design loads. A seismic hazard analysis should be performed and documented based on this local code. The minimum total design lateral seismic base shear force should be determined and applied to the component. The direction of application of seismic forces would be applied at the CG of the component that will produce the most critical load effect, or separately and independently in each of the two orthogonal directions. A qualitative seismic performance goal defines component functionality as: the component will remain anchored, but no assurance it will remain functional or easily repairable. Therefore, a seismic design factor of safety F.S. > 1 based on the Ultimate Strength of component materials would satisfy this goal. Where an anticipated load path as designed above, allows the material to be subjected to stresses beyond the yield point of the material, redundancy in the support mechanism must be addressed in order to prevent a collapse mechanism of the structure from being formed. The potential for buckling should also be evaluated. It should be noted that a rigorous analytical seismic analysis may be performed in lieu of the empirical design criteria. This work should be properly documented for review by the Chief Engineer and appropriate Safety Committee personnel. \subsection{Materials Data} All materials selected for component design must have their engineering data sources referenced along with those material properties used in any structural or engineering analysis; this includes fastener certification. There are many sources of data on materials and their properties that aid in the selection of an appropriate component material. There are many national and international societies and associations that compile and publish materials test data and standards. Problems frequently encountered in the purchasing, researching and selection of materials are the cross-referencing of designations and specification and matching equivalent materials from differing countries. It is recommended that the American association or society standards be used in the materials selection and specification process, or equivalency to these standards must be referenced. Excellent engineering materials data have been provided by the American Society of Metals or Machine Design Handbook Vol. 1-5 and Mechanical Industry Publishing Co. in PRC, Vol 1-6, 2002, which are worth investigating. \subsection{Analysis Methods} The applicable factors of safety depend on the type of load(s) and how well they can be estimated, how boundary conditions have been approximated, as well as how accurate your method of analysis allows you to be. \paragraph{Bounding Analyses:} Bounding analysis or rough scoping analyses have proven to be valuable tools. Even when computer modeling is in your plans, a bounding analysis is a nice check to avoid gross mistakes. Sometimes bounding analyses are sufficient. An example of this would be for the case of an assembly fixture where stiffness is the critical requirement. In this case where deflection is the over-riding concern and the component is over-designed in terms of stress by a factor of 10 or more, then a crude estimation of stress will suffice. \paragraph{Closed-Form Analytical Solutions:} Many times when boundary conditions and applied loads are simple to approximate, a closed-form or handbook solution can be found or developed. For the majority of tooling and fixture and some non-critical experimental components, these types of analyses are sufficient. Often, one of these formulas can be used to give you a conservative solution very quickly, or a pair of formulas can be found which represent upper and lower bounds of the true deflections and stresses. Formulas for Stress and Strain by Roark and Young is a good reference handbook for these solutions. \paragraph{Finite Element Analysis:} When the boundary conditions and loads get complex, or the correctness of the solution is critical, computer modeling is often required. If this is the case, there are several rules to follow, especially if you are not intimately familiar with the particular code or application. \begin{enumerate} \item Always bound the problem with an analytical solution or some other approximate means. \item If the component is critical, check the accuracy of the code and application by modeling a similar problem for which you have an analytical or handbook solution. \item Find a qualified person to review your results. \item Document your assumptions and results. \end{enumerate} \subsection{Failure Criteria} The failure criterion depends upon the application. Many factors such as the rate or frequency of load application, the material toughness (degree of ductility), the human or monetary risk of component failure as well as many other complications must be considered. Brittle materials (under static loads, less than 5\% yield prior to failure), includes ceramics, glass, some plastics and composites at room temperature, some cast metals, and many materials at cryogenic temperatures. The failure criterion chosen depends on many factors so use your engineering judgment. In general, the Coulomb-Mohr or Modified Mohr Theory should be employed. Ductile materials (under static loads, greater than 5\% yield prior to failure), includes most metals and plastics, especially at or above room temperature. The failure criterion chosen again ultimately rests with the cognizant engineer because of all the adverse factors that may be present. In general, the Distortion- Energy Theory, or von Mises-Hencky Theory (von Mises stresses), is most effective in predicting the onset of yield in materials. Slightly easier to use and a more conservative approach is the Maximum-Shear-Stress Theory. \subsection{Factor of Safety} Some institutions may have published guidelines which specifically discuss factors of safety for various applications. For the case where specific guidelines do not exist, the following may be used. Simplistically, if F is the applied load (or S the applied stress), and $F_f$ is the load at which failure occurs (or $S_s$ the stress at which failure occurs), we can then define the factor of safety (F.S.) as: \begin{displaymath} F.S. = F_{f} / F \quad\mathrm{or}\quad S_{s} / S \end{displaymath} The word failure, as it applies to engineering elements or systems, can be defined in a number of ways and depends on many factors. Discussion of failure criteria is presented in the previous section, but for the most common cases it will be the load at which yielding begins. \subsection{Specific Safety Guidelines for JUNO} Lifting and handling fixtures, shipping equipment, test stands, and fabrication tooling where weight, size and material thickness do not affect the physical capabilities of the detector, the appropriate F.S. should be at least 3. When life safety is a potential concern, then a F.S. of 5 may be more appropriate. Note that since the vast majority of this type of equipment is designed using ductile materials, these F.S.'s apply to the material yield point. Experimental hardware that does not present a life safety or significant cost/schedule risk if failure occurs, especially where there is the potential for an increase in physics capabilities, the F.S. may be as low as 1.5. Many factors must be taken into account if a safety factor in this low level is to be employed: a complete analysis of worst case loads must be performed; highly realistic or else conservative boundary conditions must be applied; the method of analysis must yield accurate results; reliable materials data must be used or representative samples must be tested. If F.S.'s this low are utilized, the analysis and assumptions must be highly scrutinized. Guidelines for F.S. for various types of equipment are: \begin{center} \begin{tabular}{|p{3.5cm}|c|p{3.5cm}|} \hline Type of Equipment & Minimum F.S. & Notes \\ \hline Lifting and handling & 3 - 5 & Where there is a risk to life safety or to costly hardware, choose F.S closer to 5. \\ \hline Test stands, shipping and assembly fixtures. & 3 & \\ \hline Experimental hardware & 1.5 - 3 & 1.5 is allowable for physics capability and analysis where method is highly refined \\ \hline \end{tabular} \end{center} \subsection{Documentation} It is not only good engineering practice to document the analysis, but it is an ESH\&Q requirement for experimental projects. For this reason all major components of JUNO Experiment will have their engineering analyses documented as Project Controlled Documents. Utilize your institutional documentation formats or use the following guidelines. Calculations and analyses must \begin{itemize} \item Be hard copy documented. \item Follow an easily understood logic and methodology. \item Be legible and reproducible by photocopy methods. \item Contain the following labeling elements. \begin{itemize} \item Title or subject \item Originators signature and date \item Reviewers signature and date \item Subsystem WBS number. \item Introduction, background and purpose. \item Applicable requirements, standards and guidelines. \item Assumptions (boundary conditions, loads, materials properties, etc.). \item Analysis method (bounding, closed form or FEA). \item Results (including factors of safety, load path and location of critical areas). \item Conclusions (level of conservatism, limitations, cautions or concerns). \item References (tech notes, textbooks, handbooks, software code, etc.). \item Computer program and version (for computer calculations) \item Filename (for computer calculations). \end{itemize} \end{itemize} \subsection{Design Reviews} All experimental components and equipment whether engineered or procured as turn-key systems will require an engineering design review before procurement, fabrication, assembly or installation can proceed. The L2 subsystem manager will request the design review from the project management office, which will appoint a chair for the review and develop a committee charge. The selected chair should be knowledgeable in the engineering and technology, and where practical, should not be directly involved in the engineering design effort. With advice from the L2, the technical board and project office chair appoints members to the review committee that have experience and knowledge in the engineering and technology of the design requiring review. At least one reviewer must represent ESH\&Q concerns. The committee should represent such disciplines as: \begin{itemize} \item Science requirements \item Design \item Manufacturing \item Purchasing \item Operational user \item Maintenance \item Stress analysis \item Assembly/installation/test \item Electrical safety \item ESH\&Q \end{itemize} For the JUNO project there will be a minimum of two engineering design reviews: Preliminary and Final. The preliminary usually takes place towards the end of the conceptual design phase when the subsystem has exhausted alternative designs and has made a selection based on value engineering. The L2 along with the chair ensures that the preliminary design review package contains sufficient information for the review along with: \begin{itemize} \item Agenda \item Requirements documents \item Review committee members and charge \item Conceptual layouts \item Science performance expectations \item Design specifications \item Supportive R\&D or test results \item Summaries of calculations \item Handouts of slides \end{itemize} A final design review will take place before engineering drawings or specifications are released for procurement or fabrication. The L2 along with the chair ensures that the final design review package contains sufficient information for the final review along with: \begin{itemize} \item Changes or revisions to preliminary design \item Completed action items \item Final assembly and detailed drawings \item Final design specifications \item Design calculations and analyses \end{itemize} The committee chair will ensure that meeting results are taken and recorded along with further action items. The committee and chair shall prepare a design review report for submittal to the project office in a timely manner which should include: \begin{itemize} \item Title of the review \item Description of system or equipment \item Copy of agenda \item Committee members and charge \item Presentation materials \item Major comments or concerns of the reviewers \item Action items \item Recommendations \end{itemize} The project office will file the design review report and distribute copies to all reviewers and affected groups. \section{On-site management} According to the experience from the Daya Bay Experiment, an effective technology review system has been well practiced. In JUNO, we will take it as a good reference and carry out for standardized review from the start of all system design, scheme argument, technology checks, we should also establish management system, to cover engineering drawing control, engineering change control procedure, and mechanical design standards, guidelines and reviews, etc. To control on-site work process, proper process should be introduced as well as a management structure. Safety training and safety management structure should be also introduced. \section{Equipment and Tools Preparation on Site} All of key equipment, devices and tools should be in place and get acceptance for installation on site, including \begin{itemize} \item Cranes: \begin{itemize} \item 2 sets of bridge crane with 12T/5T load capacity, and lifting range should cover any point for components delivery and installation in EH; \item Several lifting equipment are needed in Surface Assembly areas, Assembly Chamber underground, Storage area for Liquid Scintillation, and other Chambers accordingly; \end{itemize} \item Removable lifting equipment: Forklifts, manual hydraulic carriages, carts, etc. \item Mobile vehicle for smaller items: Pickup van, battery car, and so on; \item Several work aloft platform: Scissor lifts, Boom lifts, Scaffoldings and Floating platforms, etc. \item Common Tooling: complete kits of tool, fixtures, jigs, models, work-bench, cabinets, and tables, etc. \item Machine shop, Power equipment rooms, Control rooms are ready to be put into use. \item Safety measures and relevant equipment in place. \end{itemize} \section{Main Process of Installation} Since there still are several optional design versions which could be decided later, and different design could have different requirements for installation, therefore, the specific installation procedures will be established and developed later, here a very rough installation procedure is given. \begin{itemize} \item Mounting, debugging, testing of basic utilities, such as, cranes, vertical elevator, scaffoldings, test devices, etc. \item Survey, alignment, and adjustment for tolerances of form and position, include, embedded inserts, anchor plates, positioning holes, etc. \item Mounting Bottom Structure of Veto Cherenkov detector on the floor of Water Pool. \item Mounting anchor-plate for support poles of CD on the floor of Water Pool \item Pre-mounting of Top Structure of Veto system along wall of Water Pool \item Mount Substrate and Rails for Bridge over the water Pool edges \item Mount Tyvek with support structure for Water Cherenkov Detector, and with PMT positioning in the Water Pool \item Installation for CD with its PMT system and cabling in Water Pool details will be established until design final selection. \item Mount Calibration System \item Mount Geomagnetic Shielding system around Central Detector in the Pool \item Final installation for Tyvek and PMT of Veto system in the Pool \item Mount a temporary protective cover on top of Pool, to prevent from anything falling into Pool \item Mount Bridge on the rails, and drive it closed \item Mount Top Track Detector on Bridge and establish Electronic Room there \item Have Cabling completion, and test to get normal signals \item Check Grounding Electrode \item Dry Run \item Final Cleaning for Water Pool, then, drain out \item Filling Water, LS, LAB (or Oil) evenly into Water Pool and CD \item Mount Final Pool Cover, closed, and get fixation \end{itemize} \chapter{AssemblyAndInstallation} \label{ch:AssemblyAndInstallation} \input{AssemblyAndInstallation/chapter12.1.tex} \input{AssemblyAndInstallation/chapter12.2.tex} \input{AssemblyAndInstallation/chapter12.3.tex} \input{AssemblyAndInstallation/chapter12.4.tex} \input{AssemblyAndInstallation/chapter12.5.tex} \chapter{The Central Detector Calibration System} \label{ch:Calibration} \section{Requirements on the Calibration System} \label{sec:cal_req} The neutrino oscillation signal in the JUNO experiment will be quantified by the measurement of the positron energy in the inverse beta decay (IBD) interaction. The oscillation patterns are driven by the atmospheric mass-squared splitting $\Delta m^2_{32}$ and the third neutrino mixing angle $\theta_{13}$, and the solar mass-squared splitting $\Delta m^2_{21}$ and the solar mixing angle $\theta_{12}$. The interplay of the two contains the neutrino mass hierarchy (MH) information through the energy dependence of the effective atmospheric mass-squared splitting. Therefore, one needs to determine the positron energy spectrum with high precision in order to determine the neutrino MH and to carry out the precision measurements of the oscillation parameters. As stressed in Chapter~\ref{chap:intro} as well as in independent studies~\cite{Qian:2013}, the JUNO central detector energy resolution needs to be better than 3\% at 1~MeV (1.9\% at 2.5 MeV where the MH signal lies), and the absolute energy scale uncertainty to be much better than 1\% for the MH determination. On the other hand, the requirement on the energy resolution and the energy scale uncertainty is more relaxed for the sub-percent precision measurements of the oscillation parameters. There are two main contributors for the energy resolution of the liquid scintillator (LS) detectors. The first is the number of collected scintillation photons, which is influenced by the intrinsic light yield and attenuation length of the LS, the photocathode coverage of the detector, and the quantum and collection efficiencies of the photon detectors. The second is the non-uniformity of the detector response, governed by the detector geometry as well as the optical properties of the relevant detector components (e.g., the attenuation of the liquid scintillator and acrylic vessel). Nevertheless, the non-uniformity can in principle be reliably evaluated by deploying radioactive sources with known energies inside the detector during the calibration data taking. Based on experiences from the Daya Bay experiment, liquid scintillator detector has two major sources of energy non-linearity. The first one is the intrinsic energy non-linearity of the liquid scintillator, which includes the quenching of the scintillation light and the amount of the Cerenkov light within the energy ranges of interest. The second one is the energy non-linearity introduced by the electronics. A precise determination of both effects to sub-percent level is the primary challenge that the JUNO calibration system has to confront. The most reliable handle on the non-linearity effects is through the \emph{in-situ} calibrations, since the dependence on particle types as well as the Cerenkov radiation is difficult to evaluate in a bench setup elsewhere. To accurately address both the the non-uniformity and non-linearity in the detector energy response, the calibration system is required to deploy multiple sources (light or radioactive) to a wide range of positions inside the detector. Specifically, a summary of different types of detector response calibration envisioned is shown in Table~\ref{table:calib}. In addition to actively deploying sources, the reconstructed background events can also provide critical inputs to the detector response. \begin{table} \caption{\label{table:calib}Summary of the different types of calibration goals, methods and requirements} \begin{tabular}{|p{1.25in}|p{2.1in}|p{2.1in}|} \hline Calibration Goal & Calibration Method & Requirements\tabularnewline \hline \hline PMT gain & Low intensity light sources at the center & Periodical, low intensity, automatic\tabularnewline \hline Light yield (energy scale) & Radioactive sources at the center & Periodical and automatic\tabularnewline \hline PMT timing, time walk & Light sources at the center & Periodical and variation in the light intensity\tabularnewline \hline Optical property of scintillator & Radioactive sources at various positions & Periodical with optimized positions\tabularnewline \hline Boundary effect of energy response & Pre-installed guide tube system with radioactive sources & Scan the detector boundary and 1-2 times during whole period compare source and cosmogenic data\tabularnewline \hline Detector response non-uniformity & Radioactive sources at various positions and full volume cosmogenic data & Need based calibration frequency, targeted positions, compare source and cosmogenic data\tabularnewline \hline Capture time non-uniformity & Neutron source at various positions & Need based calibration frequency, targeted positions, compare source and spallation neutrons \tabularnewline \hline Energy non-linearity & Various radioactive sources & Full calibration with various sources 1-2 times during the entire period, compare source and cosmogenic data, periodical check for 2-3 types of sources at targeted positions, \tabularnewline \hline Energy non-linearity of positrons & Mono-energetic positrons to the center and various position along the central axis & Need based calibration frequency \tabularnewline \hline Position-dependent energy non-linearity & Various radioactive sources at fine position coverage & 1-2 times during the whole period, compare source and cosmogenic data, periodic check at targeted positions with 1-2 sources\tabularnewline \hline \end{tabular} \end{table} Based on the considerations above, we set the following requirements to the calibration system, \begin{enumerate} \item In order to carry out periodical calibrations, the most frequently used calibration system should be fully automated (similar to the ACU system in the Daya Bay experiment). It should be simple and extremely reliable. Such a subsystem should be able to deploy multiple radioactive and light sources into the detector, likely along the central axes; \item The calibration system is required to access the off-center positions in order to control the position non-uniformity. The source deployment to such locations is likely to occur, on the other hand, less frequently compared to that of the central axis; \item Gamma source energies should cover a significant part of the IBD energy spectrum (1-7~MeV). Positron and neutron sources are also needed to calibrate the detection efficiencies; \item A dedicated deployment system that has nearly $\sim$ 4$\pi$ coverage to the central detector should be considered; \item The absolute source positions should be controlled to better than 5~cm (preliminary Monte Carlo simulation has shown that the event vertex reconstruction accuracy can be $\sim$ 10~cm); \item An accelerator that can supply mono-energetic $e^{+}$ and $e^-$ beams is important to measure the positron non-linearity directly; \item Multiple types of particles such as $\alpha$, $e^-$, $\gamma$, and $e^{+}$ should be deployed to understand the LS quenching and Cerenkov light contribution to the energy non-linearity; \item With the photo-luminescence effect in liquid scintillator, a UV laser and optical fiber system has superior properties than those of an LED. Such more stable and versatile laser systems should be considered as light sources to calibrate the nonlinearity of the photo-sensors and the electronics. \end{enumerate} Considering all these requirements and the mechanical constraints, we selected the following mutually complementary calibration options for further investigations. \begin{enumerate} \item A central automated calibration unit (ACU) for vertical source deployment; \item A rope loop system for off-center source deployment; \item A pelletron system which can provide mono-energetic positron beams; \item Remotely operated under-liquid-scintillator vehicles (ROV) for ``4$\pi$'' coverage; \item Pre-installed guide tubes; and \item A diffuse system that can introduce short-lived radioactive isotopes into the central detector. \end{enumerate} The ACU, the rope loop system and the ROV can all scan the interior of the central detector but with different ranges and frequencies - we will describe them together. The pre-installed guide tube system is uniquely designed to understand the boundary effects of the central detector. The pelletron calibration system provides an excellent bench mark of the energy non-linearity for the IBD. A diffuse source system can validate the energy non-linearity and its position dependence. We will describe these three systems separately. \section{Conceptual Designs} \subsection{Housing and Interfaces} Calibration sources are deployed into the detector through the top chimney. To avoid Rn contamination, the entire calibration system should be enclosed in a clean and Rn free volume (``calibration house''), as shown schematically in Fig.~\ref{fig:calib_house}. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.7\textwidth]{Calibration/figs/calib_house.png} \par\end{centering} \caption{\label{fig:calib_house}Illustration of a seal calibration house on top of the detector.} \end{figure} A concept of the calibration house is shown in Fig.~\ref{fig:cal_overview}. The ACU system is located at the top. Two rope loop systems are placed inside the house. The source changing area can be accessed via two glove boxes to allow manual operation. The ROV system is also kept inside the house and moved along the rails attached to the roof. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.7\textwidth]{Calibration/figs/overview.jpg} \par\end{centering} \caption{\label{fig:cal_overview}Concept of a calibration house.} \end{figure} \subsection{Sources} \subsubsection{Routine Sources} To set the energy scales and to calibrate the positron and neutron detection efficiencies, radioactive sources for JUNO include routine gamma and positron sources together with neutron sources with correlated high energy gamma ray emissions, shown in Table~\ref{tab:sources}. \begin{table} \caption{\label{tab:sources}Radioactive sources under consideration in JUNO.} \begin{tabular}{ccc} \hline Source & Type & Radiation \\\hline $^{40}$K & $\gamma$ & 1.461 MeV\\ $^{54}$Mn & $\gamma$ & 0.835 MeV \\ $^{60}$Co & $\gamma$ & 1.173 + 1.333 MeV \\ $^{137}$Cs & $\gamma$ & 0.662 MeV\\ $^{22}$Na & e$^{+}$ & annil + 1.275 MeV\\ $^{68}$Ge & e$^{+}$ & annil 0.511 + 0.511 MeV\\ $^{241}$Am-Be & n, $\gamma$ & neutron + 4.43 MeV \\ $^{241}$Am-$^{13}$C or $^{241}$Pu-$^{13}$C & n, $\gamma$ & neutron + 6.13 MeV \\ $^{252}$Cf & multiple n, multiple $\gamma$ & prompt $\gamma$'s, delayed n's\\ \hline \end{tabular} \end{table} To minimize the risk due to contamination of the radioactivities, the enclosure of these sources should be thin-walled ($\sim$ mm) stainless steel (SS) capsule, enclosed by a round-headed acrylic shell to ensure chemical compatibility with the liquid scintillator. The deployment rope, made with SS, shall be robustly attached to the SS source enclosures. Two steel weights are attached above and below the source to maintain a minimum tension in the rope. A typical source/weight assembly used in Daya Bay is shown in Fig.~\ref{fig:source}. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.5\textwidth]{Calibration/figs/source_weight.png} \par\end{centering} \caption{\label{fig:source}Illustration of a typical source/weight assembly used in Daya Bay, envisioned for JUNO as well.} \end{figure} Simulations are underway to optimize the geometry in order to minimize impacts of the source assembly to the energy scale. In Fig.~\ref{fig:Co60} a typical $^{60}$Co spectrum is shown. The ``dead volume'' of the source assembly affects the low energy shoulder, but has a much smaller effect on the full absorption peak. The optical shadowing of the source assembly, on the other hand, will bias the location of the full absorption peak. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.5\textwidth]{Calibration/figs/Co60.png} \par\end{centering} \caption{\label{fig:Co60}an example JUNO Monte Carlo spectrum in photoelectrons for the $^{60}$Co source with source geometry.} \end{figure} \subsubsection{Mini-balloon} In order to study the liquid scintillator response to charged particles (electrons, positrons, $\alpha$'s) thereby to control the quenching effects, we are pursuing a ``mini-balloon'' concept (Fig.~\ref{fig:balloon}). \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.5\textwidth]{Calibration/figs/balloon.png} \par\end{centering} \caption{\label{fig:balloon}The mini-balloon source concept.} \end{figure} Radioactive isotope is loaded into the liquid scintillator inside a small, thin-walled, and transparent balloon ($\sim$ 10 cm OD). The balloon is further enclosed by an acrylic cylinder ($\sim$ 20 cm OD) filled with undoped liquid scintillator. Such a design will minimize the energy loss of the charged particles across the balloon wall to simulate real events, while maintaining a double encapsulation of the radioactive isotopes. A comparison of the $^{40}$K-loaded 10~$\mu$m thick balloon and that without the balloon from the Monte Carlo is shown in Fig.~\ref{fig:K40}, in which the bias in beta energy is estimated to be less than 0.3\%. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.5\textwidth]{Calibration/figs/K40.png} \par\end{centering} \caption{\label{fig:K40}Comparison of $^{40}$K spectra with a 10 $\mu$m balloon and without.} \end{figure} One important advantage is that the choice of the dopant in the balloon is very versatile, for example $^{222}$Rn ($\alpha$'s), $^{40}$K ($\gamma$'s and $\beta$'s), $^{137}$Cs ($\gamma$'s, $\beta$'s and conversion electrons), $^{68}$Ge (positrons), and even short-lived isotopes. The source deployment system can then take the source into given locations inside the detector for calibration. \subsubsection{Pulsed Light Source} Unlike the radioactive sources, the pulsed light source can generate photons at a given time with tunable intensity. These features are important to calibrate the gains of the photomultipliers as well as the response of the electronics system. Based on the experience in the Daya Bay calibration system, the stability of the pulses from light emitting diodes or LEDs ($\sim$5 ns timing) is not as good as pulsed laser system (which could achieve a ns level timing precision). Furthermore, a UV laser of $\sim$260~nm wavelength can excite the LAB molecules thereby producing photons with similar timing characters as those generated by real particle interactions. In addition, commercial light sensors, e.g., Si diode can monitor the intensity of the laser pulses to very high precision. Based on these considerations, we are developing a UV laser system coupled to a flexible optical fiber with a diffuser sphere at the end to achieve a stable, uniform, tunable, and deployable light source for JUNO. \subsection{Source Deployment Systems for the Detector Interior} As mentioned in Sec.~\ref{sec:cal_req}, three complementary subsystems are under consideration, the ACU system, the rope loop system, as well as the ROV system. They cover different ranges inside the central detector, and are envisioned to be used at different frequencies. In addition to the mechanical means to position a source, independent source locating systems will be critical to ensure the required high position accuracy. \subsubsection{Source Locating Systems} Although the rope loop can position itself via the lengths of different rope sections, mechanical uncertainty such as deformation of the ropes would introduce uncertainty to the absolute source position. Similar uncertainty arises in the ROV system as well. A dedicated source locating system is required to determine the source position to better than 5~cm. Two options are being considered. The first option is an ultrasonic system with an array of ultrasonic receivers and an ultrasonic emitter attached to the source. The emitter generates pulsed sonar waves and the receivers reconstruct the origin via the timing or phase difference. The wavelength of 150~kHz sonar wave is $\sim$ 1~cm, introducing negligible uncertainty. The dominating systematics is the positioning accuracy of the receivers as well as the speed of sound in LAB under different temperature/pressure conditions. Reflections at the interface are also systematics under study. To allow package of the sound waves, the receivers have to be installed in the inner wall of the vessel if the central detector housing is acrylic, and not so critical if the balloon option is adopted. In principle, the larger the receiver array, the higher the positioning accuracy. Studies are undergoing to optimize the array. A typical arrangement is shown in Fig.~\ref{fig:ultra}, with 9 redundant receivers mounted on the wall. \begin{figure} \begin{centering} \includegraphics[height=2.5in]{Calibration/figs/ultra.png} \par\end{centering} \begin{centering} \caption{\label{fig:ultra}A typical placement of the ultrasonic receivers.} \par\end{centering} \end{figure} An alternative technology is to locate the source via optical cameras (CCD) and one reconstruct source position via image analysis. Such a system is used by the Borexino experiment. The infrared lighting of the CCDs poses minimal risk to the PMTs. Alternatively, a battery driven LED can be attached to the source assembly to ease the analysis. \subsubsection{ACU} The Automated Calibration Unit (ACU) system is a unit very similar to that used in Daya Bay, capable of deploying a few different sources along the central axis of the detector. As illustrated in Fig.~\ref{fig:ACU}, four source deployment units will be arranged on a turntable, which can select the source to be deployed. A hole on the bottom plate of the ACU will be aligned to the center of the chimney to the central detector. Three of the sources (one light source and two radioactive sources) will be attached permanently to three deployment units. The fourth unit will have a changeable source fixture, providing flexibility for the attachment of special sources. The operation of the system will be fully automated to ensure routine (weekly) deployment of calibration sources with high reliability. Based on the experience from the Daya Bay experiment, a weekly deployment of the light source allows PMT gain calibration to subpercent precision. A weekly deployment of a gamma source, e.g., $^{60}$Co, is sufficient to track the slow drift ($\sim$1\% per year) of the overall energy scale due to changes in detector properties. Lastly, a routine deployment of the neutron source can ensure the stability of the detection efficiency. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.6\textwidth]{Calibration/figs/ACU.png} \par\end{centering} \caption{\label{fig:ACU}A sketch of the ACU system.} \end{figure} \subsubsection{Rope Loop System} The concept of a rope loop system is similar to that used in the SNO experiment, illustrated in Fig.~\ref{fig:Rope-loop-solution}. The source deployment position can be controlled by adjusting the lengths of sections A and B of the rope. In this design, one of the hanging points is through the central chimney, and the other anchor is located at around 30$\circ$ latitude, allowing a theoretically 90\% 2-dimensional coverage of a half vertical plane. \begin{figure}[!htbp] \begin{centering} \includegraphics[height=2in]{Calibration/figs/rope-SphereInside.png} \includegraphics[height=2in]{Calibration/figs/rope-SphereSide.png} \par\end{centering} \caption{\label{fig:Rope-loop-solution}Rope loop solution and its 2-dimensional coverage} \end{figure} A sketch to illustrate a deployment sequence is shown in Fig.~\ref{fig:dep_seq}. With section A going through the central chimney sources can be taken out of the detector and get changed. We are developing a scheme which allows both an automated source change while maintaining the option of manually changing the source via glove box. \begin{figure}[!htbp] \begin{centering} \includegraphics[height=2in]{Calibration/figs/dep_seq.png} \par\end{centering} \caption{\label{fig:dep_seq}3-point hang design and volume coverage} \end{figure} The overall layout of the rope loop system on top of the central detector is shown in Fig.~\ref{fig:rope_layout}. Two independent rope loop systems are being considered, each covering a half vertical plane to allow some control of the azimuthal symmetry of the detector response. Each system has two spool drives A and B to adjust the lengths of the A/B rope in a synchronous fashion. Rope A goes through a pulley, which is attached at an end of an extendable lever arm. During the deployment, the level arm is extended to move the pulley towards the center of the chimney to lower the source into the detector. Once a deployment is completed, the source is retrieved by extending B rope and shortening A rope. Once the source is out, the pulley arm gets retracted to the side so that the source change operation will be performed away from the chimney. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.8\textwidth]{Calibration/figs/rope_layout.png} \par\end{centering} \caption{\label{fig:rope_layout}An overview sketch of the rope loop system.} \end{figure} Alternative designs allowing three-dimensional coverage are also under considerations. For example, if the anchor point is allowed to move on the latitude circle, a 3-dimensional deployment can be made possible. However, with a large diameter circle attached to the inner surface of the detector, maintaining a smooth rail and reliable drive of the anchor point is a non-trivial engineering challenge. Another alternative 4-point rope design is shown in Fig.~\ref{fig:alter_rope_system}. Three hang points, separated by 120$\circ$, are placed around the equator of the central detector. Three ropes, each going through one hanging point and the central calibration port, will connect to the source. By adjusting the length of these three ropes, one can achieve a volume calibration in the bottom half of the sphere. With a fourth rope connecting the source directly from the central calibration port, the volume calibration on the upper sphere can be performed. In general, three dimensional rope system would introduce significantly more mechanical complexity. In such a case, a ROV system described below would become a viable alternative. \begin{figure}[!htbp] \begin{centering} \includegraphics[height=3in]{Calibration/figs/alternative_rope.png} \par\end{centering} \caption{\label{fig:alter_rope_system}Illustration of a 4-rope design.} \end{figure} Details of a single spool drive in the rope loop system are shown in Fig.~\ref{fig:spool}. The spool in the ACU will be under an identical design. The spool will be made either with acrylic or PTFE. Helical grooves with $\sim$mm separation will be machined on the spool, and the deployment cable of less than 1.5~mm diameter will be wound into the grooves without overlap. The rope capacity of each spool is set to be 50~m. Several additional measures will be implemented to avoid rope slipping out of the groove, including a spring loaded Teflon press to constrain the rope, a load cell to constantly monitor the tension in the cable, as well as a co-moving spooling tracker so that the rope always gets unwound perpendicular to the spool. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.6\textwidth]{Calibration/figs/spool.png} \par\end{centering} \caption{\label{fig:spool}An illustration of a spool drive.} \end{figure} To achieve automated source changing, a key component in the rope system is a ``quick connection'' connector, which allows an automated system to attach and detach a source to the rope. One of the options under considerations is illustrated in Fig.~\ref{fig:quick1}. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.6\textwidth]{Calibration/figs/quick1.png} \par\end{centering} \caption{\label{fig:quick1}One of the quick connection option.} \end{figure} The upper piece is attached to the rope loop permanently. We envision that an ultrasonic emitter can be attached to it for accurate positioning with the SS rope being the core of a coaxial cable carrying the electrical signals. The middle piece is the female connector, which is connected permanently to the upper piece with a rotary joint. Gravity will keep the middle piece horizontal to ease automated grabbing. The lower piece is the male connector, permanently attached to the source and the lower weight. The lock between the male and female is achieved with a spring-loaded key. During the attachment (detachment), both the male and female pieces have to be held by mechanical hands and pushed (pulled) relatively. One additional degree of freedom is needed to actuate the key in order to lock or unlock the connector. A manual key is located on the female connector to easy the manual operation through a glove box. An alternative option of the connection is illustrated in Fig.~\ref{fig:quick2}, where the upper piece is the same and omitted in the drawing. The connection is made doubly safe with two spring loaded locks. The head of the male key has a dumb bell shape, which can be pushed in and pulled out of the upper lock with force. The lower lock requires a hand to press the side buttons in order to get unlocked. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.4\textwidth]{Calibration/figs/quick2.png} \par\end{centering} \caption{\label{fig:quick2}An alternative quick connection option.} \end{figure} All sources will be placed in a circular storage ring surrounding the chimney (Fig.~\ref{fig:source_ring}), which can be driven automatically in order to select a source to get attached/detached to the rope loop. Such a design allows the two rope loops to share the same source storage. For illustration, a source assembly placed in the storage is also shown in the figure. For each rope loop, all sources share the same mechanical hand system to perform the source changes. \begin{figure}[!htbp] \begin{centering} \includegraphics[width=0.7\textwidth]{Calibration/figs/source_ring.png} \par\end{centering} \caption{\label{fig:source_ring}The source storage ring.} \end{figure} \subsubsection{Remotely Operated Under-liquid-scintillator Vehicles} A ROV system can deploy the source to nearly everywhere inside the central detector. Such a system is used in the SNO experiment to deploy the $^3$He neutron tubes. A conceptual design of the ROV is shown in Fig.~\ref{fig:ROV}. It can be made into an egg-shaped capsule with a diameter less than 300~mm and a height less than 500~mm. The ROV is driven by pump jet propulsion without external propeller in order to maintain a simple external geometry and a compact size. The ROV motion speed is set to be about 1 m per min, and the actual position is feedback via an ultrasonic emitter. The power as well as signals are transmitted to the ROV via a umbilical cord adjusted to nearly zero buoyancy. The ROV can also carry lighting, CCD camera and a magnet to allow emergency rescue or monitoring the interior of the detector. Radioactive source assembly is attached below the ROV, with a quick connection which would allow automated source change. For material compatibility with LAB, the enclosure of the ROV will be made with Teflon, which is also highly reflective to avoid photon losses. Monte Carlo studies are underway to optimize the ROV geometry in order to minimize impact due to dead materials or optical shadowing. \begin{figure} \begin{centering} \includegraphics[height=3in]{Calibration/figs/ROV.png} \par\end{centering} \caption{\label{fig:ROV}ROV, its umbilical cable and the loaded radioactive source} \end{figure} Based on its current location and destination location, the ROV software will automatically select the optimal trajectory. The pump jet propulsion engine controls the speed and direction of the device. The ROV should approach the designated locations with a speed less than 1~mm/s or less, then shuts off the engine. A piston driven buoyancy adjustment mechanism is used as the depth control. In case of system failure such as power failure, the control will shut down the pump propeller and pull the piston of the depth control via a spring to increase the buoyant force to ensure the ROV float to the surface. \subsection{Conceptual Design of the Guide Tube System} Based on positive experiences of the Double Chooz experiment, a guide tube system is important to understand detector response from anti-neutrino interactions occurring at the boundaries of the central detector volume. Guide tubes will be positioned along the surface of the central detector vessel. The tubes could be installed at the inner and/or outer surfaces of the vessel wall. Figure~\ref{fig:guide_tube} shows the conceptual design of the guide tube system with the tubes at the outer surface (left), and with tubes at the inner surface (right) of the central detector vessel. The number of tubes shown is for illustrative purposes only, and the final number and distribution of the tubes will be based on detailed simulations of the guide tube system and studies of detector energy response, currently under development. The tube can be made with stainless steel to provide a good mechanical strength, or it could be manufactured from an acrylic tube to be compatible with the central detector vessel materials. Radioactive sources as small as a few mm in diameter and $\sim$2~cm in length could be positioned within the tube with use of a stainless steel wire driven by a stepper motor. Figure~\ref{fig:guide_tube1} shows one considered design of source capsule inside the stales steel guide tube, with the source connected on both sides. Similar designs are considered with the acrylic guide tube solution. Source positioning accuracy of $\sim$1~cm is anticipated. \begin{figure}[!htbp] \begin{centering} \includegraphics[height=3.0in]{Calibration/figs/juno-outer-GT.png}\includegraphics[height=3.0in]{Calibration/figs/juno_inner_GT-1.pdf} \par\end{centering} \centering{}\caption{\label{fig:guide_tube}The guide tube system. } \end{figure} \begin{figure}[~!htbp] \begin{centering} \includegraphics[height=3.0in]{Calibration/figs/guide_tube1.png} \par\end{centering} \centering{}\caption{\label{fig:guide_tube1} The conceptual design of the source capsule inside the guide tube.} \end{figure} \subsection{A Pelletron-based Beam Calibration System} A pelletron is one type of electrostatic accelerator. The electric charge is transported mechanically to its high voltage terminals by a chain of pellets, which are connected with insulating materials (such as nylon). Figure~\ref{fig:Diagram-of-Pelletron} shows a diagram of a pelletron, taken from the website of National Electrostatics Corporation (NEC)~\cite{NEC:2014}. Compared with a LINAC, the electrostatic accelerator has the advantage of being more stable. This is crucial for the intended application in JUNO, as the calibration of the Pelletron energy and the JUNO detector response calibration cannot be performed at the same time. Furthermore, below \textasciitilde{}5 MeV in kinetic energy~\cite{Hinterberger:1997ur}, the pelletron solution is compact and economical enough for an underground installation. \begin{figure} \begin{centering} \includegraphics[width=0.8\textwidth]{Calibration/figs/chargingsys} \par\end{centering} \caption{\label{fig:Diagram-of-Pelletron}Diagram illustrating the principle of a pelletron. As the drive pulley and the terminal pulley rotate clockwise, the chain transports the positive charge to the terminal shell and builds up a high voltage that could be higher than 25 MV. By reversing the polarities of the charging voltages, the pelletron can easily switch to accelerate electrons and negative ions.} \end{figure} Expected positron spectra is shown in Fig.~\ref{fig:Espec}. We see that most events are between 1 and 6~MeV prompt energy. Due to finite energy resolution, the mass hierarchy signal only shows up above 2~MeV prompt energy. Therefore, a pelletron system, which can provide 1-5~MeV kinetic energy corresponding to 2-6~MeV prompt energy, can satisfy the requirement of precision energy calibration for mass hierarchy determination. Figure~\ref{fig:sys-diagram} shows the conceptual design of the pelletron system, following the ideas from Refs.~\cite{Bauer:1990zz,Huomo:1988jw}. The high purity Ge (HPGe) detector, which is calibrated by mono-energetic gamma sources, will be used to control the beam energy precisely. A precision of $10^{-4}$, which is well below the 0.1\% goal of the JUNO experiment, has been achieved with existing facilities~\cite{Bauer:1990zz,Berg:1992jr,Huomo:1988jw}. \begin{figure} \begin{centering} \includegraphics[width=0.85\textwidth]{Calibration/figs/calibsys-diagram} \par\end{centering} \caption{\label{fig:sys-diagram} The conceptual design of the pelletron system} \end{figure} \begin{figure} \begin{centering} \includegraphics[height=4.0in]{Calibration/figs/beammonitors-2} \par\end{centering} \caption{\label{fig:endcap-concept}A conceptual design of the beam pipe endcap inside the LS detector. The transparent acrylic pipe and the Mylar window make sure the scintillation light can largely pass through. Two beam monitors near the top and bottom of the beam will allow diagnostics of beam position and area when the beam first enters the detector vertically downward and near the exit point of the beam out of the beam.} \end{figure} The calibration of the JUNO detector with a pelletron requires the delivery of the positron beam into the detector center through an evacuated beam pipe. The positron beam will exit the beam pipe through a window designed to minimize energy losses in the beam while maintaining the vacuum of the beam pipe at the depth of the detector. The design and instrumentation of the pelletron beam pipe inside the detector is an engineering and R\&D challenge. \underline{Detector beam pipe requirements:} The technical requirements for the positron beam pipe inside the detector can be summarized as follows: \begin{itemize} \item Retractable, telescoping beam pipe that can be deployed into the detector during the time of calibration and retracted during normal data taking to avoid shadowing effects in the detector. During normal data taking the beam pipe is stored completely out of the detector volume. \item Beam pipe is compatible with the scintillator. To first order, the exposure of the scintillator to the beam pipe materials is limited to the duration and time period of the calibration. However, one has to account for possible long-term interaction of the scintillator with the beam pipe material after the beam pipe has been retracted from the detector. Excess scintillator may drip back into the detector or lead to unwanted contamination that may be introduced during the time of the next calibration. Compatibility issues include possible interaction with the materials of the beam pipe, damage of the deployment mechanism, leaching out of beam pipe materials into the scintillator, and degradation of the scintillator itself. The amount of beam line material to be deployed into the detector poses a challenge for the purity of the scintillator and to JUNO physics goals that require ultra low backgrounds. \item Beam pipe and its deployment mechanism have to be leak tight against the liquid scintillator to maintain the vacuum inside the beam pipe. \item Beam pipe deployment mechanism is failsafe and can always be retracted out of the detector. Since the beam pipe does interfere with the normal data taking the deployment mechanism must allow the retraction of the beam pipe even under unusual circumstances such as power failure etc. The system has to be tested to be mechanically reliable for the 10+ year lifetime of the experiment. \item It is desirable to be able to deploy the beam pipe to different depths. This would allow calibration of the detector with positrons at different points along the z-axis. The interaction region of the beam with the scintillator detector is defined by the exit point of the beam out of the mylar window. A continuous deployment along the z-axis may not be necessary, discrete steps in z are sufficient. \item Deployment of the beam pipe is automated as much as possible to allow for regular calibration and minimize the demands on the on-site personnel. \end{itemize} Inside the beam pipe instrumentation to monitor the beam position and profile are required to help determine the exit point of the beam and the interaction region of the calibration beam inside the detector. This monitoring instrumentation will also aid in diagnostics of the beam during tuning and setup. One of the challenges of the proposed calibration scheme is that a positron beam has to be delivered some 17~m into the detector center without active steering. We expect that from the entrance point of the beam into the detector region (at the top of the detector) to the point where it exits the beam pipe through the mylar window no active steering components can be used. As a result the beam will spread and its final position at the end of the beam pipe will depend on its direction when it enters the detector and possible divergence. As a result, we obtain the requirements for the monitoring instrumentation inside the detector beam pipe (see Fig.~\ref{fig:endcap-concept}): \begin{itemize} \item \underline{Top beam monitor:} Very precise beam position monitor at the top of the detector where the beam enters the detector to determine its direction and profile into the detector right after the last active magnetic steering and focusing. This beam position monitor can be fixed along the z-axis. It can be flipped or moved into the beam during tuning and setup and removed from the beam path during the beam delivery and calibration. The position resolution requirements for this beam monitor are set by the distance of the beam delivery into the detector and the width of the beam pipe. Depending on the position of the 90-degree bending magnet of the beam into the detector this monitoring setup can still be outside the detector but already in the vertical region of the beam pipe. \item \underline {Exit beam monitor:} A second beam position monitor near the end of the beam pipe that determines the position and area of the beam spot before it exits through the mylar window. Again, this beam monitor may be designed to be movable so that it can be placed in the beam during monitoring and setup and removed (rotated or flipped) during calibration. It may be desirable to make this beam monitor movable along the z-axis to allow for a vertical scan of the beam position. Since this beam monitor will be located towards the end of the telescoping beam pipe all cabling and supplies have to be routed through the retractable beam pipe structure. This be may not required for the beam monitor at the top of the detector. With a beam position monitor near the top and the bottom of the beam pipe it will be possible to diagnose potential issues in the delivery of the position beam. More than two position monitors will likely lead to unnecessary and increased complexity. \end{itemize} In the following, we summarize various factors that can induce a bias in the beam energy calibration. The first one is the energy loss in beam window. At the center of the detector, the pressure is about 3 atm. A 76~$\mu$m Mylar window is more than enough to handle this pressure. The resulting energy loss is about 12~keV and can be calibrated with the HPGe detector. The variation of the energy loss due to curved window is well below the 0.1\% of the prompt energy. The second factor is the shadowing effect of the calibration pipe and endcap. In order to minimize the impact of these two factors, a transparent endcap design using acrylic pipe (Fig.~\ref{fig:endcap-concept}) is proposed. The residual shadowing effect is about 1-2\%. The range of the shadowing effect is due to the kinetic energy range of positrons. A high energy positron would in general travel longer inside the liquid scintillator, thus has less shadowing effect due to smaller solid angle. This shadowing effect can be calibrated by injecting both electron and positron beams into the central detector and a bench setup. The principle is illustrated in the following: \begin{itemize} \item The energy responses of electron and positron are strongly correlated. The positron energy response is essentially the sum of the electron energy response and positron annihilation energy. The latter contains two 0.511~MeV gammas and can be calibrated using dedicated $^{68}$Ge radioactive source. \item The shadowing effect for the positron annihilation energy can be calibrated with the dedicated $^{68}$Ge source deployed together with the calibration tube. \item The shadowing effect can be calibrated by comparing a dedicated bench measurement with the results from the central detector calibration. \item The light pattern observed by JUNO detector provides additional handle for the shadowing effect. \item The obtained shadowing correction for positron ionization can then be applied to obtain the positron energy response together with the shadowing of the annihilation gammas. \item The resulting positron energy non-linearity model can then compared with that of the electron to further validate the energy model. \end{itemize} The following formula summarizes the correction strategy. \begin{equation} E_{e+}^{corr} = (E_{e+}^{rec} - E_{1.022}\cdot S_{gamma}^{shadowing})\cdot S_{ionization}^{shadowing} + E_{1.022} + C_{window}^{eloss}. \end{equation} Here, $E_{e+}^{rec}$ and $E_{e+}^{corr}$ are the energies before and after the corrections, respectively. The $E_{1.022}$ is the annihilation gamma energy for positron at rest, which can be calibrated directly with dedicated positron source. The $S_{gamma}^{shadowing}$ is the shadowing correction for annihilation gamma, which can be calibrated by combining dedicated source with the calibration tube. The $C_{window}^{eloss}$ represents the energy loss correction in the window, it can be calibrated directly with the HPGe detector. The $S^{shadowing}_{ionization}$ represents the shadowing effect for the scintillation light for positron ionization energy. This piece will be calibrated by i) comparing the bench data vs. calibration data vs. MC, and ii) comparing the observed light pattern vs. MC. The initial simulation shows that the residual uncertainty can be controlled to about 0.1\% with this strategy. Simulation of Pelletron calibration has been carried out in the general JUNO simulation framework to demonstrate the feasibility of reaching 0.1\% targeted energy scale uncertainty. First, for a 20 m long calibration tube inside the liquid scintillator, the buoyancy force needs to be taken into account. The left panel of Fig.~\ref{fig:ct_force} shows the required wall thickness of the stainless steel calibration tube to balance the buoyancy force. The right panel of Fig.~\ref{fig:ct_force} shows the weight of the 20~m calibration tube. Figure~\ref{fig:tube} shows the implemented geometry of calibration tube including i) $\sim$17.3~m long stainless steel calibration tube, ii) 30~cm long acrylic endcap, and iii) 7.6 $\mu$m thick curved Mylar window. The energy loss of the electrons and positrons inside the Mylar window was simulated and shown in Fig.~\ref{fig:ct_eloss}. For positrons at off-center locations, the energy loss is slightly higher than those at the center of calibration tube. This is consistent with the fact that the vertical thickness of the window at off-center location is bigger than that in the center. The energy loss of electron is slightly bigger than that of positron due to additional annihilation process that positrons can go through. The difference is much smaller than the targeted 0.1\% energy scale uncertainty band. A beam of positrons are shooting through the calibration tube to study the energy response. The positron beam cross section is assumed to be a circle with radius of 1~cm. The spread in momentum is assumed to be 0.1\%. The results are compared with those of injecting positrons directly inside the detector center without any calibration tube (Fig.~\ref{fig:ct_result1}). The increase of the difference between two cases with respect of the true prompt energy reflects the shadowing effect of the calibration tube. The bias without any correction (top panel of Fig.~\ref{fig:ct_result2}) indicate the magnitude of bias is about 1-2\%. The bias becomes smaller at high energy, because high-energy positrons can penetrate into liquid scintillator further so that the shadowing effect is smaller. The bottom panel of Fig.~\ref{fig:ct_result2} shows the residual bias after the energy loss and shadowing correction. The residual bias can be controlled to below 0.1\%. \begin{figure} \begin{centering} \includegraphics[height=2.5in]{Calibration/figs/calibration_tube_force} \par\end{centering} \caption{\label{fig:ct_force} (Left panel) Calculated required wall thickness of the stainless steel calibration tube is plotted as a function of the radius of the calibration tube to balance the buoyancy force. (Right panel) The corresponding weight of the 20 m long calibration tube is plotted.} \end{figure} \begin{figure} \begin{centering} \includegraphics[height=1.5in]{Calibration/figs/tube1} \includegraphics[height=1.5in]{Calibration/figs/tube2} \includegraphics[height=1.5in]{Calibration/figs/tube3} \par\end{centering} \caption{\label{fig:tube} (Left) Illustration of calibration tube of Pelletron inside the JUNO central detector. (Middle) Curved Mylar window for the calibration tube. (Right) Endcap of the calibration tube with the curved Mylar window. } \end{figure} \begin{figure} \begin{centering} \includegraphics[height=3.5in]{Calibration/figs/eloss} \par\end{centering} \caption{\label{fig:ct_eloss} Energy loss of the electrons and positrons inside the curved Mylar window. For electron, the prompt energy was shifted to match that of positron. The traipzoid represents 0.1\% band of the energy scale.} \end{figure} \begin{figure} \begin{centering} \includegraphics[height=1.6in]{Calibration/figs/ct_result1} \par\end{centering} \caption{\label{fig:ct_result1} (Left) The energy response of positron injecting through the calibration tube (red) is compared to that of the positron in the detector without the calibration tube(black). (Right) The difference in energy is shown with respect of the true positron prompt energy. The increase of difference reflects the shadowing effect of the calibration tube} \end{figure} \begin{figure} \begin{centering} \includegraphics[height=2.5in]{Calibration/figs/ct_result2} \includegraphics[height=2.5in]{Calibration/figs/ct_result3} \par\end{centering} \caption{\label{fig:ct_result2}(Top) The bias of energy response (1-2\%) due to the shadowing effect of the calibration tube is shown with respect of the true prompt energy. (Bottom) The residual bias of energy response ($<$ 0.1\%) after the correction of energy loss and the shadowing effect is shown with respect of the true prompt energy. } \end{figure} \subsection{Diffused Short-lived Isotope Calibration} Homogeneous energy calibration of a large liquid scintillator detector throughout the entire volume can only be achieved with distributed sources in the scintillator. The use of uniformly distributed sources was successfully demonstrated by the Sudbury Neutrino Observatory using $^{16}$N, $^{8}$Li, as well as \emph{in-situ} spikes from $^{24}$Na and $^{222}$Rn. Injection and deployment of such sources in a 20 kt-detector pose challenges including the production and injection of isotopes over long distances and their distribution throughout the detector volume. We will survey suitable short-lived isotopes, their production method and develop concepts for the injection with a gas or scintillator stream into the detector. We will pay particular attention to the possible use of positron sources. We will identify possible isotopes and develop a technical concept for the injection and distribution of sources. The principles of the isotope injection will be studied with the test chamber developed for various R\&D purposes or a re-purposed Daya Bay detector. \section{R\&D Status and Plans} \subsection{Sources} Source selection and conceptual design of the geometry need to be determined via Monte Carlo. Mini-balloon source requires prototype both mechanically as well as test loading of radioactive isotopes. A robust and simple UV laser system needs to be developed allowing $\sim$ns level timing and sub-percent intensity control. \subsection{Rope Loop System} We have learned much from the Daya Bay calibration experiences. The challenges of the rope loop approach lie in the automatic source swapping devices, interfacing with the central detector and the installation procedure. A 1:1 prototype for the mechanical system is required before the final design. We shall also find high bay area to performance deployment tests similar to that in the real detector. Additional key considerations are: \begin{itemize} \item {The rope loop subsystem will be installed after the cleaning of the central detector, or even after LS filling, thus installation procedures should be simple and they must be carried out quickly to avoid contamination;} \item {LS detector with such a large size is unprecedented. Calibration system should be designed so it can accommodate potential structural distortions; and} \item {The reliability of the source changing devices and life span should be understood well with destructive tests.} \end{itemize} \subsection{ROV} To achieve good physics performance, we need to optimize the shape and the dimension of the ROV via extensive simulations. The ROV shall be designed with high cleanliness and be fully compatible with liquid scintillator. Prototype studies are needed in a realistic deployment condition. \subsection{Source Positioning System} An ultrasonic emitter's diameter is roughly 30~mm and its working frequency 150~kHz with a pulse width of $\sim$1 ms. The emitter can be installed either on the source or on the ROV. The receiver array is mounted on the detector sphere and at least three receivers can realize 3D positioning. The actual arrangement of the array need to be optimized. To increase the reliability, we plan to have $\sim$10 receivers so when some malfunction or break down, the system can still function. Longer array baselines means better positioning precision, which implies receivers are better placed near the equator from the perspective of positioning. However, this also means the cables serving these receivers are longer and it might cause more potential issues. The material compatibility between cables and LS needs to be checked and addressed. Compared with the oceanic environment, JUNO central detector is an isolated environment. Impact to positioning uncertainty due to reflections and refraction shall be well studied and understood. \subsection{Pre-installed Guide Tubes} How to mount the guide tubes next to the acrylic panels is quite challenging due to the distortion after the detector is filled. The maximal distortion of the acrylic panel is expected to be $\sim$5~cm thus if the guide tubes are fully attached to acrylic panels, this level of distortion is highly harmful for the smoothness of the tube even the joint does not get destroyed. We need to allow certain level of distortion thus probably the two should not be fully coupled. Instead we should allow the relative movement to some level while still keep the positioning accurate enough. In addition to the mechanical concerns, the impact of these additional, tube materials to the detector performance needs to be understood better. It should not make much difference they are deployed either inside or outside of the acrylic sphere if refraction indices match well with LS and/or mineral oil. Acrylic, stainless steel, and Teflon tubes tubes are considered. Right now, the system realization and impact to the detector performance are undergoing discussion. Argonne National Lab has a high bay area, suitable for performing deployment tests of the rope loop system, and tests of the guide tube systems. \subsection{Pelletron System} The need of space and transportation underground have been checked and no obvious obstacles were observed. Currently, we need to detail out the beam line design and the interfacing with the central detector. \subsection{Diffused Short-lived Isotope Calibration} Various sources are being studied. The concept of source injection is being investigated. \section{System Reliability and Safety Concerns} It would be difficult to make large scale repair to the calibration system during the life span of the experiment. This places very stringent challenges on the reliability and longevity of the calibration system, especially that there should absolutely be no source dropping. For the automatic source swapping component of the rope loop system, we have considered the following multi-layer safety protection mechanism, \begin{enumerate} \item {The position and the clamping force of the robotic hands are constantly monitored after each motion;} \item {Robotic hands and other clamping devices are designed to be normally-close to make sure that sources are still secured during power or control signal loss;} \item {To make sure of the reliability of the quick-connects, they are pull-tested each time they engage;} \item {A fall-prevention shutter is installed beneath the source-swapping device. During source swapping, the shutter is closed; and} \item {A piece of annealed iron will be installed inside the source. In case of dropping, a ROV with a strong magnet can be deployed to recover it.} \end{enumerate} The ROV umbilical cable can take certain load besides functioning as a power and control signal cable. Thus in case of malfunction, it can be pulled out of the central detector. In addition, the design of the ROV also has some built-in safety considerations and the ability of self test. The effective density of the ROV should be roughly 95\% of LAB. During power loss or other situations of losing motion, the ROV would float automatically. The liquid dragging force shall be chosen so that the rising speed is not too high, avoiding damaging the vessel. The internal pressure of the ROV should be either positive or negative and should be constantly monitored. Larger-than-normal fluctuations would indicate a leak might have developed and an alarm signal shall be sent out. \section{Schedule} A major contribution to the calibration system and additional contributions in support of central detector, liquid scintillator and PMT QA are being investigated for US JUNO deliverables. \clearpage \begin{longtable}{|p{1.5cm}|p{4cm}|p{3cm}|p{5cm}|} \caption{Yearly targets} \\ \hline Year & Goals & Deadlines & Milestones \tabularnewline \endfirsthead \multicolumn{4}{r}{(See Next Pages.)} \endfoot \bottomrule \endlastfoot \hline 2013 &Form a working group, start conceptual designs and simulation &Form a few preliminary mechanical designs &Form 3 alternative conceptual mechanical designs before November\tabularnewline \hline 2014 &Continue design and gradually converge; Technique R\&D; Design light sources and build mock-ups &Finish the simulation of source selection; finish the design of a stable light source &Finish the stable light source design before June; Finish the simulation and analyses of the 3 alternative designs; Select 2 out 3 in September\tabularnewline \hline 2015 &Build and test a prototype; finalize both the mechanical and the electronic designs &Finish the R\&D of selecting radioactive sources; finalize the choice of radioactive sources & Finalize the mechanical design before June; finalize the construction and testing of the prototype, and the radioactive source designs before December\tabularnewline \hline 2016 &Bid, order and start the construction of calibration systems & finish one complete calibration system and testing & finish bidding before June and complete one calibration system and finish its testing before December\tabularnewline \hline 2017 &Continue the construction of the calibration system & Complete the remaining half of the project &Complete the remaining half of the project \tabularnewline \hline 2018 &Complete the construction of the calibration system; order all radioactive sources; complete DAQ and the interface with DCS &Compelete the last half of the project & Complete all remaining constructions, DAQ and the interface between DAQ and DCS before December\tabularnewline \hline 2019-2020 &Installation and commissioning &Complete the installation and commissioning of the system & complete the calibration system before the completion of the central detector installation\tabularnewline \hline \end{longtable} \chapter{Central Detector} \label{ch:Central Detector} \section{Introduction and requirement of the central detector} The central detector of JUNO aims to measure the neutrino energy spectrum using $\sim$20~kt liquid scintillator (LS) and $\sim$17,000 PMTs. An inner sphere with a diameter of around $\sim$35.4~m is designed to contain the huge volume of LS, and an outer structure with a diameter of around $\sim$40~m is needed to support the inner sphere as well as PMTs. In order to get the $3\%/\sqrt{E}$ energy resolution, the central detector is required to maximize the collection of optical signals from LS meanwhile minimize the background from a variety of radioactive sources. Since it is not possible to repair during the operation, this detector must have a long life time and a high reliability. Since the LS and PMTs are separately documented in Chapter 4 and Chapter 6, this chapter is mainly focused on the challenges: 1) Construction of the large structure of the sphere and its support; 2) Test and installation of the $\sim$17,000~PMTs; 3) Filling of LS and long term operation of the detector. Throughout the design effort, the strength and stability of the structure is the drive. In addition, the working conditions in the underground hall, safety issues during the construction, cost, and the overall time schedule should be taken into account. The ideal design of the JUNO central detector is a stainless-steel tank plus an acrylic sphere, where the stainless-steel tank is used to separate the shielding liquid (mineral oil or LAB, to be decided) from the pure water in the water pool, and the acrylic sphere is to hold the $\sim$20~kt LS, as indicated in Fig.~\ref{fig:intro:det}. Due to the limitation of space, it is very difficult to simultaneously construct the two spheres and there is also a high risk and significant schedule delay if the two spheres were constructed in series. As a result, alternative options could be: 1) the stainless-steel tank is replaced by an open space truss or other steel supporting structures; 2) the acrylic sphere is replaced by an off-site fabricated balloon. Currently, the baseline is the first one, i.e. the so called acrylic sphere plus stainless-steel support option, which will be detailed in this document. The second option of the balloon plus stainless-steel sphere is a backup, and will be described also in this chapter. The main requirements for the JUNO central detector are the following: 1) The detector should meet the physics requirements of the JUNO experiment, with $\sim$20~kt high purity LS and $\sim$17000 high Q.E. PMTs to reach an energy resolution of $3\%/\sqrt{E}$; 2) The detector should minimize the radioactive background from different sources, including environment, structure, materials and the pollution from the construction process; 3) The detector structure should be reliable and its design should meet the standards in fields of large vessel, civil architecture and engineering. Leakage is not allowed between different spheres. The structure should be safe up to a seismic intensity of Richter scale 5.5, and should not be sensitive to temperature variation. The structure must have no single point of failure; 4) Materials used for the detector should have a long-term compatibility (about $\sim$30~years) with the liquid scintillator and pure water; 5) The central detector should provide a proper interface to other systems, i.e. VETO, calibration, electronics and so on; 6) The total construction duration should be reasonably short and not longer than 18~months; 7) The lifetime of the detector should be longer than 20~years, and during this period, no significant repair is needed; 8) The cost of the central detector should also be at a reasonable level. \section{Baseline option: Acrylic Sphere with Stainless-Steel Support} \subsection{Introduction and the design requirements} In the baseline design the inner sphere is made of acrylic with a thickness of $\sim$12~cm and an inner diameter of 35.4~m. Surrounding the acrylic sphere there is a stainless-steel structure which supports the acrylic sphere and also the PMTs. Currently the stainless steel support structure has three options: the double layer truss, the single layer truss and the equator supporting method. For the first two options, the truss can also be used to fix the PMTs, but for the equator supporting method, an extra frame is needed for PMT installation. The inner diameter of the stainless-steel truss is ~40~m in the present design, supported by a number of columns which are built on the base of the water pool. For the acrylic sphere, a chimney with an inner diameter of $\sim$1~m will extend up from the top of the sphere, serving as the interface for the calibration system. The relative pressure between the LS and the water has a great effect on the sphere. The chimney will be a few meters higher than the water level in order to provide flexibility in setting the LS relative height and the resulting stress in the sphere. Due to the height of the chimney above the sphere, optical isolation is needed to block the background induced light from being detected by the PMTs. Between the acrylic sphere and the truss, there are about $\sim$17000~ inward-facing PMTs to collect optical signals produced by the LS. There is an opaque layer behind the PMTs to separate the central detector from outside veto detector. As an example, Fig.~\ref{fig:cd3-2} shows the acrylic sphere and the double layer stainless-steel truss. \begin{figure}[!htbp] \begin{center} \includegraphics[width=15cm]{CentralDetector/figures/CD3-2.png} \caption[Logo in CentralDetector]{Schematic view of acrylic sphere plus stainless-steel double layer truss} \label{fig:cd3-2} \end{center} \end{figure} Many factors need to be considered for the baseline design, such as: $\bullet$ Stability and load status of the structure at different stages, such as during the construction, after the construction and installation of PMTs, during LS and water filling, and long-term operation of the detector. $\bullet$ The acrylic sphere under different conditions such as earthquake, temperature variation and so on. $\bullet$ The maximum stress on the sphere for long term operation < 5~MPa and for shorter durations < 10~MPa. $\bullet$ Structure of the stainless-steel support should be reliable, and should meet the specifications of related fields. $\bullet$ The joint of the acrylic and stainless steel should be reliable, and the impact induced by local failures should be minimal. \subsection{Stainless-steel Supporting Structure} The acrylic sphere is the most critical part of the whole central detector and is supported by the stainless-steel structure on its outer sphere. The design of the joint of the acrylic and stainless steel structure is critical and the ball head connection for flexibility and a rubber layer for buffering are adopted in the design. Currently, there are three design options under consideration for the stainless steel structure: double layer truss, single layer truss and equator support. The stainless steel is chosen to be the 316 type and the main parameters are the following: density of 8.0~$g/cm^2$, elastic modulus of 200~GPa, poisson ratio of 0.33 and the yield stress of 240~MPa. \subsubsection{Double Layer Stainless-Steel Truss} The space truss has been a popular choice in the civil construction domain given its properties of light-weight, small-size truss member, ease of handling and transportation, high rigidity, short construction time, low cost and good seismic resistance,etc. For the central detector, the truss is selected to be the outer structure which will not only support the acrylic sphere but also the PMTs. For the design of the truss, load-carrying capacity and stability were considered under different conditions. The truss will be the square pyramid space grid which is a multiple statically indeterminate structure. This type of truss can withstand the load coming from different directions, and has better load-carrying capability than the plane truss. The truss members have good regularity and high stiffness, and they are connected to each other by bolt-sphere joints which transmit only axial compression or tension, hence there are no moments or torsional resistance. Since the truss can be built up from simple, prefabricated units of standard size and shape which will be mass-produced industrially, the units can be assembled on-site easily and rapidly which greatly reduces the time for construction. With the present design, the diameter of the inner layer truss is $\sim$38.5~m, while that of the outer layer is $\sim$42.5~m at the two poles and $\sim$40.5~m at the equator. This shrinkage from poles to equator can significantly reduce the size of experimental hall hence the cost and time for civil construction. There are hundreds of supporting rods between the acrylic sphere and the inner layer of the truss. These rods will be connected to the truss at one end and inserted into the acrylic sphere at the other end. The acrylic sphere is supported by those rods directly. To reduce the axial load on each rod and hence the stress in acrylic, some other rods are added between the sphere and the outer layer of the truss. The truss itself will installed on the bottom of the water pool by supporting columns at the lower hemisphere. Fig.~\ref{fig:cd3-3} shows the truss structure and some of its details. \begin{figure}[!htbp] \begin{center} \includegraphics[width=12cm]{CentralDetector/figures/CD3-3.png} \caption[Logo in CentralDetector]{Double layer stainless-steel truss} \label{fig:cd3-3} \end{center} \end{figure} According to JGJ7-2010, which is the technical specification for space frame structures, the truss members will be designed as the compression members and the slenderness ratio should be less than 150. Referring to the preliminary finite element analysis, the size of the chord members is selected to be $\phi$273~mm $\times$ 8~mm, and the ventral members are $\phi$219~mm $\times$ 8~mm. The supporting rods (brace members) between acrylic and truss are $\phi$102~mm $\times$ 12~mm, and the columns for supporting truss are designed to be $\phi$400~mm $\times$ 20~mm. The final size will be determined in engineering design after further stress analysis. \subsubsection{Single Layer Stainless-Steel Truss} As shown in Fig.~\ref{fig:subfig:siglelayer} (a), this single layer truss is made of I-shaped unistrut in both longitudinal and latitudinal directions. Similar to the double layer truss, the supporting rods are also used to connect the sphere to the truss. The truss itself is supported on the base of the water pool by a number of columns. To improve the stability of the single layer truss and to avoid the possible torsion, a ring of spiral bracings are added in the truss grids to prevent any occurrence of torsional vibration shape. In addition, due to space limitation in the pole region of the truss, the square shaped structure is replaced by a triangle shaped one, so the number of truss members is reduced, as sketched in Fig.~\ref{fig:subfig:siglelayer}(b). This optimization gives more space for PMT installation and keeps the grid size of the truss in a reasonable range. \begin{figure}[!htbp] \centering \subfigure[Single layer stainless-steel truss with spiral support]{ \label{fig:subfig:a} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-4a.png}} \subfigure[Reduction of the number of truss members in pole region]{ \label{fig:subfig:b} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-4b.png}} \caption{Single layer steel truss} \label{fig:subfig:siglelayer} \end{figure} Compared to the double layer truss, the single layer truss can save a significant amount of space and hence the civil construction cost. Another advantage of this option is that PMT installation is much easier since there is no interference caused by the truss members. \subsubsection{Equator Supporting Method} This supporting structure is indicated in Fig.~\ref{fig:subfig:equator}. The acrylic sphere is supported at the equator area with stainless-steel rings which connect to the wall or bottom of the water pool by rods. A radial extension growing from the sphere at the equator functions as supporting points where the stainless-steel rings are clamped to it. A rubber layer is used as buffer between the acrylic and steel ring to avoid any stress concentration, and this buffer layer is also very useful to improve the stability of the whole structure, especially under seismic conditions. The rings are separated into several parts along the circumference, each part anchored to the water pool by an H-shaped unistrut and a tilted rod. \begin{figure}[!htbp] \centering \subfigure[Overall view of acrylic sphere supported at equator]{ \label{fig:subfig:a} \includegraphics[width=7.5cm,height=6.5cm]{CentralDetector/figures/CD3-5a.png}} \subfigure[Detailed view of the support at equator by steel ring]{ \label{fig:subfig:b} \includegraphics[width=7.5cm,height=7.5cm]{CentralDetector/figures/CD3-5b.png}} \caption{Acrylic sphere supported at equator} \label{fig:subfig:equator} \end{figure} To minimize the light blocking by the steel ring, another method is using a number of separated supporting points which are distributed at the equator. In this design there are 32~blocks cast into the sheets that are assembled at the equator. Into each of these blocks will be cast a steel plate. The sphere is supported at each of these steel plates so there is no direct bearing on the acrylic. Fig.~\ref{fig:subfig:equator-design} shows details of this concept. The 32 support points on the sphere will be supported by an external steel structure that is made of columns braced against the concrete wall. A beam will project forward from the columns and through the layer of PMTs which will surround the sphere. At the end of this beam a plate is welded to the profile. The height and depth of the "C" will be much larger than the plate that is cast into the acrylic sphere. The resulting gap between the plate and the "C" accommodates the assembly tolerances. The sphere will be built up from the bottom using temporary supports until the equator ring with the 32 support points is completed. The external steel structure will be in place and the "C" plate will be surrounding the plate cast into the sphere. At this point shims will be placed around the "C" which will lock the cast plate into place. Once shims have been installed on all 32 support points the bottom temporary supports can be removed and the entire load of the sphere will be transferred to the "C" plates and external steel structure. The top half of the sphere can then be assembled. \begin{figure}[!htbp] \centering \subfigure[]{ \label{fig:subfig:a} \includegraphics[width=7.5cm,height=7cm]{CentralDetector/figures/CD3-6a.png}} \subfigure[]{ \label{fig:subfig:b} \includegraphics[width=7.5cm,height=7cm]{CentralDetector/figures/CD3-6b.png}} \caption{Detailed view of the support at equator by separated points} \label{fig:subfig:equator-design} \end{figure} \subsection{The Acrylic Sphere} The inner structure of the central detector is a transparent acrylic sphere with $\sim$35.4~m inner diameter and 120~mm in shell thickness. This acrylic sphere will be assembled and bonded with bulk polymerization by a number of acrylic sheets, as shown in Fig.~\ref{fig:cd3-7}. Considering the production capacity and transportation limit, the acrylic shell will be divided into more than 170 sheets, each is about 3~m~$\times$~8~m in dimension. The final number and size of the sheet may be modified after further consideration and discussion with manufacturers. To reduce the stress on the sphere at the supporting points, some appended acrylic pieces will be bonded on top of the sheets. The supporting structure of the sphere will be connected to these appended acrylics. A chimney of about $\sim$1~m diameter is designed on top of the acrylic sphere, which will be used as the filling port and interface to the calibration system. An outlet may be designed on the bottom of the sphere for cleaning of the sphere and LS recycling during detector running. \begin{figure}[htp] \begin{center} \includegraphics[width=10cm,height=11cm]{CentralDetector/figures/CD3-7.png} \caption{Schematic view of the acrylic sphere} \label{fig:cd3-7} \end{center} \end{figure} \subsection{Joint of the Acrylic and Truss} In the baseline design for the truss the concept of the joint between the acrylic sphere and the steel truss comes from the idea of the connecting structure used in the glass curtain wall. For each acrylic sheet, there will be one or two stainless steel disks which are embedded in it as a connecting structure, as shown in Fig.~\ref{fig:cd3-8}. At the location of each steel disk, there is an appended acrylic piece with $\sim$100~mm thickness on top of it and bonded to the sphere, and a rubber layer is placed between the acrylic and steel for buffering. The steel disk is designed to have a ball-type head, and a supporting rod will connect it to the truss. The load on the acrylic sphere will be transferred to the truss through these rods. The dimensions of the appended acrylic and the steel disk may be changed according to a further analysis of the stress. \begin{figure}[!htbp] \begin{center} \includegraphics[width=14cm]{CentralDetector/figures/CD3-8.png} \caption[Logo in CentralDetector]{Supporting joint of the acrylic sphere and steel truss} \label{fig:cd3-8} \end{center} \end{figure} In addition, there is another type of joint structure under design, as shown in Fig.~\ref{fig:cd3-9}. Instead of using the steel disk which is heavy and of large size, a steel ring is embedded into the acrylic, and connected to a steel plate on top of the appended acrylic by bolts. Rubber is placed in between the different components, and the bolt is surrounded by plastic bushing to avoid direct contact with the acrylic. This new structure is easier for construction and also can reduce the amount of stainless steel and hence the radioactive background. \begin{figure}[!htbp] \begin{center} \includegraphics[width=14cm]{CentralDetector/figures/CD3-9.png} \caption[Logo in CentralDetector]{Structure of the joint with bolt} \label{fig:cd3-9} \end{center} \end{figure} Currently the two design options are both under optimization, with a main goal of reducing the stress level to less than 5~MPa. Several prototype tests have been finished for the first option, and test for the second option will be done in the near future. \subsection{Finite Element Analysis and Testing} \subsubsection{Global FEA} We have finished a global FEA for the central detector under the following conditions. 1) sphere is empty after the completion of acrylic sphere and steel truss: the load is just the self-weight of the structure. 2) All PMTs are installed on the truss: the weight of total $\sim$17000 PMTs is about $\sim$1700 KN, which will load on the truss. 3) Sphere is filled with liquid and running for long term: the linear liquid pressure will be loaded on the sphere, and the total buoyance of the sphere is about $\sim$3100~t; the PMT's buoyancy is about $\sim$1000~t, which will be distributed on the joint nodes of the truss. During the analysis a load factor of 1.35 was used for the dead load. The FEA results include stress, deflection and stability for the three loading conditions. The effect caused by seismic load, temperature or relative liquid level difference has also been analyzed. Following are the main conclusions: (1) Stress and deflection From Table~\ref{table1}, Fig.~\ref{fig:cd3-10} to Fig.~\ref{fig:cd3-12}, we can see that the maximum stress on the acrylic sphere is $\sim$8.5~MPa in condition (3). If the load factor of 1.35 is removed, the stress is $\sim$6.3~MPa. The maximum displacement is $\sim$35~mm, occurring at the bottom of the structure, which is only $\sim$1/1000 of the sphere span and meets the technical specification for space frame structures (JGJ7-2010). \begin{table}[!htbp] \centering \caption{Maximum stress and deformation for each loading condition\label{table1}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{3cm}<{\centering}|p{3cm}<{\centering}|p{3cm}<{\centering}|p{3cm}<{\centering}} \hline Loading condition & Max. stress on sphere(MPa) & Max. stress on truss(MPa) & Max. general deflection (mm) \\ \hline \tabincell{c}{Condition 1} & 1.1 & 23.7 & 8.2 \\ \hline \tabincell{c}{Condition 2} & 1.3 & 16.3 & 6.3 \\ \hline \tabincell{c}{Condition 3} & 8.5 & 83.4 & 35.1 \\ \hline \end{tabular} \end{table} \begin{figure} \centering \subfigure[condition 1]{ \label{fig:subfig:a} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-10a.png}} \subfigure[condition 2]{ \label{fig:subfig:b} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-10b.png}} \subfigure[condition 3]{ \label{fig:subfig:c} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-10c.png}} \caption{Stress distribution on the acrylic sphere} \label{fig:cd3-10} \end{figure} \begin{figure} \centering \subfigure[condition 1]{ \label{fig:subfig:a} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-11a.png}} \subfigure[condition 2]{ \label{fig:subfig:b} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-11b.png}} \subfigure[condition 3]{ \label{fig:subfig:c} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-11c.png}} \caption{Stress distribution of the steel truss} \label{fig:cd3-11} \end{figure} \begin{figure} \centering \subfigure[condition 1]{ \label{fig:subfig:a} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-12a.png}} \subfigure[condtion 2]{ \label{fig:subfig:b} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-12b.png}} \subfigure[condition 3]{ \label{fig:subfig:c} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-12c.png}} \caption{Deflection of the central detector} \label{fig:cd3-12} \end{figure} (2) Analysis of the seismic load Based on the local geological condition, a seismic fortification intensity at Richter scale 5.5 was considered, and a seismic load of 0.1~g was taken in our analysis. Adding 0.1~g seismic load into the finite element model, the detector structure was analyzed for condition 3 and the result is shown in Fig.~\ref{fig:cd3-13}. We can see that the stress on the acrylic sphere is increased by only $\sim$1.6\% and on the steel truss by only $\sim$4.4\%. The small change shows that the structure of the detector is safe under the seismic load. \begin{figure} \centering \subfigure[Stress on the acrylic sphere under seismic condition]{ \label{fig:subfig:a} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-13a.png}} \subfigure[Stress on the steel truss under seismic condition]{ \label{fig:subfig:b} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-13b.png}} \caption{Stress on the acrylic sphere and steel truss under seismic condition} \label{fig:cd3-13} \end{figure} (3)Stability analysis The safety factor for the buckling of the acrylic sphere is increased by the support from the steel truss. The double non-linearity of the material and the geometry were taken into account for the global stability analysis, and the first elastic buckling mode of the structure was taken as the initial imperfection in calculation. Fig.~\ref{fig:cd3-14}(a) shows the results of the instability mode. Fig.~\ref{fig:cd3-14}(b) shows the deflection-load coefficient curve and we can see from it the stability factor is 2.61, which meets the specification of JGJ7-2010. \begin{figure} \centering \subfigure[Instability mode]{ \label{fig:subfig:a} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-14a.png}} \subfigure[Deflection-load coefficient curve]{ \label{fig:subfig:b} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-14b.png}} \caption{Stability analysis of the central detector} \label{fig:cd3-14} \end{figure} (4) Effect of liquid level difference Liquid level of the LS in the acrylic sphere will affect the stress status of the sphere. We did the analysis for several cases of the relative liquid level. A difference of -2~m, -1~m, 0~m, 1~m, 2~m and 3~m between the LS and water( LS level - water level) were taken into account. The results are shown in Table \ref{table2}. We can see that the variation of the relative liquid level impacts stress significantly. If the LS level in the sphere is higher than the level of water outside, the stress will be reduced significantly. Raising the LS level may be considered in our final design. \begin{table}[!htbp] \centering \caption{Impact analysis of liquid level\label{table2}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{4cm}<{\centering}|p{4cm}<{\centering}|p{4cm}<{\centering}} \hline Liquid level difference H(m)& Maximum stress on sphere (MPa) & Stress on supporting rod (MPa) \\ \hline \tabincell{c} {-2} & 10.2 & 67.7 \\ \hline \tabincell{c} {-1} & 9.4 & 61.6 \\ \hline \tabincell{c} { 0} & 8.4 & 57.6 \\ \hline \tabincell{c} { 1} & 7.4 & 53.7 \\ \hline \tabincell{c} { 2} & 6.4 & 49.7 \\ \hline \tabincell{c} { 3} & 5.6 & 46.5 \\ \hline \end{tabular} \end{table} (5) Analysis of temperature impact The detector will be constructed in a hall $\sim$728~m below the ground and the yearly temperature variation is rather small. Since the thermal expansion coefficients of the stainless steel and acrylic is different, the impact of temperature on the stress still needs to be considered. The impact of increasing or decreasing temperature by 10$^{\circ}$C has been analyzed. Table \ref{table3} shows the results. For the stainless steel truss, the stress is always less than 100~MPa and the safety level will not be affected by temperature change. For the acrylic sphere, temperature elevation is helpful while decreasing of temperature will lead to a larger stress.So avoiding temperature decrease is necessary during detector running period. \begin{table}[!htbp] \centering \caption{Temperature impact on stress\label{table3}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{5cm}<{\centering}|p{3.5cm}<{\centering}|p{3.5cm}<{\centering}} \hline Condition& Maximum stress on sphere (MPa) & Maximum stress on truss (MPa) \\ \hline \tabincell{c} {Room temperature} & 8.5 & 83.4 \\ \hline \tabincell{c} {Room temperature +10$^{\circ}$C} & 8.0 & 77.5 \\ \hline \tabincell{c} {Room temperature -10$^{\circ}$C} & 10.1 & 95.0 \\ \hline \end{tabular} \end{table} (6) Failure analysis The central detector is required to have a life-time of at least 20 years, and any crucial failure is not allowed during this period. For the acrylic sphere and steel truss, the key component is the support points on the sphere. The effect of one or more failure (for example, if a rod is broken) was evaluated. Four kinds of possible failures were analyzed, which are: $\bullet$ condition 1: failure appeared at the maximum stress point; $\bullet$ condition 2: failure appeared at all points on the same latitudinal layer where the maximum stress is located; $\bullet$ condition 3: failure appeared at the points located in one longitudinal band in the lower hemisphere; $\bullet$ condition 4. failure appeared at several points randomly distributed. The analysis results of those cases are shown in Fig.~\ref{fig:cd3-15}. \begin{figure}[!htbp] \centering \includegraphics[width=14cm]{CentralDetector/figures/CD3-15.png} \caption{Failure analysis of the central detector for (a) condition 1, (b)condition 2, (c)condition 3 and (d)condition 4.} \label{fig:cd3-15} \end{figure} In the four failure cases, the maximum stress on the acrylic sphere is 10.8~Mpa (8~Mpa in dead load), an increase of 27.3\% after failure. In fact, the possibility is very small for failures appearing at the same latitudinal or longitudinal band. One point failure is the most probable case and the stress variation is 7\% in this case. The above analysis shows that the risk to the central detector is controllable. \subsubsection{Local FEA of the Supporting Point of the Acrylic Sphere} The grid size of the FEA model affects the accuracy of the calculation, especially for a singular point and the area around it. The above analysis shows that the maximum stress of the acrylic sphere appears at the joint of the stainless steel supporting rod and the acrylic sphere. Since this node is an intersection of the beam element and the shell element, it's a typical singular point. For FEA, a local fine model around the singular point is often employed in order to improve the accuracy. A validation test is necessary to check the analysis. Fig.~\ref{fig:cd3-16} shows the structure of the connecting point between the acrylic sphere and stainless steel truss. The structure at this joint involves the base acrylic shell, the appended acrylic piece and the stainless steel part. \begin{figure}[!htbp] \centering \includegraphics{CentralDetector/figures/CD3-16.png} \caption{the structure of the connecting joint} \label{fig:cd3-16} \end{figure} For the local finite model with 3D solid elements, the base acrylic shell is 2.6~m $\times$ 2.6~m with a thickness of 120~mm, the appended acrylic is 100~mm in thickness and 900~mm in outer diameter, and the diameter of the stainless-steel disk is about 400~mm. The axial load of 14~t in the stainless-steel rod was taken into account for local FEA. Two different boundary conditions around the base acrylic were applied for analysis respectively, one being a fixed constraint and the other a simple constraint. Fig.~\ref{fig:cd3-17} shows the analysis results of the fixed constraint and Fig.~\ref{fig:cd3-18} the results of the simple constraint. \begin{figure} \centering \subfigure[stress distribution on the base acrylic. (Max: 4.2~MPa)]{ \label{fig:subfig:a} \includegraphics[width=7cm,height=6cm]{CentralDetector/figures/CD3-17a.png}} \subfigure[stress distribution on the appended acrylic. (Max: 7.2~MPa)]{ \label{fig:subfig:b} \includegraphics[width=7cm,height=6cm]{CentralDetector/figures/CD3-17b.png}} \subfigure[stress distribution on the steel disk. (Max: 68.4~MPa)]{ \label{fig:subfig:c} \includegraphics[width=7cm,height=6cm]{CentralDetector/figures/CD3-17c.png}} \subfigure[deformation distribution on the whole prototype. (Max: 2.6~mm)]{ \label{fig:subfig:c} \includegraphics[width=7cm,height=6cm]{CentralDetector/figures/CD3-17d.png}} \caption{Analysis results for fixed constraint} \label{fig:cd3-17} \end{figure} \begin{figure} \centering \subfigure[stress distribution on the base acrylic. (Max: 11.6~MPa)]{ \label{fig:subfig:a} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-18a.png}} \subfigure[stress distribution on the appended acrylic. (Max: 7.9~MPa)]{ \label{fig:subfig:b} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-18b.png}} \subfigure[stress distribution on the steel disk. (Max: 69.8~MPa)]{ \label{fig:subfig:c} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-18c.png}} \subfigure[deformation distribution on the whole prototype. (Max: 4.1~mm)]{ \label{fig:subfig:c} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-18d.png}} \caption{Analysis results for simple constraint} \label{fig:cd3-18} \end{figure} The results show that there is no large stress difference for the appended acrylic and stainless steel disk under the two boundary conditions for the base acrylic. The maximum stress in the simply supported model is 63.6\% higher than that for the fixed constraint. The main reason for this is the stress concentration at boundary line due to the applied constraint. Removing these singular points, the stress difference for the two conditions is going to be small. For the central detector, the real loading status is between the fixed and simple constraint, so the maximum stress on the acrylic should be less than 10~Mpa. \subsubsection{Prototype and Testing of the Supporting Structure of the Acrylic Sphere} A prototype of the supporting structure was constructed to study the technique of fabricating the acrylic with steel built-in. Tests of the prototype were performed to understand the load-carrying capacity of the structure and to measure the stress and deformation to check FEA results. The prototype and the test can provide a reference and useful parameters for further design. Fig.~\ref{fig:cd3-19} shows the prototype and some pictures taken during the test. \begin{figure} \centering \subfigure[the joint prototype]{ \label{fig:subfig:a} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-19a.png}} \subfigure[Installation of the prototype]{ \label{fig:subfig:b} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-19b.png}} \subfigure[Preparation of the prototype test]{ \label{fig:subfig:c} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-19c.png}} \subfigure[load-carrying test of the prototype]{ \label{fig:subfig:d} \includegraphics[width=7.5cm]{CentralDetector/figures/CD3-19d.png}} \caption{the prototype and test} \label{fig:cd3-19} \end{figure} The stress and deformation on the measured points in this test are shown in Fig.~\ref{fig:cd3-20}. The measured points are distributed in the edge of the base acrylic, the border between the base acrylic and the appended acrylic, and the opening area of the appended acrylic. Fig.~\ref{fig:cd3-21} is the load-strain curve of each measured point, which shows that all points have good elastic characteristics. All the measured strain data were analyzed to get the stress status of the prototype. \begin{figure} \begin{minipage}[t]{0.5\textwidth} \centering \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-20.png} \caption{layout of the measuring points} \label{fig:cd3-20} \end{minipage} \begin{minipage}[t]{0.5\textwidth}\centering\includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-21.png} \caption{load-strain curve of the measurement points} \label{fig:cd3-21} \end{minipage} \end{figure} Under the axial load of 14~t, the maximum stress of the measured points is 8.6~Mpa which is located on the top of the appended acrylic. The measured stress was compared with the result from FEA as shown in Fig.~\ref{fig:cd3-22}. The deviation is less than 8\%, which demonstrates that the FEA is reasonable and can simulate the real stress on the joint by using the local refinement method. \begin{figure}[!htbp] \centering \includegraphics{CentralDetector/figures/CD3-22.png} \caption{Comparison of stress between measurement and FEA} \label{fig:cd3-22} \end{figure} The above prototypes and tests have verified that the supporting structure of the sphere is safe under 14~t load as well as the reliability of FEA. Since defects existed in fabrication of those prototypes(such as polymerization not being completely finished due to the too low working temperature, the ball joint not working as expected due to being accidentally glued to the appended acrylic), a third prototype has been designed and tested with the known problems fixed. In addition, to test the maximum load-carrying capacity of the joint structure, an orthogonal load was applied to the prototype. The new prototype and test are shown in Fig.~\ref{fig:cd3-23}. \begin{figure} \centering \subfigure[the third prototype of joint]{ \label{fig:subfig:a} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-23a.png}} \subfigure[Test of the third prototype]{ \label{fig:subfig:b} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-23b.png}} \subfigure[The crack pattern of the third prototype]{ \label{fig:subfig:c} \includegraphics[width=7cm,height=6cm]{CentralDetector/figures/CD3-23c.png}} \caption{Third prototype test} \label{fig:cd3-23} \end{figure} Given the improvements in the new prototype, the maximum load that the prototype can bear is up to 51 tons, which is significantly larger than the 28 tons obtained from the previous prototypes. Since the maximum design load is 14 tons, a 3.6 safety factor is reached by the third prototype. Additionally, the measurements confirm the FEA results, and the maximum stress is about 8.4~MPa, as shown in Table~\ref{table4}: \begin{table}[!htbp] \centering \caption{Comparison of FEA and measurement for the third prototype\label{table4}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{2.5cm}<{\centering}|p{3.5cm}<{\centering}|p{2cm}<{\centering}|p{2cm}<{\centering}|p{2cm}<{\centering}} \hline Load & Properties for comparison & Measured point & data & FEA\\ \hline \tabincell{c}140KN & Stress/MPa & A7-3 & 8.467 &8.392 \\ \hline \tabincell{c}{ } &{ } & B7-3 & 3.479 & 2.963 \\ \hline \tabincell{c}{ } &{ } & C1-3 & 3.262 & 3.093 \\ \hline \tabincell{c}{ }& Displacement /mm& W4-3 & 0.512 & 0.630 \\ \hline \end{tabular} \end{table} Since we require that the maximum stress on the acrylic should not be larger than 5~MPa, optimization of the supporting structure is still needed to further reduce the stress and get a larger safety factor. \subsubsection{Optimization of the Detector Structure} From the global and local FEA of the detector structure as well as the prototype test, we can see that the maximum stress in the acrylic shell (not including the appended acrylic) is less than 5~MPa and meet our requirement, but stress in the joints (on the appended acrylic) is still larger than 8~MPa and further optimization is needed. The optimization and modification can be done by: $\bullet$ reducing the axial load on the connecting rod; $\bullet$ raising the LS liquid level in the sphere; $\bullet$ improving the uniformity of the whole structure to get a better loading status; $\bullet$ reducing the number of connecting rods in the top and bottom area of the sphere to simplify the construction (see Fig.~\ref{fig:cd3-24}). The results after those optimizations are shown in Table~\ref{table5}. \begin{figure} \centering \subfigure[Inner layer truss before sparsifying the truss members]{ \label{fig:subfig:a} \includegraphics[width=6cm,height=6cm]{CentralDetector/figures/CD3-24a.png}} \subfigure[Inner layer truss after sparsifying]{ \label{fig:subfig:b} \includegraphics[width=7.4cm,height=6cm]{CentralDetector/figures/CD3-24b.png}} \caption{Inner layer truss after sparsifying the truss members} \label{fig:cd3-24} \end{figure} \begin{table}[!htbp] \centering \caption{Results from FEA after optimization\label{table5}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{3cm}<{\centering}|p{2.7cm}<{\centering}|p{2.7cm}<{\centering}|p{1.3cm}<{\centering}|p{1.3cm}<{\centering}|p{2.5cm}<{\centering}} \hline Conditions & Max-stress on sphere(Singular point excluded) /MPa& Max-stress on sphere(Singular point included) /MPa & Max-stress on truss /MPa & Max-axial load /KN & Max-displacement /mm\\ \hline \tabincell{c}{Optimization 1 }& 9.6 &6.4 & 74.7 & 164.7 & 47.4 \\ \hline \tabincell{c} {Optimization 2 }& 6.9 &5.2& 115.1 & 135.9 & 27.3\\ \hline \tabincell{c} {Optimization 3 }& 6.5 &4.9& 91.3 & 126.1 & 28.4\\ \hline \tabincell{c} { No 1.35 factor}& 4.8& 3.6 & 67.6 & 93.4 & 21.0 \\ \hline \end{tabular} \end{table} Note for Table~\ref{table5}: $\bullet$ In all conditions, the liquid level difference is $\sim$3~m; $\bullet$ Before optimization: the thickness of the upper acrylic hemisphere is 8~cm and the lower hemisphere is 12~cm. The number of supporting rods is 503 and no support comes from the outer layer of the truss to the upper hemisphere. $\bullet$ Optimization step 1: the thickness of the sphere is changed to 10~cm. The number of the supporting rods is increased to 646 and symmetrically placed on the sphere; $\bullet$ Optimization step 2: the supporting rods at both poles are sparsified, and the total number is 610. From Table~\ref{table5}, we can see the stress in the acrylic is 3.6~MPa if the support point is excluded, and the maximum axial load in the supporting rods is only 93.4~KN. Since the stress is proportional to the load, it means that the stress should be also reduced a lot, and the FEA modeling to confirm this is currently underway. \subsection{Study of the Acrylic Properties} Acrylic is the main material used for the central detector construction. Since the detector will involve LS, LAB and high-purity water during long-time operation, the acrylic materials in the liquids should be studied. At present, a number of tests have been finished or are in progress. (1) Normal mechanical properties test Referring to the standards of ASTM-D638, ASTM-D695 and ASTM-D790, the tensile strength, compressive strength and flexural strength of the acrylic had been tested and the results are 68.2~MPa, 108.3~MPa and 88.0~MPa, respectively. Since the thickness of the acrylic sphere is 120~mm, it needs 2 or 3 layers of thinner sheet to be bonded together. In order to know the strength after bonding, we also did the test for the double-layer material. The test result is 66.8~MPa for the tensile strength and 101.7~MPa for the flexural strength. (2) Aging test in LS The acrylic samples were submerged into LS at temperatures of 50$^{\circ}$C, 60$^{\circ}$C, and 70$^{\circ}$C. The testing time lasted 30, 60, 120, and 180 days for each temperature point. After that the mechanical properties of the aged acrylic was measured to compare with the fresh one. Results are shown in Fig.~\ref{fig:cd3-25}. Testing shows that the material has different aging behavior for different mechanical properties. Here, we mainly focus on the tensile strength since it is the weakest strength compared to others. It is known that 60 days at 70 degrees corresponds to 16 years at 20 degrees, so we can predict that the tensile strength will drop by 18\% after 16 years of running. To validate this conclusion, further tests are in progress. \begin{figure} \centering \subfigure[Tensile strength]{ \label{fig:subfig:a} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-25a.png}} \subfigure[Compressive strength]{ \label{fig:subfig:b} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-25b.png}} \subfigure[Flextual strength]{ \label{fig:subfig:c} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-25c.png}} \caption{Performance of the acrylic after aging} \label{fig:cd3-25} \end{figure} (3) Test of the old acrylic pieces To study the influence of the surrounding environment such as ultraviolet radiation, sunlight, temperature and humidity, measurements have been made for old acrylic pieces after 14 and 17 years in outdoor. Results are shown in Table~\ref{table6}. \begin{table}[!htbp] \centering \caption{Strength of the acrylic after outdoor exposure\label{table6}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{4cm}<{\centering}|p{4cm}<{\centering}|p{4cm}<{\centering}} \hline Explosure time& 14 years & 17 years \\ \hline \tabincell{c} {Elastic modulus /MPa} & 803 & 942 \\ \hline \tabincell{c} {Poisson ration} & 0.405& 0.405 \\ \hline \tabincell{c} { Tensile strength} & 56.6 & 53.57 \\ \hline \end{tabular} \end{table} The tensile strength of the normal acrylics is about 68~MPa. From the table, we can see that it drops by 17\% and 21\% respectively, for 14 years and 17 years in outer door. (4) Test for the new bonding techniques The construction of the acrylic sphere will require a large amount of bonding. Since the normal bonding technique needs a long time for curing, and therefore a very long construction time, the manufacturers are studying two new bonding techniques(ultraviolet irradiation and fast bonding) to reduce the curing time. The tensile strength with those new techniques has been tested and compared with the normal ones. The strength for ultraviolet irradiation is 46.5~MPa and for fast bonding is 55.3~MPa. The latter corresponds to 80\% of the full strength of the acrylic. Details are shown in Table~\ref{table7}. Fig.~\ref{fig:cd3-26} shows the cross-section where the break happens; it can be seen that the cross-section of normal bonding (A) and fast bonding (B) are similar, while for the other two (C and D) it has a very flat surface, indicating that the bonding syrup is not completely adhering with the original acrylic, hence giving a lower strength. \begin{table}[!htbp] \centering \caption{Comparison of the different bonding techniques\label{table7}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{4cm}<{\centering}|p{2.5cm}<{\centering}|p{3cm}<{\centering}|p{2.5cm}<{\centering}} \hline Bonding techniques& time /hours & Tensile strength /MPa & Ratio \\ \hline \tabincell{c} {Without bonding} &{ }& 68.2 & 100\% \\ \hline \tabincell{c} {Normal bonding} & 12& 56.4& 82.7\% \\ \hline \tabincell{c} {Ultraviolet irradiation \\ without annealing} &3 & 37.2 & 54.5\% \\ \hline \tabincell{c} {Ultraviolet irradiation \\ with annealing} &7 & 46.5 & 68.2\% \\ \hline \tabincell{c} {Fast bonding} &4 & 55.3 & 81.1\% \\ \hline \end{tabular} \end{table} \begin{figure}[!htbp] \centering \includegraphics[width=12cm]{CentralDetector/figures/CD3-26.png} \caption{cross section of the broken acrylic} \label{fig:cd3-26} \end{figure} (5) Crazing test Users and manufacturers of plastics have been aware of a phenomenon that glassy plastics will develop micro-cracks, which start at the surface of the material and grow perpendicularly to the direction of stress. Studies have shown that there are at least five factors which may cause glassy plastics to craze, as listed in the following: $\bullet$ Applied or residual tensile stress; $\bullet$ Stress generated on the surface of the acrylic due to differential temperature; $\bullet$ Stress generated on the surface of the acrylic due to absorption and desorption of moisture; $\bullet$ Weathering; $\bullet$ Contact with organic solvents. Crazes are initiated and propagated only when the tensile stress on the material surface exceeds some critical value. For the central detector of JUNO, the acrylic sphere will be immersed in LS and water under stress loading. The problems of crazing become important concerning the reliability and longevity of the structural parts manufactured from acrylic. We did the following tests on acrylic samples in laboratory. Samples with 4.8~MPa stress applied were immersed in the LS and the liquid mixture of 80\% LS+20\% C$_{9}$H$_{12}$, at the room temperature and 60$^{\circ}$C. The results are summarized as the following: 1) No crazing appeared by submerging in LS for 90 days at the room temperature; 2) No crazing appeared by submerging in the mixture of 80\% LS+20\% C$_{9}$H$_{12}$ for 90 days at the room temperature; 3) By submerging samples in LS at 60$^{\circ}$C, two of the samples showed initial crazing after 30 days, and the other two had no crazing. After 90 days, the initial crazing had grown along the direction of thickness. More tests and studies on crazing will be made in future. (6) Test of light transmittance The test results indicate that the light transmittance of the acrylic with 4~cm-8~cm thickness is over 96\%, and the thickness has no significant effect on it. The measured attenuation length in air is $\sim$110~cm at 410~nm and $\sim$278~cm at 450~nm. (7) Creep test The creep test has been partially finished up to now for the acrylic submerged in LS at 20 degrees under a stress of 30~MPa, 25~MPa, 20~MPa, and 15~MPa. The results are shown in Fig.~\ref{fig:cd3-27}. This measured curve is consistent with the theoretical prediction, and the time before breaking increases dramatically at lower stress level. To understand the allowable maximum stress, more data are needed. Currently the creep tests are ongoing, including the test in pure water. \begin{figure}[!htbp] \centering \includegraphics{CentralDetector/figures/CD3-27.png} \caption{Creep curve of the acrylic under different stress } \label{fig:cd3-27} \end{figure} \section{Backup option: Balloon in a steel tank} \subsection{Introduction} The backup option of the central detector is a balloon with a stainless-steel tank. The tank, located in the center of the experimental hall, is about $\sim$40~m in diameter. A balloon of $\sim$35.4~m in diameter is installed in the steel tank. This balloon, welded from transparent films, is filled with $\sim$20~kt of LS as a target. The buffer liquid between the tank and balloon is LAB or MO. The density of LS in the balloon is 0.3-0.5\% larger than the buffer liquid. The soft balloon needs a reliable supporting structure to keep it spherical and reduce the stress, so a thin acrylic layer surrounding the balloon is designed for this purpose. Both the balloon and the acrylic structure have the same diameter. The inner wall of the steel tank is covered by more than $\sim$17000 20~inch PMTs. Fig.~\ref{fig:cd3-28} shows the structural sketch of this option. \begin{figure}[[!htbp]] \centering \includegraphics[width=14cm]{CentralDetector/figures/CD3-47.png} \caption{Schematic view of stainless-steel tank and the balloon} \label{fig:cd3-28} \end{figure} \subsection{Balloon} The balloon is the most important structure in this option. The candidate materials for the balloon include PA, ETFE, PET, PE/PA and FEP. There are two flanges with a pipe on the top and bottom of the balloon, which will provide the interface for filling and recycling. As the balloon is supported by the acrylic layer, the stress of the balloon is very small, and we do not have to worry about the balloon's shape. Therefore, the main study on the balloon is as follows: $\bullet$ the compatibility of the balloon material in the liquid; $\bullet$ the radioactivity of the material; $\bullet$ the transparency and aging properties of the material; $\bullet$ the cleanliness of the balloon during its production and installation; $\bullet$ sealing of the balloon. The primary task is to understand the material properties. Present studies show that the yield strength of ETFE film is about 18~MPa. This film has a good compatibility with all the liquids, good aging resistance, strength and impact resistance. It is self-cleaning, but its light transmittance is poor. For example, the transmittance of a 50 micron ETFE film is only 93\%. The test result shows that PE/PA film is compatible with MO and the compatibility test with liquid scintillator is ongoing. PET film has good compatibility and light transmittance, but its flexibility is slightly worse. FEP film has good light transmittance, but there is no wide application for this film so far. In order to keep the cleanness and minimize the dust, the balloon must be made and installed in a clean room. The cleanliness of the workshop in the factory should hopefully be at the 1000 level. Constant temperature and humidity are also needed during the production in the workshop. Providing clean gas to remove radon is necessary during the balloon production. The balloon should be carefully packaged, not only for safe transportation and ease of opening in the experimental hall, but also for avoiding contamination and damage. Leak tests of the balloon should be done in the workshop and at the experimental site. Assuming that the concentration of PPO penetrating into the buffer liquid is less than 10~ppm in 10~years, the leakage of the balloon should be about 5~$\times$~$10^{-2}$~cc/s at 3~mbar hydrostatic pressure. In the workshop, the method of leak test can refer to that of the Borexino experiment based on SF6 gas \cite{Borexino}.Sealed into another balloon, pumping the SF6 gas into the inner balloon and getting it into every corner, we can measure and obtain the curve of concentration-time of SF6 in the outer balloon and calculate the leakage. A vacuum leak test is also very useful. After the transportation to the underground hall, the balloon will be installed into the stainless-steel tank and the final leak test will be done onsite. The balloon can be filled with purified gas mixed with a certain concentration of SF6. The balloon's leakage is determined by testing the concentration of SF6 in the volume between tank and balloon. Further studies and tests should be done to get a detailed solution for the leak test. \subsection{Acrylic Structure for the Balloon Support} Acrylic here is designed as the supporting structure for the balloon. It is a sphere with $\sim$35.4~m diameter, fixed on the steel tank by supporting legs, as shown in Fig.~\ref{fig:cd3-29}. The acrylic layer is made up of many acrylic sheets and it is not necessary to seal it. In other words, it will reduce the difficulty and working time for onsite installation. The thickness of acrylic sheets in this option is 30~mm, which allows us to have a wide choice from commercial products. \begin{figure}[[!htbp]] \centering \includegraphics[width=8cm]{CentralDetector/figures/CD3-48.png} \caption{The acrylic supporting structure for the balloon} \label{fig:cd3-29} \end{figure} The density of liquid inside and outside the balloon has a small difference, expected to be 0.3-0.5\%. In stress analysis, we set a 0.5\% density difference as a starting condition, and the others conditions being hydrostatic pressure towards the internal surface of the balloon and external surface of the supporting structure. In this preliminary analysis, we used tetrahedral finite elements to analyze a 9$^{\circ}$ partial spherical layer with 25~mm thickness, and acrylic supporting legs were set as fixed ends. The results showed that the stress of the main part is below 2~MPa and the total displacement is about 9~mm, as shown in Fig.~\ref{fig:cd3-30}. Since the element density affects the local stress on the supporting points, further analysis with more detailed supporting structure will be done. \begin{figure} \centering \subfigure[the stress distribution]{ \label{fig:subfig:a} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-49a.png}} \subfigure[the deflection distribution]{ \label{fig:subfig:b} \includegraphics[width=7cm,height=7cm]{CentralDetector/figures/CD3-49b.png}} \caption{the stress and deformation of the acrylic support} \label{fig:cd3-30} \end{figure} \subsection{Stainless-steel Spherical Tank} In the backup option, the outer structure is a spherical stainless-steel tank. It will not only bear the weight of the liquid inside and support the balloon and the acrylic structure, but also provide an installation structure for the PMTs. The tank is not the same as a conventional pressure tank; its design should consider the installation process and on-site conditions as well as issues like cost and reliability. Based on the requirements above, the preliminary structure of the tank is designed as shown in Fig.~\ref{fig:cd3-31}. \begin{figure}[[!htbp]] \centering \includegraphics[width=8cm]{CentralDetector/figures/CD3-50.png} \caption{A preliminary structure of the stainless-steel tank} \label{fig:cd3-31} \end{figure} The tank is supported by landing legs, which are welded to the tank at its equatorial belt. The tank has a chimney on the top, and its size should satisfy requirements such as the installation of a calibration facility, the entrance for lifting the PMTs, ventilation and lighting of the tank, arrangement of the cables, etc. The chimney reaches the liquid level, so it is easy to seal. At the bottom of the tank, there is a hole to be used to dispose the waste liquid from cleaning and as a manhole during the installation stage. The tank is made of $\sim$170 welded steel plates, which are located in 9 belts of the spherical tank. These plates can be produced in the factory and then be assembled and welded after being transported to the site. Based on our investigation, there are many professional companies who have experience in making large spherical tanks. In addition, the largest spherical tank in China is up to 50~m in diameter, while the largest stainless-steel spherical tank is only 17 meters in diameter. Although there is a certain degree of difficulty and challenge for the $\phi$40~m stainless-steel tank in this option, there are many examples to follow. \subsection{Installation Issues} After the construction of the stainless-steel tank in the experimental hall, the PMTs and acrylic support sheets will be installed. To reduce the time for installation, several PMTs will be pre-assembled as modules. The bottom of the PMT module is made of stainless steel plate and the PMTs can be mounted on it. The ribs on the bottom plate are designed to increase its rigidity, as each PMT suffers about $\sim$70~kg buoyancy. There has to be about 10~cm clearance between the bottom plate of the PMT module and the inner wall of the stainless-steel tank, which will provide space for cable arrangement. During the installation, scaffolding will be installed in the stainless-steel tank. After the tank is completed, the scaffolding can be used to wash the tank from the top to the bottom. If the cable and PMT modules can be installed separately, the pre-layout of the cables can be finished in the inner wall of the tank before the PMT installation is started. A sling can also be hung along the inner wall of the tank for lifting the PMT module. After the tank cleaning and cable arrangement, the scaffolding can be removed. The pipe for filling the liquid scintillator and LAB will be installed outside the tank, and the pipe outlet, which is also the entrance into the steel tank, is at the bottom of the tank. The PMT modules will be installed from bottom to top one by one. The PMT modules can be moved by a manual chain which is hung by the sling. The cable can be connected to each PMT after it is installed. The PMTs in each module will be tested after the installation of each layer. After the PMT modules are installed, acrylic rods will be installed at the back plate of each module, and the acrylic sheet will be mounted on the rod to support the balloon. Once PMTs are tested, water can be filled into the tank. When water reaches the top of that layer of PMTs, installation of the next layer of PMTs can be started. After installing all PMTs and acrylic sheets, the water can be drained from the bottom of the tank. During this process, it is necessary to guarantee that the air coming into the tank is clean. After all water is removed, the balloon will be lifted into the tank from the top, and the filling pipe can be connected to the bottom flange of the balloon through the manhole. The top hole and bottom manhole will then be sealed to prevent dust getting into the tank. The flange on the top of the balloon will be sealed, and the gas for the leak test will go through the pipeline for filling liquid scintillator into the balloon to replace the air in it. Once the balloon leak test completed, the manhole at the bottom will be sealed by the flange. After making sure that every part works well, liquid scintillator will be filled into the balloon. During the process, the filling speed of LS, LAB and water outside the tank should be controlled strictly to keep the level of different liquids the same. \subsection{Conclusions} Since the balloon can be made in the factory and then delivered to underground lab for the final installation and leak check, the time needed onsite is relatively short. The cost of fabricating a balloon is much lower compared to the acrylic sphere so more than one balloon can be made. If one balloon has trouble in the leak checking or any other damage, it can be replaced in a short time. On the other hand, the balloon is a soft structure and not easily damaged under a shock load (such as seismic load). The use of the acrylic layer as support for the balloon reduces the stress and greatly improves the safety. Since the acrylic sheets in this option do not need to be sealed, the difficulty and time needed for on-site construction is reduced. Because the detector construction in this option involves the tank welding onsite, installation of the acrylic sheets and the balloon in the tank, cleaning and leak test of the balloon, the procedure will be more complex. Therefore this option is considered as a backup. \section{PMT assembly} \subsection{PMT testing and validation} The JUNO central detector will use 20~inch PMTs with highest possible quantum efficiency and large photo-cathode. The main specifications of the PMT include: $\bullet$ Response to single photoelectrons $\bullet$ Quantum efficiency of photocathode $\bullet$ Gain-HV function and gain nonlinearity $\bullet$ Transit time spread, ratio of after pulse and time structure $\bullet$ Dark noise rate $\bullet$ Uniformity of detection efficiency To precisely measure those parameters, a test bench integrated together with software and hardware needs to be setup. Based on the requirements, a dark room with shielding to earth's magnetic field is necessary. Efficient testing can be implemented by using a light source with multiple optical fiber coupling. Testing of PMTs consists of several steps: single sample test, test of a group of samples and batch test. Details will be finalized later. \subsection{High Voltage System of PMTs} A high voltage system is needed to provide stable power to the PMTs. This includes high voltage power supply, cable and water-proof structure. The three options below are under consideration: 1) One to one option: one HV channel supplies one PMT, and a divider is needed to produce different voltages for each electrode. This option can have a control to each PMT HV independently but a huge number of cables are needed. 2) Multi-core cable option: a group of HV channels are carried by a single cable with multiple cores, via an exchange box close to the PMTs, the HV supplies a bunch of PMTs. In this option: the number of cables are decreased, while each PMT is still controlled independently. However, a dry box is needed for HV exchange and distribution, and a connector with water-proof design is needed. 3) Front-end HV module option: for this option, the low voltage is firstly delivered to the front end near the detector, then the low voltage is transformed to high voltage which is supplied to the PMTs. The advantage of this option is to save cost and significantly reduce the number of cables. The voltage of the PMTs is provided by a Cockroft Walton capacitor diode, which converts the low voltage carried by a flat cable to high voltage and finally supplies the whole power grid. Compared to the traditional HV power based on dividers, this method of voltage multiplier has some clear advantages: 1) the expensive high voltage cable and connector is replaced by a cheap flex cable. This is an efficient way to save cost, reduce the number of cable and improve the overall quality. 2) the high efficiency of the voltage multiplier can reduce heat power consumption significantly. In general this kind of device will consume power less than 50~mW without light load, while under maximum light load, the power consumption is less than 200~mW. As a comparison, the option using dividers will consume 3.6~W in the form of heat loss. 3) the high voltage is very stable due to a deep negative feedback designed in the module, and complete load characters are provided. 4) the cost of the high voltage power supply with the multiplier itself is low, about 3-4 times lower than the other options. The main parameters of the JUNO 20'' PMT high voltage power supply are shown in Table~ref{table8}, Table~ref{table9} and Table~ref{table10} below: \begin{table}[!htbp] \centering \caption{USB bus adapter\label{table8}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{7cm}<{\centering}|p{6cm}<{\centering}} \hline Maximum number of channel& 127 \\ \hline \tabincell{c} {Power supply} & computer USB port \\ \hline \tabincell{c} {BUS voltage/ V} & LV - 5V,BV - 24 \\ \hline \tabincell{c} {Interface} & RS-485 \\ \hline \tabincell{c} {Communication line to computer} & USB-2.0 \\ \hline \tabincell{c} {Working temperature/degree} & (- 10) - (+40) \\ \hline \tabincell{c} {Working humidity/ \%} & 0 - 80 \\ \hline \tabincell{c} {size/ mm x mm x mm} & 70x50x22 \\ \hline \tabincell{c} {Weight/ kg} & 0.15 \\ \hline \end{tabular} \end{table} \begin{table}[!htbp] \centering \caption{20'' PMT HV module\label{table9}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{7cm}<{\centering}|p{6cm}<{\centering}} \hline PMT connection, grounding& cathode \\ \hline \tabincell{c} {Range of HV for anode / V} & 1500-2500* \\ \hline \tabincell{c} {Operating voltage/ V} & +2300* \\ \hline \tabincell{c} {Precision of anode HV/ V} & ~ 0.25 \\ \hline \tabincell{c} {Error of HV output} & 3\% \\ \hline \tabincell{c} {Stability of HV / \%} & 0.05\% \\ \hline \tabincell{c} {Temperature coefficient/ ppm/degree} & 100 \\ \hline \end{tabular} \end{table} \begin{table}[!htbp] \centering \caption{HV value of PMT dynode\label{table10}} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{8cm}<{\centering}|p{5cm}<{\centering}} \hline focus / V& +300 \\ \hline \tabincell{c} {D1/ V} & +400 \\ \hline \tabincell{c} {D2/ V} & +1200 \\ \hline \tabincell{c} {D3/ V} & +1300 \\ \hline \tabincell{c} {D4/ V} & +2100\\ \hline \tabincell{c} {Anode/ V} & +2300 \\ \hline \tabincell{c} {Up limit of anode current/ mA} & 100 \\ \hline \tabincell{c} {Cross talk/ mV} & 20 \\ \hline \tabincell{c} {Nominal voltage/ V} & +5,+24 \\ \hline \tabincell{c} {Power consumption of single module /W} & 0.1 \\ \hline \tabincell{c} {Communication protocol } & RS-485 \\ \hline \end{tabular} \end{table} \subsection{PMT related Structure Design} PMT related structures include PMT voltage divider, implosion protection structure, potting structure and mechanical supporting structure. JUNO will use 20'' PMTs, each one weighting about 10kg and the buoyancy in water of 70~kg. The PMTs are mounted on the stainless-steel truss. A preliminary design and test of the MCP-PMT divider has been carried out. The method of positive HV powering has been adopted, to reduce the number of cables and connectors, and the signal is readout by a splitter. A potting structure is needed since the PMTs are working in water, and a conceptual view is shown as Fig.~\ref{fig:cd3-32}. The potting will use a plastic cover and the water-proof adhesive to seal the PMT base, electronics and cable from contact with water. The waterproof design needs to consider long-term compatibility with high purity water and 4~atm hydrostatic pressure. After finishing the design and related tests, we will set up a batch test system to test each of the PMTs before installation. \begin{figure}[!htbp] \centering \includegraphics{CentralDetector/figures/CD3-28.png} \caption{conceptual view of PMT potting structure} \label{fig:cd3-32} \end{figure} The following sections will discuss the implosion protection structure of the PMTs. Currently there are three ideas under consideration to protect the PMTs from implosion in 40~m depth of water: A. the PMT is covered by an acrylic shell for the upper half sphere and the lower part by another shell made from FRP, PE or ABS. A few small holes are opened on the shell to allow air flow and liquid circulation. When the PMT collapses, the amount of ingoing water is minimized by those holes, hence the strength of the induced shock wave is suppressed; B. an air chamber is placed around the lower half of the PMT. When collapse happens, the air chamber will quickly expand due to a sudden drop of the local water pressure, hence the amount of ingoing water is also reduced and the shock wave is weakened; C. a thin layer of film is attached to the PMT surface, this film will prolong the PMT collapse time, and hence the intensity of inrushing water. The film is transparent so no light will be blocked. (1) Study of PMT implosion mechanism The shell of the PMT is made of fragile glass, so an implosion could happen when damaged glass shell is immersed in water. As was experienced in Super-K, implosion of one PMT could lead to a cascaded implosion and eventually all the PMTs would be destroyed. A simulation of the PMT implosion is shown in Fig.~\ref{fig:cd3-33}. The PMT glass shell crashes in 5~ms due to an outer pressure of 5 atm, the water then rushes inward suddenly. After about 40~ms, due to the collision of inrushing water, the implosion happens, and a shock wave is produced which propagates outwards. \begin{figure}[!htbp] \centering \subfigure[]{ \label{fig:subfig:a} \includegraphics[width=6.3cm,height=7cm]{CentralDetector/figures/CD3-29a.png}} \subfigure[]{ \label{fig:subfig:b} \includegraphics[width=8.4cm,height=7cm]{CentralDetector/figures/CD3-29b.png}} \caption{Simulation of PMT implosion} \label{fig:cd3-33} \end{figure} A preliminary study of the implosion behaviors such as shock wave strength as a function of time, distance, PMT volume and water pressure has been performed. As shown in Fig.~\ref{fig:cd3-34}, the strength from simulation is consistent with the real test. The strength follows a 1/r relation to distance as seen from Fig.~\ref{fig:cd3-35}, and is proportional to PMT volume and water pressure as shown in Fig.~\ref{fig:cd3-36} and Fig.~\ref{fig:cd3-37}, respectively. \begin{figure}[!htbp] \begin{minipage}[t]{0.5\textwidth} \centering \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-30.png} \caption{Peak pressure of the shock wave} \label{fig:cd3-34} \end{minipage} \begin{minipage}[t]{0.5\textwidth}\centering\includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-31.png} \caption{strength of the shock wave as a function of distance} \label{fig:cd3-35} \end{minipage} \end{figure} \begin{figure}[!htbp] \begin{minipage}[t]{0.5\textwidth} \centering \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-32.png} \caption{strength of the shock wave as a function of PMT volume} \label{fig:cd3-36} \end{minipage} \begin{minipage}[t]{0.5\textwidth}\centering\includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-33.png} \caption{strength of the shock wave as a function of pressure} \label{fig:cd3-37} \end{minipage} \end{figure} (2) PMT implosion test and simulation The purpose of the PMT implosion test is to measure the shock wave parameters, including production, propagation and relation to space and time. In addition, the test can verify the function of the acrylic cover. Fig.~\ref{fig:cd3-38} shows the schematic view of the implosion test and the pressurized tank. The tests are currently underway. \begin{figure}[!htbp] \centering \subfigure[Schematic view of the implosion test]{ \label{fig:subfig:a} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-35a.png}} \subfigure[the pressure tank for implosion test]{ \label{fig:subfig:b} \includegraphics[width=7.5cm,height=6cm]{CentralDetector/figures/CD3-35b.png}} \caption{the setup for implosion test} \label{fig:cd3-38} \end{figure} \begin{figure}[!htbp] \centering \subfigure[the 2D model for simulation]{ \label{fig:subfig:a} \includegraphics[width=7cm,height=4cm]{CentralDetector/figures/CD3-36a.png}} \subfigure[the pressure-time curve for point 1]{ \label{fig:subfig:b} \includegraphics[width=7cm,height=4cm]{CentralDetector/figures/CD3-36b.png}} \subfigure[Pressure-time curve of different points]{ \label{fig:subfig:c} \includegraphics[width=7cm,height=4cm]{CentralDetector/figures/CD3-36c.png}} \subfigure[Peak pressure as a function of distance]{ \label{fig:subfig:d} \includegraphics[width=7cm,height=4cm]{CentralDetector/figures/CD3-36d.png}} \caption{Simulation of the PMT implosion} \label{fig:cd3-39} \end{figure} To get the correct implosion model, simulation have been done following the real boundary conditions in the test. Fig.~\ref{fig:cd3-39}(a) shows a 2D simulation of single PMT without protection enclosure, in which the diameter of the water tank is 1.5~m, the diameter of the air bubble (PMT volume) is 0.5~m, the hydrostatic pressure is 4 atms, and the gas pressure in PMT is about $10^{-5}$ Pa. The boundary is set as a rigid border. There are 6 points in simulation which are uniformly distributed from the center to the border, in which point 2 is floating and movable to measure the size variation of the air bubble, and the other 5 points are fixed to measure the dynamic pressure induced by the shock wave. Fig.~\ref{fig:cd3-39}(b) shows that the peak pressure appears at t=35.2~ms. From Fig.~\ref{fig:cd3-39}(c), multiple shock waves are emitted during the process of compression and rebound of the air bubble, and the waves frequently happen at beginning or the end of the compression. The air bubble is finally compressed to a minimum size with the pressure reaching a maximum value of 20~MPa inside of the bubble, then the strongest shock wave is initiated. Fig.~\ref{fig:cd3-39}(d) shows the pressure for different points, where the pressure drops along with the distance. At the border of the PMT (d=25cm), the strength of the shock wave is about 11~MPa. The data from the real test will validate the simulation described above. Tests of multiple PMT implosion, and simulation of the effect of the acrylic cover are also in progress. \section{Prototype} \subsection{Motivations of the Prototype Detector} Following the progress and schedule of the JUNO experiment, considering the requirements of each sub-system, a prototype detector was proposed to test key technical issues: 1) Test and study of the PMT candidates: The JUNO experiment is designed to use large area, high quantum efficiency PMTs to realize the highest photo-cathode coverage. Candidate PMTs include Hamamatsu, Hainan Zhan Chuang (HZC), and a newly developed MCP-PMT. It is necessary to compare their performance in a real scintillator detector. From the prototype, we expect to extract the effective PMT parameters and compare them. 2) Test and study of the liquid scintillator: Liquid scintillator is another key component of the JUNO detector, a lot of experiences show that the final LS parameters in a detector, such as light yield, attenuation length, stability (or aging) and background level, are different from what were measured in the lab. With the prototype, many of those parameters can be extracted and optimized. 3) Test and study of electronics: The planned PMT waveform readout electronics will be very flexible for different physics requirements. The electronics system should be tested in a real LS detector. 4) Waveform data analysis algorithms and detector performance study: With the prototype, algorithms of PMT waveform readout scheme will be developed and detector performances will be understood. \subsection{Prototype Design} The prototype detector is shown in Fig.~\ref{fig:cd3-40}, which re-uses the Daya Bay prototype as the main container. An acrylic sphere locates at the center of the stainless-steel tank (SST) as a LS vessel, and is viewed by 51 PMTs dipped in the pure water. The expected photo-cathode coverage is ~55\%. The water tankand the PP/Lead layer is designed to have 1 m.w.e. shielding, aiming to reduce the radioactivity coming from the outside of SST. The expected trigger rate is less than [email protected], and the energy resolution is ~4\%@1MeV. \begin{figure}[!htbp] \centering \subfigure[]{ \label{fig:subfig:a} \includegraphics[width=6cm,height=5cm]{CentralDetector/figures/CD3-37a.png}} \subfigure[]{ \label{fig:subfig:b} \includegraphics[width=5cm,height=5cm]{CentralDetector/figures/CD3-37b.png}} \caption{the prototype design} \label{fig:cd3-40} \end{figure} As shown in Fig.~\ref{fig:cd3-41}, the diameter of the acrylic sphere is 50 cm with a thickness of ~1 cm, and a tube with a diameter of 5 cm and a length of 70 cm is located at the top of the acrylic sphere for filling and calibration. The 51 high quantum efficiency PMTs, in the diameter of 8'', 9'' and 20'', are uniformly arrayed in 4 layers facing to the center of the acrylic sphere. As follows is the detailed arrangement of the PMTs: $\bullet$ Top layer: which includes 4 MCP-PMTs of 20'' and 2 Hamamatsu PMTs of 20''; $\bullet$ Middle layers: which are divided into two layers and each layer including 8 MCP-PMTs of 8'', 4 HZC-PMTs of 9'' and 4 Hamamatsu PMTs of 8''; $\bullet$ Bottom layers: which are divided into two rings, the inner ring includes 3 MCP-PMTs of 8'', 2 HZC-PMTs of 9''and 2 Hamamatsu PMTs of 8'', and the outer ring has the same arrangement as the top layer. \begin{figure}[[!htbp]] \centering \includegraphics[width=6cm,height=6cm]{CentralDetector/figures/CD3-38.png} \caption{The side view of the prototype detector} \label{fig:cd3-41} \end{figure} Fig.~\ref{fig:cd3-42} is the shielding system. The bottom and top of this system are covered by 10 cm thick lead plus 10 cm thick pp plate, while each of the other 4 sides is shielded by a customized water tank in 1m x 5m x 3m dimension constructed by standard stainless-steel elements which are widely used by water and conditioning system. One of the 4 sides is movable to allow the internal detector installation. \begin{figure}[!htbp] \centering \subfigure[schematics of the shielding structure]{ \label{fig:subfig:a} \includegraphics[width=5.5cm,height=4cm]{CentralDetector/figures/CD3-39a.png}} \subfigure[one side of the water tank opened]{ \label{fig:subfig:b} \includegraphics[width=5.5cm,height=4cm]{CentralDetector/figures/CD3-39b.png}} \subfigure[bottom and top shielding design with lead and pp plate]{ \label{fig:subfig:c} \includegraphics[width=5.5cm,height=4cm]{CentralDetector/figures/CD3-39c.png}} \caption{the shielding system} \label{fig:cd3-42} \end{figure} \subsection{Expected Performance} With preliminary Geant4 simulation (simulation geometry is shown in Fig.~\ref{fig:cd3-43}), we confirm that the detector will have a good measurement of electron and alpha for LS study. Gammas have a large energy leakage due to the limited detector dimension, we will need a spectral fitting to extract more parameters. Measurements of Muon and calibration will provide more opportunities for detailed study. According to the prototype design, we expect to measure PMTs' parameters, liquid scintillator, electronics and background, to achieve the goals. \begin{figure}[[!htbp]] \centering \includegraphics[width=8cm]{CentralDetector/figures/CD3-40.png} \caption{Geant4 simulation geometry} \label{fig:cd3-43} \end{figure} The prototype is now under preparation and will be operational soon. \section{Construction of the central detector} \subsection{Assembly of the Stainless-steel Truss} The stainless-steel truss used in the central detector has a bolted connection, which makes the on-site assembly easy and safe without any welding. The design and the construction of the truss should follow the technical standards, design specifications and procedures. In China, the space truss technique is very popular for building construction. Many companies have capability and rich experience in design and production of steel trusses. There are two main stages for the construction of the stainless-steel truss. The first stage is the off-site preparation in the factory, and the second stage is assembly onsite. All the truss members, bolts and joints will be finished for the machining and production in factory before transportation, and the machining precision and quality should be ensured at this stage. When all parts are transported to the experimental hall, some important tests need to be done onsite to make sure that all conditions meet the requirements for construction before truss assembly. Theses tests include: checking the foundation of the truss area; double checking the parts of the truss; checking the height mark, axial and gradient of the embedded part, and so on. Since the complexity of truss assembly depends on the accuracy and roundness of the first three circles, the beginning preparation of axis positioning and height level is more important for the whole assembly of the truss. To control and check the coordinate position of each truss joint, some reference and checking points should be installed at different areas of the hall. When all conditions for the construction meets the requirements of design and standards, the truss assembly can be started following the detailed construction procedure. Theodolites or other survey instruments will be used during the truss assembly to control the error. For the central detector, the stainless-steel truss and acrylic sphere are not independent and need overall consideration of the construction sequence and procedure. It should be better to make the truss and acrylic sphere layer by layer alternately. The detailed procedure needs further discussion with the factory. In general, the whole procedure needs to ensure the safety and reliability of the sphere and truss. \subsection{Construction of the Acrylic Sphere} As Fig.~\ref{fig:cd3-7} shows, the acrylic sphere is made up of many acrylic sheets which are bonded together to form the sphere. Satisfying the production capacity and transportation limit, the size of each acrylic sheet should be as big as possible to reduce the total bonding length onsite. For the preliminary consideration and design of the sphere, there are a total of $\sim$170 acrylic sheets and divided into 17 layers. Each sheet is less than 3~m $\times$ 8~m in size. The total bonding length of the sphere is about 1.8~km. Production and machining of the acrylic sheets will be finished in the factory before construction onsite. In the factory, the first step is to produce a flat sheet of acrylic with 120~mm thickness. The next step is to thermoform the flat acrylic into a spherical sheet with 35.4~m inner diameter in a concave spherical mold, and these thermo-formed sheets should be stored carefully to prevent damage to the acrylic surface. After that are the steps of machining, milling, polishing and annealing. Some measurements should be done for each step to control the quality in the factory. One such step is the thickness measurement of each flat sheet used as the thermoforming blank. The second measurement is for the thickness of the sheet after thermoforming. In general, the acrylic blank increases in thickness around its border and decreases in thickness at the center after the thermoforming operation in the female mold. The third measurement is the acrylic thickness after annealing. The curvature of the acrylic blank should also be measured to ensure the assembly requirement \cite{Handbook}. In the factory, the fixtures for positioning and assembly should be machined. After all the part production and fixture preparation is finished by the manufacturer, they will be delivered to the underground to do the construction of the acrylic sphere onsite. The acrylic sphere can be constructed with the help of the truss. The acrylic sectors can be supported by the truss joints during assembling, positioning and bonding layer by layer. The truss can also be used as the scaffolding or platform for sphere construction. The acrylic sphere will be built from the bottom to the top, layer by layer. The quality control and check of dimension, positioning, roundness, bonding status and annealing need to be ensured for each layer throughout the period of construction. Inspections of the finished spherical sphere consist of careful visual checks, dimensional measurements, and stress checking. For the construction of the detector, two cranes will be needed in the hall. One is to carry and lift the components of the detector, and the other is a lift to carry people to access the working area in the pool. The working platforms should be designed and installed inside and outside the acrylic sphere for the processes of positioning, assembly, sanding and polishing. Safety nets should be installed surrounding the sphere during construction to prevent any dropping of objects which will damage the detector or people. Special requirements may be needed for the ventilation system of the assembly cavern. There are special rules and regulations governing working underground to which the project must conform. \subsection{PMT Installation} When the acrylic sphere and stainless-steel truss are finished, about 16000 PMTs need to be installed on the truss facing inwards to receive signal from the LS. These PMTs should be arranged as close to each other as possible to provide more than 75\% optical coverage. The PMT support structure must provide a stable and accurate platform for the mounting and positioning of the PMT assemblies to view the active detector regions. The PMT support structure should also provide a mechanism to protect adjacent PMTs from the risks of cascaded implosion and permit the replacement of the PMT assemblies. To reduce the installation time, PMTs will be installed by PMT module. Several PMTs are assembled to be one module in the surface assembly building. The PMT modules will then be delivered underground for on-site installation. At present, two options for PMT installation are considered. One option is to use stainless steel as the support panel of the PMT module, and each PMT with its protecting cover will be mounted using a clamp on the panel as Fig.~\ref{fig:cd3-44} shows. Some slots on the truss should be left for the lifting and installation of the PMT module. The circular rails along the latitude will be installed on the joint of the inner layer of the truss. There are pulleys at the back of the stainless-steel panel to allow the PMT element to slide to the expected location and finish the mounting. Fig.~\ref{fig:cd3-45} shows the sketch of this option. The design of the module size and shape should consider the location and number of the joints and supporting rods between acrylic sphere and truss to get the maximum optical coverage for the central detector. \begin{figure}[!htbp] \begin{minipage}[t]{0.5\textwidth} \centering \includegraphics[width=8cm]{CentralDetector/figures/CD3-43.png} \caption{PMT module with the stainless-steel backplate} \label{fig:cd3-44} \end{minipage} \begin{minipage}[t]{0.5\textwidth}\centering\includegraphics[width=8cm]{CentralDetector/figures/CD3-44.png} \caption{Installation of the PMT module} \label{fig:cd3-45} \end{minipage} \end{figure} The other option is to use profile stainless steel as the supporting structure. PMTs will be mounted on the profile steel one by one. This profile steel will be pre-installed on the inner layer of the truss. Fig.~\ref{fig:cd3-46} shows the sketch of this option. This option allows many work groups to do the installation at different working areas simultaneously. \begin{figure}[[!htbp]] \centering \includegraphics[width=10cm]{CentralDetector/figures/CD3-45.png} \caption{Using profile steel for PMT installation} \label{fig:cd3-46} \end{figure} \subsection{Filling system for the Liquid Scintillator} Filling will start when the construction of central detector is finished. The filling liquid involves LS and water for the main option and LAB needs to be considered for the backup option. The veto system will be responsible for water filling and the central detector system is responsible for LS and LAB filling. The two systems should coordinate their filling speed to balance the pressure. During filling, cleanness, background, temperature, liquid storage tank, pipe design and process monitoring should be considered and controlled. 1) Controlling the Cleanness and low background During filling, the detector and LS need to be kept clean, especially to prevent the radioactive background and radon gas from pollution. The central detector will be cleaned using purified water before filling. For the acrylic sphere plus stainless-steel option, a special cleaning robot will be designed and used for cleaning the acrylic sphere and the other parts such as the steel truss. The PMTs will be washed first. The acrylic sphere can be washed with LS at the end. For the backup option, the inner wall of the stainless-steel tank and the outer surface of balloon will be water flushed and the inner surface of the balloon can be cleaned by filling and pumping water repeatedly. The filling pipes and pump, which will come into contact with LS, need special care and cleaning, and they should be compatible with LS. Before filling, the whole filling system will be cleaned with LS several times and the used LS will be disposed of. Spaces where the LS is exposed to air, such as inside the acrylic or balloon and the LS storage tank, will be filled with nitrogen to prevent radon pollution and contact with oxygen. 2) Pure water exchange or nitrogen exchange for filling Pure water exchange means we need to firstly fill the acrylic sphere or balloon with water and then fill LS into the vessel while draining the water out. Because of its lower density, the LS will be float at the top of the water. This process filters the LS and is helpful to remove the background. The other method is exchanging with pure nitrogen, which means filling the acrylic sphere or balloon with nitrogen first and then putting pure LS into the vessel to replace the nitrogen. 3) Controlling the liquid temperature The LS temperature will be high if it is produced and stored on the ground. Since we hope the liquid for filling can be kept at 20$^{\circ}$C, the liquid needs to be cooled before filling the detector. The place for cooling should be selected on ground or underground according to the general deployment of the civil construction, and the cooling method can be water cooling or air-conditioning. 4) The preliminary plan for the LS pipe, storage tank, overflow tank and LS cycling system. As shown in Fig.~\ref{fig:cd3-47}, the filling port of the central detector is designed at the bottom of the sphere. The filling pipe comes from the LS storage tank which sits in the storage room, and goes down along the water pool wall to connect to the filling port. When the LS filling is finished, the pipes between the storage tank and central detector make them communicating vessels and then the storage tank can be used as the overflow tank. If the LS needs recycling during running, we will remove the LS from top of the detector and transport it to the LS recycling equipment, and then pump it back into LS storage tank. Finally the LS will be filled back into the central detector through filling pipes from the storage tank. 5) Filling process control Filling process control involves the speed control, monitoring of liquid levels, monitoring of stress and deformation of the detector, monitoring of stress, temperature, and so on. The design of filling speed and capacity of the filling system should consider the requirements of total filling time, the capacity of LS storage and production, and cost. The design of the relative liquid level difference should consider the load stress on the detector. \begin{figure}[!htb] \centering \includegraphics[width=14cm]{CentralDetector/figures/CD3-46.png} \caption{Sketch of the LS filling system} \label{fig:cd3-47} \end{figure} \section{Reliability and Risk analysis} $\bullet$ Structural reliability analysis The baseline option takes large glass buildings and techniques used for glass curtain walls as an important reference, given that those buildings and techniques have been matured for long time in the field of architecture. In our case, the material of glass is replaced by acrylic, and the crucial parts include bulk-polymerization of the acrylic sheets, different liquid located inside and outside of the acrylic and long term water pressure and buoyancy. Reliability of the stainless-steel structure is less important than that of the acrylic sphere in this option. We should consider: 1) The aging properties of acrylic under stress and submersion in organic or inorganic liquid needs further tests, to make sure it will be stable during the 20-year running. 2) Acrylic is a brittle material, if there is a crack it will grow under stress. Cracks can be repaired during the construction but will be fatal after the liquid filling is finished. 3) The stress is generally larger in the joints of the acrylic and truss than other areas, so needs to be controlled to keep it low enough. 4) As for the requirements of the SNO and Daya Bay, the stress on the acrylic should be less than 5~MPa, and more work is needed to reach this goal. $\bullet$ Construction reliability Fabrication of a thick acrylic sheet in a spherical shape is not a mature technique, and on-site construction of the large acrylic sphere also needs more R$\&$D effort. $\bullet$ The reliability of FEA The analysis needs to be compared with test results, and the accuracy of analysis should be verified by joint tests and small prototype tests. Maximum errors should be taken as the input parameters for the analysis. $\bullet$ Construction risk Construction of the truss should be under control since there is a mature specification. For construction of the acrylic sphere, since there will be a large mount of work done onsite, a standard procedure should be followed and strict quality control should be applied especially on the key parameters of the structure such as stress, deformation and stability. $\bullet$ Local failure Local failure means a failure of the joint between the acrylic sphere and truss, so the stress of the whole structure would need to be re-analysed. According to the current analysis, if the failure happens on the truss, it will not be too serious. If it happens on the acrylic, however it will result in severe damage, especially when the acrylic sphere has been filled with liquid scintillator. This kind of failure should be avoided at all costs. $\bullet$ Long term compatibility and leakage problem It has been proved that acrylic is compatible with liquid scintillator, and bulk-polymerization will not lead to leakage problems. $\bullet$ Long term performance The lifetime of acrylic is up to 30 years and the acrylic sphere for the SNO \cite{SNO} experiment has already been running for nearly 20 years. In our case, however the aging effect of the acrylic still needs more study. \section{Schedule} The schedules of the central detector are the following: $\bullet$ 2013 1) Conceptual design of the detector, including FEA and prototype test 2) Consideration of the central detector schedule, budget, manpower related issues. $\bullet$ 2014 1) Review existing options, select one main option and one candidate option. Perform extensive study on the two options, solve the key design parameters by analysis and prototype. 2) Preliminary design and test of implosion-proof structure of PMT. 3) Start to study PMT potting and LS filling. 4) Start to study batch test for PMT. $\bullet$ 2015 1) Optimize the detector design, conduct prototype design, fabrication and test. Start to consider on-site construction of the detector. 2) Further study the water-proof of PMT 3) Finish a preliminary design of the implosion-proof structure of PMT 4) Study the batch test system of PMT, including electronics performance test, water-proof and pressure test. 5) Determine the final central detector structure. $\bullet$ 2016 1) Continue detector structure design and on-site construction study 2) Determination of the batch test design for PMT 3) Start to design the liquid filling system and monitoring system. 4) Determination of the high voltage option for PMT 5) Start to manufacture the implosion-proof structure for PMT 6) Finish the engineering design of central detector. $\bullet$ 2017 1) Finish bidding for the central detector structure, determine the construction company and start to produce the components. 2) Determine the high voltage option and water-proof design of PMT. 3) Start PMT batch test. 4) Determine liquid filling option, bidding for the system. 5) Design of nitrogen protection system for liquid scintillator. $\bullet$ 2018 1) Start to construct the central detector. 2) Continue PMT batch test. 3) Start to install detector monitor and other necessary devices. $\bullet$ 2019 1) Finish central detector structure construction. 2) Finish PMT installation. 3) Finish detector monitor installation. 4) Finish installation of liquid filling system, start to fill the liquid. $\bullet$ 2020 1) Finish liquid filling. 2) Finish detector commissioning and start data taking. \vbox{} The total time for the central detector construction is estimated about 18 months, as follows: $\bullet$ Construction of the stainless-steel truss: 3 months (January 2018 - April 2018). This includes interface(column supports on the base of the water pool), fixtures, hoisting, measurement, fixing, surveys etc. $\bullet$ Construction of the acrylic sphere: 8 months (May 2018 - November 2018). This includes acrylic sheet hoisting, positioning, fixture installation, joint connection and adjustment, bulk-polymerization, sanding, stress monitoring, installation of the LS filling device and the interface of calibration. $\bullet$ Survey of the acrylic sphere and leak test: about 1 month (December 2018) $\bullet$ Cleaning of the acrylic sphere. The total area of the inner and outer surface of the sphere is about 8000 $m^{2}$. A high cleanliness level is needed for the inner surface and it is hard for humans to clean, so a robot cleaner is under consideration. An outlet need to be reserved in the bottom of the sphere. It will be closed after cleaning. $\bullet$ Installation of PMTs and monitors: 5.5 months (February 2019 - June 2019). A number of work groups are considered to work in parallel at a rate of 2 PMTs/hour and 8 working hours/day, so 144 PMTs can be installed in one day. In total this requires about 120 working days, meaning 5.5 months, for the $\sim$17000 PMTs. $\bullet$ Cleaning of the whole detector after installation. This is not included in the timetable of the central detector since it can be done together with VETO. \chapter{Civil Design and Facility} \label{ch:CivilConstruction} \section{ Experimental Site Location and Layout} The location of the JUNO experiment and its supporting facilities should meet the following criteria: \begin{enumerate} \item The site should have an equal distance from both the Yangjiang and the Taishan Nuclear Power Station with a largest possible overburden; \item The site plan and support facilities should be comprehensive for safety, access to the site, and ease in managing the logistics of equipment and personnel above and below the ground for construction and operational considerations of the experiment; \item The site engineering design and construction plan should be comprehensive in its consideration for energy conservation and emissions reduction, along with minimizing the impact on the local ecological environment; \item The site should have readily available access to adequate utility water and electric power; \item The surrounding transportation infrastructure of road should be reasonable for transportation of experimental equipment to the site. The road on site should be comprehensive in its consideration for easy equipment transportation; \item The site plan for the above ground campus should be comprehensive in its design consideration for seamless and harmonious integration with the local topography, landforms and foliage for civil construction, experimental installation and onsite workforce; \item All the design should meet applicable national building and construction codes and specifications for civil construction should be followed. \end{enumerate} \subsection{ Experiment Site Location} The site location is driven by the physics requirements to be optimally equidistant from both nuclear power complexes, along a central axis determined by the latitudes and longitudes of the (6) reactors at Yanjiang Station and (4) reactors in Taishan Station. The site, with surface and underground facilities, and experimental detector should be located in an area approximate 2,000~m in length and 200~m in width along this central axis at a distance approximate 53~km from the Yangjiang and Taishan Nuclear Power stations for the baseline physics requirements. In addition, the experimental hall should be at a depth greater than 700~m. The lithology and rock formation should be carefully considered for optimum stability, along with mountain topography to maximize rock overburden. Prospective site locations along the central axis are shown in Fig.~\ref{Fig10-1} and Fig.~\ref{Fig10-2}, at points (1-3) with elevations of 230~m, 210~m and 205~m respectively. In view of the changes of boundary condition of the Mesozoic granite stocks intruding into the Paleozoic rock mass, point (4) at 268.8 m elevation away from this boundary of granite was selected. This preliminary location for the experimental hall is far enough away from the granite boundary to offer adequate safety margin. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{CivilConstruction/figures/Fig10-1.jpg} \caption{ The site region (red dash line) required by Physics, orange shadow is granite area } \label{Fig10-1} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{CivilConstruction/figures/Fig10-2.jpg} \caption{ Experimental Hall Location (position 4) in the granite area with green dash line} \label{Fig10-2} \end{center} \end{figure} \subsection{ Experiment Hall Layout} With site location determined, access options to the underground experimental hall must be considered. Life safety and emergency egress require that at least two independent access paths from the underground experimental hall to the surface should be available. With this requirement in mind, the site plan layout for underground access is either: "vertical shaft and inclined shaft" or "vertical shaft and normal tunnel". Both options would have the same vertical shaft with cage-lift, but the normal tunnel with an 8\% grade slope would be 6~km in length as compared to an inclined shaft at a grade slope of 42.6\% and 1.34~km in length. The normal tunnel construction cycle is too long and costly with transportation safety issues using motorized vehicles for access. The inclined shaft option was selected for its short construction cycle, lower cost, and the use of rail cars for equipment transport underground provides a safer and controlled environment. Fig.~\ref{Fig10-3} is a general layout of the site and facilities with the "vertical shaft + inclined shaft" scheme. Here the vertical shaft is 581~m in depth; the inclined shaft is 1,340~m in length with a slope of 42.6\%; and the overburden at the experimental hall is 728.8~m. The access entry to the inclined shaft will be close to the experimental hall, located in the granite mountain area. All of these factors are favorable for civil construction and stability of the rock. A site survey was conducted to determine the optimal location of both access shafts. The vertical shaft is located at the abandoned quarry. Impact of this site during the construction and the operation to the surrounding area, traffic conditions and local residents, have to be studied well, as well as the construction cycle. To facilitate the construction and reduce the impact to the local residents, the construction yard and rock disposal area should be arranged near the entry of shaft construction site. With full consideration of these site conditions and other factors, such as difficulties in land acquisition, etc., the entrance portal of the inclined shaft was determined to be at Shenghe Village, Jinji Town, Kaiping City. \begin{figure}[Fig10-3] \begin{center} \includegraphics[width=0.8\textwidth]{CivilConstruction/figures/Fig10-3.jpg} \caption{ Layout of the Site } \label{Fig10-3} \end{center} \end{figure} The entrance portal of the inclined shaft (Fig.~\ref{Fig10-3}) includes permanent buildings, a construction yard area, and a waste disposal area. All stone debris will be disposed in the waste disposal area, along with that from the vertical shaft. The waste transportation route to the waste disposal area does not pass through neighboring villages or poultry farms, to limit the impact to the local life and economy. \subsubsection{ Underground Cavities } The layout of the underground facilities (Fig.\ref{Fig10-4}) includes a cavity for the main experimental hall, ancillary rooms with connecting tunnels. \begin{figure}[Fig10-4] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-4.jpg} \caption{ Layout of the Underground Facility } \label{Fig10-4} \end{center} \end{figure} The experimental hall contains a spherical detector tank 38.5~m in diameter located in a cylindrical water pool with inner diameter of 42.5m, as shown in Fig\ref{Fig10-5}. Shape of the experimental hall ceiling, either a dome or a vault scheme, was studied. Criteria such as workflow, construction cost and schedule, rock stability, experimental equipment installation and operation were considered. Vault scheme was selected with two crane bridges on a common rail to span installation hall to water pool. A 2.75~m wide walkway on both side of pool is chosen, and the span of upper vault structure becomes 48~m. A 10~m wide area at the entry of experiment hall will be used for equipment un-loading, resulting in a hall 55.25~m long. The total height of the hall from vault apex to the bottom of the water pool will be 69.5m. The underground assembly hall, pump room, and electrical equipment room are arranged at the same elevation as the inclined shaft entry into the experimental hall. The assembly hall is right next to the experimental hall. This hall is 40~m in length, 12~m span, and 12.5~m height and connected with the inclined shaft through a tunnel with cross-section of 5.7~m in width and 6.5~m in height. The rail car can transport the equipment directly to the experimental hall or to the door of assembly hall. Two bridge cranes on the same rail, each equipped with two 12.5~ton hooks, are arranged in the experimental hall, for handling of experimental equipment. An additional 10~ton crane is located in the assembly hall, for handling of the cargo and equipment from the inclined shaft rail car. Liquid scintillator storage room, processing room, and filling room are located at the same elevation and opposite to the assembly hall towards the vertical shaft at a distance far enough away from the experimental hall for fire safety consideration. Electronics room, water purification room, gas room, and refuge room are also arranged around the experimental hall. A single access/drainage tunnel with 4.5~m in width that encompasses the experimental hall interconnects these ancillary rooms. The inclined shaft being 1,340~m in length, 5.7~m in width, and 5.6~m height, bottoms out at elevation -460~m. From there it proceeds at the same level to the experimental hall by way of an 85~m long interconnecting tunnel with a turning radius of 35~m. The slope of the inclined shaft at 42.6\% ($23.08^{\circ}$) is consistent with requirements for debris transportation during the excavation and rail car vehicle specification, while minimizing the length of the inclined shaft. The transport equipment in the inclined shaft is a single-drum cable winch, which operates at a speed of 4~m/s. A walkway is located on one side of the inclined shaft for the maintenance personnel to carry out patrol check and for underground evacuation in case of an emergency. Utilities such as cable bridges, fire water pipeline, air conditioning water pipeline, and liquid scintillator pipeline, etc. are secured on the side walls of the inclined shaft, while air conditioning conduit and air supply conduit, etc., are located under the upper crown of the shaft. The access entry to the vertical shaft is at 130~m elevation. The shaft is 581~m in depth, with the bottom access entry connected to the experimental hall through an interconnection tunnel of 300~m in length at 3\% slope. The cross-section of this tunnel is 4.5~m $\times$ 5.2~m. A lift-cage elevator is used in the vertical shaft with a maximum lifting speed of 6~m/s, and a load capacity of 8-10 people. In addition, a vertical ladder stairway is installed for personnel evacuation in case of emergencies and for equipment maintenance access. Vent duct and cable utilities also pass through this vertical shaft. \begin{figure}[Fig10-5] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-5.jpg} \caption{ Sideview of the Experimental Hall } \label{Fig10-5} \end{center} \end{figure} \subsubsection{ Layout of Surface Constructions } The ground campus comprises inclined shaft entry area and vertical shaft entry area. The two areas are connected by a 2.5 km road. Exhaust room, electrical power room, and an access room are arranged near the vertical shaft entry. The inclined shaft entry area is a comprehensive area for assembly, office, dorms, and other service. A rock disposal area is also arranged in this area. The landscape at the rock disposal area will be recovered after the construction is completed. The living area is near the entrance of the campus, and is provided with two houses and one dorm building with 54 rooms. There is a canteen on site to provide food service. The experimental office building consists of a control room, a computer room, meeting rooms, and offices. The assembly area is arranged at the inclined shaft entry, and comprises one assembly building (3000 $m^{2}$ ) and temporary storage buildings. The assembly building is air-conditioned and provided with a crane to facilitate equipment transportation. In addition, power station, utility buildings for pure water, LN2, et al are all in the area. The liquid scintillation area is arranged away from the living area in the valley area at the access entry. It includes liquid scintillation storage tanks, purification and nitrogen facilities. \input{CivilConstruction/section2} \section{Utility Systems} \subsection{\bf Power distribution and grounding system} To meet the requirements of having redundant, reliable electrical power to the site, two 10~kV commercial power transmission lines will be introduced to the power distribution centers at the entrance to the inclined shaft and the vertical shaft from 110~kV transformer substations in Jinji Town and in Chishui Town, respectively. They belong to different public grids. An underground 10~kV power equipment room is provided to distribute power through the experimental hall and ancillary rooms. In addition to these commercial power sources, the surface power distribution center will be equipped with a self-starting diesel generator, which will supply power for emergency lighting, ventilation or waste water sump-pumps during disruptions of the main power supply. The total expected demand for power distribution is approx. 5,000~kVA from the combined power supply sources, each of which can provide 4,000~kVA of power. Steady-state power loads include ventilation, air-conditioning, and lighting systems along with operational experimental equipment. However, peak power capability is required during the LS distillation process and its transfer to the detector. An isolated transformer will be installed in the underground power equipment room to distribute clean electrical power on a separate grid to experimental equipment to minimize system noise. The electrical grounding systems include both a clean ground grid for experimental equipment and a safety ground grid for utility systems. The clean ground grid for experimental equipment on clean electrical power will be connected to the secondary winding of the source isolation transformer. The two grounding grids will have no electrical connection between them. The surface safety ground grid is distributed and transferred to underground with power transmission cables, and is connected to the underground safety ground grid. The clean ground for experimental equipment is distributed locally in the underground experimental hall. It is understood that the underground geological formation mainly consists of granite, having a high electrical resistivity, not conductive for electrical grounding. Based on our experience, the electrical grounding systems should include appropriate measures, such as high-efficiency electrolytic ion grounding electrodes to mitigate electrical safety concerns. This grid should also connect to the water pool steel-bar structure, to improve the connectivity with the ground and to increase the overall capacity, hence further stabilize the ground. \subsection{\bf Ventilation and air-conditioning system} The temperature in the underground experimental hall shall be maintained at a nominal 22$^{\circ}C$ and the relative humidity shall be lower than 70\%, with adequate ventilation to sustain at least 50 people working 24/7 underground. Air quality in the experimental hall can be normal room air that meets cleanliness requirements for personnel safety. The cleanliness in ancillary halls related to liquid scintillator storage or processing should meet class-100,000-level which is comparable to the requirements for the Experimental Hall-5 at Daya Bay. Special measures (e.g., high efficiency filters) can be taken during liquid scintillator operations to improve the cleanliness to class-10,000-level. The air exchange rate in the experimental hall is designed for 6~volumes/day, to reduce radon in the air and maintain a comfortable environment for underground workers. Water chillier units and fresh air handling systems for supplying dehumidified cool air underground are located at the entrances to both the inclined and vertical shafts. This conditioned air supply flows through insulated ducts in each shaft for dispersion in the experimental hall. Air is then drawn from the experimental hall via exhaust ducts to a main duct in the inclined shaft, and then exhausted to the atmosphere using exhaust fans. \subsection{\bf Water supply and drainage system} The underground utility water supply comes from the fire-fighting water system. Normal water usage is obtained from the fire-fighting water system main pipeline, so both water systems are from a common source. Ultra-pure water for experimental use either in the detector water pool or in scintillator processing will have its own supply by stainless steel pipes from the surface to underground facilities. The underground wastewater comes from natural rock seepage, utility water or fire-fighting water. It is collected in a series of grate-covered tunnel floor trenches and diverted to a number of sump-pits for pump-out to the surface. Since most of wastewater during normal operations comes from rock seepage, five sump-pit drainage systems have been strategically located through the underground facilities, based on the distribution of water seepage. Sump-pit-1, the main sump-pit, is located at the bottom of the inclined shaft, and is distributed through the curved interconnecting tunnel to the experimental areas. Sump-pit-2 is located at elevation of -200~m in a side portal area of the inclined shaft. Sump-pit-3 is located at the bottom area of the water pool, and is used during pool cleaning. Sump-pit-4 is located at bottom of vertical shaft. Sump-pit-5 is located at an elevation of -100~m in a side portal area of the inclined shaft where the rock formation is poor and water seepage is severe. The water from whole underground facilities is pumped up to surface via sump-pit-1 and then sump-pit-2. This design may choose less power pump for sump-pit-1, which have less impact on the experiment during its operation. Water pool draining will install additional dedicated pump system to sump-pit-1 temporarily. \subsection{\bf Fire control system} The fire control system consists of automatic fire alarms and interlocks, fire-fighting water distribution to hydrant-hose units, numerous portable fire extinguishers, emergency smoke exhaust ventilation, emergency lighting, personnel evacuation planning, and refuge-rooms. The automatic fire alarm system covers smoke detection and alarming for all underground cavities, tunnels and shafts, along with surface facilities. All alarm signals will be sent to the surface fire control center to be monitored by dedicated fire-watch personnel around the clock. Relevant zones are interlocked to respond to alarm signals by taking the following actions: cut off main power supply sources and start up emergency power generator, close fire partition doors and relevant air exhaust valves to isolate the ignition source, and start emergency exhaust fans. Fire hydrant-hose units and portable fire extinguishers are the preferred method for fighting fires in the underground complex over that of sprinklers systems. The emergency smoke exhaust ventilation systems can be energized based on the location of the incident and the personnel evacuation planned route. For example, smoke can be exhausted through the vertical shaft while personnel escape via the inclined shaft; alternatively, smoke can be exhausted through the inclined shaft while personnel escape via the vertical shaft. In the case of a fire/smoke emergency the main power feeds will be shut off automatically by interlocks, and the emergency power generator will start, to supply power for smoke exhaust ventilation and lighting to ensure safe personnel evacuation. \subsection{\bf Communication system} A variety of personnel communication systems will be employed throughout the surface and underground facilities of the experimental complex. All fire emergency communication systems are directly connected to the surface fire control center monitored by dedicated fire-watch personnel around the clock. An intercom communication system that has broadcast, point-to-point calling, and central command functionality will be used for working communications throughout the experimental complex. Local telephone systems are furnished at several key locations at the surface construction site for off-site communication. High-bandwidth and private optical fiber channels will be acquired from local network providers for the experimental complex, to meet the demand for transmission of experimental data and general office communication. \section{Risk Analysis and Measures} The risk assessment for the project depends on the results of the detailed geological survey (section 10.2). Special risk assessment of the civil construction project include: risk assessment of geological hazards, seismic studies, water and soil conservation, and ecological impact on the surrounding environment. Geological survey and recommendations will be fully complied. Only major risks during the construction and operation phase, especially unpredictable risks and mitigation measures, are discussed here. The underground cavities in this project are deep with large open spans, and the inclined shaft passes through complex geological area. Therefore, unpredictable collapse of the cavities, shafts or tunnels, along with water seepage are the major risks in the construction phase, and they will have direct impact on the project schedule and construction cost. The geological survey also indicates that faults may exist in the interconnecting tunnels at the experimental hall elevation from the inclined shaft or the vertical shaft, to the experimental hall. This represents the greatest uncertainty to the construction and worker safety. Possible mitigation measures include: perform geological forecast in advance to make appropriate engineering plans, and prepare structural design solutions based on geological conditions as the project is implemented. Another major risk in the construction phase is rock burst. Although the rock burst intensity in the experimental hall predicted by the geological survey is not high, plans should be there to prepare for the worst. During the operational phase of the project, underground collapse and flooding are unpredictable risks that must be considered. It is recommended that a real-time rock monitoring system, such as a 3-D visual early warning system, should be established. This would monitor the stress and deformation of the cavity enclosing rock in real-time, to provide automatic early warning so appropriate and timely measures can be taken. Additionally, rock-safety consultants can be employed to analyze the data taken from monitoring system to provide hazard analysis and timely safety warnings. Dual-egress design also provides safety assurance for underground evacuation of personnel. Underground water conditions will be monitored and alarmed for accumulation and daily discharge to the surface. In conjunction with patrol checks of underground cavities and tunnels, any water abnormalities can be acted upon. \section{Schedule} The total civil construction period for the entire site complex is 36 months. This work has started on January 2015, and the civil construction phase of the project should be finished and handed over to the experimental users by December 2017. \section{Geological Survey} \subsection{Geographical Conditions of the Region} The site is 5~km away to Jinji Town, which is 40 km away to Kaiping City (Fig. (\ref{Fig10-6})). There are provincial road S367, S275 and village roads connecting the towns and villages. A new road will be built to connect the site to the outside. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-6.jpg} \caption{ Map of the Project Region } \label{Fig10-6} \end{center} \end{figure} The underground facility is underneath the Dashi Mountain area in Jinji Town, which has an elevation of 326 meters. The main mountain ridge extending approximately in south-north direction. The elevation of the hill area is about 30m${\sim}$300m while the topographic slope is usually 10$^{\circ}$ to 20$^{\circ}$. The main gulch exists and extends approximately in east-west direction. The depth of gulch cuts is 50 meters to 100 meters, and there is perennial water in the gulch. The largest river in the area, Dongkeng River, originates from Dongkeng Forest Farm and flows across the project area and Shenghe Village. Rivers in the region usually flow approximately in south-north direction. All of rivers are branches of Chishui River. Niuao Water Reservoir (a small reservoir) and several ponds are distributed at the southeast side of the mountain. The project area is of Southeast Asian tropical maritime monsoon climate, with frequently strong typhoons in the transition period from summer to autumn, which bring abundant rainfall. Kaiping City is surrounded by the plenty of rivers with wide water area. It is not severely cold in winter and very hot in summer. The climate is mild and abundant in rain. The annual mean air temperature is 21.7$^{\circ}$C while the annual mean humidity is as high as 82\%. Annual rainfall varies from 1,700mm to 2,400mm. The rainy season mainly concentrates from April to September while the dry climate distributes from October to February. The perennial dominant wind direction is east. Affected by the subtropical monsoon, the period from June to October in every year is a strong wind season with a force of 6-9 from east. The main active fracture zone in the region is the Enping-Xinfeng zone, which is at approx. 60 km distance to the project area. A magnitude-6 earthquake once happened near Mingcheng long fracture zone. Magnitude-4 or lower earthquakes happened near Kaiping, and a magnitude-4 earthquake happened in Hecheng area. The rmoluminescence survey value of the fracture near Gaoming is 245,200 years, and the the rmoluminescence survey value of the Hecheng-Jinji fracture is 153,300 to 350,000 years. The fractures belong to Mid-Pleistocene active faults. According to the earthquake catalogue of China Seismic Network (CSN), although there were more than 64 earthquakes happened within 150 km range of project area with magnitude-3.0 or higher since 1970, wherein, only 3 moderate earthquakes higher than magnitude 4.75 happened. No earthquake higher than 4.0 happened near the project area (within 25 km range). The earthquake distribution map in the project region shows that the earthquakes are mainly distributed in Yangjiang region in southwest. The highest earthquake is a magnitude-4.9 in Nanhai region on Mar. 26, 1995, which is 100 km away from the project area. Therefore, the impact of earthquakes on the project area is low. According to the "Seismic ground motion parameter zonation map of China", the seismic ground peak acceleration is 0.05g in 50 years. The characteristic time of the seismicresponse spectrum is 0.35 s, and the corresponding basic seismic intensity is degree VI. \subsection{Geological Survey} A detailed geological survey for the inclined shaft, vertical shaft, and experimental hall areas of the experimental station was carried out in 2013 to ascertain the geological conditions of the main constructions in the project region and provide a geological basis for engineering design and construction and raise proposals for handling main geological problems. The survey result is summarized in Fig. {\ref{Fig10-9}}. There are Yanshanian granite stocks at the location of the experimental hall. The granite mass intruded into the Palaeozoic sand rocks. The contact zone between them is a hornfelsic zone. Details are described in the following. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-9.png} \caption{ Engineering Geological survey in the Experiment Region } \label{Fig10-9} \end{center} \end{figure} \subsubsection{Engineering Geological Conditions of the Inclined Shaft} The access entry of the inclined shaft is in a mild gulch with width of 400~m in northeast to Shenghei Village. Vegetation near the access entry is dense. The gulch in front of the access entry is open and favorable for organization of construction and arrangement of surface buildings (see Fig. \ref{Fig10-10}). The formation penetrated by the inclined shaft mainly consists of Ordovician Xinchang ($O_{1x}$), Cambrian Bacun (${\in}bc^c$), hornfels and granitic intrusions. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-10.png} \caption{ Topography near the Access Entry of Inclined Shaft. The underground facility is at left. The entry of inclined shaft is at right. } \label{Fig10-10} \end{center} \end{figure} There is a water reservoir (Shantan Water Reservoir) at the upstream of the main gulch at the access entry of the inclined shaft (see Fig. \ref{Fig10-10}). The water reservoir has only $5\times10^4 m^3$ storage capacity, which has no feeding river at its upstream. The water is mainly supplied by rainfall and ground water discharge. The catchment area of the water reservoir is approximately $0.3 km^2$. Therefore, the water inflow into the water reservoir is not dangerous even in the rainy season, and will not cause severe flood discharge. There are two floodways behind the dam of the water reservoir. The minimum distance from the floodways to the access entry of the inclined shaft is approx. 130 meter. The elevation of the floodways is lower than that of the access entry by approx. 5 meter. The area at the downstream of the flood discharge area is a lower and wider plain terrain. In summary, the flood discharge from Shantang Water Reservoir in the rainy season will have no impact on the access entry of the inclined shaft. From the geological survey, the covering layer at the access entry is silty clay gravel soil in $1m\sim3m$ thickness. The bedrocks are mainly Xinchang Fm($O_{1x}$) gray and dark gray siltstone mixed with feldspar-quartz sandstones and clay shale stones in $N6^{\circ}E$, $SE{\angle}58^{\circ}$ occurrence. The bedrocks are completely weathered, locally intensely weathered, and fractured in fractured-loose structure. The condition is not perfect for shaft construction. Slope protection and water drainage should be considered. The classification of the tunnel enclosing rocks is predicted and evaluated as follows (see Fig. \ref{Fig10-11}): \begin{enumerate} \item 1413m${\sim}$1350m section: The enclosing rocks near the tunnel entrance are mainly silt rocks with locally distributed quartz sandstones and clay shale stones. They are fully or intensely weathered while some are moderately weathered. Joints and fractures are developed and filled with silt. The rocks are mainly in silt filled cataclastic structure while some are in mosaic cataclastic structure. The crown and side walls have water seepage or water dripping. The quality rating of the enclosing rocks is grade V. The cavity stability is very poor without self-stabilization time or with very short self-stabilization time. \item 1350m${\sim}$1283m section: the tunnel enclosing rocks in the region are mainly silt rocks and quartz sandstones with moderately or slightly weathered Joints and fractures. The rocks are mainly in mosaic cataclastic structure while some are in layer structure. The crown and side walls have water seepage or water dripping. The quality rating of the enclosing rocks is grade IV. The crown enclosing rocks are unstable, and the self-stabilization time is short. Various large-size deformations and damages may occur. \item 1283m${\sim}$808m section: The tunnel enclosing rocks are mainly quartz sandstones, with silt rocks distributed locally, and are slightly weathered. Joints and fractures are developed. The rocks are mainly in mosaic cataclastic structure while some are in layer structure. The crown and side walls have water seepage or water dripping. The quality rating of the enclosing rocks is grade III. The crown enclosing rocks have poor local stability. \item 808m${\sim}$418m section: The rocks are hornstones and slightly weathered. Joints are developed or undeveloped. The rocks are mainly in layer structure. The crown and side walls have water seepage or water dripping. the quality rating of the enclosing rocks is mainly grade III, and grade II locally. The cavity enclosing rocks have high overall stability. \item 418m${\sim}$0m section: the rocks are granite rocks and slightly weathered. The joints are undeveloped. The rocks are mainly in block structure. The crown and side walls have water seepage or water dripping. The quality rating of the enclosing rocks is grade II. The cavity enclosing rocks have high stability and will not have plastic deformation. \end{enumerate} Attention should be paid to water abundance in the sandstones as indicated on the cross section of geophysical prospecting along the route (A) near the ditch of Dongkeng River in the original inclined shaft route (see Fig. \ref{Fig10-11}). \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-11.jpg} \caption{ Geophysical Investigation of Inclined Shaft } \label{Fig10-11} \end{center} \end{figure} \subsubsection{ Engineering Geological Conditions of the Vertical Shaft} The vertical shaft is located in the Dongkeng Quarry. It is 589 m in depth and 6 meter in diameter. \begin{figure}[Fig10-12] \begin{center} \includegraphics[width=0.8\textwidth]{CivilConstruction/figures/Fig10-12.jpg} \caption{ Topography at the Entry of the Vertical Shaft } \label{Fig10-12} \end{center} \end{figure} The vertical shaft is located in a quarry and at the end of the gulch (see Fig. \ref{Fig10-12}). The upstream catchment area of the gulch is approx. $0.1 km^2$. In the rainy season, severe torrential flood will be produced in the gulch. Water drainage and special care must be considered during construction and operation. The vertical shaft is in granite mountain. An exploring borehole ZK1 was drilled at the location of vertical shaft for the geological survey. The borehole was deeper than bottom of water pool in the experimental hall. Various borehole tests were carried out, including terrestrial stress test, high pressure water test, specific resistance test, natural gamma test, borehole televiewer test, acoustic wave test, and ground temperature test, etc. The vertical shaft is near the granite intrusion boundary. In the area of vertical shaft, the joints, fractures, and faults are developed. 5 exposed faults in different sizes are distributed nearby. Eleven or more faults or compresso-crushed zones in different sizes were unveiled in the borehole. The classification of the vertical shaft enclosing rocks (Fig. \ref{Fig10-13}) is predicted and evaluated as follows: The rock quality of vertical shaft entry section with depth of 10 meters is grade IV. The granite rocks are moderately weathered. Joints and fractures are well developed. The section is in mosaic cataclastic structure or block structure. The shaft depth to the -118 meter, the rocks are mainly slightly weathered or moderately weathered rocks in some regions. As unveiled in the borehole surveying work, there are several small developed faults and compressive structural planes (grade IV structural planes). The rocks are mainly in sub-block structure or in block structure locally at some places. The quality rating of the enclosing rocks is grade III. The rocks of shaft bottom section are slightly weathered or fresh. The complete section is in perfect block structure. The quality rating of the enclosing rocks is grade II. The shaft wall enclosing rocks have high overall stability and will not have plastic deformation. The water inflow in the vertical shaft is predicted with Dupuit formula. According to the water pressure test result, the permeability factor of the rock mass of the vertical shaft is approximately 0.009 m/d. The shaft is approx. 589 meters in depth and 3 meters in radius. The calculated water inflow in the vertical shaft is estimated $1473m^3/d$. There is only one concern that a fault of F2 may cross the horizontal tunnel between bottom of vertical shaft and experimental hall. The rock in the F2 region is grade IV. According to the survey, the F2 fault is a compressional-shear shifted reversed fault, in $N20^{\circ}{\sim}55^{\circ}E$ and $SE{\angle}59^{\circ}{\sim}90^{\circ}$ occurrence. The occurrence variation is severe. The fault is mainly developed in granite mass, in 0.5 meters${\sim}$3.0 meters width, and extending approximately 400 meters. On the ground surface, the fault zone mainly consists of flake rocks, mylonite, cataclastic rocks, and lens with inclined scratches. The cementation is poor to moderate. As unveiled in the ZK1 borehole, the deep fault zone is mainly filled with quartz. The cementation is good to very good and is usually in healed form. The buried depth of the F2 fault is approx. 660 meters. It is speculated that the cementation in the deep part of the fault is good to very good. The F2 fault area has no surface water and is close to the dividing ridge in which the rock mass is thin and has weak water permeability. In summary, it is speculated that the F2 fault zone has weak water permeability while the water inflow is low. The fault will not have severe water inflow at the level the tunnel is excavated to the F2 fault. \begin{figure}[Fig10-13] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-13.png} \caption{ Some Core Samples from the exploring drill well in the Vertical Shaft } \label{Fig10-13} \end{center} \end{figure} \subsubsection{ Engineering Geological Conditions of the Experimental Hall } The experimental hall is located under a small hill at the southeast side of the vertical shaft. The surface elevation is 268 m, and the buried depth is approx. 763 m. The profile of hill is a mountain ridge. The experimental hall is in the granite mass, which consists of gray medium-fine grained adamellite. In the plan view, the experimental hall is relatively close to the central zone of the intruding small stocks. A geophysical prospecting diagram of the experimental hall and the vertical shaft area obtained with high-frequency magnetotelluric method is shown in Fig.~\ref{Fig10-14}. Both the vertical shaft area and the experimental hall area consist granite mass. It is speculated that there are faults in the horizonal connecting tunnel between the vertical shaft and the experimental hall. \begin{figure}[Fig10-14] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-14.jpg} \caption{ Geologic Diagram Obtained with the High-Frequency Magnetotelluric Method near the Experimental Hall } \label{Fig10-14} \end{center} \end{figure} The major engineering geological issues include: {\bf Terrestrial Stress and Rock Burst Analysis} According to the laboratory test, the volumetric weight of slightly weathered (or fresh) granite is approximately $26.6{\times}10^3kN/m^3$, the saturated uniaxial compressive strength is 80 MPa${\sim}$120 MPa as usual. Since the maximum buried depth of the experimental hall is 763 meter, according to the terrestrial stress test, it is estimated that the terrestrial stress at the experimental hall is not higher than 20MPa and the rock strength/stress ratio is 4.85. The deep core obtained in the borehole ZK1 has high integrality in whole. The core near the borehole bottom (651m buried depth) is still in long column shape. Since the experimental hall is relatively close to the center of the intruding rock stocks, it is speculated that the rock mass there has high integrality. Through comprehensive analysis, it is speculated that the stress in the experimental hall is moderated, and will not result in severe rock burst. {\bf External Water Pressure Analysis} According to the water pressure test and the high-pressure water test, the permeable rate of the granite rock is usually lower than 1 Lu and considered as weak permeability. The difference between the ground water level and the bottom of experimental hall is 577.79 meter. According to the "Code for Engineering Geological Investigation of Water Resources and Hydropower", the external water pressure reduction factor of weak permeable rocks is 0.1${\sim}$0.2. Thus, the effective external water pressure of the experimental hall is only 58 meters${\sim}$116 meters, corresponding pressure is 0.58MPa${\sim}$1.16MPa. The high-pressure water test shows that most of the joints and fracture structure of the granite did not change under the water pressure and are mainly of turbulence type while some of them are of filling type. Dilatation or erosion phenomenon is not found. In summary, the external water pressure at the experimental hall is relatively low, and its impact on the cavity stability of the experimental hall is trivial. {\bf Cavity Water Inflow Analysis} The water inflow in the experimental hall is predicted with underground water dynamics methods. The maximum water inflow in the early stage and the long-term stable water inflow are predicted with several methods. According to the water pressure test, the deep granite is slightly permeable. The maximum water inflow in early stage calculated with HirishiOshima empirical equation is approximately $395 m^3/d$. It is $527 m^3/d$ with the empirical formula method. The long-term stable water inflow was estimated to be $244 m^3/d$ and $120 m^3/d$ respectively with Ochiai Toshiro formula and empirical formula methods. In summary, the maximum water inflow in early stage and the long-term water inflow in the experimental hall are low. {\bf Underground Temperature} According to the underground temperature measurement in the borehole, ZK1, the underground temperature at 450 meter depth (-314 meter elevation) or deeper is $31{\sim}32 ^{\circ}C$ (see Fig. \ref{Fig10-15}). Therefore, temperature is not a big issue for the experiment. \begin{figure}[Fig10-15] \begin{center} \includegraphics[width=\textwidth]{CivilConstruction/figures/Fig10-15.png} \caption{ Measured Underground Temperature} \label{Fig10-15} \end{center} \end{figure} {\bf pH value} The pH value of underground water in borehole ZK1 at the location of the vertical shaft is analyzed to be 7.25${\sim}$7.75 while the pH value of ground water is near 7. It indicates that the deep ground water is medium corrosive. {\bf Rock Radioactivity} Table~\ref{tab:rockrad} shows the result of rock radioactivity measurement at different depths in borehole ZK1. All sample results meet the requirements for main building materials (as per "Code for Radionuclides Limits in Building Materials" (GB6566-2010)) except for the external exposure index Ir of three samples, which are slightly greater than 1.0. The rock material meets the criteria for Class B products according to the "Code for Radiation Protection Classification and Control of Natural Stone Material Products". \begin{table}[H] \caption{Result of Rock radioactivity Test in Borehole ZK1} \label{tab:rockrad} \centering \scalebox {0.8} { \begin{tabular}{|c|c|p{1.8cm}|p{1.8cm}|p{1.8cm}|p{1.8cm}|p{1.8cm}|p{1.8cm}|} \hline Site No. & Sampling & \multicolumn{6}{|c|}{Radioactivity} \\ \cline{3-8} & Depth & Specific Activity CRa of $^{226}Ra$ (Bq/kg) & Specific Activity CTh of $^{232}Th$ (Bq/kg) & Specific Activity CK of $^{40}K$ (Bq/kg) & Radium Equivalent Concentration C[e]Ra & Internal Irradiation Index IRa & External Irration Index Ir \\ \hline FSYY-1 & 299.18$\sim$300.28 & 166.6 & 122.1 & 1066.4 & 425.2782 & 0.8 & 1.2 \\ \hline FSYY-2 & 403.11$\sim$404.11 & 155.9 & 110.0 & 1106.3 & 401.7544 & 0.8 & 1.1 \\ \hline FSYY-3 & 557.28$\sim$558.38 & 88.6 & 83.6 & 1239.7 & 310.5536 & 0.4 & 0.9 \\ \hline FSYY-4 & 616.47$\sim$617.12 & 173.1 & 129.5 & 536.6 & 395.1458 & 0.9 & 1.1 \\ \hline FSYY-5 & 633.02$\sim$633.77 & 116.4 & 111.2 & 342.2 & 296.6336 & 0.6 & 0.8 \\ \hline FSYY-6 & 650.23$\sim$650.93 & 130.1 & 113.0 & 1062.4 & 376.1412 & 0.7 & 1.0 \\ \hline \hline \end{tabular} } \end{table} \subsubsection{ Stability Assessment of the Enclosing Rock } The experimental hall is 48 meters wide, and can be divided into two parts: the upper cavity is enclosed by the crown and high side walls (vertical section) that connect the crown part, and is mainly used to accommodate lifting appliances and other facilities (e.g., cable channel, etc.), and, the lower cavity is a cylindrical pool, 42.5 meters in diameter and 42.5 meter in depth. Utility halls are arranged around the experimental hall. The experimental hall is large and complex in structure. The rock types and long term stabilibty of underground facility should be assessed thouroughly. The length of the experimental hall is 55.25 meters. The axial is directed in $N37^{\circ}W$. The cavity enclosing rocks are granite, slightly weathered, in block structure. The developed joints are mainly steeply inclined while locally developed joints are slowly inclined and moderately inclined. The joint planes are straight, coarse, and are usually enclosed without filled material. Some joints are slightly open or fully open and filled with rock debris or kaolin film. There exists local water seepage or water dripping at some places. The quality rating of the enclosing rock is considered as grade II. Overall, the cavity enclosing rocks are stable. However, local unstable blocks may be produced under the inter-cutting action of the three groups of steeply inclined joints because of the large span of the experimental hall. According to the terrestrial stress test, the direction of maximum horizontal stress is $N10^{\circ}W{\sim}N54^{\circ}W$. Since the upper cavity is approximately in a square shape, the terrestrial stress has little influence on the selection of the axial direction. The terrestrial stress difference among different directions are small and has little impact on the stability of the side walls. Since the cavity span is large, the terrestrial stress is in static state while the redistributed stress of the cavities is complex. Stability verification must be carried out and appropriate engineering measures must be taken. The geological conditions (e.g., lithology of enclosing rocks, joint development, and underground water, et al) in the pool are the same as those of the upper cavity. However, because the main joints are steeply inclined, water pool may form an unfavorable combination with the side walls. However, the pool enclosing rocks are overall stable. The geological conditions of the ancillary cavities are the same as those of the experimental hall. The cavity enclosing rocks have high overall stability. Although the utility cavities are small, at intersections among the experimental hall and ancillary cavities, care must be taken for the blasting control in order to keep the integrality and stability of structure. \chapter{DAQ and DCS} \label{ch:DAQAndDCS} The data acquisition (DAQ) system and detector control system (DCS) are introduced separately as followed. \section{Data Acquisition System} The main task of the data acquisition (DAQ) system is to record antineutrino candidate events observed in the antineutrino detectors. In order to understand the backgrounds, other types of events are also recorded, such as cosmic muon events, low energy radioactive backgrounds etc. Therefore, the DAQ must record data from the electronics of antineutrino and muon detectors with precise timing and charge information. DAQ should build event with separated data fragment from all electronics devices. Then DAQ needs to analyze and process data to compress data and monitor data quality. At the last step, DAQ should save most relevant data to disk. \subsection{System Requirements} This section presents the DAQ main design requirements. \subsubsection{Event Rate} The central detector (CD) of JUNO is composed of about $\sim$17,000~PMTs. According to MC simulation, order of tens antineutrino candidates per day can be acquired by the CD. Most events of the CD come from radioactive backgrounds of PMT glasses, steel and liquid scintillator material. The trigger system could reduce PMT dark noise and radioactive backgrounds event rate to order of 1~kHz. DAQ system will be designed to handle a maximum trigger rate of 1~kHz. The water shielding detector is composed of about 1500~PMTs and adopt the same electronics design. But the trigger rate could be less than 100~Hz with radioactive backgrounds and muon events. The top muon detector is under design and the bandwidth requirements should be small and ignored. The calibration system could bring several hundreds to 1~kHz additional event rate. But the trigger rate could be set to a reasonable level. A supernova explosion will generate many neutrino events over a short time scale. For example, about $\sim$10k events will be generated in $\sim$10~seconds from a 10~kpc distance supernova explosion. \subsubsection{Readout Data Rate Estimation} The electronics system plans to use 12 bits FADC at 1 GHz sample rate with 2 ranges to acquire signal waveform in maximum 1~$\mu$s time window. In case of 16 bits occupation of one sample, the data size of single PMT signal could be estimated about 2k bytes including bytes of time stamp and header. Table \ref{DAQRate} shows DAQ readout data estimation, and the total data rate of CD is about $\sim$2.4~GBytes/s which means data rate per PMT is $\sim$140~kBytes/s in Physics mode. The high energy cosmic muon events of CD are estimated to be about 3 Hz and will fire all $\sim$17,000~PMTs of the CD. Neutrino and background low energy events are estimated to be less than 1 kHz with less than 3,000 and 1,000 fired PMTs in average. Water shielding detector is estimated to be with $\sim$100~Hz event rate and $\sim$10\% channels fired per event. We can set up both high and low energy calibration with similar data rate. The data rate of calibration run modes is 4.6 GB/s and means 273 KByte/s per PMT, which is about 2 times of the data rate of normal physics run mode. According to the table, 3,000~PMTs are assumed to be fired by one supernova event. The additional supernova event data rate is about 3 times larger than that of normal physics running modes. So DAQ was required 11.6 GB/s of total readout bandwidth at least and means 675~KB/s per PMT as 5 times physics mode when supernova happens with calibration together. But it will make remarkable increase for closed supernova explosion. Supernova events will only continue about 10~seconds, LED based calibration run can be paused to improve the performance. So DAQ only needs to keep about 4.6 GB/s data processing performance for calibration run mode and buffer additional 60 GB supernova data in memory to process latter. Due to the signal width is very small than 1~$\mu$s data window, we can assume to compress waveform to about 50~ns effective width with about 100~Bytes per signal. As estimated in the table about 95\% low energy event data will be removed. Then only about 200~MB/s data in Physics run mode and 300~MB/s in calibration run mode need to be built to full events and save to disk. The maximum data storage requirements is 600`MB/s when supernova happens during calibration run. \begin{table}[!hbp] \caption{JUNO DAQ Data Rate Estimation} \label{DAQRate} \newcommand{\tabincell}[2]{\begin{tabular}{@{}#1@{}}#2\end{tabular}} \begin{tabular}{p{3.5cm}|p{1.5cm}|p{2cm}|p{2cm}|p{2cm}|p{2cm}} \hline & Channels & Data size per channel(Bytes) & Fired nPMT & Trigger rate(Hz) & Data Rate (MBytes/s) \\ \hline \tabincell{c}{CD\\(Low energy events)} & 17,000 & 100(50ns)/ 2,000(1us) & 1,000 & 1k & 100/2,300 \\ \hline \tabincell{c}{CD\\(High energy events)} & 17,000 & 2,000(1us) & 17,000 & 3 & 96 \\ \hline Water Shelf Detector & 1,500 & 100(50ns)/ 2,000(1us) & 150 & 100 & 1.5/30 \\ \hline \tabincell{c}{Sum\\(Physics mode)} & & & & & 197.5/2426 \\ \hline \hline \tabincell{c}{CD\\(Calibration events)} & 17,000 & 100(50ns)/ 2,000(1us) & 1,000-17,000 & 1k &100/2,300\\ \hline \tabincell{c}{CD\\(Supernova events)} & 17,000 &100(50ns)/ 2,000(1us) & 3000 & 1k & 300/6,900\\ \hline \tabincell{c}{Sum\\(Maximum)} & & & & & 597.5/11,626\\ \hline \end{tabular} \end{table} \subsubsection{Data Process Requirements} There are several types of read out data links between DAQ and electronic. Ethernet with TCP/IP protocol could be a simple option. DAQ could organize readout branch with groups of PMTs. To deal with tens GB/s readout bandwidth, DAQ could use multiple 10 Gbps link to readout. Then DAQ should assemble completed events with data fragment from different links by trigger number or time stamp. DAQ needs to finish waveform compression if electronics cannot do that. If we want to reduce data rate further, we need software trigger to remove more background events. DAQ also needs to merge different detectors' events together and sorts events by time stamp. \subsubsection{Other Functional Requirements} DAQ also needs provide common functions like run control, run monitoring, information sharing, distributed process manager, software configure, bookkeeping, Elog, data quality monitoring, remote monitoring and so on. \subsection{Conceptual Design Schema} The JUNO experiment has many similar cases with BESIII and the Daya Bay experiments in DAQ part. We can design and develop JUNO DAQ based on BESIII, Daya Bay and ATLAS DAQ\cite{Bes3TDR,DybTDR,DybDaqTDR}. Figure \ref{daqArc} is the conceptual design schema of JUNO DAQ, DAQ readout electronics data through network. Except that network switches are placed in underground experiment hall, all other DAQ computers are deployed at ground computer room. DAQ constructs a blade servers' farm to process data, and the farm is connected with underground switches through multiple 10 gigabit fibers. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{DAQAndDCS/figures/daqArc.jpg} \caption{DAQ Conceptual Architecture Design Diagram} \label{daqArc} \end{center} \end{figure} Due to electronics systems are not easy to dynamically distribute different DAQ hosts, it is better to design a two level event building at DAQ. The first level (ROS, read out system) reads out front-end electronics (FEE) data and builds event fragments and then the second level finishes full event building. It is better that DAQ compresses waveform data at the first level to reduce unuseful data transferring. The simplest waveform compression algorithm is zero compression with a configurable threshold. Only sample data exceeding the threshold can be reserved. JUNO DAQ can use the same two level event building data flow scheme refer to BESIII event building as shown in Fig. \ref{daqEB} \cite{LifeiThesis}: \begin{enumerate} \item front-end electronics (FEE) send data to ROS(read out system) through network. \item ROS receive all data slice of one event and send event id to TSM(trigger synchronizing manager). \item TSM send event id to DFM(data flow manager) when TSM get all same event id from all ROSs. \item DFM assign event id to a free EBN(event building node). \item EBN send data request to each ROSs. \item ROSs send requested data to EBN. \item EBN receive all ROSs data fragments of one event and finish full event building, then send event id back to DFM. \item DFM send event id to ROSs to clear data buffer. \ldots \end{enumerate} \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{DAQAndDCS/figures/daqEB.pdf} \caption{DAQ Event Building Collaboration Diagram} \label{daqEB} \end{center} \end{figure} Software trigger can be performed as BESIII event filter\cite{LiuyjThesis}. The figure \ref{daqEF} shows event filter collaboration diagram, each EF (event filter) node requests events from EB node, then sends them to PT (process task) to analyze data for software trigger and data quality monitoring, at the end sends triggered event to DS (data storage) node for storage. Another option is performing software trigger before full event building as Atlas Level 2 trigger. It is better if DAQ needs compress waveform. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{DAQAndDCS/figures/daqEF.pdf} \caption{DAQ Event Filter Collaboration Diagram} \label{daqEF} \end{center} \end{figure} All three detectors can share one DS node to save data to disks. Then event merging and sorting can be performed at DS node. \subsection{R$\&$D Plan} JUNO DAQ needs the following technical research for detail design: \begin{enumerate} \item Waveform compression algorithms and read out host performance. \item Software trigger algorithms. \item Event building network and performance. \item Event sorting and storage. \item Big scale computer farm computing and software developing. \item Integrating and test with dedicated electronics and detector system. \end{enumerate} \subsection{Manufacture, Developing and Installation} DAQ will not rely on custom made DAQ hardware, so there are no manufacture issues. But we need to investigate and survey different brands and types to integrate design schema. Hardware installation will be outsourced to vendors or system integration company. DAQ software will be designed and developed based on BESIII, Daya Bay and ATLAS DAQ software. \subsection{Risk Analysis and Safety Evaluation} The waveform data compression ratio could not reach expectation of detectors and electronics real performance. There are another two challenges for DAQ up to now. One is that electronics might adopt none hardware trigger schema, the other is that supernova explosion could happen in a closer location. \subsection{Schedule} Major technical issues research according to electronics design. Finish DAQ technical design in 2013-2015. Software developing, test and debug with electronics in 2016-2017. Installation, deployment and integration in 2018-2020. \section{Detector Control System} \subsection{Requirements} The main task of the Detector Control System is to establish long-term monitoring of the parameters affecting the performance of the experimental equipment. The parameters include pressure, temperature, humidity, liquid level of the scintillation, electronics, gas pressure and the pressure in the lobby and the entire electrical and mechanical environment working status of the devices. Some subsystems need to provide device control such as calibration system, gas system, water cycle system and power system. The real time operation states of the devices will be monitored and recorded into database. When an exception occurs the system can issue timely warnings at the same time through a secure interlock. Meanwhile, the devices can be protected by the safety interlock automatically to prevent equipment damage and personal protection. \subsubsection{System Requirement} The system will meet the requirements of an effective mass of 20,000 tons of central detector, which contains about 17,000 photomultiplier tubes. Due to the large scale of the detector there are about $\sim$1000 temperature and humidity monitoring points, about $\sim$20,000 channels of high voltages, and thousands of power supplies etc. There is also pure water system, gas system as well. The collection of the system requirements is based on the design of the detector, electronics and trigger system. The design requirements include the hardware and software design, the test bed of the subsystem design and the system integration. \subsubsection{Function Requirement} According to the experimental requirements the general software framework will be designed. The high voltage, the temperature monitoring module, the hardware interface, and the hardware and the software protocols will be defined. Drivers and data acquisition framework of the exchange process will be developed. The function modules of the DCS system are shown below. \begin{figure}[htb] \label{dcsFW} \begin{center} \includegraphics{DAQAndDCS/figures/dcsFW.JPG} \caption{Framework Design of Detector Control System} \end{center} \end{figure} \subsubsection{System Schema Design} \begin{itemize} \item Content According to the actual hardware and software requirements of the experiment the system uses a hierarchical design framework. The system will be divided into the global control layer, the control layer and the data acquisition layer. The global control layer will realize the overall experimental equipment information collection, safety interlock, remote control, data storage, network information release and permissions management. Local control layer will realize the local control of the local equipment in the experiment hall. And support the local device monitoring, data recording, data upload and alarm. The data acquisition layer is response for the various hardware interfaces. It can support for embedded systems such as ARM, FPGA and standard industry interface hardware such as PLC, USB and serial RS232 interface and acquisition, which based on the network devices such as the TCP/IP data acquisition interface. The system will be built on a distributed system development method. According to the experimental equipment distribution characteristics the distributed data exchange platform will be used for the development. Global control systems share the data and the interactively control commands by information sharing pool. Configuration files will use the text format specification which can realize the remote configuration, distributed sharing and management. The system uses a module based approach of development. This method can achieve rapid integration of complex systems. From a functional view the system will be divided into data acquisition, control module, alarm module, memory module, data sharing and processing, system configuration, privilege management and user interface. Each subsystem can choose module assembly interface based on the actual system requirements. The modules can be divided by subsystem as following: \begin{enumerate} \item High voltage monitoring system, including the central detector PMT high voltage system, RPC high voltage etc. \item Detector electronics chassis monitoring system, including the central detector electronics, RPC detector electronics, both inside and outside the pool detector electronics, monitoring content includes electronic temperature, power supply, fan, over current protection, low-voltage power supply. \item Temperature and humidity monitoring system, including liquid temperature detector, temperature monitoring room. \item Gas monitoring system, including the gas support system and center detector cover gas system \item Center detector overflow tank monitoring, oil monitoring, liquid level clarity, camera monitoring and the calibration system of monitoring center. \item Experimental hall of environmental temperature, humidity, pressure monitoring system, the radon monitoring system, video monitoring. \item Control room monitoring system, database system and Webpage remote monitoring system. \item Water system \end{enumerate} \item Key Technology The key technology of detector control system is the integration framework development of the software. The system will develop a set of integrated management module according to the detector hardware requirements. The system functions will be realized by the conceptual design and the detailed modular design. For the hardware system using commercial framework or special design system the data acquisition module will be designed for the integration. The commercial software will interface with the DCS framework by the integration. The software design is completed after repeated assembly test and performance test before put into use. The user feedback will be fully considered during the software design and the commissioning period. The software will enter the formal running when the stage of training finished. The traditional design of the JUNO central detector is a stainless-steel tank plus an acrylic sphere, where stainless-steel tank is used to separate the shielding liquid (mineral oil or LAB, to be decided) from the pure water in the water pool, and the acrylic sphere is to hold the 20kt LS. The design principle and key technology of the Daya Bay neutrino experiment could be a reference. The LS separated by an organic glass container is 35.4~meters in diameter. For the quality of the target substance which is one of the most important factors affecting the precision of experiment the quality change of the target material this requires for long-term supervisory. At the same time, the temperature inside the detector needs to be monitored. Environmental temperature variation not only impact on detector performance, the detector energy scale but also on photomultiplier tube noise, amplification and electronics precision. This requires many point of monitoring from the experimental hall. Therefore, the design of reliable, compatible of monitoring system is needed. Various experimental parameters such as temperature, humidity, voltage, electronics currents, radon density accurately level need to be achieved in the long-term and stable running of the experiment. Software tools and applications which provide a software infrastructure for use in building distributed control systems will be used to operate devices. Such distributed control systems typically comprise tens or even hundreds of computers, networked together to allow communication between them and to provide control and feedback of the various parts of the device from a central control room, or even remotely over the internet. System will use Client/Server and Publish/Subscribe techniques to communicate between the various computers. Most servers (called Input/Output Controllers or IOCs) perform real-world I/O and local control tasks, and publish this information to clients using the Channel Access (CA) network protocol. \item Development Process Top development platform can be used to configure software such as LabVIEW or open source platforms such as EPICS (a widely used software platform for the large-scale control system physical device). The device driver can be developed by hardware I/O controller (IOC) server. Subsystems in the pre-design stage will build a testbed for the development of hardware drivers to test the hardware and software functions. Each detector subsystem model will be tested in testbed platform, as well as data acquisition interface. For the hardware which has no testbed device simulation models should be made with the software development team and a common definition of the interface specification, including data format, the transmission frequency, control flow, the interface distribution, database table and so on. \end{itemize} \chapter{ExecutiveSummary} \label{ch:ExecutiveSummary} \chapter{Introduction} \label{chap:intro} \section{Neutrino Physics: Status and Prospect} \subsection{Fundamentals of Massive Neutrinos} \label{subsec:mixing} In the Standard Model \cite{Weinberg} of particle physics, the weak charged-current interactions of leptons and quarks can be written as \begin{eqnarray} {\cal L}^{}_{\rm cc} = &-&\frac{g}{\sqrt{2}} \left[ \overline{\left(e \hspace{0.3cm} \mu \hspace{0.3cm} \tau \right)^{}_{\rm L}} \ \gamma^\mu \ U \left(\begin{matrix} \nu^{}_1 \cr \nu^{}_2 \cr \nu^{}_3 \cr \end{matrix}\right)^{}_{\rm L} W^-_\mu \right. \nonumber \\ &+& \left.\overline{\left(u \hspace{0.3cm} c \hspace{0.3cm} t \right)^{}_{\rm L}} \ \gamma^\mu \ V \left(\begin{matrix} d \cr s \cr b \cr \end{matrix}\right)^{}_{\rm L} W^+_\mu \right] + {\rm h.c.} \; , \end{eqnarray} where all the fermion fields are the mass eigenstates, $U$ is the $3\times 3$ Maki-Nakagawa-Sakata-Pontecorvo (MNSP) matrix \cite{MNS}, and $V$ denotes the $3\times 3$ Cabibbo-Kobayashi-Maskawa (CKM) matrix \cite{CKM}. Given the basis in which the flavor eigenstates of the three charged leptons are identified with their mass eigenstates, the flavor eigenstates of the three active neutrinos and $n$ sterile neutrinos read as \begin{eqnarray} \left(\begin{matrix} \nu^{}_e \cr \nu^{}_\mu \cr \nu^{}_\tau \cr \vdots \cr \end{matrix} \right) = \left(\begin{matrix} U^{}_{e1} & U^{}_{e2} & U^{}_{e3} & \cdots \cr U^{}_{\mu 1} & U^{}_{\mu 2} & U^{}_{\mu 3} & \cdots \cr U^{}_{\tau 1} & U^{}_{\tau 2} & U^{}_{\tau 3} & \cdots \cr \vdots & \vdots & \vdots & \ddots \cr \end{matrix} \right) \left(\begin{matrix} \nu^{}_1 \cr \nu^{}_2 \cr \nu^{}_3 \cr \vdots \cr \end{matrix} \right) \; . \end{eqnarray} where $\nu^{}_i$ is a neutrino mass eigenstate with the physical mass $m^{}_i$ (for $i=1,2, \cdots, 3+n$). Equation (1.1) tells us that a $\nu^{}_\alpha$ neutrino can be produced from the $W^+ + \ell^-_\alpha \to \nu^{}_\alpha$ interaction, and a $\nu^{}_\beta$ neutrino can be detected through the $\nu^{}_\beta \to W^+ + \ell^-_\beta$ interaction (for $\alpha, \beta = e, \mu, \tau$). So oscillation may happen if the $\nu^{}_i$ beam with energy $E \gg m^{}_i$ travels a proper distance $L$. In vacuum, the oscillation probability of the $\bar\nu^{}_e \to \bar\nu^{}_e$ transition turns out to be \begin{eqnarray} P(\overline{\nu}^{}_e \to \overline{\nu}^{}_e) = 1 - \frac{4}{\displaystyle \left(\sum_i |U^{}_{e i}|^2 \right)^2} \sum^{}_{i<j} \left(|U^{}_{e i}|^2 |U^{}_{e j}|^2 \sin^2 \frac{\Delta m^{2}_{ij} L}{4 E} \right) \; , \end{eqnarray} with $\Delta m^2_{ij} \equiv m^2_i - m^2_j$ being the mass-squared difference. Note that the denominator on the right-hand side of Equation (1.3) is not equal to one if there are heavy sterile antineutrinos which mix with the active antineutrinos but do not take part in the flavor oscillations. Note also that the terrestrial matter effects on $P(\overline{\nu}^{}_e \to \overline{\nu}^{}_e)$ are negligibly small, because the typical value of $E$ is only a few MeV and that of $L$ is usually less than several hundred km for a realistic reactor-based $\overline{\nu}^{}_e \to \overline{\nu}^{}_e$ oscillation experiment. If the $3\times 3$ MNSP matrix $U$ is exactly unitary, it can be parameterized in terms of three flavor mixing angles and three CP-violating phases in the following standard way \cite{PDG}: \begin{eqnarray} U & \hspace{-0.2cm} = \hspace{-0.2cm} & \left( \begin{matrix} 1 & 0 & 0 \cr 0 & c^{}_{23} & s^{}_{23} \cr 0 & -s^{}_{23} & c^{}_{23} \cr \end{matrix} \right) \left( \begin{matrix} c^{}_{13} & 0 & s^{}_{13} e^{-{\rm i}\delta} \cr 0 & 1 & 0 \cr -s^{}_{13} e^{{\rm i}\delta} & 0 & c^{}_{13} \cr \end{matrix} \right) \left( \begin{matrix} c^{}_{12} & s^{}_{12} & 0 \cr -s^{}_{12} & c^{}_{12} & 0 \cr 0 & 0 & 1 \cr \end{matrix} \right) P^{}_\nu \nonumber \\ & \hspace{-0.2cm} = \hspace{-0.2cm} & \left( \begin{matrix} c^{}_{12} c^{}_{13} & s^{}_{12} c^{}_{13} & s^{}_{13} e^{-{\rm i} \delta} \cr -s^{}_{12} c^{}_{23} - c^{}_{12} s^{}_{13} s^{}_{23} e^{{\rm i} \delta} & c^{}_{12} c^{}_{23} - s^{}_{12} s^{}_{13} s^{}_{23} e^{{\rm i} \delta} & c^{}_{13} s^{}_{23} \cr s^{}_{12} s^{}_{23} - c^{}_{12} s^{}_{13} c^{}_{23} e^{{\rm i} \delta} & -c^{}_{12} s^{}_{23} - s^{}_{12} s^{}_{13} c^{}_{23} e^{{\rm i} \delta} & c^{}_{13} c^{}_{23} \cr \end{matrix} \right) P^{}_\nu \; , \end{eqnarray} where $c^{}_{ij} \equiv \cos\theta^{}_{ij}$ and $s^{}_{ij} \equiv \sin\theta^{}_{ij}$ (for $ij = 12, 13, 23$) are defined, and $P^{}_\nu = {\rm Diag}\{e^{{\rm i}\rho}, e^{{\rm i}\sigma}, 1\}$ denotes the diagonal Majorana phase matrix which has nothing to do with neutrino oscillations. In this case, \begin{eqnarray} P(\overline{\nu}^{}_e \to \overline{\nu}^{}_e) &=& 1 - \sin^2 2\theta^{}_{12} c^4_{13} \sin^2 \frac{\Delta m^{2}_{21} L}{4 E}\nonumber \\ &-& \sin^2 2\theta^{}_{13} \left[ c^2_{12} \sin^2 \frac{\Delta m^{2}_{31} L}{4 E} + s^2_{12} \sin^2 \frac{\Delta m^{2}_{32} L}{4 E} \right] \; , \end{eqnarray} in which $\Delta m^2_{32} = \Delta m^2_{31} - \Delta m^2_{21}$. The oscillation terms driven by $\Delta m^2_{21}$ and $\Delta m^2_{31} \simeq \Delta m^2_{32}$ can therefore be used to determine $\theta^{}_{12}$ and $\theta^{}_{13}$, respectively. \subsection{Open Issues of Massive Neutrinos} \label{subsec:openissue} \begin{table}[t] \vspace{-0.25cm} \begin{center} \caption{The best-fit values, together with the 1$\sigma$ and 3$\sigma$ intervals, for the six three-flavor neutrino oscillation parameters from a global analysis of current experimental data \cite{GF1}.} \vspace{0.5cm} \begin{tabular}{c|c|c|c} \hline \hline Parameter & Best fit & 1$\sigma$ range & 3$\sigma$ range \\ \hline \multicolumn{4}{c}{Normal neutrino mass hierarchy $(m^{}_1 < m^{}_2 < m^{}_3$)} \\ \hline $\Delta m^2_{21}/10^{-5} ~{\rm eV}^2$ & $7.54$ & 7.32 --- 7.80 & 6.99 --- 8.18 \\ $\Delta m^2_{31}/10^{-3} ~ {\rm eV}^2$~ & $2.47$ & 2.41 --- 2.53 & 2.27 --- 2.65 \\ $\sin^2\theta_{12}/10^{-1}$ & $3.08$ & 2.91 --- 3.25 & 2.59 --- 3.59 \\ $\sin^2\theta_{13}/10^{-2}$ & $2.34$ & 2.15 --- 2.54 & 1.76 --- 2.95 \\ $\sin^2\theta_{23}/10^{-1}$ & $4.37$ & 4.14 --- 4.70 & 3.74 --- 6.26 \\ $\delta/180^\circ$ & $1.39$ & 1.12 --- 1.77 & 0.00 --- 2.00 \\ \hline \multicolumn{4}{c}{Inverted neutrino mass hierarchy $(m^{}_3 < m^{}_1 < m^{}_2$)} \\ \hline $\Delta m^2_{21}/10^{-5} ~{\rm eV}^2$ & $7.54$ & 7.32 --- 7.80 & 6.99 --- 8.18 \\ $\Delta m^2_{31}/10^{-3} ~ {\rm eV}^2$~ & $2.34$ & 2.28 --- 2.40 & 2.15 --- 2.52 \\ $\sin^2\theta_{12}/10^{-1}$ & $3.08$ & 2.91 --- 3.25 & 2.59 --- 3.59 \\ $\sin^2\theta_{13}/10^{-2}$ & $2.40$ & 2.18 --- 2.59 & 1.78 --- 2.98 \\ $\sin^2\theta_{23}/10^{-1}$ & $4.55$ & 4.24 --- 5.94 & 3.80 --- 6.41 \\ $\delta/180^\circ$ & $1.31$ & 0.98 --- 1.60 & 0.00 --- 2.00 \\ \hline\hline \end{tabular} \end{center} \end{table} Although we have known quite a lot about massive neutrinos, where the current status of neutrino oscillation measurements can be summarized in Table~2.1, we have many open questions about their fundamental properties and their unique roles in the Universe \cite{XZ}. In the following we concentrate on some intrinsic flavor issues of massive neutrinos which may related to future neutrino experiments: (a) The nature of neutrinos and their mass spectrum; {\it Question (1): Dirac or Majorana nature?} {\it Question (2): Normal or inverted mass hierarchy?} {\it Question (3): The absolute mass scale?} (b) Lepton flavor mixing pattern and CP violation; {\it Question (4): The octant of $\theta^{}_{23}$?} {\it Question (5): The Dirac CP-violating phase $\delta$?} {\it Question (6): The Majorana CP-violating phases $\rho$ and $\sigma$?} (c) Extra neutrino species and unitarity tests; {\it Question (7): Extra light or heavy sterile neutrinos?} {\it Question (8): Direct and indirect non-unitary effects?} \section{JUNO Experiment} \label{sec:juno} The Jiangmen Underground Neutrino Observatory (JUNO) is a multi-purpose neutrino experiment. It was proposed in 2008 for neutrino mass hierarchy (MH) determination by detecting reactor antineutrinos from nuclear power plants (NPPs)~\cite{zhanl2008, yfwang2008,caoj2009}. The site location is optimized to have the best sensitivity for mass hierarchy determination, which is at 53~km from both the Yangjiang and Taishan NPPs. The neutrino detector is a liquid scintillator (LS) detector with a 20~kton fiducial mass, deployed in a laboratory 700 meters underground. The JUNO experiment is located in Jinji town, Kaiping city, Jiangmen city, Guangdong province which is shown in Fig.~\ref{fig:intro:location}. The thermal power and baselines are listed in Table~\ref{tab:intro:NPP}. \begin{figure}[htb!] \centering \includegraphics[width=0.7\textwidth]{Introduction/Figs/JUNO_location.pdf} \caption{Location of the JUNO site. The distances to the nearby Yangjiang NPP and Taishan NPP are both 53 km. Daya Bay NPP is 215 km away. Huizhou and Lufeng NPPs have not been approved yet. Three metropolises, Hong Kong, Shenzhen, and Guangzhou, are also shown. \label{fig:intro:location} } \end{figure} \begin{table}[htb] \centering \begin{tabular}{|c|c|c|c|c|c|c|}\hline\hline Cores & YJ-C1 & YJ-C2 & YJ-C3 & YJ-C4 & YJ-C5 & YJ-C6 \\ \hline Power (GW) & 2.9 & 2.9 & 2.9 & 2.9 & 2.9 & 2.9 \\ \hline Baseline(km) & 52.75 & 52.84 & 52.42 & 52.51 & 52.12 & 52.21 \\ \hline\hline Cores & TS-C1 & TS-C2 & TS-C3 & TS-C4 & DYB & HZ \\ \hline Power (GW) & 4.6 & 4.6 & 4.6 & 4.6 & 17.4 & 17.4 \\ \hline Baseline(km) & 52.76 & 52.63 & 52.32 & 52.20 & 215 & 265 \\ \hline \end{tabular} \caption{Summary of the thermal power and baseline to the JUNO detector for the Yangjiang (YJ) and Taishan (TS) reactor cores, as well as the remote reactors of Daya Bay (DYB) and Huizhou (HZ).\label{tab:intro:NPP}} \end{table} JUNO consists of a central detector, a water Cherenkov detector and a muon tracker (shown in Fig.~\ref{fig:intro:det}). The central detector is a LS detector of 20~kton target mass and $3\%/\sqrt{E{\rm (MeV)}}$ energy resolution. The central detector is submerged in a water pool to be shielded from natural radioactivities from the surrounding rock and air. The water pool is equipped with PMTs to detect the Cherenkov light from muons. On top of the water pool, there is another muon detector to accurately measure the muon track. \begin{figure}[htb!] \centering \includegraphics[width=0.7\textwidth]{Introduction/Figs/JUNO_generic_detector.png} \caption{A schematic view of the JUNO detector. \label{fig:intro:det} } \end{figure} It is crucial to achieve a $3\%/\sqrt{E{\rm (MeV)}}$ energy resolution for the determination of the MH. A Monte Carlo simulation has been developed based on the Daya Bay Monte Carlo. The photoelectron yield has been tuned according to the real data of Daya Bay. The required energy resolution can be reached with the following improvements from Daya Bay \cite{DYB}: \begin{itemize} \item The PMT photocathode covergage $\geq 75$\%. \item The PMT photocathode quantum efficiency $\geq 35$\%. \item The attenuation length of the liquid scintillator $\geq 20$~m at 430~nm, which corresponds to an absorption length of 60~m with a Rayleigh scattering length of 30~m. \end{itemize} For the real experimental environments, there are many other factors beyond the photoelectron statistics that can alter the energy resolution, including the dark noise from the PMTs and electronics, the detector non-uniformity and vertex resolution, and the PMT charge resolution. A generic parametrization for the detector energy resolution is defined as \begin{equation} \frac{\sigma_{E}}{E} = \sqrt{\left(\frac{a}{\sqrt{E}}\right)^2+b^2+\left(\frac{c}{E}\right)^2}\,\;,\label{eq:mh:abcterms} \end{equation} where the visible energy $E$ is in MeV. Based on the numerical simulation as shown in the next section, a requirement for the resolution of ${a}/{\sqrt{E}}$ better than $3\%$ is approximately equivalent to the following requirement, \begin{equation} \sqrt{\left({a}\right)^2+\left({1.6\times b}\right)^2+\left(\frac{c}{1.6}\right)^2}\leq 3\%\;.\label{eq:mh:abc} \end{equation} \section{Physics Potentials} \label{sec:potentials} \subsection{Mass Hierarchy} \label{subsubsec:MH} \begin{figure \begin{center} \begin{tabular}{c} \includegraphics*[width=0.6\textwidth]{Introduction/Figs/Espec.pdf} \end{tabular} \end{center} \caption{Neutrino energy spectra (upper panel) and their ratio (lower panel) with the true (NH) and fit (IH) MHs.} \label{fig:Espec} \end{figure} The neutrino mass hierarchy (MH) answers the question whether the third generation ($\nu_3$ mass eigenstate) is heavier or lighter than the first two generations ($\nu_1$ and $\nu_2$). The normal mass hierarchy (NH) refers to $m_3 > m_1$ and the inverted mass hierarchy (IH) refers to $m_3 < m_1$. JUNO is designed to resolve the neutrino MH using precision spectral measurements of reactor antineutrino oscillations, where the general principle is shown in Fig.~\ref{fig:Espec}. In the JUNO simulation, we assume a 20~kton LS detector and a total thermal power of two reactor complexes of 36~GW. We assume nominal running time of six years, 300 effective days per year, $80\%$ detection efficiency and a detector energy resolution $3\%/\sqrt{E{\rm (MeV)}}$ as a benchmark. The simulation details can be found in Ref.~\cite{JUNO}. In Fig.~\ref{fig:Espec}, the expected neutrino energy spectrum of the true normal MH, and best-fit neutrino spectrum of the wrong inverted MH are shown in the upper panel, and the ratio of two spectra is illustrated in the lower panel. One can observe that the distinct feature of two MHs lies in the fine structures of the neutrino spectrum. To quantify the sensitivity of MH determination, we define the following quantity as the MH discriminator, \begin{equation} \Delta \chi^2_{\text{MH}}=|\chi^2_{\rm min}(\rm NH)-\chi^2_{\rm min}(\rm IH)|, \label{eq:mh:chisquare} \end{equation} where the minimization process is implemented for the oscillation parameters and systematics. The discriminator defined in Eq.~(\ref{eq:mh:chisquare}) can be used to obtain the optimal baseline, which is shown in the left panel of Fig.~\ref{fig:mh:baseline}. An optimal sensitivity of $\Delta \chi^2_{\text{MH}}\simeq16$ can be obtained for the ideal case with identical baseline at around 50~km, where the oscillation effect of $\delta m^2_{21}$ is maximal. \begin{figure \begin{center} \begin{tabular}{cc} \includegraphics*[bb=20 20 290 232, width=0.43\textwidth]{Introduction/Figs/MH_Baseline.pdf} & \includegraphics*[bb=20 16 284 214, width=0.42\textwidth]{Introduction/Figs/MH_disL.pdf} \end{tabular} \end{center} \caption{The MH discrimination ability as the function of the baseline (left panel) and as the function of the baseline difference of two reactors (right panel).} \label{fig:mh:baseline} \end{figure} The impact of the baseline difference due to multiple reactor cores is shown in the right panel of Fig.~\ref{fig:mh:baseline}, by keeping the baseline of one reactor unchanged and varying that of the other. A rapid oscillatory behavior is observed, which demonstrates the importance of baseline differences for the reactor cores. The worst case is at $\Delta L \sim 1.7$~km, where the $|\Delta m^2_{ee}|$ related oscillation is canceled between the two reactors. Taking into account the reactor power and baseline distribution of the real experimental site of JUNO, we show the reduction of the MH sensitivity in Fig.~\ref{fig:mh:ideal}, which gives a degradation of $\Delta \chi^2_{\text{MH}}\simeq5$. There are also reactor and detector related uncertainties that affect the MH sensitivity. Rate uncertainties are negligible, because most of the MH sensitivity is derived from the spectral information. On the other hand, the energy-related uncertainties are important, including the reactor spectrum error, the detector bin-to-bin error, and the energy non-linearity error. By considering realistic spectral uncertainties and taking into account the self-calibration of oscillation patterns of reactor antineutrino oscillations, we obtain the nominal MH sensitivity of JUNO as shown with the dashed lines in Fig.~\ref{fig:mh:chi2eemumu}. In addition, due to the difference between $|\Delta m^2_{ee}|$ and $|\Delta m^2_{\mu\mu}|$, precise measurements of the two mass-squared difference can provide additional sensitivity to MH, besides the sensitivity from the interference effects. In Fig.~\ref{fig:mh:chi2eemumu}, we show with the solid lines the improvement by adding a $|\Delta m^2_{\mu\mu}|$ measurement of $1\%$ precision, where an increase of $\Delta \chi^2_{\text{MH}}\simeq8$ is achieved for the MH sensitivity. \begin{figure \begin{center} \begin{tabular}{c} \includegraphics*[bb=26 22 292 222, width=0.5\textwidth]{Introduction/Figs/MH_Real_Ideal.pdf} \end{tabular} \end{center} \caption{The comparison of the MH sensitivity for the ideal and actual distributions of the reactor cores and baselines. The real distribution gives a degradation of $\Delta \chi^2_{\text{MH}}\simeq5$.} \label{fig:mh:ideal} \end{figure} \begin{figure \begin{center} \begin{tabular}{c} \includegraphics*[bb=25 20 295 228, width=0.5\textwidth]{Introduction/Figs/MH_Prior_NMH_10.pdf} \end{tabular} \end{center} \caption{the reactor-only (dashed) and combined (solid) distributions of the $\Delta\chi^2$ function, where a $1\%$ relative error of $\Delta m^2_{\mu\mu}$ is assumed and the CP-violating phase ($\delta$) is assigned to be $90^\circ/270^\circ$ ($\cos\delta=0$) for illustration. The black and red lines are for the true (normal) and false (inverted) neutrino MH, respectively. \label{fig:mh:chi2eemumu}} \end{figure} \subsection{Precision Measurement of Mixing Parameters} \label{subsec:prec} JUNO is a precision experiment in terms of huge statistics (${\cal O}(100\rm k)$ inverse beta decay (IBD) events), optimal baseline, unprecedented energy resolution ($3\%/\sqrt{E}$) and accurate energy response (better than $1\%$). Therefore, besides the determination of the neutrino mass hierarchy (MH) \cite{JUNO}, the ${\cal O}(100\rm k)$ IBD events allow JUNO to probe the fundamental properties of neutrino oscillations and access four additional parameters $\theta_{12}$, $\theta_{13}$, $\Delta m^2_{21}$, and $|\Delta m^2_{ee}|$. \begin{table}[!htb]\footnotesize \begin{center} \begin{tabular}[c]{l|l|l|l|l|l} \hline\hline & $\Delta m^2_{21}$ & $|\Delta m^2_{31}|$ & $\sin^2\theta_{12}$ & $\sin^2\theta_{13}$ & $\sin^2\theta_{23}$ \\ \hline Dominant Exps. & KamLAND & MINOS & SNO & Daya Bay & SK/T2K \\ \hline Individual 1$\sigma$ & 2.7\% \cite{KLloe} & 4.1\% \cite{MINOS} & 6.7\% \cite{SNO} & 6\% \cite{DBloe} & 14\% \cite{SKth23,T2Kth23} \\ \hline Global 1$\sigma$ & 2.6\% & 2.7\% & 4.1\% & 5\% & 11\% \\ \hline\hline \end{tabular} \caption{\label{tab:prec:current} Current precision for the five known oscillation parameters. The dominant experiments and their corresponding 1$\sigma$ accuracy and global precision from global fitting groups \cite{GF1} are shown in the first, second and third row, respectively.} \end{center} \end{table} \begin{table}[!htb] \begin{center} \begin{tabular}[c]{l|l|l|l|l|l} \hline\hline & Nominal & + B2B (1\%) & + BG & + EL (1\%) & + NL (1\%) \\ \hline $\sin^2\theta_{12}$ & 0.54\% & 0.60\% & 0.62\% & 0.64\% & 0.67\% \\ \hline $\Delta m^2_{21}$ & 0.24\% & 0.27\% & 0.29\% & 0.44\% & 0.59\% \\ \hline $|\Delta m^2_{ee}|$ & 0.27\% & 0.31\% & 0.31\% & 0.35\% & 0.44\% \\ \hline\hline \end{tabular} \caption{\label{tab:prec:syst} Precision of $\sin^2\theta_{12}$, $\Delta m^2_{21}$ and $|\Delta m^2_{ee}|$ from the nominal setup to those with more systematic uncertainties. The systematics are added one by one from the left cell to right cell.} \end{center} \end{table} Current precision for the five known oscillation parameters is summarized in Table~\ref{tab:prec:current}, where both the results from individual experiments and from global analyses \cite{GF1} are presented. We notice that most of the oscillation parameters have been measured to better than $10\%$. Among all the four parameters that are accessible in JUNO, the $\theta_{13}$ measurement from JUNO is less accurate than that of Daya Bay because the designed baseline is much larger than the optimized one ($\sim2\,\rm km$) and only one single detector is considered in the design concept. The ultimate $\sin^22\theta_{13}$ sensitivity of Daya Bay will be about $3\%$, and would lead this precision level in the foreseeable future. Therefore, we shall consider the precision measurements of $\theta_{12}$, $\Delta m^2_{21}$ and $|\Delta m^2_{ee}|$\footnote{There will be two degenerate solutions for $|\Delta m^2_{ee}|$ in case of undetermined MH. We consider the correct MH in the following studies.}. With the nominal setup as in the study of the MH measurement \cite{JUNO}, we estimate the precision of the three relevant parameters, $\sin^2\theta_{12}$, $\Delta m^2_{21}$ and $\Delta m^2_{ee}$, which can achieve the level of $0.54\%$, $0.24\%$ and $0.27\%$, respectively. Moreover, the effects of important systematic errors, such as the bin-to-bin (B2B) energy uncorrelated uncertainty, the energy linear scale (EL) uncertainty and the energy non-linear (NL) uncertainty, and the influence of background (BG) are presented. As a benchmark, a $1\%$ precision for all the systematic errors considered is assumed. The background level and uncertainties are the same as in the previous chapter for the MH determination. In Table~\ref{tab:prec:syst}, we show the precision of $\sin^2\theta_{12}$, $\Delta m^2_{21}$ and $|\Delta m^2_{ee}|$ from the nominal setup to those with more systematic uncertainties. We can see that the energy-related uncertainties are more important because most of the sensitivity is derived from the spectrum distortion due to neutrino oscillations. In summary, we can achieve the precision of $0.5\% - 0.7\%$ for the three oscillation parameters $\sin^2\theta_{12}$, $\Delta m^2_{21}$ and $|\Delta m^2_{ee}|$. Precision tests of the unitarity of the lepton mixing matrix and mass sum rule are possible with the unprecedented precision of these measurements. \subsection{Supernova Neutrinos} \label{subsec:sn} Measuring the neutrino burst from the next nearby supernova (SN) is a premier goal of low-energy neutrino physics and astrophysics. According to current understanding and numerical simulation, one expects a neutrino signal with three characteristic phases as shown in Fig.~\ref{fig:sn:SNburst}. In a high-statistics observation one should consider these essentially as three different experiments, each holding different and characteristic lessons for particle and astrophysics. \begin{itemize} \item[\em 1.]{\em Infall, Bounce and Shock Propagation}.---Few tens of ms after bounce. Prompt $\nu_e$ burst, emission of $\bar\nu_e$ at first suppressed and emission of other flavors begins. \item[\em 2.]{\em Accretion Phase (Shock Stagnation)}.---Few tens to few hundreds of ms, depending on progenitor properties and other parameters. Neutrino emission is powered by accretion flow. Luminosity in $\nu_e$ and $\bar\nu_e$ perhaps as much as a factor of two larger than each of the $\nu_x$ fluxes \item[\em 3.]{\em Neutron-star cooling}.---Lasts until 10--20~s, powered by cooling and deleptonization of the inner core on a diffusion time scale. No strong asymmetries. Details depend on final neutron-star mass and nuclear equation of state \end{itemize} \begin{figure \centering \includegraphics[width=1\textwidth]{Introduction/Figs/SNburst.pdf} \caption{Three phases of neutrino emission from a core-collapse SN, from left to right: (1)~Infall, bounce and initial shock-wave propagation, including prompt $\nu_e$ burst. (2)~Accretion phase with significant flavor differences of fluxes and spectra and time variations of the signal. (3)~Cooling of the newly formed neutron star, only small flavor differences between fluxes and spectra. \label{fig:sn:SNburst}} \end{figure} In order to estimate the expected neutrino signals at JUNO, we assume a fiducial mass of 20~kton LS. For a typical galactic SN at 10~kpc, we take the time-integrated neutrino spectra as $f_\nu(E_\nu) \propto E^\alpha_\nu \exp[-(1+\alpha)E_\nu/\langle E_\nu \rangle]$ with a nominal index $\alpha = 3$ and $\langle E_\nu \rangle$ being the average neutrino energy. Furthermore, a total neutrino energy of $E_{\rm tot} = 3 \times 10^{53}~{\rm erg}$ is assumed to be equally distributed among neutrinos and antineutrinos of three flavors. As average neutrino energies are both flavor- and time-dependent, we calculate the event rates for three representative values $\langle E_\nu \rangle = 12~{\rm MeV}$, $14~{\rm MeV}$ and $16~{\rm MeV}$, and in each case the average energy is assumed to be equal for all flavors. The total numbers of neutrino events in JUNO are summarized in Table~\ref{table:events}. Some comments on the detection channels are presented as follows. (1) The inverse beta decay (IBD) is the dominant channel for supernova neutrino detection at both scintillator and water-Cherenkov detectors. In the IBD reaction \begin{equation} \overline{\nu}_e + p \to e^+ + n \; , \label{eq: IBD} \end{equation} the neutrino energy threshold is $E^{\rm th}_\nu \approx 1.8~{\rm MeV}$. The deposition of positron kinetic energy and the annihilation of the positron give rise to a prompt signal. In addition, the neutron is captured on free protons with a lifetime of about $200~{\rm \mu s}$, producing a $2.2~{\rm MeV}$ $\gamma$. Hence the coincidence of prompt and delayed signals increases greatly the power of background rejection. \begin{table}[!t] \centering \begin{tabular}{ccccccccc} \hline \multicolumn{1}{c}{\multirow {2}{*}{Channel}} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{\multirow {2}{*}{Type}} & \multicolumn{1}{c}{} & \multicolumn{5}{c}{Events for different $\langle E_\nu \rangle$ values} \\ \cline{5-9} \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$12~{\rm MeV}$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$14~{\rm MeV}$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$16~{\rm MeV}$} \\ \hline \multicolumn{1}{l}{$\overline{\nu}_e + p \to e^+ + n$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{CC} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$4.3\times 10^3$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$5.0\times 10^3$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$5.7\times 10^3$} \\ \multicolumn{1}{l}{$\nu + p \to \nu + p$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{NC} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$6.0\times 10^2$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$1.2\times 10^3$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$2.0\times 10^3$} \\ \multicolumn{1}{l}{$\nu + e \to \nu + e$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{NC} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$3.6\times 10^2$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$3.6\times 10^2$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$3.6\times 10^2$} \\ \multicolumn{1}{l}{$\nu +~^{12}{\rm C} \to \nu +~^{12}{\rm C}^*$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{NC} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$1.7\times 10^2$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$3.2\times 10^2$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$5.2\times 10^2$} \\ \multicolumn{1}{l}{$\nu_e +~^{12}{\rm C} \to e^- +~^{12}{\rm N}$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{CC} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$4.7\times 10^1$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$9.4\times 10^1$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$1.6\times 10^2$} \\ \multicolumn{1}{l}{$\overline{\nu}_e +~^{12}{\rm C} \to e^+ +~^{12}{\rm B}$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{CC} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$6.0\times 10^1$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$1.1\times 10^2$} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{$1.6\times 10^2$} \\ \hline \end{tabular} \caption{Numbers of neutrino events in JUNO for a SN at a galactic distance of 10 kpc.} \label{table:events} \end{table} (2) As an advantage of the LS detector, the charged-current interaction on $^{12}{\rm C}$ takes place for both $\nu_e$ and $\overline{\nu}_e$ via \begin{eqnarray} && \nu_e +~^{12}{\rm C} \to e^- +~^{12}{\rm B} \; , \label{eq: CCnue}\\ && \overline{\nu}_e +~^{12}{\rm C} \to e^+ +~^{12}{\rm N} \; . \label{eq: CCnueb} \end{eqnarray} The energy threshold for $\nu_e$ is $17.34~{\rm MeV}$, while that for $\overline{\nu}_e$ is $14.39~{\rm MeV}$. The subsequent beta decays of $^{12}{\rm B}$ and $^{12}{\rm N}$ with a $20.2~{\rm ms}$ and $11~{\rm ms}$ half-life, respectively, lead to a prompt-delayed coincident signal. Hence the charged-current reactions in Eqs.~(\ref{eq: CCnue}) and (\ref{eq: CCnueb}) provide a possibility to detect separately $\nu_e$ and $\overline{\nu}_e$. (3) Neutral-current interaction on $^{12}{\rm C}$ is of crucial importance to probe neutrinos of non-electron flavors, i.e., \begin{equation} \nu +~^{12}{\rm C} \to \nu +~^{12}{\rm C}^* \; , \label{eq: NCnux} \end{equation} where $\nu$ stands for neutrinos or antineutrinos of all three flavors. A $15.11~{\rm MeV}$ $\gamma$ from the de-excitation of $^{12}{\rm C}^*$ to its ground state is a clear signal of SN neutrinos. Since neutrinos of non-electron flavors $\nu_x$ have higher average energies, the neutral-current interaction is more sensitive to $\nu_x$, providing a possibility to pin down the flavor content of supernova neutrinos. However, it is impossible to reconstruct the neutrino energy in this channel. (4) Elastic scattering of neutrinos on electrons will carry the directional information of incident neutrinos, and thus can be used to locate the SN. This is extremely important if a SN is hidden by other stars and the optical signal is obscured by galactic dust. The elastic scattering \begin{equation} \nu + e^- \to \nu + e^- \label{eq:ESe} \end{equation} is most sensitive to $\nu_e$ because of the larger cross section. However, it is difficult to determine the direction of the scattered electron in the scintillator detector. At this point, large water-Cherenkov detectors, such as Super-Kamiokande, are complementary to scintillator detectors. (5) Elastic scattering of neutrinos on protons has been proposed as a promising channel to measure supernova neutrinos of non-electron flavors~\cite{Beacom:2002hs,Dasgupta:2011wg}: \begin{equation} \nu + p \to \nu + p \; . \label{eq: ESp} \end{equation} Although the total cross section is about four times smaller than that of IBD reaction, the contributions from all the neutrinos and antineutrinos of three flavors will compensate for the reduction of cross section. As shown in Refs.~\cite{Beacom:2002hs,Dasgupta:2011wg}, it is possible to reconstruct the energy spectrum of $\nu_x$ at a large scintillator detector, which is very important to establish flavor conversions or spectral splits of SN neutrinos. For a realistic measurement of $\nu_x$ spectrum, a low-energy threshold and a satisfactory reconstruction of proton recoil energy are required. Taking into account the quenching effect, the total number of events is about $2240$ above a threshold $0.2~{\rm MeV}$. In summary, we show the time-integrated neutrino event spectra of SN neutrinos with respect to the visible energy $E^{}_{\rm d}$ in the JUNO detector for a SN at 10 kpc, where no neutrino flavor conversions are assumed for illustration and the average neutrino energies are $\langle E^{}_{\nu_e}\rangle = 12~{\rm MeV}$, $\langle E^{}_{\overline{\nu}_e}\rangle = 14~{\rm MeV}$ and $\langle E^{}_{\nu_x}\rangle = 16~{\rm MeV}$. The main reaction channels are shown together with the threshold of neutrino energies: (1) IBD (black and solid curve), $E_{\rm d} = E^{}_\nu - 0.8~{\rm MeV}$; (2) Elastic $\nu$-$p$ scattering (red and dashed curve), $E_{\rm d}$ stands for the recoil energy of proton; (3) Elastic $\nu$-$e$ scattering (blue and double-dotted-dashed curve), $E_{\rm d}$ denotes the recoil energy of electron; (4) Neutral-current reaction ${^{12}{\rm C}}(\nu, \nu^\prime){^{12}{\rm C}^*}$ (orange and dotted curve), $E_{\rm d} \approx 15.1~{\rm MeV}$; (5) Charged-current reaction ${^{12}{\rm C}}(\nu_e, e^-){^{12}{\rm N}}$ (green and dotted-dashed curve), $E_{\rm d} = E_\nu - 17.3~{\rm MeV}$; (6) Charged-current reaction ${^{12}{\rm C}}(\overline{\nu}_e, e^+){^{12}{\rm B}}$ (magenta and double-dotted curve), $E_{\rm d} = E_\nu - 13.9~{\rm MeV}$. \begin{figure}[!t] \centering \includegraphics[width=0.7\textwidth]{Introduction/Figs/spectra.pdf} \vspace{-0.4cm} \caption{The neutrino event spectra with respect to the visible energy $E^{}_{\rm d}$ in the JUNO detector for a SN at 10 kpc, where no neutrino flavor conversions are assumed for illustration and the average neutrino energies are $\langle E^{}_{\nu_e}\rangle = 12~{\rm MeV}$, $\langle E^{}_{\overline{\nu}_e}\rangle = 14~{\rm MeV}$ and $\langle E^{}_{\nu_x}\rangle = 16~{\rm MeV}$. (See the text for details)} \label{fig:spectra} \end{figure} \subsection{Diffuse Supernova Neutrino Background} \label{subsec:DSNB} The integrated neutrino flux from all past core-collapse events in the visible universe forms the diffuse supernova neutrino background (DSNB), holding information on the cosmic star-formation rate, the average core-collapse neutrino spectrum, and the rate of failed supernovae. The Super-Kamiokande water Cherenkov detector has provided the first limits \cite{Malek:2002ns,Bays:2011si} and eventually may achieve a measurement at the rate of a few events per year, depending on the implementation of its gadolinium upgrade. The JUNO detector has the potential to achieve a measurement comparable to Super-Kamiokande, benefiting from the excellent intrinsic capabilities of liquid scintillator detectors for antineutrino tagging and background rejection. Sources of correlated background events must be taken into account when defining the energy window, fiducial volume and pulse shape selection criteria for the DSNB detection. Reactor and atmospheric neutrino IBD signals define an observational window reaching from 11 to $\sim$30\,MeV. Taking into account the JUNO detector simulation of the expected signal and background rates, a $3\sigma$ signal is conceivable after 10~years of running for typical assumptions of DSNB parameters. A non-detection would strongly improve current limits and exclude a significant range of the DSNB parameter space. \begin{table}[!htb] \begin{center} \begin{tabular}{|c|cc|cc|} \hline Syst. uncertainty BG & \multicolumn{2}{c|}{5\,\%}& \multicolumn{2}{c|}{20\,\%}\\ \hline $\mathrm{\langle E_{\bar\nu_e}\rangle}$ & rate only & spectral fit & rate only & spectral fit \\ \hline 12\,MeV & $2.3\,\sigma$ & $2.5\,\sigma$ & $2.0\,\sigma$ & $2.3\,\sigma$\\ 15\,MeV & $3.5\,\sigma$ & $3.7\,\sigma$ & $3.2\,\sigma$ & $3.3\,\sigma$\\ 18\,MeV & $4.6\,\sigma$ & $4.8\,\sigma$ & $4.1\,\sigma$ & $4.3\,\sigma$\\ 21\,MeV & $5.5\,\sigma$ & $5.8\,\sigma$ & $4.9\,\sigma$ & $5.1\,\sigma$\\ \hline \end{tabular} \caption{The expected detection significance after 10 years of data taking for different DSNB models with $\langle E_{\bar\nu_e} \rangle$ ranging from 12\,MeV to 21\,MeV ($\Phi=\Phi_0$). Results are given based on either a rate-only or spectral fit analysis and assuming 5\% or 20\% for background uncertainty.} \label{tab:snd:det_sig} \end{center} \end{table} \begin{figure} \centering \includegraphics[width=0.66\textwidth]{Introduction/Figs/exclusion_plot_10y.pdf} \caption{Exclusion contour at 90\%~C.L.\ as a function of the mean energy of the SN spectrum $\langle E_{\bar\nu_e}\rangle$ and the flux amplitude (SN rate times total emitted energy per SN). We assume 5\% background uncertainty and no DSNB signal detection ($N_ {\rm det}=\langle N_{\rm bg} \rangle$) after~10\,yrs.} \label{fig:snd:exclusion} \end{figure} Following the method described in Ref.~\cite{Rolke:2004mj}, we show in Tab.~\ref{tab:snd:det_sig} the expected detection significance after 10 years of data taking for different DSNB models with $\langle E_{\bar\nu_e} \rangle$ ranging from 12\,MeV to 21\,MeV. Results are given based on either a rate-only or spectral fit analysis and assuming 5\% or 20\% for the background uncertainty. For each DSNB model it was assumed that the number of detected events equals the sum of the expected signal and background events. After 10\,years the DSNB can be detected with $>3\,\sigma$ significance if $\mathrm{\langle E_{\bar\nu_e} \rangle}\ge 15\,$MeV. If there is no positive detection of the DSNB, the current limit can be significantly improved. Assuming that the detected event spectrum equals the background expectation in the overall normalization and shape, the upper limit on the DSNB flux above 17.3\,MeV would be $\sim 0.2\,{\rm cm}^{-2}{\rm s}^{-1}$ (90\%~C.L.) after 10 years for $\langle E_{\bar\nu_e}\rangle=18$\,MeV. This limit is almost an order of magnitude better than the current value from Super-Kamiokande~\cite{Bays:2011si}. In Fig.~\ref{fig:snd:exclusion} we show the corresponding exclusion contour at 90\%~C.L. as a function of the mean energy with $\langle E_{\bar\nu_e} \rangle$ ranging from 12\,MeV to 21\,MeV. \subsection{Solar Neutrinos} \label{subsec:solar} \begin{figure} \begin{center} \begin{tabular}{c} \centerline{\includegraphics[width=0.6\linewidth]{Introduction/Figs/solar-fig-singles_clean2.pdf}} \\ \centerline{\includegraphics[width=0.6\linewidth]{Introduction/Figs/solar-fig-singles_clean1.pdf}} \end{tabular} \end{center} \caption{(a)On the top, the expected singles spectra at JUNO with the radio purity assumption $10^{-16} {\rm g/g}$ for $^{238}{\rm U}$ and $^{232}{\rm Th}$, $10^{-17} {\rm g/g}$ for $^{40}{\rm K}$ and $^{14}{\rm C}$. (b) On the bottom, assuming a background level one order of magnitude better for $^{238}{\rm U}$, $^{232}{\rm Th}$ and for $^{40}{\rm K}$, $^{14}{\rm C}$.} \label{fig:solar:simul2} \end{figure} More than forty years of experiments and phenomenological analyses brought to a crystal clear solution of the Solar neutrino problem in terms of oscillating massive neutrinos interacting with matter, with only the LMA region surviving in the mixing parameter space. The rather coherent picture emerging from all of these solar neutrino experiments is in general agreement with the values of the mixing parameters extracted from KamLAND data~\cite{KL-recent}. A global three flavor analysis, including all the solar neutrino experiments and KamLAND data and assuming CPT invariance, gives the following values for the mixing angles and the differences of the squared mass eigenvalues~\cite{Borexino-fase-I}: $ {\rm tan}^2 \, \theta_{12} = 0.457^{+0.038}_{-0.025} \, ; {\rm sin}^2 \, \theta_{13} = 0.023^{+0.014}_{-0.018} \, ; \Delta m_{21}^2 = 7.50^{+0.18}_{-0.21} \, \times \, 10^{-5} \, {\rm eV}^2\, . $ The measured values of the neutrino fluxes produced in different phases of the $pp$ chain and of the CNO cycle are consistent with the Standard Solar Model (SSM). But they are not accurate enough to solve the metallicity problem~\cite{Villante-Serenelli-2014}, discriminating between the different (high Z versus low Z) versions of the SSM. Because most of the experimental results fall somehow in the middle between the high and the low Z predictions and the uncertainties are still too high. In order to operate this discrimination, a future experimental challenge would be an even more accurate determination of the $^8{\rm B}$ and $^7{\rm Be}$ fluxes, combined with the measurement of CNO neutrinos. We present in Fig.~\ref{fig:solar:simul2}(a) the expected singles spectra at JUNO with the radio purity assumption of $10^{-16} {\rm g/g}$ for $^{238}{\rm U}$ and $^{232}{\rm Th}$, $10^{-17} {\rm g/g}$ for $^{40}{\rm K}$ and $^{14}{\rm C}$, and Fig.~\ref{fig:solar:simul2}(b) the expected singles spectra at JUNO with an ideal radio purity assumption of $10^{-17} {\rm g/g}$ for $^{238}{\rm U}$ and $^{232}{\rm Th}$, $10^{-18} {\rm g/g}$ for $^{40}{\rm K}$ and $^{14}{\rm C}$. The $^7{\rm Be}$ spectrum clearly stands out all backgrounds. More studies on the capability sensitivity deserve further simulations. \subsection{Atmospheric Neutrinos} \label{subsubsec:atm} \begin{figure} \begin{center} \begin{tabular}{c} \centerline{\includegraphics[width=0.6\linewidth]{Introduction/Figs/Atm_NH.pdf}} \\ \centerline{\includegraphics[width=0.6\linewidth]{Introduction/Figs/Atm_IH.pdf}} \end{tabular} \end{center} \caption{The MH sensitivities of atmospheric neutrinos as a function of livetime for the true NH (upper panel) and IH (lower panel) cases.} \label{fig:atm:MHyears} \end{figure} Atmospheric neutrinos are one of the most important neutrino sources for neutrino oscillation studies. Atmospheric neutrinos have broad ranges in the baseline length (15~km $\sim$ 13000~km) and energy (0.1 ~GeV $\sim$ 10~TeV). When atmospheric neutrinos pass through the Earth, the matter effect will affect the oscillations of the neutrinos. The Mikheyev-Smirnov-Wolfenstein resonance enhancement may occur in the normal mass hierarchy (NH) for neutrinos and inverted mass hierarchy (IH) for antineutrinos. Therefore, JUNO has the capability to measure the neutrino MH through detecting the upward atmospheric neutrinos. This is complementary to the JUNO reactor antineutrino results. JUNO's ability in MH determination is at the 1$\sigma$-2$\sigma$ level for 10 years of data taking. An optimistic estimation of the MH sensitivity is obtained based on several reasonable assumptions: \begin{itemize} \item[$\bullet$] An angular resolution of $10^\circ$ for the neutrino direction and $\sigma_{E_{vis}} = 0.01 \sqrt{E_{\rm vis}/{\rm GeV}}$ for the visible energy of neutrino events are assumed. \item[$\bullet$] The $\nu_e/\bar{\nu}_e$ CC events are identified and reconstructed in the $e^\pm$ visible energy range with $E^e_{\rm vis} >1$ GeV and $Y_{\rm vis} < 0.5$, where $Y_{\rm vis}$ is defined as the ratio of the hadron visible energy in the visible energy of an atmospheric neutrino event. \item[$\bullet$] The selected $\nu_e/\bar{\nu}_e$ events are divided into two samples in terms of the numbers of Michael electrons with $N_e = 0$ ($\bar\nu_e$-like events) or $N_e \geq 1$ ($\nu_e$-like events). \item[$\bullet$] The $\nu_\mu/\bar{\nu}_\mu$ CC events are selected with $L_{\mu} > 3 $ m, where $L_{\mu}$ is the track length inside the LS region for the charge lepton $\mu^\pm$. \item[$\bullet$] The selected $\nu_\mu/\bar{\nu}_\mu$ events are divided into four samples. First, these events are grouped as partially contained (PC) and fully contained (FC) events depending on whether $\mu^\pm$ can escape from the LS region or not. Second, FC events with $N_e \geq 2$ or $\mu^{-}$ capture on $^{12}$C or $Y_{\rm vis}>0.5$ are defined as FC $\nu_\mu$-like events, and the residual FC events are defined as FC $\bar\nu_\mu$-like events. Finally, PC events with $N_e \geq 1$ or $Y_{\rm vis}>0.5$ are classified as PC $\nu_\mu$-like events, and all the other PC events are PC $\bar\nu_\mu$-like events. \end{itemize} The MH sensitivity of the separate contribution of muon and electron neutrino events and their combinations in the cases of both normal and inverted MHs are presented in Fig.~\ref{fig:atm:MHyears}. We can observe from the figure that the $\nu_e/\bar{\nu}_e$ events have better sensitivity than $\nu_\mu/\bar{\nu}_\mu$ events because more $\nu_e/\bar{\nu}_e$ events with higher energies can deposit their whole visible energy in the LS region. The combined sensitivity can reach 1.8$\sigma$ (2.6$\sigma$) after 10 (20) years of running. Finally, we want to mention that atmospheric neutrinos can also be used to search for CP violation effects and precisely measure the atmospheric mixing angle $\theta_{23}$. \subsection{Geo-Neutrinos} \label{subsubsec:geo} For half a century we have established with considerable precision the Earth's surface heat flow $46 \pm 3 $~TW ($10^{12}$~watts), however we are vigorously debating what fraction of this power comes from primordial versus radioactive sources. This debate touches on the composition of the Earth, the question of chemical layering in the mantle, the nature of mantle convection, the energy needed to drive Plate Tectonics, and the power source of the geodynamo, the magnetosphere that shields the Earth for the harmful cosmic ray flux. Over the last decade particle physicists have detected the Earth's geoneutrino flux, the planetary emission of electron anti-neutrinos that are derived from naturally occurring, radioactive beta-decay events inside the Earth \cite{bib3,bib4}. Matter, including the Earth, is mostly transparent to these elusive messengers that reveal the sources of heat inside the Earth. By detecting a few events per years we are now measuring the geoneutrino flux from Thorium and Uranium inside the planet, which in turn allows us to determine the amount of radiogenic power driving the Earth's engine. Predicting the geoneutrino signal at JUNO demands that we accumulate the basic geological, geochemical and geophysical data for the regional area surrounding the detector. Experience tells us that in the continents the closest 500~km to the detector contributes half of the signal and it is this region that needs to be critically evaluated \cite{bib5}. This goal demands that the physical (density and structure) and chemical (abundance and distribution of Th and U) nature of the continent must be specified for the region. The main tasks include surveys and descriptions of the geology, seismology, heat flow, and geochemistry of the regional lithosphere. The dominate background for the geo-neutrino observation is the reactor antineutrinos. In addition, other backgrounds include the cosmic-muons spallation products ($^{9}$Li-$^{8}$He isotopes, fast neutrons), accidental coincidences of non-correlated events, and backgrounds induced by radioactive contaminants of scintillator and detector materials (i.e., $^{13}$C ($\alpha$, n)$^{16}$O). In Tab.~\ref{tab:geo:Nev}, we summarize the predicted geo-neutrino signal and backgrounds considered in the sensitivity study. \begin{table \label{tab:exp} \vspace{0.4cm} \begin{center} \begin{tabular}{ll} \hline \hline Source & Events/year \\ \hline Geoneutrinos & $408\pm 60$ \\ U chain & $311\pm 55$ \\ Th chain & $92\pm 37$ \\ Reactors & $16100\pm 900$ \\ Fast neutrons & $3.65\pm 3.65$ \\ $^{9}$Li - $^{8}$He & $657\pm 130$ \\ $^{13}$C$(\alpha,n)^{16}$O & $18.2\pm 9.1$ \\ Accidental coincidences & $401\pm 4$ \\ \hline \hline \end{tabular} \end{center} \caption{Signal and backgrounds considered in the geoneutrino sensitivity study: the number of expected events for all components contributing to the IBD spectrum in the 0.7 - 12 MeV energy region of the prompt signal. We have assumed 80\% antineutrino detection efficiency and 17.2\,m radial cut (18.35\,kton of liquid scintillator).} \label{tab:geo:Nev} \end{table} Precision of the reconstruction of geoneutrino signal is shown in Tab.~\ref{tab:geo:Fit}, for the running time in 1, 3, 5, and 10 years. Different columns refer to the measurement of geo-neutrino signal with fixed Th/U ratio, and U and Th signals fit as free and independent components. The given numbers are the position and root mean square (RMS) of the Gaussian fit to the distribution of the ratios between the number of reconstructed and generated events. It can be seen that while RMS is decreasing with longer data acquisition time, there are some systematic effects which do not depend on the acquired statistics. With 1, 3, 5, and 10 years of data, the statistical error amounts to 17\%, 10\%, 8\%, and 6\% respectively with the fixed chondritic Th/U ratio. \begin{table \footnotesize \label{tab:exp} \vspace{0.4cm} \begin{center} \begin{tabular}{cccc} \hline \hline Number of years & U + Th (fixed chondritic Th/U ratio) & U (free) & Th (free) \\ \hline 1 & $0.96\pm 0.17$ & $1.01\pm 0.32$ & $0.79\pm 0.66$ \\ 3 & $0.96\pm 0.10$ & $1.03\pm 0.19$ & $0.80\pm 0.37$ \\ 5 & $0.96\pm 0.08$ & $1.03\pm 0.15$ & $0.80\pm 0.30$ \\ 10 & $0.96\pm 0.06$ & $1.03\pm 0.11$ & $0.80\pm 0.21$ \\ \hline \hline \end{tabular} \end{center} \caption{Precision of the reconstruction of geoneutrino signal. See the text for details.} \label{tab:geo:Fit} \end{table} \subsection{Nucleon Decay} \label{subsubsec:nucleon} Being a large underground LS detector, JUNO is in an excellent position in search for nucleon decays. In particular, in the SUSY favored decay channel $p \to K^+ + \bar\nu$, JUNO will be competitive and complementary to other experiments using water Cherenkov and liquid argon detectors ~\cite{kearns:isoup}. The protons in the JUNO detector are provided by both the hydrogen nuclei and the carbon nuclei. Using the Daya Bay liquid scintillator as a reference, the H to C molar ratio is 1.639. For a 20~kt fiducial volume detector, the number of protons from hydrogen (free protons) is $1.45\times10^{33}$ and the number of protons from carbon (bound protons) is $5.30\times10^{33}$. If the proton is from hydrogen, it decays at rest. The kinetic energy of the $K^+$ is fixed by kinematics to be 105~MeV, which gives a prompt signal in the liquid scintillator. The $K^{+}$ has a lifetime of 12.4 nanoseconds and would quickly decay into five different channels, $K^{+} \to \mu^+ \nu_{\mu}$ and $K^{+} \to \pi^+ \pi^0$ are two most probably decay modes. In either case, there is a short-delayed ($\sim$12~ns) signal from the $K^{+}$ daughters. If the $K^+$ decays into $\mu^+ \nu_{\mu}$, the delayed signal comes from the $\mu^+$, which has a fixed kinetic energy of 152~MeV from the kinematics. The $\mu^+$ itself decays 2.2~$\mu s$ later into $e^+ \nu_e \bar\nu_{\mu}$, which gives a third long-delayed signal with well known (Michel electron) energy spectrum. If the $K^+$ decays into $\pi^+ \pi^0$, the $\pi^+$ deposits its kinetic energy (108~MeV) and the $\pi^0$ instantaneously decays ($\tau = 8.4\times10^{-17} s$) into primarily two gamma rays (98.80\%) with the sum of deposited energy equal to the total energy of $\pi^0$ (246~MeV). The delayed signal includes all of the aforementioned deposited energy. Then, the $\pi^+$ decays ($\tau=26$ ns) primarily into $\mu^+ \nu_{\mu}$ (99.99\%). The $\mu^+$ itself has very low kinetic energy (4.1 MeV), but it decays 2.2~$\mu s$ later into $e^+ \nu_e \bar\nu_{\mu}$, which gives the third long-delayed decay positron signal. The simulated hit time distribution of a $K^{+} \to \mu^+ \nu_{\mu}$ event is shown in Fig.~\ref{fig:nd:pdk_event}, which displays a clear three-fold coincidence: \begin{itemize} \item A prompt signal from $K^+$ and a delayed signal from its decay daughters with a time coincidence of 12 ns. \item Both the prompt and delayed signals have well-defined energy. \item One and only one decay positron with a time coincidence of 2.2 $\mu$s from the prompt signals. \end{itemize} The time coincidence and the well-defined energy provide a powerful tool to reject background, which is crucial in the proton decay search. \begin{figure \centering \includegraphics[width=0.7\textwidth]{Introduction/Figs/pdk_event.pdf} \caption{The simulated hit time distribution of photoelectrons (PEs) from a $K^{+} \to \mu^+ \nu_{\mu}$ event at JUNO.} \label{fig:nd:pdk_event} \end{figure} The sensitivity to proton lifetime, can be calculated as $\tau (p \to K^+ + \bar\nu) = N_p T R \epsilon / S$, where $N_p$ ($6.75\times10^{33}$) is the total number of protons, $T$ is the measuring time where we assumed 10 years, $R$ (84.5\%) is the $K^+$ decay branching ratio included in the analysis, and $\epsilon$ (65\%) is the total signal efficiency. $S$ is the upper limit of number of signal events at certain confidence interval, which depends on the number of observed events as well as the expected number of background events. The expected background is 0.5 events in 10 years. If no event is observed, the 90\% C.L upper limit is $S = 1.94$. The corresponding sensitivity to proton lifetime is $\tau > 1.9\times10^{34}$ yr. This represents a factor of three improvement over the current best limit from Super-Kamiokande, and starts to approach the region of interest predicted by various GUTs models. In a real experiment, the sensitivity may decrease if background fluctuates high. In the case that one event is observed (30\% probability), the 90\% C.L upper limit is $S = 3.86$. The corresponding sensitivity to proton lifetime is $\tau > 9.6\times10^{33}$ yrs. If two events are observed (7.6\% probability), the sensitivity is further reduced to $\tau > 6.8\times10^{33}$ yrs. \begin{figure \centering \includegraphics[width=0.7\textwidth]{Introduction/Figs/pdk_sens.pdf} \caption{The 90\% C.L. sensitivity to the proton lifetime in the decay mode $p \to K^+ + \overline\nu$ at JUNO as a function of time. In comparison, Super-Kamiokande's sensitivity is also projected \cite{kearns:isoup}.} \label{fig:nd:pdk_sens} \end{figure} In Fig.~\ref{fig:nd:pdk_sens} we plot the 90\% C.L. sensitivity to the proton lifetime in the decay mode $p \to K^+ + \overline\nu$ at JUNO as a function of the running time. Due to the high efficiency in measuring this mode, JUNO's sensitivity will surpass Super-Kamiokande's after 3 yrs of data taking. \subsection{Light Sterile Neutrinos} \label{subsubsec:sterile} Motivated from the anomalies of LSND \cite{Aguilar:2001ty}, MiniBooNE \cite{Aguilar-Arevalo:2013pmq}, the reactor antineutrino anomaly \cite{Mention:2011rk}, and Gallium anomaly \cite{Giunti:2010zu}, the light sterile neutrino \cite{Giunti:2013aea,Kopp:2013vaa} is regarded as one of the most promising possibilities for new physics beyond the three neutrino oscillation paradigm. Therefore, future experimental oscillation searches at short baselines are required to test the light sterile neutrino hypothesis \cite{Lasserre:2014ita}. Several possible methods of sterile neutrino studies are considered at JUNO. The first one is the use of existing reactor antineutrinos, which can test the active-sterile mixing with the mass-squared difference ranging from $10^{-5}$ to $10^{-2}$ eV$^{2}$. This parameter space is irrelevant to the short baseline oscillation, but could be tested as the sub-leading effect of solar neutrino oscillations. The direct test of short baseline oscillations at JUNO requires additional neutrino sources placed near or inside the detector. Using 50 kCi $^{144}$Ce-$^{144}$Pr as the antineutrino source at the detector center, JUNO can reach the $10^{-2}$ level of active-sterile mixing at 95\% C.L. after 1.5 yrs of data taking. On the other hand, a cyclotron-driven $^8$Li source can be employed as the decay-at-rest (DAR) neutrino source near the detector. The high DAR flux coupled with the large size of the JUNO detector allows us to make an extremely sensitive search for antineutrino disappearance in the region of short baseline oscillation anomalies. With a 60 MeV/amu cyclotron accelerator, JUNO can reach the $10^{-3}$ level of active-sterile mixing at 5$\sigma$ C.L. after 5 yrs of data taking. IsoDAR$@$JUNO will be able to map out definitively the oscillation wave. There is no other planned experiment that matches this sensitivity. \subsection{Indirect Dark Matter Search} \label{subsubsec:dark} The existence of non-baryonic dark matter (DM) in the Universe has been well established by astronomical observations. DM can also be detected indirectly by looking for the neutrino signature from DM annihilation or decays in the Galactic halo, the Sun or the Earth. In particular, the search for the DM-induced neutrino signature from the Sun has given quite tight constraints on the spin-dependent (SD) DM-proton scattering cross section $\sigma^{\rm SD}_{\chi p}$~\cite{Aartsen:2012kia}. \begin{figure \centering \includegraphics[width=0.7\textwidth]{Introduction/Figs/LS-SD.pdf}\\ \caption{ The JUNO $2\sigma$ sensitivity to the spin-dependent cross section $\sigma^{\rm SD}_{\chi p}$ after 5 yrs of data taking. The constraints from the direct detection experiments are also shown for comparison. } \label{fig:SD} \end{figure} In general, DM inside the Sun can annihilate into leptons, quarks and gauge bosons. The neutrino flux results from the decays of such final-state particles. Here we consider two annihilation modes $\chi\chi\to \tau^+\tau^-$ and $\chi\chi\to \nu\bar{\nu}$ as a benchmark. The sensitivity calculations for $\sigma^{\rm SD}_{\chi p}$ are given in Fig.~\ref{fig:SD}, where one can see that the JUNO sensitivity is much better than current direct detection constraints set by COUPP~\cite{Behnke:2012ys}, SIMPLE~\cite{Felizardo:2011uw} and PICASSO~\cite{Archambault:2012pm} experiments. Notice that the sensitivity for $m_{\chi}< 3$ GeV becomes poor due to the DM evaporation from the Sun. \subsection{Other Exotic Searches} \label{subsubsec:exotic} The standard three-neutrino mixing paradigm can describe most of the phenomena in solar, atmospheric, reactor, and accelerator neutrino oscillations \cite{PDG}. However, other new physics mechanisms can operate at a sub-leading level, and may appear at the stage of precision measurements of neutrino oscillations. Therefore, other exotic searches can be performed at JUNO using reactor neutrinos, solar neutrinos, atmospheric neutrinos and supernova neutrinos. With the reactor neutrino oscillation, JUNO can: \begin{itemize} \item test the nonstandard neutrino interactions at the source and detector \cite{ohlsson:14a,Li:2014mlo}; \item test the Lorentz and CPT invariance violation \cite{Li:2014rya}; \item test the mass-varying properties of neutrino propagation \cite{Schwetz:2005fy}. \end{itemize} Meanwhile, using solar and atmospheric neutrino oscillations, JUNO can \begin{itemize} \item test the nonstandard neutrino interactions during the propagation \cite{Ohlsson:2012kf,Bolanos:2008km}; \item test the long range forces \cite{Joshipura:2003jh}; \item test the super light sterile neutrinos \cite{deHolanda:2010am}; \item test the anomalous magnetic moment \cite{Arpesella:2008mt,Giunti:2014ixa}. \end{itemize} Therefore, with precision measurements JUNO will be a powerful playground to test different exotic phenomena of new physics beyond the Standard Model. \chapter{Liquid Scintillator} \label{ch:LiquidScintillator} \section{Introduction} The innermost part of the JUNO detector is formed by 20,000 tons of liquid scintillator (LS) contained inside an acrylic sphere of 35\,m diameter. The LS serves as target medium for the detection of neutrinos and antineutrinos. The primary reactions for reactor electron antineutrinos ($\bar\nu_e$) is the inverse beta decay on free protons, $\bar\nu_e+p\to n+e^+$, resulting in a prompt positron and a delayed signal from the neutron capture on hydrogen ($\tau\sim200\,\mu$s). The scintillator itself is a specific organic material containing molecules featuring benzene rings that can be excited by ionizing particles. Compared to other materials, this process is very efficient: About $10^4$ photons in the near UV and blue are emitted per MeV of deposited energy. The time profile of the light emission is dictated by the decay constants of the excited states of the organic molecules but also depends on the type and energy of the incident particle. The LS is composed of several materials: The solvent liquid, linear alkyl benzene (LAB), forms the bulk of the target material and is excited by the ionizing particles. This passes on the excitation to a two-component system of a fluor (PPO) and a wavelength-shifter (Bis-MSB), that are added at low concentration (a few g/l resp.~mg/l) and by subsequent Stokes' shifts increase the wavelength of the emitted photons to $\sim 430$\,nm. This shift is crucial as the long wavelength avoids spectral self-absorption by the solvent and allows the photons generated by a reaction at the very center of the target volume to reach the photosensors (PMTs) mounted on a scaffolding outside the target volume. At 430\,nm, the transparency of the compound liquid will be largely governed by the Rayleigh scattering of photons on the solvent molecules. It can be reduced by the presence of unwanted organic molecules featuring absorption bands in the wavelength region of interest. The energy resolution of JUNO is specified with 3\,\% at 1\,MeV, corresponding to at least 1,100 photoelectrons (pe) per MeV of deposited energy. Compared to Borexino, this corresponds to more than twice the pe yield. To meet this demanding requirements, both the initial light yield and the transparency of the liquid have to be optimized simultaneously. In spite of the large number of neutrinos crossing the detector and the huge target mass, the minuscule cross-section of weak interaction allows the detection of only a few tens of neutrino events per day. Low-background conditions are therefore absolutely crucial. From the point of view of the LS, this means that the concentration of radioactive impurities inside the liquid should result in an activity on the same level or below the rate of neutrino events. This is less critical in case of the inverse beta decay channel for $\bar\nu_e$ detection. The reaction provides a fast coincidence signature that can be used to suppress single-event background. As a consequence, light yield, fluorescence time profile, transparency and radiopurity are the key features of the LS. These quantities will be mainly defined by the original materials, but as well by the presence of organic and radioactive impurities, temperature and aging effects. For the next years, we foresee a broad spectrum of laboratory measurements to characterize different brands of LAB and wavelength shifters. Multiple experimental methods will be used, like GC-MS, LS-MS, UV-VIS, and ICP-MS, amongst others. A close contact with the producing companies has been established which will allow the optimization of the production quality within certain limits. Moreover, we will study a variety of purification techniques to optimize both optical performance and radiopurity of the LS. Finally, a large-scale test will be conducted in one of the Antineutrino Detectors (ADs) of the Daya-Bay experiment, which will allow a to study the effect of purification on a sample of about 20 tons of LS. The final aim is to set up a scheme for the mass production of 20,000 tons of LS. This study will comprise all steps from the initial production, the transport on site, the local facilities for liquid-handling, plants for purification and instrumentation for quality control as well as safety and cleaning methods~\cite{An:2012eh}. \section{Specifications} The basic specifications and requirements for the liquid scintillator are: \begin{itemize} \item {\bf Total mass:} 20,000 tons. \item {\bf Composition:}\\ Scintillator solvent: Linear alkyl benzene (LAB)\\ Scintillating fluor: PPO (2,5- diphenyloxazole) at 3 g/L\\ Wavelength shifter: Bis-MSB at 15~mg/L \item {\bf Light yield:}\\ Minimum of 1,100 photoelectrons (pe) per MeV \item {\bf Transparency:}\\ Attenuation length @ 430\,nm: > 22~m\\ Absorption coefficient of less than $3\times10^{-3}$ in spectral range 430$-$600\,nm \item {\bf Radiopurity:}\\ $\bar\nu_e$ detection: Concentrations of $^{238}$U, $^{232}$Th $\leq10^{-15}$\,g/g, $^{40}$K $\leq10^{-16}$\,g/g. \\* $\nu_e$ detection: Concentrations of $^{238}$U, $^{232}$Th $\leq10^{-17}$\,g/g, $^{40}$K $\leq10^{-18}$\,g/g. \end{itemize} \section{Study of Optical Properties} \subsection{Light Yield} The light yield and emission time profile of the LS are key aspects of JUNO. For a given solvent (LAB), they depend mainly on the the concentrations of the fluor PPO and the wavelength-shifter Bis-MSB. The resulting light yield and fluorescence times of the composite LS will be measured as a function of these concentrations. While for both quantities higher concentrations will in general increase the performance, the optimum concentration of PPO and Bis-MSB will be determined taking into regard the detrimental effects of self-absorption on the attenuation length and the radioimpurities introduced into the scintillator. The studies presented in the following section mainly concentrate on the characterization of the light yield as a function of fluor concentration and temperature. Moreover, the non-linearity of the light yield as a function of the energy of incident electrons will be studied as it has implications on the linearity in the energy response of positrons. Beyond the setups described here, first studies of the fluorescence properties in laboratory experiments have already been performed at the INFN institutes in Milano and Perugia and will be intensified in the future. \subsubsection{Effect of Fluor Concentrations} \label{sec:ls:ly-fluor} The light yield (LY) of an LS sample depends greatly on the concentration of fluor(s) added to the solvent. At low concentrations, it will increase almost linearly with the fluor concentration. However, the increase will become less steep once a concentration of 1$-$2\,g/L is reached. To optimize the amount of collected light with regard to self-absorption effects which are expected to play a role for concentrations of several g/L and more, the LY of a variety of LAB samples featuring varying concentrations of PPO and Bis-MSB will be studied. This will be done by a setup based on a Compton coincidence detector, an established technique for a comparative measurement of the LY of LS samples. Figure~\ref{fig:ls:ly} shows the basic principle of the experimental setup: A PMT is attached to a vessel containing the LS sample that is excited by gamma-rays from a nearby source. Without a coincidence detector, the pulse height spectrum observed by the PMT will correspond to a Compton shoulder, which will require a relatively complicated fit to evaluate the light out. Adding a coincidence detector to record the scattered gamma-rays allows to define a fixed energy deposition in the LS sample by fixing the scattering geometry. In these circumstances, the pulse height spectrum can be fitted by a simple Gaussian, providing a more accurate result. \begin{figure}[htb] \begin{center} \includegraphics[width=5cm]{LiquidScintillator/figures/LS4-1.png} \caption[Schematic of the setup for light-out measurements]{Schematic setup for the relative determination of the light out} \label{fig:ls:ly} \end{center} \end{figure} The experimental setup shown in Fig.~\ref{fig:ls:ly2} has been designed to determine the LY with less than 2\,\% uncertainty. The panels on the right display the pulse-height spectra obtained in a sample measurement and demonstrate the impact of the coincidence measurement. Beyond fluor concentrations, this setup will be used in the long term to test the effects of purification techniques, LS temperature and aging. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-2.png} \caption[Foto of the setup for LY measurements]{Setup for measuring the LS light out and preliminary result} \label{fig:ls:ly2} \end{center} \end{figure} \subsubsection{Temperature Dependence} An increase in scintillation LY has been observed when lowering the temperature of the LS. This will be studied in a somewhat modified setup shown in Fig.~\ref{fig:ls:ly3} that has been adjusted for the application of different temperature levels to the LS sample. The $^{137}$Cs $\gamma$-source, the LS vessel, the PMT and the coincidence detector are put into an enclosed thermostat, allowing to vary the temperature from -40~$^{\circ}$C to 30~$^{\circ}$C. The signals from the PMT and the coincidence detector are recorded by a CAEN DT5751 FADC unit. A temperature-resistant optical fiber is used to monitor the temperature response of the PMT. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{LiquidScintillator/figures/LS4-3.png} \caption[Setup for measuring temperature-dependence of LY]{Experimental setup to determine the temperature dependence of the LY} \label{fig:ls:ly3} \end{center} \end{figure} \subsubsection{Energy Non-linearity of Electron Signals} While the light output of LS is an almost linear function of the deposited energy, additional small non-linear terms will introduce an additional systematic uncertainty for the energy reconstruction in JUNO. Of special relevance is the non-linearity in case of positron signals that might affect the sensitivity of the mass hierarchy measurement. To take these effects correctly into account, it is important to study the energy response in small-scale laboratory setups. For practical reasons, these experiments will rely on electrons instead of positrons. A corresponding experimental setup has been realized at IHEP, designed to measure the energy non-linearity of electron signals below 1\,MeV at a precision of better than 1\,\%. The measurement is performed by inducing low-energy recoil electrons by Compton scattering of $\gamma$-quanta in a small LS sample. As the incident energy $E_\gamma$ is known, a specific electron energy $E_e$ can be selected by fixing the scattering angle $\theta$ (see above). It holds \begin{equation} E_e=E_\gamma\bigg[1-\frac{1}{1+\alpha(1-\cos\theta)}\bigg], \end{equation} where $\alpha=\frac{E\gamma}{m_e}$, with $m_e$ the electron mass. The non-linearity of the energy response is obtained by determining the ratio of observed light output and computed electron energy as a function of $\theta$. \begin{figure}[htb] \begin{center} \includegraphics[width=14cm]{LiquidScintillator/figures/LS4-4.png} \caption[Conceptual drawing of the electron non-linearity setup]{Conceptual drawing of the electron energy non-linearity setup} \label{fig:ls:enl} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-5.png} \caption[Photograph of the electron non-linearity setup]{Photograph of the electron energy non-linearity setup} \label{fig:ls:enl2} \end{center} \end{figure} Figure~\ref{fig:ls:enl} shows a conceptual drawing of the experiment, while Fig.~\ref{fig:ls:enl2} displays the actual laboratory setup.The gamma source used in the setup is $^{22}$Na with an activity of 0.3\,mCi. From $\beta^+$ decay to an excited state, it provides two $\gamma$-rays of 0.511\, MeV and one at 1.275 MeV. After passing through a lead collimator of 9\,mm diameter, the $\gamma$-rays scatter in the LS sample, which is held in a cylindrical silica cup of 5\,cm diameter and height. The light generated by the recoil electron is collected by a PMT (XP2020) attached to the LS cup. Seven coincidence detectors are placed at various angles ($20^{\circ}$, $30^{\circ}$, $50^{\circ}$, $60^{\circ}$, $80^{\circ}$, $100^{\circ}$, $110^{\circ}$) at 60\,cm distance from the LS cup.These coincidence detectors consist of an inorganic crystal scintillator (LaBr) and a PMT (XP2020). The signals acquired simultaneously by all PMTs pass a Fan-In-Fan-Out unit (CAEN N625) before being sent to a trigger board (CAENN405) and a FADC (CAEN N6742). Using an array of PMTs rather than a single PMT mounted to a rotatable arm is reducing the required acquisition time and avoids variations of the PMT response induced by the rotation of the dynode chain in magnetic stray fields. Preliminary results are expected for late 2015. At the present state, the uncertainty on electron energy non-linearity is estimated to be smaller than 2\,\%. The results will be used as well in the data analysis of the Daya Bay experiment. \subsection{Optical Transparency} Compared with other neutrino experiments, the optical transparency of the LS in JUNO is absolutely crucial. By scaling up the target volume from several hundreds to 20,000 tons, the diameter of the target volume in JUNO will be 35.4~m. If light transmission in the liquid scintillator is too low, scintillation photons from a neutrino interaction in the center of the detector will be absorbed by the liquid itself before reaching the photomultipliers positioned at the verge of the volume. The light transmission is described by the attenuation length which quantifies the distance over which the original light intensity is reduced to a fraction of $1/e$. This attenuation is due to two kind of processes: Light absorption on organic impurities, in which case the photon can either be fully absorbed or re-emitted, and light scattering off the solvent molecules. In the latter case, light is not lost but merely re-directed and may still be detected. Several laboratory scale experiments will be conducted to characterize LS samples regarding their transparency: Measurements include the wavelength-dependent attenuation spectrum of LS in relatively short cells (10\,cm), precise measurements of the LS attenuation length over large samples (several meters) and independent measurements of the scattering length of LS. Moreover, the fraction of organic impurities of different samples will be determined and the long-term stability of LS samples tested. \subsubsection{Characterization of Attenuation Spectrum} The attenuation spectrum of a material corresponds to the fraction of incident radiation absorbed by the sample over a certain thickness as a function of the wavelength of the incident light. The measurement is performed by commercial UV-Vis spectrometers that usually are sensitive to a wavelength range from 190 to 900\,nm. These measurements require only a low amount of time and a relatively small amount of LS as typical sample cells contain only 30\,ml or less. Therefore, this method is very suitable for the real-time monitoring of solvent or LS samples. Periodic measurements can be used to study the long-term stability of LS samples. A first characterization of samples can be done based only on the solvent LAB (without adding the fluors). It is found that the resulting UV-Vis spectra vary for LAB samples from different providers. Fig.~\ref{fig:ls:specsolvent} shows the attenuation curves of LAB samples produced by several companies that apply different manufacturing technologies. Substantial divergences are present in the wavelength region below 500\,nm which is most important for the propagation of scintillation light. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-6.png} \caption[UV-Vis attenuation spectra of LAB samples]{The UV-Vis attenuation spectra of LAB from different manufacturers} \label{fig:ls:specsolvent} \end{center} \end{figure} When loading the solvent with the fluors necessary to form the liquid scintillator, the attenuation spectrum changes. The shift of the absorption band to longer wavelengths is illustrated in Fig.~\ref{fig:ls:specfluors}. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-7.png} \caption[UV-Vis attenuation spectrum: Impact of fluors]{The UV-Vis spectra of LAB and LAB-based liquid scintillators} \label{fig:ls:specfluors} \end{center} \end{figure} \subsubsection{Long-range Measurement of Attenuation Length} As stated above, the light intensity generated by particle interactions in the bulk of the LS volume will be attenuated by scattering and absorption before reaching the PMTs. The relation between the intensity $I(d)$ remaining after propagation by a distance $d$ in the LS and the initial light intensity $I_0$ can be expressed by \begin{equation} \label{eq:ls:att} I(d)=I_0 \exp(-d/\ell_{\rm att}). \end{equation} The length scale $\ell_{\rm att}$ quantifying the distance after which $I(d)$ is reduced to $1/e$ of the initial value $I_0$ is the attenuation length. The accuracy that can be reached by UV-vis spectrometers on the attenuation length is limited if $\ell_{\rm att}\geq10$\,m. This is due to the very small decrease of the light intensity when passing through a sample cell of less than 10\,cm in length and thus considerably shorter than the expected attenuation length of the LS ($\ell_{\rm att}>20$\,m). Considering the importance of exact knowledge of $\ell_{\rm att}$ for light propagation in the large LS volume, dedicated measurements are performed in the wavelength range around 430\,nm which corresponds to the emission band of Bis-MSB and at the same time represents the optimum intersection of scintillator transparency and PMT photosensitivity. Figure~\ref{fig:ls:att} shows the corresponding laboratory device at IHEP and a schematic sketch of the setup. A similar device with an even longer sample cell is currently developed at TU Munich (Germany). Measurements of this kind will be especially valuable to assess the effect of purification techniques on the optical transparency of the LS. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-8.png} \caption[Setup for attenuation length measurement]{Setup for a long-range measurement of the attenuation length} \label{fig:ls:att} \end{center} \end{figure} \subsubsection{Determination of Scattering Lengths} The attenuation of the initial scintillation light can be caused by a variety of processes: Absorption by organic impurities, self-absorption by the fluors (primarily Bis-MSB), Mie scattering off dust or particulates in suspension in the LS, or Rayleigh scattering of the solvent molecules. In general, the relation between the corresponding optical length scales can be given as: \begin{equation} \frac{1}{\ell_{\rm att}}=\frac{1}{\ell_{\rm abs}}+\frac{1}{\ell_{\rm are}}+\frac{1}{\ell_{\rm scat}}, \end{equation} where $\ell_{\rm abs}$ describes the absorption length without re-emission, $\ell_{\rm are}$ means absorption followed by re-emission, and $\ell_{\rm scat}$ is the Rayleigh scattering length. Only the light absorbed without re-emission will be fully lost for event reconstruction. All other processes merely divert a photon from its original path so that it can still be detected by a PMT. For an ideal scintillator without impurities, $\ell_{\rm att}\approx \ell_{\rm Ray}$ as Rayleigh scattering is the only irreducible process. In comparison, a conceivable contribution from Mie scattering on dust particles is expected to be negligible because the LS will be virtually dust-free after purification. Absorption-reemission might play a role as the long-wavelength tail of the Bis-MSB absorption band reaches into its emission spectrum centered around 430\,nm and might thus be relevant for light propagation. The physics requirements of JUNO necessitate to come at least close to this ideal state in order to maximize the light output of the target LS. On the other hand, light emitted in a neutrino interaction and scattered while traveling to the PMTs will be deflected from a straight line of sight, influencing both spatial reconstruction and pulse-shaping capabilities of the experiment. Therefore, a precise measurement of the Rayleigh scattering length will allow to determine the ultimate transparency limit for an LS based on the given solvent (LAB in case of JUNO), and at the same time provide valuable input for MC simulations that study the expected detector performance. Moreover, the impact of different purification techniques on the composition of the scintillator can be monitored by studying the combined results of attenuation and scattering lengths measurements (see below). The Rayleigh scattering length expected for a given material can be calculated based on the Einstein-Smoluchowski-Cabannes (ESC) formula, \begin{equation} \frac{1}{l_{ray}}=\frac{8\pi^3}{3\lambda^4}kT\bigg[\rho\bigg(\frac{\partial\epsilon}{\partial\rho}\bigg)_T\bigg]\beta_T\bigg(\frac{6+3\delta}{6-7\delta}\bigg) \end{equation} where $\lambda$ is the wavelength of scattered light, $\rho$ is the density and $\epsilon$ the average dielectric constant of the liquid, $k$ is the Boltzmann constant, $T$ is the temperature, $\beta_T$ is the isothermal compressibility and $\delta$ is the depolarization ratio. Based on the ESC formula, the Rayleigh scattering of LAB can be obtained by measuring the reflectivity, isothermal compressibility, and the depolarization. Comparing to direct measurements of the scattered intensity (see below), this relative measurement can lower the systematic error on the depolarization ratio and directly derive the scattering length as a function of the incident wavelengths $\lambda$ of the scattered lights. Figure~\ref{fig:ls:scat1} shows a schematic drawing of the setup located at Wuhan University. Complementarily to the described measurements, a direct measurement of the amount of scattered light will be performed at the University of Mainz (Germany). By investigating the dependence of the scattered intensity on the incident wavelength, polarization and scattering angle, contributions from different scattering processes as well as absorption/re-emission can be resolved. On the one hand, this will provide an important cross-check of the Rayleigh scattering length measurement. On the other hand, it will allow to scrutinize purified samples for residual dust particles and organic impurities and, more importantly, to assess the impact of Bis-MSB self-absorption on light propagation. Measurements of this kind have been already performed in Ref.~\cite{Wurm:2010ad}. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-10.png} \caption[Schematic setup for scattering length measurements]{Schematic diagram of Rayleigh scattering measuremens} \label{fig:ls:scat1} \end{center} \end{figure} \subsubsection{Organic Impurities from GC-MS Measurements} \label{sec:ls:gcms} Gas chromatography combined with mass spectrometry (GC-MS) is a commonly used technique to analyze the composition of complex organic mixtures. GC-MS analysis combines the high resolution of gas chromatography with the high sensitivity of mass spectrometry. Type and concentration of impurities in LAB as well as the primary LAB components can be resolved and identified by the GC-MS method. The results will provide useful information for improving the purification scheme. \begin{figure}[htb] \begin{center} \includegraphics[width=14cm]{LiquidScintillator/figures/LS4-9.png} \caption[GC-MS Chromatogram of LAB]{Total ion chromatogram of LAB and impurities in LAB from GC-MS measurements} \label{fig:ls:gcms} \end{center} \end{figure} We have analyzed a sample of commercially available LAB produced by the Nanjing LAB factory. In Fig.~\ref{fig:ls:gcms}, the black curve indicates the components of LAB. The remaining curves stem from the impurities. The result suggests that \begin{enumerate} \item There are many impurities in LAB, which can be divided into several categories: fluorene, naphathelene derivatives, biphenyl derivatives, diphenyl alkane, and small amounts of alcohol, ketone, and ester. \item The retention times of impurities coincide with LAB components. The impurities in LAB must be separated and enriched before GC-MS analysis. \item The resolution of the chromatogram is not satisfying. The analytical conditions of GC-MS must be optimized in order to perform precise qualitative and quantitative analyses. \end{enumerate} Therefore, both the LAB pre-treatment methods and instrumental conditions have to be optimized in order to establish the GC-MS method for studying type and level of impurities in LAB samples. GC-MS is an important addition to the characterization of the LS as it provides data on the chemical composition complementary to the optical measurements. We expect that following the adaptation of the method we will be able to clearly identify the main organic impurities in different brands of LAB. This will allow us to cooperate with the LAB manufacturers in order to achieve an optimization of the chemical purity of the LS raw materials. \subsection{Long-term Stability} JUNO is supposed to be in operation for at least 10 years. This poses the question of the long-term stability of the composite liquid and especially of its scintillation and optical properties. A common method of investigation often applied in chemistry are aging tests relying on heating the samples. According to the Van't Hoff equation, the rate of possible deteriorating processes will increase by a factor 2 to 4 for every 10~C$^{\circ}$ over room temperature (25~$^{\circ}$C). Therefore, a 6-months aging test performed at 40~$^{\circ}$C is equivalent to 1.5 - 4 years at 25~$^{\circ}$C. The temperature-treated samples will be characterized by their light out, attenuation length and absorption spectrum. During the tests, the LS samples will be filled into containers of stainless steel. This will allow the performance of a material compatibility test with the LS at the same time. Different types of stainless steel (SS316 or SS304) will be investigated. Some ageing experiments are being conducted on LS. Figure~\ref{fig:ls:aging} shows two 15-liter stainless steel vessels. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-11.png} \caption[Setup for LS aging tests]{Laboratory setup for LS aging tests} \label{fig:ls:aging} \end{center} \end{figure} \subsection{Selection of Raw Materials} \subsubsection{Solvent} LAB is a family of organic compounds with the formula C$_6$H$_5$C$_n$H$_{2n+1}$. When used as a detergent, $n$ lies typically between 10 and 16. For JUNO, a specific product selected for $n=10-13$ will be chosen as it shows better optical properties. In China, the list of LAB manufacturers includes the Nanjing LAB factories of Jinling Petrochemical Company, the Fushun wash \& chemical factory of the Fushun Petrochemical Company, and the Jintung Petrochemical Corporation Ltd. The annual output of each factory is very large. Only several months will be needed to produce 20 ktons of LAB for JUNO. In general, LAB production in China is based on a HF catalyst (HF-acid method). Outside China, an alternative technology based on a non-corrosive solid acid catalyst (Solid-acid method) is more common. The advantages of the HF acid method are the high conversion efficiency and the long-term stability in operation. The disadvantage is that it generates more by-products because the syntheisizing reaction is more violent than in case of the solid acid method. Therefore, the general expectation is that production by the solid-acid method should result in better optical properties of LAB. Instead, our investigations based on the attenuation spectra of LAB samples suggest that the method of synthesis is not the critical factor for the resulting optical properties. Figure~\ref{fig:ls:optpur1} compiles the attenuation spectra of LAB samples produced both by the HF-acid method in the Nanjing LAB factory (specially ordered by optimizing production flow, not commercially available) and by the solid-acid method in an Egyptian plant (Helm AG). The sample from Nanjing shows much lower absorption for the spectral range below 410\,nm. As the spectral measurement is not sufficiently sensitive at the most interesting wavelength of 430\,nm, the sample was also tested in the long cell at IHEP. Again, the Nanjing sample showed a much longer attenuation length $\ell_{\rm att}\approx 20$\,m, while the LAB by Helm featured only $\ell_{\rm att}\approx12$\,m. Based on these results, it seems that Chinese manufacturers using the HF-acid method should be able to provide LAB of sufficient quality for JUNO as long as the quality of the raw materials is assured and the production flow is optimized. For this, we foresee a cooperation with domestic LAB factories. In addition, we will further investigate samples of LAB from foreign companies. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-12.png} \caption[UV-Vis attenuation spectra of LAB from different providers]{UV-Vis attenuation spectra of LAB produced by HF acid method and solid acid method} \label{fig:ls:optpur1} \end{center} \end{figure} \subsubsection{Fluor (PPO) and Wavelength Shifter (bis-MSB)} As described above, a system of two fluors will be used in JUNO: LAB will non-radiatively pass on the excitations to the primary fluor PPO, which will in turn perform a mainly non-radiative transfer to the secondary wavelength shifter Bis-MSB. As a consequence, the effective light emission (including optical propagation effects) will be shifted to 430\,nm. Due to the self-absorption of the fluors, the absorption spectra of the complete LS will differ significantly from pure LAB. This might be further enhanced by optical impurities introduced along with the fluors. Figure~\ref{fig:ls:fluor} illustrates the influence of PPO on the absorption spectrum of LS. If PPO concentration is increased from the baseline value of 3\,g/L to 10\,g/L, an effect induced by the tails of the absorption band of PPO becomes visible that extends up to 460\,nm. This clearly demonstrates that the optimum PPO concentration must be balanced between light yield and optical transparency in the 430\,nm range, taking re-emission processes into account. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-16.png} \caption[UV-vis attenuation spectra of LAB doped with PPO]{UV-Vis attenuation spectra of LAB doped with PPO} \label{fig:ls:fluor} \end{center} \end{figure} Organic impurities introduced with PPO will as well have an impact on attenuation. Figure~\ref{fig:ls:fluor2} shows the attenuation spectra of LAB samples containing 10\,g/L PPO. The PPO was obtained from two different sources: one from rpi (Research Products International Corp), one from the Ukraine. For the latter, spectra for both the original and purified PPO (by Haiso pharmaceutical chemical Co., Ltd., Hubei, China) are shown separately. The positive impact of purification on transparency is clearly visible. The original quality of PPO from rpi seems to be comparable to purified PPO from Ukraine. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-17.png} \caption[UV-vis attenuation spectra with PPO from selected providers]{UV-vis attenuation spectra of LAB with PPO from selected providers} \label{fig:ls:fluor2} \end{center} \end{figure} A similar or greater impact on the attenuation spectra is expected from the addition of the secondary fluor Bis-MSB. Like in the case of PPO, the chosen concentration has to be optimized and the impact of impurities studied. Two measured attenuation spectra of Bis-MSB in LAB are given in Fig.~\ref{fig:ls:fluor3}. While the same concentration of Bis-MSB was realized and the effective light yield of both solutions is identical, the absorption spectra differ. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-18.png} \caption[UV-vis attenuation spectra of LAB with Bis-MSB]{UV-vis attenuation spectra of Bis-MSB in LAB solutions} \label{fig:ls:fluor3} \end{center} \end{figure} \subsection{Purification Techniques} \label{sec:ls:opt_purification} Beyond the choice of suitable raw materials, optical properties can be further improved by dedicated purification methods. Several methods of purification have been investigated at IHEP, including distillation, column purification, water extraction and nitrogen stripping. Figure~\ref{fig:ls:optpur2} displays a compilation of attenuation spectra of LAB samples by different manufacturers. For comparison, the spectrum of Nanjing LAB after purification in an aluminum column is shown. It clearly shows the best performance in the wavelength region of interest. Similar studies are currently performed at INFN and German institutes~\cite{Ford:2011zza}. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-13.png} \caption[UV-vis attenuation spectra before and after purification]{UV-vis attenuation spectra of LAB samples from different providers. In case of Nanjing LAB spectra taken before and after purification in an aluminum column are shown.} \label{fig:ls:optpur2} \end{center} \end{figure} \subsubsection{Fractional Distillation of LAB} Distillation relies on the (partial) evaporation of the initial liquid to separate its constituents during vapor condensation. However, simple one-stage distillation usually does not meet the requirements for separating complex mixtures. This is usually achieved by multi-stage or fractional distillation: The technique relies on a heat exchange between upward-streaming vapor and downward-streaming condensate inside the distillation column. By this re-heating, the concentration of volatile components in the upper vapor increases, while less volatile components are enriched in the lower condensate. As the vapor keeps rising and the condensate keeps falling, several equilibrium states between gas and liquid are established along the column. Therefore, distillation occurs in multiple stages. If the fractional column is sufficiently long, an efficient separation of volatile from solid components is achieved. In a first step, a setup including a single-stage distillation column was built up at IHEP. However, the resulting purification efficiency shown in Fig.~\ref{fig:ls:dist} proved insufficient. Therefore, the setup was upgraded to fractional distillation by the addition of a Vigreux column. This resulted in a substantial improvement due to the increased number of theoretical plates. \begin{figure}[htb] \begin{center} \includegraphics[width=10.3cm]{LiquidScintillator/figures/LS4-14.png} \caption[UV-Vis spectra of LAB before and after distillation]{UV-Vis spectra of LAB before and after distillation} \label{fig:ls:dist} \end{center} \end{figure} Compared to single-stage distillation, fractional distillation is much more efficient in the removal of the impurities causing absorption lines in the region from 330 to 410\,nm. However, in case of the IHEP setup, neither method provides a visible improvement in region around 430\,nm. Possible explanation comprise: \begin{itemize} \item The boiling points of the organic impurities influencing absorption at 430\,nm are too close to LAB to be removed by distillation. \item New organic impurities are generated by distillation, e.g.~by the break-up of LAB molecules. \item UV-vis spectroscopy is not sufficiently accurate to resolve an improvement. \end{itemize} While the unexpected absence of a beneficial affect of distillation in the 430-nm range requires further exploration, this method clearly improves transparency in the wavelength region below 410~nm. However, First results from a similar test setup at INFN Perugia in fact show an increase of the attenuation length in the region of interest. The Vigreux column proved to be simple and easy to operate for laboratory-scale operations. However, a system based on fractional columns will be necessary for the distillation of large amounts of LAB for JUNO. \subsubsection{Purification of LAB by Aluminum Columns} The impact of purification in an aluminum (Al$_2$O$_3$) column on the optical properties of LAB has been studied in detail at IHEP. The results show that column chromatography is very effective in removing optical impurities. The attenuation spectra of Fig.~\ref{fig:ls:column} clearly demonstrate that column purification does not only remove the impurities featuring absorption in the region from 330 to 410\,nm, but also improves transparency in the critical range around 430~nm. These findings have been confirmed by long-range attenuation length measurements. Column purification increased the attenuation length of an LAB sample originally featuring $\ell_{\rm att}\approx 9$\,m to 20\,m. Moreover, the most transparent raw product, Nanjing special LAB, featured an attenuation length of 25\,m after purification. Provided the manufacturer can produce this high-quality LAB in large quantities, aluminum column purification will be sufficient to obtain a LS meeting the experimental specifications. \begin{figure}[htb] \begin{center} \includegraphics[width=10.1cm]{LiquidScintillator/figures/LS4-15.png} \caption[UV-Vis spectra of LAB before and after column purification]{UV-Vis spectra of LAB before and after purification in an aluminum column} \label{fig:ls:column} \end{center} \end{figure} Further laboratory studies are foreseen for the future: \begin{enumerate} \item Optimization of the Al$_2$O$_3$ column purification \item Study of alternative column packings, such as a molecular sieve \item Regeneration of the used Al$_2$O$_3$ \item Types and concentration of impurities in LAB which are removed by the Al$_2$O$_3$ column. \end{enumerate} \subsubsection{Further Purification Techniques Applied to LS} Distillation cannot completely remove the radioactive contaminants dissolved in LAB. Thus, water extraction and nitrogen stripping are applied after mixing of distilled LAB, PPO and Bis-MSB. \medskip\\ {\bf Water extraction} utilizes the polarity of water molecules to extract polarized impurities and radioactive free-state metal ions from LAB. The method efficiently removes metallic radionuclides such as $^{238}$Th, $^{232}$U, $^{210}$Bi and $^{40}$K. \medskip\\ {\bf Gas stripping} is used for purging the LS from radioactive gases, mainly $^{85}$Kr and $^{39}$Ar, oxygen (which will decrease the light yield due to photon quenching) and water (introduced by water extraction). \subsubsection{Purification of Fluors} The purification of the fluors has to be carefully considered for the JUNO experiment. In the Daya Bay experiment, purification of PPO was performed in cooperation with a domestic chemical company. A domestic producer has also been identified for Bis-MSB: The provided sample meets the requirements concerning light yield, while it will have to be purified to obtain the necessary optical transparency. Moreover, experience from Borexino suggests that the radiopurity of the fluors is an important issue that has to be studied by laboratory experiments. \section{Radiopurity Studies} An important prerequisite for the detection of low-energy neutrinos is the radiopurity of the detector and especially the target material, i.e.~the LS. The residual contamination aimed for is on the level of the neutrino event rate, so in the order of hundreds of counts per day within the target volume. In case of antineutrino detection, the inverse beta decay coincidence signal somewhat relaxes the requirement on radiopurity as the fast double signature allows to discriminate most of the single event backgrounds. \begin{table} \begin{center} \begin{tabular}{ccc} \hline Isotope & Antineutrinos & Solar neutrinos\\ \hline $^{232}$Th & $\leq 10^{-15}$\,g/g & $\leq 10^{-17}$\,g/g \\ $^{238}$U & $\leq 10^{-15}$\,g/g & $\leq 10^{-17}$\,g/g \\ $^{40}$K & $\leq 10^{-16}$\,g/g & $\leq 10^{-18}$\,g/g \\ \hline \end{tabular} \caption[Requirements for LS radiopurity]{Requirements for LS radiopurity concerning uranium, thorium and potassium, listed both for antineutrino and for solar neutrino detection.} \label{tab:ls:radiopurity} \end{center} \end{table} Radiopurity levels are usually specified by the concentration of $^{232}$Th, $^{238}$U and $^{40}$K in the LS. The basic requirements for the JUNO LS are listed in Tab.~\ref{tab:ls:radiopurity}: The baseline scenario assumes a contamination on the level of $10^{-15}$ gram of U/Th and $10^{-16}$ gram of $^{40}$K per gram of LS, which will be sufficient for the detection of reactor antineutrinos. More stringent limits have to be met for the detection of solar neutrinos by elastic neutrino-electron scattering. Here, $10^{-17}$\,g/g resp.~$10^{-18}$\,g/g should be reached. These requirements are more than a factor 1000 resp.~10 above the current world-record held by the LS of the Borexino experiment. However, due to the scale of the project, achieving these radiopurity levels is a demanding task and will have to be studied and planned in considerate detail. \medskip\\ While members of the natural $^{232}$Th and $^{238}$U decay chains are the most common contaminants, also other sources of radioactive impurities for the LS have to be taken into account. Moreover, the contamination may arise from different sources that have to be avoided or at least controlled: \begin{itemize} \item {\bf Dust particles} containing elements of the natural U/Th decay chains as well as radioactive potassium $^{40}$K. \item {\bf Radon emanation}, especially the more long-lived $^{222}$Rn that is released from building materials (such as granite, brick sand, cement, gypsum, etc.) but also plastics, cables etc. \item The {\bf radioactive noble gases $^{85}$Kr and $^{39}$Ar} can be introduced as residual contamination of the nitrogen used to create an inert atmosphere in the liquid handling system or by exposure to the ambient air \item {\bf Surface depositions of $^{210}$Pb:} $^{222}$Rn will settle on surfaces exposed to air and decay to the long-lived isotope $^{210}$Pb ($\tau \sim 30$\,yrs), which in turn decays into $^{210}$Po and $^{210}$Bi. \item {\bf $^{210}$Po} from surface contaminations, distilled water or other unknown sources. \item {\bf Radioimpurities of the fluors}, mostly $^{40}$K. \item {\bf Radioactive carbon $^{14}$C } intrinsic to the hydrocarbons of the LS. \end{itemize} While the majority of these studies is currently carried out at Chinese institutes, the European collaborators and especially the fraction that is involved in the Borexino experiment has a considerable experience in radio-purification. Several lab-scale studies are foreseen for the immediate future. \subsection{Experimental Setups} The aimed-for radiopurity levels of the JUNO specifications are too low to be measured by standard laboratory-scale experiments. Gamma spectrometers can only establish upper limits on the level of contamination. However, the effectiveness of purification techniques can be tested by artificially loading LS samples with greater amounts of radioactive elements and measuring the activity before and after applying a purification step. Moreover, the close relations with the Daya Bay experiment will allow a test of LS radiopurity by filling a large sample of LS ($\sim$20 tons) into one of the subvolumes of a Daya Bay Antineutrino Detector (AD). The AD will both offer the necessary capacity to hold a significant quantity of LS and a suitable low-background environment to test radiopurity to the level of the specificationsf~\cite{Ford:2011zza,Mark:2008gc,Mike:2008gc}. \subsubsection{Low-background Gamma Spectrometer} A standard laboratory technique to measure level and type of radioactive impurities inside a LS sample are gamma spectrometers. The device currently in use has been employed to measure LS samples before purification. Corresponding to the minimum sensitivity of the setup, only an upper limit can be set on the sample activity, which is smaller than 0.1\,Bq/kg. The corresponding concentration of $^{238}$U should be smaller than $8.1\cdot10^{-9}$\,g/g, so 6$-$7 orders of magnitude above the JUNO specification levels. The activity corresponding to the specified level of $10^{-15}$\,g/g is on the order of $10^{-8}$\,Bq/kg and thus clearly beyond the sensitivity of this and comparable setups. \subsubsection{Assays with Loaded Scintillator Samples} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-19.png} \caption[Laboratory setup for radon-loading of LS samples]{Laboratory setup for radon-loading of LS samples} \label{fig:ls:rnload} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=6cm]{LiquidScintillator/figures/LS4-20.png} \caption[Natural thorium decay chain]{The natural decay chain of $^{232}$Th} \label{fig:ls:rnchain} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{LiquidScintillator/figures/LS4-21.png} \caption[Setup for measuring radio-purification efficiency]{Laboratory setup for measuring the efficiency of radio-purification} \label{fig:ls:counter} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=14cm]{LiquidScintillator/figures/LS4-22.png} \caption[Sample pulse of a fast $^{212}$Bi-$^{212}$Po coincidence]{Sampel pulse of a fast $^{212}$Bi-$^{212}$Po coincidence.} \label{fig:ls:bipo} \end{center} \end{figure} In order to investigate the effect of purification schemes on LS in laboratory experiments, the only solution is an artificial pollution of the samples with radioactivity, bringing the activity to a level accessible by small-scale counting experiments. This is achieved by loading the LS with suitable radioisotopes. The experience gained in the purification campaigns carried out in Borexino suggests that the radioisotope most difficult to remove from the LS is the long-lived $^{210}$Pb. Therefore, it is crucial that purification tests will be carried out using$^{212}$Pb as a tracer. While easier to tag in lab experiments, its chemical properties correspond to $^{210}$Pb. However, direct deposition of lead in LS is difficult as metallic powders will not dissolve in the organic liquid. Instead, the method commonly chosen for loading is the exposition of an LS sample to radon which will in turn decay into lead. The corresponding setup inside a glove box is shown in Fig.~\ref{fig:ls:rnload}: A solid $^{232}$Th source contained in a small box is used to produce the noble gas $^{220}$Rn ($T_{1/2}=56\,$sec) by emanation (cf.~Fig.~\ref{fig:ls:rnchain}). This box is constantly ventilated by an air pump, creating a stream of radon-enriched air that is lead by a hose to a further container holding the LS sample. As a noble gas, radon is easily dissolved into the LS by purging the sample. It rapidly decays to $^{216}$Po and then $^{212}$Pb, which is $\beta$-unstable with a half-life of $T_{1/2}=10.6\,$h and is thus well-adjusted for performing purification tests. Its daughters $^{212}$Bi and $^{212}$Po decay in a fast coincidence that provides a further experimental tag. The activity is measured by the experimental setup depicted in Fig.~\ref{fig:ls:counter}. In a light-tight box, a pair of 2'' PMTs is placed on both sides of an LS sample cell to detect the light produced by the $^{212}$Pb decays. Gamma rays from ambient radioactivity are attenuated by a shielding of low-activity lead bricks. Thus, the background rate for the $^{212}$Pb measurement is 0.5\,s$^{-1}$. If the fast coincidence of $^{212}$Bi and $^{212}$Po is used, the background rate is further reduced to only 2\,d$^{-1}$. A sample waveform of a Bi-Po event acquired in this setup is shown in Fig.~\ref{fig:ls:bipo}. \subsubsection{Purification Test with a Daya Bay AD} Once different purification schemes have been tested on the laboratory scale, the test in a medium-sized setup containing several tons of LS will be an important next step to assure that purification techniques are efficient down to the extremely low radiopurity levels of the JUNO specifications. The Antineutrino Detectors (ADs) of the Daya Bay Experiment are ideally suited to do such a study. The low-background conditions provided by the underground labs (the Daya Bay Near Experimental Halls are more than 100 meters underground) as well as the low-background material used in constructing the ADs themselves provide favorable conditions to measure the residual radioactivity of up to 20 tons of LS filled into one of the acrylic volumes of the AD in question. Moreover, an AD test setup allows the practicability of the purification scheme on the scale of a system able to handle $\sim$100 liters of LS per hour. For this purpose, a circular system will be set up. After initial filling with an LAB-based LS replacing the original LS in the AD, purification techniques will be probed in loop mode by monitoring the change in radiopurity levels. This setup will provide important information for the design of a system able to handle the 20,000 tons of LS for the filling of the JUNO detector. Beyond radiopurity, the filled AD will also provide information on the scintillation and optical properties of the LS. The large-scale setup allows to study the effect of purification on both the light yield and the light attenuation in the scintillator under realistic conditions. \subsection{Purification Methods} \label{sec:ls:radiopurity} For the moment, the methods investigated for their purification efficiency concerning radioactive contaminants comprise aluminum columns, distillation, water extraction, and nitrogen stripping (cf.~Sec.~\ref{sec:ls:opt_purification}). \subsubsection{Aluminum Columns} Purification in aluminum Al$_2$O$_3$ columns proves to be very efficient in removing organic impurities from LAB, resulting in increased transparency. Beyond this, laboratory studies also show a decrease in radioactive impurities by adsorption on Al$_2$O$_3$. The Al$_2$O$_3$ best suited for radio-purification was selected by gamma spectroscopy. Results of a comparative study are shown in Table~\ref{tab:ls:al2o3}. The Al$_2$O$_3$ featuring the lowest activity levels in $^{238}$U, $^{232}$Th, and $^{40}$K is produced by the Zibo Juteng chemical company ($\gamma$-type). \begin{table}[!htb] \begin{center} \begin{tabular}{cccccc} \hline Nuclide & energy & XP type & $\gamma$ type & (GY)Al$_2$O$_3$ & hamamasu low \\ & & Al$_2$O$_3$ & Al$_2$O$_3$ & specific activity & background glass \\ & keV & Bq/kg & Bq/kg & Bq/kg & Bq/kg \\ \hline $^{232}$Th & 583.1 & 3.94 & 0.07 & 0.52 & 2.02 \\ $^{238}$U & 609.4 &3.93 & 0.16 & 0.60 & 5.46 \\ $^{40}$K & 1460.8 & 5.49 & 0.54 & 4.30 & 24.14 \\ \hline \end{tabular} \caption[Comparison of Al$_2$O$_3$ radioactivity levels]{Comparison of Al$_2$O$_3$ radioactivity levels.} \label{tab:ls:al2o3} \end{center} \end{table} The radiopurity assay was based on a $^{220}$Rn-loaded LAB sample of 211\,ml volume. Measurements based on counting of $^{212}$Bi/Po coincidences were performed before and after column purification. A spin-type column was used, with a fill height of 20\,cm of Al$_2$O$_3$. Based on the $^{212}$Bi/Po count rate, a radio-purification efficiency of 99.4\,\% could be determined. For the future, an extension of these studies is foreseen at European institutes, especially INFN and JINR Dubna, which will also comprise adsorbants other than aluminum oxide. \subsubsection{Distillation of LAB} Distillation relies on the difference in boiling points and volatility of the LS components and possible impurities (see above). In particular, excellent efficiency is expected for the removal of radioactive metal ions because of the vast difference in boiling points compared to LS components. However, there are also large differences between the boiling points of the solvent LAB and the solutes PPO and Bis-MSB, which make the application of this technique very demanding once the three components are mixed for forming the LS. Thus, distillation is most suited for a purifications stage before mixing. The effectiveness of distillation regarding radio-purification has been tested in the laboratory setup depicted in Fig.~\ref{fig:ls:dist2}. Radioactivity levels of a $^{220}$Rn-loaded LAB sample have been determined before and after distillation. Based on $^{212}$Bi/Po coincidence counting the efficiency has been determined to 98.4\,\%. Further tests will be performed after adding a reflux unit that should further improve the purification efficiency. \begin{figure}[htb] \begin{center} \includegraphics[width=14cm]{LiquidScintillator/figures/LS4-25.png} \caption[Setup for distillation tests]{Laboratory setup for distillation tests} \label{fig:ls:dist2} \end{center} \end{figure} In the current setup, we find that after distillation with high temperatures or long heating times the LS samples feature a reduced attenuation length, change color and release a strong smell. A possible reason is the chemical break-up of organic impurities by the heating process. Possible solutions are an optimization of the distillation process or a reduction of heat-sensitive impurities in the LAB raw materials. It should be noted that similar laboratory-scale distillation tests at INFN Perugia did not show a degradation of the LS. In the final large-scale setup for JUNO, vacuum distillation will be used for a reduction of the necessary heating temperature and the related energy consumption. To reduce the energy consumption on the experimental site, the distillation plant could be realized directly at the LAB factory. The preliminary design of the installation comprises multistage plates and a vacuum distillation tower for LAB with a processing capacity of up to 12,000 liters per hour, operated at a temperature of $220\,^{\circ}$C. These preliminary working parameters will be refined by further studies carried out at INFN institutes. \subsubsection{Water Extraction} Purification by water extraction relies on the polarity of water molecules to separate polarized impurities, e.g. free-state ions of radioactive metals, from the non-polar LAB and fluor molecules. Mixing of water and LS is performed in a purification tower that contains a counter-current water extraction column. Pure water is filled at the top while LS enters at the bottom of extraction column. The mixing of the two liquids is supported by a porous filling. As soon as a state of full mixing has been reached, the LS leaves at the top of the column while the waste water is collected by the drainage system at the bottom. The water extraction method is highly efficient for removing metal ions: Radium is removed at the level of 96.5\%. For lead and polonium, the purification efficiency of 82$-$87\,\% is somewhat lower as both can form chemical bonds with the organic molecules of the solvent that are fairly stable at room temperature. \subsubsection{Nitrogen Stripping} Purging the LS by nitrogen aims both at the removal of dissolved radioactive noble gases (argon, krypton, xenon, and radon) in the liquid and the removal of oxygen, which acts as a quencher lowering the light yield by oxidation of the fluor molecules. Beyond active purification by purging, ultrapure nitrogen will be used for establishing a nitrogen atmosphere or blanket inside the liquid handling system and in the final detector to avoid gas contamination of the LS. For laboratory purposes, industrial-grade high-purity nitrogen is mostly sufficient. However, the high solubility of noble gases in the LS makes it necessary to add a further purification step for the nitrogen before bringing it in contact with the liquid. For this, high-purity nitrogen of grade 6.0 (99.9999\,\% pure nitrogen) is passing through a cold trap at liquid argon temperatures. The resulting ultrapure nitrogen can reach radioactive background levels of 10$^{-17}$\,g/g, corresponding to 0.36\,ppm of $^{39}$Ar and 0.16\,ppt of $^{85}$Kr. Figure~\ref{fig:ls:coldtrap} shows a schematic drawing of a laboratory-scale setup for the production of ultrapure nitrogen based on an active carbon absorber submerged in a bath of liquid argon. The resulting radiopurity will be tested by GC-MS measurements (Agilent 7890A/5975, cf.~Sec.~\ref{sec:ls:gcms}). \begin{figure}[htb] \begin{center} \includegraphics[width=11cm]{LiquidScintillator/figures/LS4-26.png} \caption[Schematic drawing of ultrapure N2 production unit]{Schematic setup of the production unit for ultrapure nitrogen.} \label{fig:ls:coldtrap} \end{center} \end{figure} \section{Mass Production of 20 kton LS} {\bf Production. } The production of 20 kilotons of LS will be achieved in three or four cycles over a period of 1$-$2 years. Based on the current status of laboratory studies (see above), the solvent LAB will most likely be provided by the company from Nanjing. Criteria for quality control will be developed jointly by IHEP and the company. Similar arrangements will be made for the solutes PPO and Bis-MSB. \medskip\\ {\bf Transport.} A professional logistics company will take care of the transport of the LAB from the production line at Nanjing to the JUNO experimental site. Suitable transport containers will be manufactured by a company specifically for this purpose, meeting specifications set by IHEP concerning radiopurity, air tightness and material compatibility with LAB. \begin{figure}[htb] \begin{center} \includegraphics[width=15cm]{LiquidScintillator/figures/LS4-27.png} \caption[Schematic drawing of LS purification plants]{Schematic drawing of the system of LS purification plants} \label{fig:ls:filling} \end{center} \end{figure} \bigskip \noindent The on-site LS handling system is depicted in Fig.~\ref{fig:ls:filling}. The facilities include: \medskip\\ {\bf Storage tanks.} Since the LAB will arrive in 3 to 4 batches of 5\,kt each, the foreseen storage area comprises about 5,000 square meters, featuring several tanks with a volume of 3 ktons and another set of smaller tanks on the scale of several hundred tons. This includes not only tanks for LAB, but also for white oil and temporal containers needed during unloading. In addition, an area of 30\,m$^2$ is to be reserved for storage of the fluors. \medskip\\ {\bf Surface purification plants.} Fig.~\ref{fig:ls:filling} depicts the basic scheme for the on-site system of purification plants: The current design mainly relies on an initial purification of the individual LS components (solvents and solutes) before mixing them to obtain the final LS. The major task of purifying LAB is to be performed on surface. The foreseen purification steps include distillation, adsorption in an aluminum oxide column, water extraction and nitrogen purging (cf.~sec.~\ref{sec:ls:radiopurity}). Assuming a total mass of 20\,kt of LS and a period of 200 days reserved for continuous purification and detector filling, the required processing capacity of the purification plants is 100 tons per day. Based on the power consumption of similar (but smaller-scale) purification plants realized for the Borexino, KamLAND and SNO+ experiments, the corresponding power consumption is about 2\,MW. In addition, a supply of ultrapure nitrogen on a rate of 200\,m$^3$/h as well as ultrapure water of 6 tons per hour has to be assured. The system will require a surface clean room of 200\,m$^2$ area and 10\,m height for its installation. In addition, rooms for the production plants of ultrapure water and nitrogen are needed (30\,m$^2$ each). \medskip\\ {\bf LS mixing plants.} After purification of the individual components, LS mixing will be performed. A complication arises from the low solubility of Bis-MSB on LAB. To reduce the processing time and assure a constant amount of the fluor in the LS, batches of highly concentrated LAB-Bis-MSB solution (150\,mg/l) will be prepared beforehand and then watered down with pure LAB during mixing. The mixing tanks will be made from organic glass to ensure both the material compatibility of the LS and to allow for a visual inspection of the mixing product. The mixing system will be housed in a room of 300\,m$^2$. \medskip\\ {\bf Quality assurance and control.} Constant control of the LS leaving the purification and mixing plants is of uttermost importance for the success of JUNO. As the LS is produced in cycles, even a single low-quality batch passing the controls unnoticed and filled into the detector could potentially spoil the high-quality of all LS produced up to this point. Therefore, a careful program of QA/QC measures before and after purification and mixing has to be developed. Separate teams of operators and inspectors should be formed and be supervised by an on-site manager, following a strict protocol. Two inspection rooms of a total area of 60\,m$^2$ will house systems for QC, including GC-MS for detection of organic impurities and small-scale experiments to monitor attenuation length and spectrum as well as light out of the LS. \medskip\\ {\bf Underground installations.} Approved batches of LS will be sent by a pipeline to an underground storage area where it remains until detector filling. The pipe system should allow the flow of LS in both directions and feature high-pressure resistance to allow for a safe transport of the LS. Two storage tanks of 200 tons for LS and white oil are foreseen close to the detector cavern. Beyond the surface plants for the initial purification of the LS, a secondary system is foreseen underground to perform loop-mode purification of the LS while the experiment is running. This online purification system will rely on water extraction and nitrogen stripping. The system will require two underground halls of 500\,m$^2$ area and 10\,m in height. The necessary supply of water, electricity and gas will be roughly equivalent to the requirements of the surface system. \medskip\\ {\bf Cleanliness conditions.} Ensuring the cleanliness and purity of all materials in contact with the LS and its raw materials will be decisive for meeting the radiopurity requirements of JUNO. All pipes, pumps storage and processing vessels through which the LS passes from arrival on-site until filling of the detector will have to meet the requirements for chemical compatibility, cleanliness of all surfaces and air-tightness. All surfaces have to be thoroughly cleaned and probably rinsed with LAB before being brought into contact with the LS. Removal of dust on surfaces is mandatory: At an average activity of 200 Bq/kg in U and Th, not more than 10\,g of dust are allowed to be dissolved in the total mass of 20\,kt of LS. What is more, piping and tank materials will be containing U and Th themselves and thus will potentially emanate radon that will diffuse into the liquid. This process can be mitigated by applying suitable surface liners stopping the diffusion. The expected radon level in the underground air is about 100\,Bq/m$^3$. Contact of LS and process water to air has to be strictly avoided as the solubility of radon in LS is far greater than in water. To reach this goal, all the liquid handling system must be absolutely air-tight, and an atmosphere of ultrapure nitrogen will be established inside the system. \medskip\\ All surface and underground facilities will be equipped with cameras, fire protection, monitors for oxygen and radon as well as alarm systems for poisonous gases. \section{Risk Assessment} \begin{table} \begin{center} \begin{tabular}{lcc} \hline Properties & LAB & Mineral Oil\\ \hline Quantity & 20\,kt & 7\,kt \\ \hline Appearance & odorless, & odorless,\\ & colorless, & colorless,\\ & transparent & transparent \\ Flammable & yes & yes \\ Toxic & no & no \\ Corrosive & no & no \\ \hline Chemical formula & C$_6$H$_5$C$_n$H$_{2n+1}$ & C$_n$H$_{2n+2}$ \\ & ($n=10-13$) & \\ Density [g/cm$^3$] & 0.855$-$0.870 & 0.815$-$0.840\\ Freezing point & -50\,$^{\circ}$C & \\ Flash point & 135\,$^{\circ}$C & 145\,$^{\circ}$C\\ Vapor pressure (20$^\circ$C) & 0.1\,mmHg & 0.1\,mmHg \\ \hline \end{tabular} \caption[Table of material properties]{Table of material properties} \label{tab:ls:matprop} \end{center} \end{table} The JUNO detector will contain both LS and and mineral oil. Material properties are listed in Table~\ref{tab:ls:matprop}. \medskip\\ {\bf LAB. } The main component of LS is linear alkyl benzene (LAB), of which 20 ktons are required. In industry, it is mainly used in the production of detergents. For this, LAB is sulfonated to form linear alkylbenzene sulfonate (LAS), which in turn serves as raw material for washing powders, liquid detergents, pesticide emulsifiers, lubricants and dispersants. It is non-poisonous and odorless. For use in JUNO, the only safety concern is its flammability. \medskip\\ {\bf Mineral oil.} The buffer volume surrounding the acrylic sphere might be filled with 7,000 tons of mineral oil. The foreseen type is food grade mineral oil 10\#. It is a liquid by-product of the distillation of crude oil which is performed during the production of gasoline or other petroleum-based products. Again, the only safety concern is flammability. Mineral oil is widely used in biomedicine, cell cultures, veterinary uses, cosmetics, mechanical, electrical and industrial engineering, food preparation etc. \medskip\\ Due to the flammability of both liquids, the main risk to be taken into account is fire safety. The necessary precautions have to be taken into account during transport, on-site storage, purification and mixing of the LS components as well as during detector operation. From an experimental point of view, the intactness of the LS poses additional constraints on all operations. Cleanliness and tightness of the liquid handling system have to be considered carefully. Further risks will be induced by the installation of the LS purification plants and storage tanks in an underground cavern at a depth of 700\,m. This will result in complicated hoisting equipment, high-altitude operations, and a need for accurate positioning. Purification underground will pose additional risks because a possible spillage of nitrogen are more serious under these confined conditions. All operations of this kind will have to be carefully planned beforehand. \section{Schedule} \begin{itemize} \item[2013] Realization of laboratory setups for LS properties. First explorative measurements of LS properties. Measurement of the depolarization ratio to determine the Rayleigh scattering length. \item[2014] Realization of laboratory-scale setups for testing different purification method:\\ Low-pressure distillation, aluminum-oxide column, nitrogen stripping and water extraction. First steps towards the optimization of the purification scheme: Tests of optical purification: Measurements of light out, attenuation length and spectrum, and water content of the purified LAB. Tests of radio-purification: Establish the technology for injecting $^{212}$Pb and $^{224}$Ra in LS samples and setup of the corresponding measurement device to study the effect of purification on radioactive contamination levels. Start of joint research programs with companies providing LAB. \item[2015] Design and realization of a purification system. Final determination of purification parameters. Start of measurements for electron energy non-linearity. Design and setup of the quality control devices. Start of joint research activities with providers of PPO and Bis-MSB.\\ Conception and setup of devices for quality testing. \item[2016] Jan$-$Jun: Setup and testing of purification plants at Daya Bay. Jul$-$Dec: Installation and debugging the purification system. \item[2017] Determination of the final recipe of LS production based on the experimental data and the purification methods. Final measurement of light out and attenuation length of the LS. Signing of contracts with the companies providing the purification plants and start of the design. Signing of contracts with the providers for PPO and Bis-MSB. Start of the batch production at the end of the year. Finalization of the design of storage tanks for LAB. Design and build the LS purificaiton hall on the ground. \item[2018] Signing of all contracts for LAB storage area. Setup of the large storage tanks before the middle of the year:\\ 3$-$5 of 3000$-$1000\,m$^3$ tanks for LAB, LS and white oil Installation of the purification system until end of the year. Setup of two 200-ton underground buffer tank. Setup and testing of the LS QA/QC systems. \item[2019] Testing of all purification plants. Setup and testing of ultrapure water and nitrogen plants. Design and build the LS mixture system. Production of 8 ktons of LS of sufficient quality for the central detector. \end{itemize} \newpage \chapter{Offline Software And Computing} \label{ch:OfflineSoftwareAndComputing} \section{Design Goals and Requirements} \input{OfflineSoftwareAndComputing/requirements.tex} \input{OfflineSoftwareAndComputing/system_introduction.tex} \section{Offline Software Design} \input{OfflineSoftwareAndComputing/platform.tex} \input{OfflineSoftwareAndComputing/event_model.tex} \input{OfflineSoftwareAndComputing/geometry.tex} \input{OfflineSoftwareAndComputing/generator.tex} \input{OfflineSoftwareAndComputing/detector_sim.tex} \input{OfflineSoftwareAndComputing/digitization_evtMixing.tex} \input{OfflineSoftwareAndComputing/detector_rec.tex} \input{OfflineSoftwareAndComputing/event_display.tex} \input{OfflineSoftwareAndComputing/database_service.tex} \section{Computing System Design} \input{OfflineSoftwareAndComputing/computing_network.tex} \input{OfflineSoftwareAndComputing/data_trans.tex} \input{OfflineSoftwareAndComputing/com_storage.tex} \input{OfflineSoftwareAndComputing/management_and_service.tex} \input{OfflineSoftwareAndComputing/dev_flow_quality_control.tex} \subsection{Computing Network and Public Network Environment} A 1 Gbps dedicated link, named IHEP-JUNO-LINK, is planned between the JUNO experiment site and IHEP. It is proposed that this link be used for data transfer between the experiment site and the IHEP data center as well as for connecting the onsite office network and the external network. The 1 Gbps bandwidth is sufficient to fully meet the experiment's requirements for stable data transfer, based on an estimated data volume of 2 PB per year, which should take an estimated average bandwidth of 544 Mbps. The remaining bandwidth can be used for network crash recovery, experiment remote control and office network connectivity for JUNO onsite researchers. \begin{figure}[htb] \begin{center} \includegraphics[width=.8\textwidth]{OfflineSoftwareAndComputing/figures/network.png} \caption[WAN topology]{\label{fig:storage}WAN Topology } \end{center} \end{figure} The JUNO network link is planned to be provided by the Chinese Science and Technology Network (CSTNet) \cite{CSTNet}. The data will first be transferred from the experiment site to IHEP through IHEP-JUNO-LINK, then relayed to collaborating sites through CSTNet. At present, IHEP is connected to the CSTNet core network through two 10 Gbps links, one of which supports IPv4 and the other IPv6. The bandwidth from IHEP to the USA is 10 Gbps and from IHEP to Europe is 5 Gbps, both of which are through CSTNet and have good network performance. \subsubsection{Network Information System Security} Network security is becoming more and more important in all kinds of information systems. The JUNO application system and even the computing environment also face security risks, so we are planning to apply risk control and safety protection in the following aspects: physical layer, network layer, system layer, transport layer, virus threats layer and management layer. The safety of the physical layer is the foundation of the whole information system security. In general, the security risks to the physical layer mainly include: system boot failure and database information loss due to power failure, whole systematic destruction caused by natural disasters, probable information loss caused by electromagnetic radiation and so on. In order to ensure the robustness of the physical layer, a powerful UPS should be deployed for the IT system and devices; proper access control policies should also be implemented for some special areas. The risks to the network layer consist of risks to the network boundaries and risks to safety risk of the data transfer. A powerful network firewall system will be deployed in the JUNO network to avoid risks to the network boundaries. To ensure the safety of data transfers, there are two things we must consider: data integrity and confidentiality. To ensure data integrity, we can add checksums and provide multi-layer data buffering during transfers.To ensure data confidentiality, the information system should use encrypted transmission. The system layer risk usually refers to the safety of operation systems and applications. In JUNO, a rigorous hierarchical control policy and auditing operation logs will be used to reduce and prevent security threats to the system layer. For virus threats, while traditional viruses mainly spread through a variety of storage media, modern viruses mainly spread through the network. When a computer in the LAN is infected, the virus will quickly spread through the network to hundreds or thousands of machines in the same network. In JUNO, rigorous access control policy and real-time detection for the network traffic will be implemented for all the network nodes and IT systems to provide a safe network environment. The risk to the management layer is usually caused by unclear responsibilities and illegal operation of the IT system, so the most feasible solution is to combine management policies and technical solutions together to achieve the integration of technology and policy. \subsection{Computation and Storage} \subsubsection{Local Cluster} About 10,000 CPU cores should be used for raw data simulation, reconstruction and data analysis to meet JUNO offline computing requirement. All the resources will be managed by the batch system of the IHEP Computing Center. Several job queues with different priorities will be established to share the resources in the most efficient way. An optimized scheduling algorithm will be developed based on the features of the the JUNO computing environment and hardware performance. The scheduler should dispatch jobs to the most suitable computing resource so that high overall computing efficiency is obtained. Job types can be defined based on the computing model, characteristics and I/O bandwidth consumption for different jobs. For this, it is necessary to study and test different types of jobs running on different machines with different CPUs. It is also necessary to analyze the effect of the distributed storage system on job running time. New scheduling optimization algorithms based on the results of this study should be implemented as a plug-in which can be integrated with the batch job system. This will be used for resource management and job scheduling to improve the efficiency of single jobs and the overall resource usage. \subsubsection{Computing Virtualization} Local clusters are managed by job management systems, with Torque \cite{Torque}, Condor \cite{HTCondor} and LSF \cite{LSF} being popular choices in HEP systems. These are responsible for scheduling jobs to a physical machine. With ever more powerful CPUs available, virtual machine clusters - cloud computing - can be used to relieve the pressure on computing resources at peak times. Cloud computing also reduces the burden of system management in applying upgrades to operating system and offline software, as well as satisfying requests for different software versions. Running jobs on virtual machines (VMs) can also avoid the trouble caused by hardware heterogeneity and give more flexibility in software configuration, without requiring changes in the job itself. Virtual clusters can therefore improve the availability and reliability of the computing resources as well as providing elastic resource expansion and flexible resource allocation. Rather than allocating a job to a node, as for physical cluster resource allocation, jobs are allocated to a CPU core. One CPU core supports one VM running a JUNO job. The lifetime of the VM depends on the job running time. This makes it possible to pre-empt jobs based on priority, and to migrate jobs. The VM can be migrated online and the job running on the VM can be hung up before the migration and released after the migration. These features of VMs will make the computing platform more robust. \subsubsection{Distributed Storage Management and Data Sharing} JUNO data processing is a typical massive data computation task. Distributed storage management is necessary for its large scale of data. The following figure shows the architecture of the storage. \begin{figure}[htb] \begin{center} \includegraphics[width=.8\textwidth]{OfflineSoftwareAndComputing/figures/com_storate_f1.png} \caption[Storage architecture]{\label{fig:storage}The storage architecture } \end{center} \end{figure} The logical device layer, which is over the physical layer, provides stable and reliable storage for the upper layers. The logical storage devices are independent of the physical devices and isolated from them. They can be elastically extended or narrowed, copied and be migrated among physical devices. The features of logical devices, including high availability, high performance and high security, make it easy to satisfy the storage demands of JUNO. The logical device layer gives the storage system strong scalability with PB level capacity. The storage management layer provides the functions of resource allocation, scheduling and load balancing. It provides a global unified name space, called the virtual storage layer. This can shield the heterogeneity of physical devices and integrate a unified virtual storage resource for users. The distribution process layer is over the storage management, providing fault tolerance, consistency, distributed task scheduling support (such as map-reduce \cite{mapreduce}) and metadata management. The application interface layer provides rich interfaces compatible with Posix semantics, and is transparent to users. This virtual distributed storage system should fully satisfy JUNO's requirement of 10 PB storage, providing a high-reliability, high-performance storage service which can be extended elastically as needed. . \subsubsection{Distributed computation} Since JUNO has a large computation scale, it needs cooperative work among the group members. Distributed computation integrates heterogeneous resources from different sites, which could be shared by all the sites located in different region. A unified user interface would be provided to receive user jobs and job is dispatched to the suitable sites by the system. Job ``pushing'' and ``pulling'' are the two main scheduling methods. \subsection{Database Service} Databases play more and more important roles in offline data processing. They are used to store many kinds of information, such as, sizes and positions of complicated detector geometry, optics information, running parameters, calibration constants, schema evolution, bookkeeping and so on. At the same time databases provide access to all kinds of information through their management services, such as creation, query, modification and deletion, which will be frequently used during offline data processing. Many types of database have been used in particle physics experiments, of which the most popular are Oracle \cite{Oracle}, Mysql \cite{MySQL}, PostgreSQL \cite{postgreSQL} and SQLite \cite{SQLite}. Many of them provide user-friendly APIs interfaces for different programming languages. Recently, NoSQL databases have become more and more popular and powerful in some other fields, so we are also investigating the possibility of using some NoSQL database. Regardless of which kind of underlying database will be chosen,the general schema of database services will be implemented in three levels. The lowest level is the APIs of the underlying database; the middle level is the new extension of the APIs according to the requirements of the JUNO experiment; the upper level is to provide user-friendly services for different applications such as detector simulation, event reconstruction and physics analysis. Since Daya Bay already has one good database interface, we will implement it, optimizing and extending the current interface to provide more flexible and more powerful access to the different types of information stored by the databases. We will also set up two types of database servers, master servers,and the slave servers. The former are used to store and manage all information and can only be accessed by the slave servers, while the latter are responsible for retrieving information from the master servers and providing access for all applications. \subsection{Data Process Requirement} Data taking period of JUNO will start from 2019 and raw data it generates each year would be about 2PB. Raw data would be transferred to computing center of IHEP via dedicate network connection. A large scale offline computing platform is necessary for the computation of simulation and reconstruction. Besides, it will support analysis computation. The platform will also provide enough storage space for the raw data, Monte Carlo data and data archive. With the rough estimation, the scale of the platform would be 10,000 cpus/cores, 10PB disk storage and 30PB archive. 40Gbps bone switching network is the connection among computing nodes and storage servers. To promote the data process speed, the platform would integrate computing resources of collaboration group members via distributed computation. Data would be shared and computation tasks would be dispatched to the sites included in the platform cooperatively. So JUNO puts forward high demands to data analysis process, data storage, data transfer and data share as the following four aspects. \subsection{Data transferring requirement} Data transfer and exchange system with features of high performance, high security and high availability is necessary. To guarantee the experiment data transferring from experiment site to IHEP data center in time, the system should monitor and trace the online status of data transferring and exchange and it should provide data transfer visualization interface. This could also guarantee the data dispatch and share in the collaborative group member sites smoothly. \subsection{Computing Environment} To guarantee JUNO data process smoothly, a 10,000 cpus/cores local cluster would be established. The collaboration group member would also provide computing resources according to the different cases. The platform includes not only local cluster, but remote resources from collaboration group member, who is responsible for the JUNO computation task either. \subsection{Data Storage} JUNO will generate huge experiment data. Data from experiment site should be transferred and stored at computing center of IHEP. The capacity of storage in IHEP would be 10 PB. It could support 10,000 cpus/cores concurrence access. Data migration between tape library and disk is transparent to users. \subsection{Maintenance Monitoring and Service Support} To guarantee the platform stable, an intelligent monitoring system would be developed and an information system aimed at infrastructure management should be established. The status of both platform and network connection should be monitored in a very fine grain control. The monitor tool could give the status assessment of the whole platform and give a tendency prediction and pre-warning for the relative services of the platform. A set of information management tools should be developed including conference support, document management and online-shift management etc. \subsection{Data transfer and sharing} Data transfer and sharing is the foundation of physical analysis in JUNO. In steady operations mode, raw data will be transferred over the network from the experiment site to the Computing Center for long term data storage, then distributed to the collaboration group members according to their requirements. The raw data from the JUNO experiment is open to all collaboration group members for local or remote processing and analysis. The raw data from JUNO will be acquired by the online data acquirement system (DAQ) which will be deployed onsite, and then stored in a local disk cache with sufficient storage to keep one month's raw data. After that, the raw data will be transferred to the IHEP Computing Center in Beijing. During the transfers, the checksum of the data will also be transferred to make sure of the data integrity. If the data integrity check is failed, the raw data should be retransferred. After the data transfer is done and the data integrity check is ok, the status of the data in the DAQ local disk cache will be marked as TRANSFERRED. Moreover, a high/low water line deletion algorithm will be used to clear the outdated data in DAQ local disk cache. Most of the data will be transferred automatically by the data transfer system, so some possible human errors can be reduced. To ensure the stability and robustness of the data transferring system, a monitoring system will be developed and deployed which will provide the data transfer and sharing status in real-time and track the efficiency of the data transfer system. Combined with the status of the IT infrastructure (including network bandwidth), the data transfer system will optimize the transfer path and recover transfer failures automatically to improve the performance and stability of the system. \subsection{Event Reconstruction} \subsubsection{PMT waveform reconstruction} The total number of photoelectrons and the first hit time of each PMT can be reconstructed based on the waveform sampling from the readout electronics with two algorithms, charge integration or waveform fitting. A fast reconstruction algorithm has been developed using the charge integration with appropriate baseline subtraction. The first hit time is determined as the time above threshold with time walk correction. A waveform fitting algorithm based on a PMT response model is under development, which is expected to give better timing and can reduce the distortion of the charge measurement caused by the PMT and electronics effects such as after-pulse and overshoot. The fast reconstruction is designed for event selection, while the waveform fitting is expected to be used for small subset samples such as IBD candidates or calibration data. \subsubsection{Vertex reconstruction in the Central Detector} Since the energy of a reactor antineutrino is usually below 10 MeV, it can be approximated as a point source in the JUNO detector. Then the event vertex can be reconstructed based on the relative arrival time from the point source to different PMTs, by minimizing a likelihood function: \begin{equation} F(x_0,y_0,z_0,T_0)=-\sum_i \log(f_t(t_i-T_0-\frac{L(x_0,y_0,z_0,x_i,y_i,z_i}{c_{eff}})), \end{equation} where $x_0,y_0,z_0,T_0$ are the event vertex position components and start time to be fitted, $x_i,y_i,z_i,t_i$ are the position components and hit time of the $i_{th}$ PMT, $c_{eff}$ is the speed of light in the detector, and $f_t$ is the probability density function of the PMT hit time with the time-of-flight being subtracted, which contains the decay time of scintillation light and the time resolution of the PMT. If the multiple photoelectrons in the same PMT can be clearly separated, $f_t$ can be expressed analytically, otherwise it relies on MC simulation. A charge center method is used to calculate the event vertex to the first order, which can be used as the initial value of the likelihood fitting. \subsubsection{Visible energy reconstruction} The visible energy in the central detector of a point source is reconstructed using the maximum likelihood fitting: \begin{equation} \mathcal{L}(E_{vis})=\prod_{no-hit}{P_{no-hit}(E_{vis},i)}\times\prod_{hit}{P_{hit}(E_{vis},i,q_i)}, \end{equation} where $E_{vis}$ is the visible energy to be fitted, $P_{no-hit}(E_{vis},i)$ and $P_{hit}(E_{vis},i,q_i)$ are the probability that the $i_{th}$ PMT has no hit and has the total charge of $q_i$, respectively. This probability is calculated based on the knowledge of the detector response model, including the light yield of the liquid scintillator, the attenuation length of the liquid scintillator and the buffer, the angular response function of the PMT, and the PMT charge resolution, as well as the reconstructed vertex as input. Another fast reconstruction algorithm has also been developed, which is based on the total number of photoelectrons collected from all PMTs with corrections using the calibration sources. \subsubsection{Muon tracking} The muon tracking in the central detector is based on the PMT hit time. Since the energy deposit from cosmic muons is at the level of GeV, the average number of photoelectrons at each PMT will be greater than 100, therefore the first hit time of the PMTs is dominated by the fast component of the scintillation light, which approximately follows a Gaussian distribution. A $\chi^2$ function is defined: \begin{equation} \chi^2=\sum_i(\frac{t_i^{exp}-t_i^{mea}}{\sigma})^2, \end{equation} where $t_i^{exp}$ and $t_i^{mea}$ are the expected and measured first hit time of the $i_{th}$ PMT, respectively, $\sigma$ is the time resolution of the PMT, and $t_i^{exp}$ is calculated based on the tracking parameters and the optical model in the detector. As shown in Fig.~\ref{fig:rec:muon_tracking}, since the scintillation light is isotropic, the intersection angle $\theta$ can be calculated analytically. For muons passing through the detector, the PMTs around the injection point and the outgoing point see more light and form two clusters, which can be used to estimate the initial tracking parameters. A muon tracking algorithm combined with the veto detectors is still under development. \begin{figure}[htb!] \centering \includegraphics[width=.5\textwidth]{OfflineSoftwareAndComputing/figures/muon_tracking.pdf} \vspace{-0.15cm} \caption{An illustration of the muon tracking. The red line is the muon track to be fitted, $R_c$ is the light source which provides the first hit to the PMT at $R_i$, determined by the intersection angle $\theta$. If $R_c$ is calculated to be outside the detector, it will be set as the injection point $R_0$.} \label{fig:rec:muon_tracking} \end{figure} \subsection{Detector Simulation} The detector simulation package for JUNO is based on Geant4. Detector simulation includes detector geometry description, physics processes, hit recording, and user interfaces. \subsubsection{Description of detector geometry} An accurate description of detector geometry is the foundation of detector simulation. In JUNO simulation, the geometry of the liquid scintillator, acrylic/nylon ball, PMTs, stainless steel tank or trusses need to be described in simulation software. About 17,000 PMTs are placed on a spherical surface. Geant4 is very sensitive to the geometry problem of volume overlaps or extrusions, which in general reduce the simulation speed or waste time in infinite loops. Some built-in or external tools can help to find such problems. At the stage of detector design, the detector geometry keeps changing. It is preferable that the geometry and event data can be exported in the same data file after simulation, so that the reconstruction job can read both information at the same time to ensure geometry consistency between simulation and reconstruction. In addition, the detector description method should allow easy modifications to the geometry with changes made to the detector design. In simulation of the liquid scintillator detector, the simulation of optical photon processes is a key issue. The optical parameters for materials such as liquid scintillator, acrylic, mineral oil and PMT should be carefully defined. The optical parameters include refractive index, absorption length, light yield factor, quantum efficiency and so on. Each optical parameter varies with the wave-length of optical photons and may be time-dependent in the future. Due to the production technique, the parameters of each individual PMT may be different. How to effectively manage these optical parameters is an important issue in the detector data description. A possible method is to combine Geant4 with GDML/XML and database technology to manage the detector geometry and optical parameters. In addition, simulation of the optical processes requires the definition of an optical surface, on which optical photons can have different behavior like refraction, reflection and absorption. \subsubsection{Physics processes} The physics processes in the JUNO detector simulation include standard electromagnetic, ion, hadronic and optical photon processes and optical model for the PMTs. Optical photon processes include scintillation, Cerenkov effect, absorption and Rayleigh scattering. When a charged particle passes through liquid scintillator, ionization energy loss is converted to scintillation light. The non-linearity between scintillation light and the ionization energy loss is called the quenching effect. During the propagation of light in the liquid scintillator, the re-emission effect needs to be considered, which is not included in standard Geant4 processes and needs to be implemented in our experiment. For such a large liquid scintillator detector as JUNO, the simulation of high energy muons in liquid scintillator can be too CPU-time consuming to be affordable. Reducing light yield or using a parameterized model for fast simulation are possible solutions to speed up the muon simulation. \subsubsection{Hit recording} The hit information, including the number of hits in each PMT and the hit time, is the output at the stage of Geant4 tracking. This information will be used as input to electronics simulation in the next step. In addition, particle history information should be saved, including the information of the initial primary particles, neutron capture time and position, energy deposition and position, direction and position of the hit on the PMT, and the relationship between track and hits. \subsubsection{User interface} A friendly user interface is necessary to use the simulation package, which should provide easy configuration of simulation parameters such as radius of LS volume, buffer thickness, vertex position, particle type and flags to switch on/off a specific physics process. \section{Software Development and Quality Control} The JUNO offline software system is composed of several sub-systems such as the framework, event generator, detector simulation, calibration, reconstruction, physics analysis and so on. Each sub-system contains one or more packages,which are divided according to their functionalities. In principle, they are relatively independent in order to allow easy dynamic loading at the run time. The relationship between the necessary dependencies are implemented using the Configuration Management Tool (CMT) which is also used to compile each package. Developers from different institutes and universities are collaborating together to write code in parallel. Subversion (SVN) \cite{svn} is deployed as the code repository and version control system to track the development of the offline software and tag it for releases. One central SVN server has been set up. All users and developers can easily check out the required release version and commit modifications according to their privileges. Trac \cite{Trac} is an enhanced wiki and issue tracking system for software development projects. Trac and SVN can be used successfully working together to provide cross searching and references. Some specific tools, testing algorithms/packages will be created with the development of the offline software in order to test and monitor its functionality and make sure that the whole software is going in the correct direction. Every time a new version is released, a complete test will be performed and a report given showing the results of some characteristic quantities, including CPU consumption and memory usage. Several kinds of documentation, such as Wiki \cite{Wiki} pages and DocDB \cite{DocDB} will be provided.The Wiki pages will generally record the objectives and progress of each subsystem, while the DocDB will store technical documentation and status reports. \subsection{Digitizer Simulation} The digitizer simulation consists of electronics simulation, trigger simulation and readout simulation. The main goal is to simulate the real response of the three systems and apply their effects to physics results. The readout simulation transforms Monte Carlo data into the same format as experimental data and simplifies the analysis procedure. The digitizer simulation will be implemented in the JUNO offline framework, SNiPER. Electronics simulation plays an important role in the digitizer simulation. The waveform from each PMT will be recorded using FADC and analyzed offline to get the final amplitude and timing information for each PMT. In electronics simulation, to transform the input hit information (given by Geant4) into electronic signals which are as real as possible, we need to build a model to exactly describe the single photo-electron response, which includes the effects of signal amplitude, rising time, pre-pulse, after-pulse, dark rate, etc. Electronics simulation should also have the ability to handle background noise, overlapping waveforms from multiple hits on the same PMT, and the effects from waveform distortion and non-linear effects. For very large signals like cosmic muons, however, it is difficult to save all FADC information so only integrated charge will be read out. At the current stage, since the detailed JUNO electronics design has not been finally determined, the electronics simulation imported from the Daya Bay experiment is implemented as a temporary substitute for the JUNO electronics simulation. It will be updated to the real case when the JUNO electronics design is finalized. Trigger simulation takes the output from the electronics simulation as its input to simulate the real trigger logic and clock system, and decides whether or not to send a trigger signal if the current event passes a trigger. Currently, we have implemented a simple nHit trigger simulation. Based on the trigger signals provided by the trigger simulation, FADC information in the readout window will be saved for each PMT and form an event. At the same time, the time stamp will be tagged. The final data format will be the same as for real data. \subsection{Background mixing} In real data, most events come from background, like natural radioactive events and cosmic muon induced events. In order to simulate the real situation, a background mixing algorithm needs to be developed to mix signal events with background events. It is an essential module if we want to make Monte Carlo data match well with real data. There are two options for background mixing: hit level mixing and readout level mixing. For hit level mixing, the hits from both signal and background are sorted by time first, and then handled by the electronics simulation. Hit level mixing is closer to the real case, but requires a lot of computing resources. Readout level mixing is much easier to implement and requires less computing resources, but it can not accurately model the overlapping between multiple hits. Since both options have advantages and disadvantages, more studies are necessary before an option is finally selected as the official JUNO background mixing algorithm. \subsection{Event Display} The event display is an indispensable tool in high energy physics experiments. It helps to develop simulation and reconstruction packages in the offline software, to analyze the physics in a recorded interaction and to illustrate the whole situation to general audiences in an imaginable way. \subsubsection{Detector description and display} The display of the detector is provided by the geometry service in the offline software. The visual attributes of every detector unit are associated and controlled by identifiers. The event display software will support two display modes -- 3D display and 2D projection display. The 2D projection will be realized by the 2D histograms that are provided by ROOT. For 3D display, several popular 3D graphic engines, like OpenGL \cite{OpenGL}, OpenInventor \cite{OpenInventor} and X3D \cite{X3D}, have been proven as practical choices for event display software for some other HEP experiments. More studies are necessary to determine which infrastructure 3D graphic engine will work best for JUNO and its software environment. Currently a solution based on OpenGL is under study to realize the 3D display mode. \subsubsection{Event data and display} The event data in the offline system has various formats, including raw data, simulation data, calibrated data, reconstruction data and data analysis results. Their data formats are uniformly defined by the JUNO offline event model. In processing the event data for display, we need to study how to display the particle trajectories, their interactions with materials and hit response in simulation data, the reconstructed event vertices, energy and display attributes, as well as the relationships between these. \subsubsection{Graphical User Interface} A Graphical User Interface (GUI) provides a convenient interface for users to execute complicated commands with simple actions such as pushing buttons, selections, sliding a ruler or pushing hot keys. Frequently used functions, like event control, 3D view rotation and translation, zooming and time controls can be realized through widgets. A ROOT-based GUI framework has been implemented to realize these functions, giving the advantages of seamless binding with the ROOT analysis platform, ROOT-based geometry service and event model. \subsubsection{Graphic-based analysis and tuning} Graphic-based analysis will further provide some useful functions such as reconstruction algorithm tuning, event trigger and event type discrimination, which can help users to quickly filter, analyze and understand the events in which they are interested. \subsection{Event Model} The event model defines the key data units to be processed in the offline data processing. It not only defines the information included in one event at different data processing steps such as event generation, detector simulation, calibration, reconstruction and physics analysis, but also provides the ways for events to be correlated between different processes. Since ROOT \cite{ROOT} has been extensively used as the data storage format in most particle physics experiments, it is decided that the JUNO event model is based on the ROOT TObject. So all the event classes are inherited from TObject. In this way, we can directly take advantage of the powerful functionalities provided by ROOT, such as schema evolution, input/ouput streamer, runtime type information (RTTI), inspection and so on. In order to quickly view event information and make the decision whether or not a given event is selected, the event information is divided into two levels: Event Header and Event. The Event Header defines the sequential information (such as RunId, EventId and Timestamp etc.) and characteristic information (such as energy of events), while Event defines more detailed information about the event. In this way, the event header can be read in from data files or other storage and its characteristic information can be used for fast event selection without taking extra CPU time or memory required to read in the contents of a full event. \begin{figure}[htb] \begin{center} \includegraphics[width=.8\textwidth]{OfflineSoftwareAndComputing/figures/DataModel_04_02.png} \caption[Event Model]{\label{fig:EventModel} Design Schema of JUNO Event Model} \end{center} \end{figure} Event correlations are very important for neutrino physics analysis since inverse beta decay (IBD) events will be regarded as two separate event signals: prompt signal and delayed signal. Both of them actually refer to the same physics event. In the JUNO experiment, we design SmartRef to fully meet this requirement and to avoid the duplication in writing these information into ROOT files. SmartRef is based on the TRef class in ROOT but has more powerful functions. It not only provides correlations between events but also provides the information for lazy loading of the data. For the JUNO event model, the class EvtNavigator is designed to organize all the event information at different processing steps and acts as an index when processing events. In Figure \ref{fig:EventModel}, we can see that EvtNavigator has several SmartRefs which refer to the headers of this event at different processes. Event model classes are usually simple. Only some quality variables and functions need to be defined. These functions can be divided into three types: Set-functions to set new values, Get-functions to return the variable's value, and streamer functions for reading/writing data from/to files. Most of the code is similar event by event, so an XOD (Xml Object Description) tool is developed and used to describe the event information with a more readable XML file. XOD can automatically produce the event header file, the event source file and the dictionary file needed by ROOT. \subsection{Generator} Generators produce simulated particles with desired distribution of momentum, direction, time, position and particle ID. The generators for the JUNO experiment include inverse beta decay (IBD) generator, radioactivity generator, muon generator, neutron generator, and calibration source generator. Some of the generators can be imported from the existing Daya Bay software, while the others need to be developed for JUNO specifically. The generators are grouped into three kinds according to their development platform. One is based on Geant4 ParticleGun for shooting particles with specified particle ID, momentum and position distribution. The second kind, such as ${}^{238}U$ and ${}^{232}Th$ radioactivity generators, is developed with existing Fortran libraries. The third kind of generators, such as IBD and GenDecay, will be developed in C++ language,. The output data of generators should be in the format of HepEvt or HepMC \cite{HepMC} for better communication with other applications. The design of particle vertex position generation need to be flexible and allow the distribution to be geometry related, for example, a random distribution on the surface of a PMT, in liquid scintillator or in acrylic. The muon generator particularly depends on the muon flux, energy spectrum and angular distribution at the JUNO site. A digitized description of the actual landform at the JUNO site need to be acquired first, and then used as input to MUSIC \cite{MUSIC} to simulate muons passing through the rock, getting the flux, energy spectrum and angular distribution of the muons finally reaching the JUNO detector. The effect of muon bundles also needs to be considered for such a large detector as JUNO. \subsection{Geometry} The geometry system describes and manages the detector data and structure in software, providing a consistent detector description for applications like simulation, reconstruction, calibration, alignment, event display and analysis. All subsystems, including the Central Detector, Top Veto Tracker, Water Cherenkov Pool and the Calibration system, will be described in the geometry system. The design of the JUNO detector also requires the geometry system to be flexible and able to handle time-dependent geometry and the co-existence of different designs. Several detector description tools have been developed and implemented in high energy physics experiments, with some of the most popular languages being XML (eXtensible Markup Language) \cite{XML}, HepRep \cite{HepRep}, VRML(Virtual Reality Modeling Language, updated to X3D) \cite{VRML} \cite{X3D} and GDML (Geometry Description Markup Language) \cite{GDML}. At the stage of conceptual design for the JUNO detector, the detector description geometry is based on GDML, including geometry data, detector structure, materials and optical parameters. In consideration of the co-existence of multiple major designs and frequent minor modifications to the designs, we adopt GDML as the main tool to transfer detector description data between different applications, because GDML requires less human work and has the advantage of automatic translation of detector data between Geant4 \cite{Geant4} and ROOT, the two most popular HEP programs, which are widely used in simulation, reconstruction and analysis. However, it is noteworthy that GDML also has some limitations. For example, its Geant4 and ROOT interfaces are not completely consistent, so some information like optical surface description and matrix elements may be lost in the detector data translation. The GDML interfaces do not support complicated (user defined) shapes and require more work to implement the extensible functions. These small problems need to be solved in future geometry software development. At the current stage, the detector geometry is described in each individual Geant4 simulation code, which also serves as the unique source of detector description. When a simulation job is running, it constructs the detector structure in Geant4 and exports it in GDML format; at the end of the simulation job, the GDML detector data is read back and converted into ROOT geometry object format, which is bundled with the output event data ROOT file while being written out. When another application such as reconstruction or event display wants to read in the geometry information, it searches the bundled ROOT file for the geometry object to initialize the geometry offline, and uses the geometry service package to retrieve the detector unit information it requires. This mechanism gives every detector designer the maximum flexibility to modify his specific design, uses a single module to handle multiple detector designs at the same time and guarantees the consistency of detector description between different applications. In the future, when the detector design is fixed, it is preferred to generate the detector description within the geometry system and provide it to all applications through interfaces, rather than the current solution of generating geometry data in Geant4. \subsection{Maintenance and service support} \subsubsection{Network support platform} Network support platform provides professional and comprehensive solutions for JUNO network and public services. On the one hand, it integrates related resources on the network side, and then transfers them to the expert of the user service, which can realize quick response to network complaints, and effectively raise user service front-end settlement ratio, reduce complaint processing links, reduce complaint processing duration, and enhance user satisfaction.On the other hand,as for the information related to network complaint received from user sides, the network side will obtain it automatically and pay active attention to it, and realize the comprehensive correlation between user complaint and network execution situation, and the quality degradation warning of terminal-to-terminal network, so as to guide network optimization, enhance network quality, and further realize active care for users and enhance service quality. \subsubsection{Fine-grained monitoring and maintenance} Since JUNO data process is a large scale task including many devices, a fine-grained monitoring system is necessary to guarantee smooth running of the platform. It is required that the monitoring system be real-time, easy to use and can recover from some unexpected errors itself. For those errors which can not be recovered, a warning should be sent in time. \begin{figure}[htb] \begin{center} \includegraphics[width=.8\textwidth]{OfflineSoftwareAndComputing/figures/monitor_arc.png} \caption[Monitoring tool architecture]{\label{fig:monitor}Fine-grained monitoring architecture } \end{center} \end{figure} \subsubsection{ Information Service Platform} The information Service Platform is an information system which provides services for the research and management activities of the JUNO experiment. The Information Service Platform will provide system services for many aspects of the experimental research and management, such as the JUNO Experiment website, wiki, conference management, document management, collaboration management, video conference, workshop live broadcasting/recording, remote shifts, personal data sharing, science advocacy and so on. \subsection{Software Platform} The offline software system is designed to meet the various requirements of data processing. As JUNO software developers are dispersed all over the world, a unified software platform is required, which provides a formal working environment. This has advantages for resource optimization and manpower integration, which can improve the software development, usage and maintenance. Based on our experience, Linux OS and GNU compiler are the first choice for the JUNO software platform. However other popular OSes and development environments should be considered for the sake of compatibility. The CMT \cite{CMT} tool, which can calculate package dependencies and generate Makefiles, is used for the software configuration and management. An automated installation tool can ease the deployment of the JUNO software, and is also helpful for daily compilation and testing. Users are able to concentrate on the implementation of software functionality, without suffering from different development environments. Nowadays, mixed programming with multiple languages is practical, so we can choose different technologies for different parts of our software. The main features of the application are implemented via C++ to guarantee efficiency. User interfaces are provided in Python for additional flexibility. Boost.Python, as a widely used library, is a good choice to integrate C++ and Python. If it is included properly at the beginning of the system design, most users will be able to enjoy the benefits of mixed programming without knowing the details of Boost.Python. We will implement a software framework, SNiPER, as the core software of the JUNO platform. SNiPER stands for "Software for Non-collider Physics ExpeRiments". As shown in Fig.\ref{fig:OfflinePlatform}, components of the JUNO software, such as simulation and reconstruction, are executed as plug-ins based on SNiPER. The principle of this new framework is simplicity and high efficiency. It depends on a minimal set of external libraries, fully meeting our requirements without over-engineering or loss of efficiency. \begin{figure}[htb] \begin{center} \includegraphics[width=.8\textwidth]{OfflineSoftwareAndComputing/figures/Offline_Software_Platform.png} \caption[Offline Software Platform]{\label{fig:OfflinePlatform}The structure of the offline software system} \end{center} \end{figure} The SNiPER kernel is the foundation of the JUNO software system, and determines the vitality of the whole platform. We design and manage SNiPER modularly. Every functional element is implemented as a module, and can be dynamically loaded and configured. Modules are designed as high cohesion units with low couplings between each other. They communicate only through interfaces. We can replace or modify one module without affecting any of the others. Since it's flexible and expandable, users can contribute to it independently. The model is inspired by other pioneering software frameworks such as Gaudi \cite{Gaudi}. Expandable modules are strategically distinguished as algorithms and services. An algorithm provides a specific procedure for data processing. A sequence of algorithms composes the whole data processing chain. A service provides useful features that can be called by users when necessary. Algorithms and services are plugged and executed dynamically. They can be selected and combined flexibly for different requirements. Data to be processed may come from different places and be in different types and formats. The results may also be stored in different ways. In the framework we reserve interfaces for multi-stream I/O support, so that the I/O service can communicate conveniently with other modules via the interface. At present, the disk access could be a serious bottleneck in massive data processing. This should be considered in I/O service developing. Many techniques, such as lazy loading, are adopted to optimize the I/O efficiency. The framework also involves many frequently used functions, such as the logging mechanism, particle property lookup, system resource loading, database access and histogram booking, etc. These contents are wrapped in services in SNiPER, which will ease the development procedure. Users can configure a job interactively with the Python command line or in batch mode with script files. The flexibility of Python will ease the job execution procedure. The JUNO experiment has a long lifetime and software technology is still evolving rapidly, so the framework has to be very expandable. It should be able to integrate with new tools and libraries in future. This will save the manpower needed for developing, and improve the system robustness. It is expected that JUNO will have a huge amount of data. In order to make use of all possible computing resources, it is necessary to consider parallel computing techniques. Modern CPUs are generally multi-core. Multi-threaded programming extracts the capacity of multi-core CPUs, and speeds up a single job significantly. In PC clusters, a server/client distributed system can be implemented and deployed, too. The latest computing technology and HEP computing techniques, including GPUs, Grid and Cloud, should be explored and used as appropriate. \subsection{Requirement Analysis} The offline software plays an important role in improving physics analysis quality and efficiency. From the point of view of physics analysis, the following requirements need to be applied to the offline software system. \begin{itemize} \item Compared to collider experiments, neutrino experiments have two specific characteristics: 1). There is time correlation between events. 2).There is only a very small fraction of signal events among a large number of backgrounds. The software framework should therefore provide a mechanism of flexible data I/O and event buffering to enable high efficiency data access and storage, as well as the capability to retrieve the events within a user-defined time window. \item An interface should be provided for different algorithms to interchange event data. There should be a unified geometry management. Software and physics analysis parameters should be managed by a conditional database. An interface to retrieve correct parameters according to software versions should also be provided. \item There are many kinds of events in the JUNO experiment. Apart from reactor neutrino events, they are supernova neutrinos, geo-neutrinos, solar neutrinos, atmospheric neutrinos, radioactivity etc. The offline software should provide accurate simulations to describe interactions of neutrinos and background events with the detector. \item The detector simulation software should be able to simulate detector performance and guide detector design. In addition, the calculation of detector efficiencies and systematic errors, which are needed by physics analysis, also relies on detector simulation. \item The detector performance depends on event reconstruction. In order to get better energy resolution, optical model in liquid scintillator detector should be constructed, using PMT charge and time information to get event vertex and energy. Event vertex can also be used to veto the natural radioactivity background events from outside the liquid scintillator. To reduce the isotope background from muons, accurate muon track reconstruction is needed. For atmospheric neutrinos, the reconstruction of short tracks inside the liquid scintillator and identification of different charged particles are needed. \item Event display software is needed in order to show the detector structure, physics processes in the detector, and reconstruction performance. \end{itemize} The JUNO detector will produce about 2 PB raw data every year, which will be transferred back to the Computing Center at the Institute of High Energy Physics (IHEP) in Beijing through a dedicated network connection. Due to the large data volume, a big scale of offline computing platform is required by Monte Carlo production, raw data processing and physics analysis, as well as data storage and archiving. A rough estimation suggests a level of 10,000 CPU cores, 10 PB disk storage and 30 PB archive in the future. The computing nodes and storage servers will be connected to each other by a 40-Gbps backbone high-speed switching network. In order to improve data processing capacity, the platform will integrate the computing resources contributed by outside members via a distributed computing environment. So the experiment data can be shared among collaboration members and computation tasks can be dispatched to all of the computing sites managed by the platform. Considering the above needs, requirements on data processing, storage, data transfer and sharing, can be summarized as follows: \paragraph{Data Transfer} A data transfer and exchange system with high performance, security and reliability is required. To guarantee the timely transfer of raw data from the experiment site to the IHEP data center, the system should monitor and trace the online status of data transfers and it should provide a data transfer visualization interface. This should also guarantee data is dispatched and shared among different data sites within the collaboration smoothly. \paragraph{Computing Environment} A computing farm with 10,000 CPU cores will be established at the IHEP Computing Center to facilitate successful processing of the JUNO data. The collaboration members will provide additional computing resources. The computing environment includes not only the local computing farm, but also those remote resources contributed by outside collaboration members. \paragraph{Data Storage} JUNO will generate a huge amount of experiment data. Data from the experiment site will be transferred and stored at the IHEP Computing Center. The capacity of storage at IHEP will be 10 PB for storing both real and simulated data. It should support concurrent access by applications running on 10,000 CPU cores. Data migration between tape repository and disk pools should be transparent to users. \paragraph{Maintenance, Monitoring and Service Support} To improve the stability of the JUNO computing resources, an intelligent monitoring system will be developed. With this system, the status of both platform and network connection can be monitored through a fine-grained control. The monitor system can not only give the status assessment of the whole platform but also give predictions and prior warnings for various platform services. A set of information management tools should be developed including conference support, document management and online-shift management etc. \subsection{Introduction} The offline software and computing system consists of two separate parts: offline software and computing. It not only closely connects data processing and physics analysis, but also it can be regarded as a bridge between detector operations and physics analysis. The primary tasks of this system are to process raw data collected by the detector and produce reconstructed data, to produce Monte Carlo data, to provide software tools for physics analysis, and to provide the networking and computing environment needed by data processing and analysis. The offline software includes software framework, event generator, detector simulation, event reconstruction and event display etc. The framework is the underlying software supporting the whole system. Based on this framework, experiment related software will be developed such as software for event data model, data I/O, event generation, detector simulation, event reconstruction, physics analysis, geometry service, event display, and database service etc. To facilitate the development, the infrastructure of offline software provides various useful tools such as tools for compiling, debugging, deployment and installation etc. The offline computing is designed to build a high efficiency and reliable computing environment for the JUNO experiment. It has the following major parts: a data transfer platform to transfer raw data from the experiment site to the Computing Center at IHEP, a large scale of computing farm for data processing and data analysis, a massive data storage subsystem and a monitoring and management platform providing fine-grained monitoring and management of the computing resources. \chapter{Assembly, Installation and Engineering Issues} \label{ch:AssemblyAndInstallation} \input{AssemblyAndInstallation/chapter12.1.tex} \input{AssemblyAndInstallation/chapter12.2.tex} \input{AssemblyAndInstallation/chapter12.3.tex} \input{AssemblyAndInstallation/chapter12.4.tex} \input{AssemblyAndInstallation/chapter12.5.tex} \section{Introduction} After the completion of civil construction, through systems integration, assembly, installation and commissioning, all subsystems will be formed as a whole JUNO detector. In this process, we need to establish various engineering regulations and standards, and to coordinate subsystems' assembly, installation, testing and commissioning, especially their onsite work. \subsection{Main Tasks} Main tasks of the integration work include: \begin{itemize} \item To prepare plans and progress reports of each phase; \item To establish a project technology review system; \item To standardize the executive technology management system; \item To have strictly executive on-site work management system; \item To develop and specify security management system on-site; \item To prepare common tools and equipment for each system, and to guarantees project progress; \item To coordinate the installation progress of each system according to the on-site situation. \end{itemize} \subsection{Contents} The work contents mainly include: \begin{itemize} \item To summarize the design, and review progress of each subsystem; \item To organize preparation work for installation in the experiment region; \item To inspect and certify Surface Buildings, underground Tunnels, and Experiment Hall with relevant utilities; \item To coordinate technology interfaces between key parts; \item To coordinate the procedure of assembly and installation both on surface and underground; \end{itemize} \section{ Design Standards, Guidelines and Reviews} \subsection{Introduction} There will be a document to outline the mechanical design standards or guidelines that will be applied to the design work. It also describes the review process of engineering design that will be implemented to ensure that experimental equipment meets all requirements for performance and safety. The following is a brief summary of the guidance that all mechanical engineers/designers should follow in the process. For reasons ranging from local safety requirements to common sense practices, the following information should be understood and implemented in the design process. \subsection{Institutional Standards} When specific institutional design standards or guidelines exist, they should be followed. The guidelines outlined are not meant to replace but instead to supplement institutional guidelines. The majority of equipment and components built for the JUNO Experiment will be engineered and designed at home institutions, procured or fabricated at commercial vendors, then eventually delivered, assembled, installed, tested and operated at the JUNO Experimental facilities, Jiangmen, China. The funding agencies in other countries as well as Jiangmen, China will have some guidelines in this area, as would your home institutions. Where more than one set of guidelines exist, use whichever are the more stringent or conservative approaches. \subsection{Design Loads} The scope of engineering analysis should take handling, transportation, assembly and operational loads into account as well as thermal expansion considerations caused by potential temperature fluctuations. For basic stability and a sensible practice in the design of experimental components an appropriate amount of horizontal load (0.10~g) should be applied. In addition, seismic requirements for experimental equipment is based on a National Standard of People's Republic of China Code of Seismic Design of Buildings GB 50011-2010. At the location of the JUNO Experiment the seismic fortification intensity is grade 7 with a basic seismic acceleration of 0.10~g horizontal applied in evaluating structural design loads. A seismic hazard analysis should be performed and documented based on this local code. The minimum total design lateral seismic base shear force should be determined and applied to the component. The direction of application of seismic forces would be applied at the CG of the component that will produce the most critical load effect, or separately and independently in each of the two orthogonal directions. A qualitative seismic performance goal defines component functionality as: the component will remain anchored, but no assurance it will remain functional or easily repairable. Therefore, a seismic design factor of safety F.S. > 1 based on the Ultimate Strength of component materials would satisfy this goal. Where an anticipated load path as designed above, allows the material to be subjected to stresses beyond the yield point of the material, redundancy in the support mechanism must be addressed in order to prevent a collapse mechanism of the structure from being formed. The potential for buckling should also be evaluated. It should be noted that a rigorous analytical seismic analysis may be performed in lieu of the empirical design criteria. This work should be properly documented for review by the Chief Engineer and appropriate Safety Committee personnel. \subsection{Materials Data} All materials selected for component design must have their engineering data sources referenced along with those material properties used in any structural or engineering analysis; this includes fastener certification. There are many sources of data on materials and their properties that aid in the selection of an appropriate component material. There are many national and international societies and associations that compile and publish materials test data and standards. Problems frequently encountered in the purchasing, researching and selection of materials are the cross-referencing of designations and specification and matching equivalent materials from differing countries. It is recommended that the American association or society standards be used in the materials selection and specification process, or equivalency to these standards must be referenced. Excellent engineering materials data have been provided by the American Society of Metals or Machine Design Handbook Vol. 1-5 and Mechanical Industry Publishing Co. in PRC, Vol 1-6, 2002, which are worth investigating. \subsection{Analysis Methods} The applicable factors of safety depend on the type of load(s) and how well they can be estimated, how boundary conditions have been approximated, as well as how accurate your method of analysis allows you to be. \paragraph{Bounding Analyses:} Bounding analysis or rough scoping analyses have proven to be valuable tools. Even when computer modeling is in your plans, a bounding analysis is a nice check to avoid gross mistakes. Sometimes bounding analyses are sufficient. An example of this would be for the case of an assembly fixture where stiffness is the critical requirement. In this case where deflection is the over-riding concern and the component is over-designed in terms of stress by a factor of 10 or more, then a crude estimation of stress will suffice. \paragraph{Closed-Form Analytical Solutions:} Many times when boundary conditions and applied loads are simple to approximate, a closed-form or handbook solution can be found or developed. For the majority of tooling and fixture and some non-critical experimental components, these types of analyses are sufficient. Often, one of these formulas can be used to give you a conservative solution very quickly, or a pair of formulas can be found which represent upper and lower bounds of the true deflections and stresses. Formulas for Stress and Strain by Roark and Young is a good reference handbook for these solutions. \paragraph{Finite Element Analysis:} When the boundary conditions and loads get complex, or the correctness of the solution is critical, computer modeling is often required. If this is the case, there are several rules to follow, especially if you are not intimately familiar with the particular code or application. \begin{enumerate} \item Always bound the problem with an analytical solution or some other approximate means. \item If the component is critical, check the accuracy of the code and application by modeling a similar problem for which you have an analytical or handbook solution. \item Find a qualified person to review your results. \item Document your assumptions and results. \end{enumerate} \subsection{Failure Criteria} The failure criterion depends upon the application. Many factors such as the rate or frequency of load application, the material toughness (degree of ductility), the human or monetary risk of component failure as well as many other complications must be considered. Brittle materials (under static loads, less than 5\% yield prior to failure), includes ceramics, glass, some plastics and composites at room temperature, some cast metals, and many materials at cryogenic temperatures. The failure criterion chosen depends on many factors so use your engineering judgment. In general, the Coulomb-Mohr or Modified Mohr Theory should be employed. Ductile materials (under static loads, greater than 5\% yield prior to failure), includes most metals and plastics, especially at or above room temperature. The failure criterion chosen again ultimately rests with the cognizant engineer because of all the adverse factors that may be present. In general, the Distortion- Energy Theory, or von Mises-Hencky Theory (von Mises stresses), is most effective in predicting the onset of yield in materials. Slightly easier to use and a more conservative approach is the Maximum-Shear-Stress Theory. \subsection{Factor of Safety} Some institutions may have published guidelines which specifically discuss factors of safety for various applications. For the case where specific guidelines do not exist, the following may be used. Simplistically, if F is the applied load (or S the applied stress), and $F_f$ is the load at which failure occurs (or $S_s$ the stress at which failure occurs), we can then define the factor of safety (F.S.) as: \begin{displaymath} F.S. = F_{f} / F \quad\mathrm{or}\quad S_{s} / S \end{displaymath} The word failure, as it applies to engineering elements or systems, can be defined in a number of ways and depends on many factors. Discussion of failure criteria is presented in the previous section, but for the most common cases it will be the load at which yielding begins. \subsection{Specific Safety Guidelines for JUNO} Lifting and handling fixtures, shipping equipment, test stands, and fabrication tooling where weight, size and material thickness do not affect the physical capabilities of the detector, the appropriate F.S. should be at least 3. When life safety is a potential concern, then a F.S. of 5 may be more appropriate. Note that since the vast majority of this type of equipment is designed using ductile materials, these F.S.'s apply to the material yield point. Experimental hardware that does not present a life safety or significant cost/schedule risk if failure occurs, especially where there is the potential for an increase in physics capabilities, the F.S. may be as low as 1.5. Many factors must be taken into account if a safety factor in this low level is to be employed: a complete analysis of worst case loads must be performed; highly realistic or else conservative boundary conditions must be applied; the method of analysis must yield accurate results; reliable materials data must be used or representative samples must be tested. If F.S.'s this low are utilized, the analysis and assumptions must be highly scrutinized. Guidelines for F.S. for various types of equipment are: \begin{center} \begin{tabular}{|p{3.5cm}|c|p{3.5cm}|} \hline Type of Equipment & Minimum F.S. & Notes \\ \hline Lifting and handling & 3 - 5 & Where there is a risk to life safety or to costly hardware, choose F.S closer to 5. \\ \hline Test stands, shipping and assembly fixtures. & 3 & \\ \hline Experimental hardware & 1.5 - 3 & 1.5 is allowable for physics capability and analysis where method is highly refined \\ \hline \end{tabular} \end{center} \subsection{Documentation} It is not only good engineering practice to document the analysis, but it is an ESH\&Q requirement for experimental projects. For this reason all major components of JUNO Experiment will have their engineering analyses documented as Project Controlled Documents. Utilize your institutional documentation formats or use the following guidelines. Calculations and analyses must \begin{itemize} \item Be hard copy documented. \item Follow an easily understood logic and methodology. \item Be legible and reproducible by photocopy methods. \item Contain the following labeling elements. \begin{itemize} \item Title or subject \item Originators signature and date \item Reviewers signature and date \item Subsystem WBS number. \item Introduction, background and purpose. \item Applicable requirements, standards and guidelines. \item Assumptions (boundary conditions, loads, materials properties, etc.). \item Analysis method (bounding, closed form or FEA). \item Results (including factors of safety, load path and location of critical areas). \item Conclusions (level of conservatism, limitations, cautions or concerns). \item References (tech notes, textbooks, handbooks, software code, etc.). \item Computer program and version (for computer calculations) \item Filename (for computer calculations). \end{itemize} \end{itemize} \subsection{Design Reviews} All experimental components and equipment whether engineered or procured as turn-key systems will require an engineering design review before procurement, fabrication, assembly or installation can proceed. The L2 subsystem manager will request the design review from the project management office, which will appoint a chair for the review and develop a committee charge. The selected chair should be knowledgeable in the engineering and technology, and where practical, should not be directly involved in the engineering design effort. With advice from the L2, the technical board and project office chair appoints members to the review committee that have experience and knowledge in the engineering and technology of the design requiring review. At least one reviewer must represent ESH\&Q concerns. The committee should represent such disciplines as: \begin{itemize} \item Science requirements \item Design \item Manufacturing \item Purchasing \item Operational user \item Maintenance \item Stress analysis \item Assembly/installation/test \item Electrical safety \item ESH\&Q \end{itemize} For the JUNO project there will be a minimum of two engineering design reviews: Preliminary and Final. The preliminary usually takes place towards the end of the conceptual design phase when the subsystem has exhausted alternative designs and has made a selection based on value engineering. The L2 along with the chair ensures that the preliminary design review package contains sufficient information for the review along with: \begin{itemize} \item Agenda \item Requirements documents \item Review committee members and charge \item Conceptual layouts \item Science performance expectations \item Design specifications \item Supportive R\&D or test results \item Summaries of calculations \item Handouts of slides \end{itemize} A final design review will take place before engineering drawings or specifications are released for procurement or fabrication. The L2 along with the chair ensures that the final design review package contains sufficient information for the final review along with: \begin{itemize} \item Changes or revisions to preliminary design \item Completed action items \item Final assembly and detailed drawings \item Final design specifications \item Design calculations and analyses \end{itemize} The committee chair will ensure that meeting results are taken and recorded along with further action items. The committee and chair shall prepare a design review report for submittal to the project office in a timely manner which should include: \begin{itemize} \item Title of the review \item Description of system or equipment \item Copy of agenda \item Committee members and charge \item Presentation materials \item Major comments or concerns of the reviewers \item Action items \item Recommendations \end{itemize} The project office will file the design review report and distribute copies to all reviewers and affected groups. \section{On-site management} According to the experience from the Daya Bay Experiment, an effective technology review system has been well practiced. In JUNO, we will take it as a good reference and carry out for standardized review from the start of all system design, scheme argument, technology checks, we should also establish management system, to cover engineering drawing control, engineering change control procedure, and mechanical design standards, guidelines and reviews, etc. To control on-site work process, proper process should be introduced as well as a management structure. Safety training and safety management structure should be also introduced. \section{Equipment and Tools Preparation on Site} All of key equipment, devices and tools should be in place and get acceptance for installation on site, including \begin{itemize} \item Cranes: \begin{itemize} \item 2 sets of bridge crane with 12T/5T load capacity, and lifting range should cover any point for components delivery and installation in EH; \item Several lifting equipment are needed in Surface Assembly areas, Assembly Chamber underground, Storage area for Liquid Scintillation, and other Chambers accordingly; \end{itemize} \item Removable lifting equipment: Forklifts, manual hydraulic carriages, carts, etc. \item Mobile vehicle for smaller items: Pickup van, battery car, and so on; \item Several work aloft platform: Scissor lifts, Boom lifts, Scaffoldings and Floating platforms, etc. \item Common Tooling: complete kits of tool, fixtures, jigs, models, work-bench, cabinets, and tables, etc. \item Machine shop, Power equipment rooms, Control rooms are ready to be put into use. \item Safety measures and relevant equipment in place. \end{itemize} \section{Main Process of Installation} Since there still are several optional design versions which could be decided later, and different design could have different requirements for installation, therefore, the specific installation procedures will be established and developed later, here a very rough installation procedure is given. \begin{itemize} \item Mounting, debugging, testing of basic utilities, such as, cranes, vertical elevator, scaffoldings, test devices, etc. \item Survey, alignment, and adjustment for tolerances of form and position, include, embedded inserts, anchor plates, positioning holes, etc. \item Mounting Bottom Structure of Veto Cherenkov detector on the floor of Water Pool. \item Mounting anchor-plate for support poles of CD on the floor of Water Pool \item Pre-mounting of Top Structure of Veto system along wall of Water Pool \item Mount Substrate and Rails for Bridge over the water Pool edges \item Mount Tyvek with support structure for Water Cherenkov Detector, and with PMT positioning in the Water Pool \item Installation for CD with its PMT system and cabling in Water Pool details will be established until design final selection. \item Mount Calibration System \item Mount Geomagnetic Shielding system around Central Detector in the Pool \item Final installation for Tyvek and PMT of Veto system in the Pool \item Mount a temporary protective cover on top of Pool, to prevent from anything falling into Pool \item Mount Bridge on the rails, and drive it closed \item Mount Top Track Detector on Bridge and establish Electronic Room there \item Have Cabling completion, and test to get normal signals \item Check Grounding Electrode \item Dry Run \item Final Cleaning for Water Pool, then, drain out \item Filling Water, LS, LAB (or Oil) evenly into Water Pool and CD \item Mount Final Pool Cover, closed, and get fixation \end{itemize} \chapter{AssemblyAndInstallation} \label{ch:AssemblyAndInstallation} \input{AssemblyAndInstallation/chapter12.1.tex} \input{AssemblyAndInstallation/chapter12.2.tex} \input{AssemblyAndInstallation/chapter12.3.tex} \input{AssemblyAndInstallation/chapter12.4.tex} \input{AssemblyAndInstallation/chapter12.5.tex} \section{Design Goals and Specifications} \subsection{Introduction} Photon detectors measure the scintillation light created by interactions of the neutrinos with liquid scintillator and are key components for accomplishing the physics goals of JUNO. Important requirements for photon detectors used in JUNO include high detection efficiency for scintillation photons, large area, low cost, low noise, high gain, stable and reliable, long lifetime. JUNO plans to use approximately 20,000 tons or 23175~m$^{3}$ liquid scintillator hosted in a spherical container of diameter 35.4~m. Photon detectors will be mounted on a sphere 40~m in diameter covering at least 75\% of the 4654~m$^{2}$ surface area of the sphere according to the reference design of the central detector. Large area vacuum photomultipliers are the only viable choice for the JUNO photon detectors. Table~\ref{tab:PMTsrequiredbyJUNO} shows the numbers and sizes of PMTs required for JUNO in two scenarios with photocathode coverage ratio of 75\% and 70\%. Assuming 508~mm (20~inch) diameter PMTs with a photocathode diameter 500~mm are used, $\sim$17,000 such PMTs are required by the central detector of JUNO to cover 75\% of the detector surface. \begin{table}[!hbp] \centering \caption{PMTs required by JUNO\label{tab:PMTsrequiredbyJUNO}} \begin{tabular}{|c|c|c|} \hline \hline PMT mounting diameter & photocathode coverage & 508 mm \\ \hline 40 m & 70\% & 16,000 \\ \hline 40 m & 75\% & 17,000 \\ \hline \end{tabular} \end{table} \subsection{PMT Spectications} The main specifications for the 508~mm PMTs in JUNO are given in Fig.~\ref{fig:specificationOfPMTs}. Bialkali (BA) photocathode will be used. The spectral response of the BA photocathode matches the light emission spectrum of the liquid scintillator well. One of the most critical requirements for JUNO PMTs is high photon detection efficiency (PDE), which is equal to the photocathode quantum efficiency (QE) multiplied by the efficiency of photoelectron collection efficiency (PCE) in the PMT vacuum. In order to achieve the required peak PDE of 35\%, the peak photocathode QE is set to 38\% by assuming the PCE to be 93\%. We hope the average PDE for scintillation photons, which have a rather broad spectrum, can be > 30\%. Note that PMTs are measured by manufacturers in air and with incident light beams perpendicular to the surface of the glass entrance while PMTs are submerged either in water or in oil when used in JUNO, and the scintillation light can penetrate the curved glass window of a PMT at any angle. Small discrepancies regarding the photocathode QE between the manufacturers, measurements and those experimentally observed can occur as a result of differences in light reflections at the medium/glass interface. JUNO PMTs will work in photon counting mode and the typical gain required is $10^{7}$. The PMT noise pulse level must be low so that the single photoelectron peak can be clearly separated from the PMT noise. In JUNO cosmic events the minimum momentum of cosmic ray muons is 218~GeV and the penetrating muons' rate is 3.8~Hz. Energy deposition by muons that do not generate showers would be as high as 7~GeV and a total of $7\times10^{7}$ scintillation photons would be produced. The required dynamic range of JUNO PMTs is given as 0.1-4000~pe's. The photocathode noise rate, rate of prepulse and after-pulses are given in Fig.~\ref{fig:specificationOfPMTs}. The radioactivity levels of the PMTs, which mostly comes from the glass, are also given. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.1.png} \caption[ Specifications of PMTs for JUNO]{\label{fig:specificationOfPMTs} Specifications of PMTs for JUNO } \end{center} \end{figure} In addition to the 508~mm diameter PMTs for measuring scintillation light generated by neutrino reactions in the detector, approximately 2000 of 203~mm (8~inch) diameter PMTs are needed for cosmic ray muon veto by detecting Cherenkov light generated by cosmic muons. These PMTs will be mounted on the outside surface of the stainless steel structure that supports the liquid scintillator vessel. The total area to be covered is approximately 5000~m$^{2}$. The requirements for these PMTs are similar to those used in the Daya Bay neutrino detector. \section{MCP-PMT R\&D } In order to ensure optimum performance of the detector, the PDE of PMTs in the detector is preferred to be $\ge$~35\%, averaging over the entire photocathode and scintillation spectrum. No commercial large PMTs can achieve this. A R\&D effort intended to accomplish such an ambitious goal was started in early 2009. This effort was led by IHEP (Beijing) and joined by research institutes, university groups and related companies in China, and a formal R\&D group was established in early 2011. Significant progress in all aspects of PMT R\&D has been made. The goal of the design and manufacturing 508~mm PMTs that meet the requirements listed in Fig.~\ref{fig:specificationOfPMTs} and in particular to achieve the 35\% PDE is not easy and cannot be taken for granted. If we are unable to accomplish our stated goal in time, we will make corresponding adjustments to the design of the experiment. Depending on the status of our R\&D effort by the time of the decision on the PMT choice, commercially available PMTs that may or may not have the desired photon detection efficiency can be considered. It is worth noting that in early 2009, a team from IHEP lead by Yifang Wang, then the deputy director of IHEP visited a former PMT manufacturer, the No. 741 factory of the Chinese Electronics Administration in Nanjing, China. The manager of that factory became interested after hearing of our effort to develop large PMTs for JUNO (then Daya bay II). He made an effort to line up financial support from two entrepreneurs in Beijing and successfully negotiated a deal with PHOTONIS France S.A.S. and bought their stalled PMT production line and all related technologies. A Chinese PMT company, SCZ PHOTONIS, was established in 2011 and have now started to make 12 inch PMTs, with a plan in place to make 20 inch PMT samples soon. Meanwhile, IHEP representatives have been in contact with Hamamatsu, Japan, which have been developing high efficiency 20 inch PMTs for the future Hyper Kamiokande experiment in Japan and made good progress in the last few years. Hamamatsu has agreed to deliver 20 inch PMT samples to IHEP for testing. We will discuss the R\&D effort for JUNO PMTs by the collaboration led by IHEP (Beijing) in this section. Options provided by the two commercial companies will be discussed in section 6.4.2. \subsection{MCP-PMT R\&D at IHEP} The R\&D effort to develop PMTs for JUNO started in early 2009 when the Daya bay II experiment was in the very early conceptual development stage. The main obstacle for achieving the physics goals of the experiment was the lack of commercially available PMTs that can reach the photon detection efficiency (PDE) required. Of the two major PMT manufacturers in China, the aforementioned No. 741 Factory in Nanjing had stopped PMT production for a number of years and another PMT manufacturer, the No. 261 Factory in Beijing was bought by Hamamatsu and the resulted Hamamatsu Photonics (Beijing) produced only low-end small PMTs. The R\&D work on PMT related technologies such as Bialkali photocathode had ceased for more than a decade in China. The quantum efficiency (QE) of a vacuum PMT is conventionally defined as the number of photoelectrons emitted from the photocathode layer into vacuum by the number of incident photons. When the QE of a PMT is measured, the device is placed in air and the incident light beam is perpendicular to its glass window action. Typically 4\% of photons are reflected by the glass surface and more light is reflected at the interface of the glass substrate and the photocathode layer, which has a high index of refraction. The cathode sensitivity and QE as functions of wavelength of the incident photons are plotted in Fig.~\ref{fig:Fig.5.2} (taken from the PMT manual published by Philips Photonics). The blue curve is QE of conventional bialkali photocathode (SbK2Cs) with a peak value of about 30\%. The spectral response of the conventional BA photocathode matches the spectrum of scintillation light from the liquid scintillators well and the estimated average QE for the entire scintillation light emission spectrum is about 25\% based on this plot. However for large PMTs, the QE is normally lower than the values shown in Fig.~\ref{fig:Fig.5.2}. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.2.png} \caption[ The photocathode sensitivity and quantum efficiencies of various photocathodes]{\label{fig:Fig.5.2} The photocathode sensitivity and quantum efficiencies of various photocathodes} \end{center} \end{figure} For the 20 inch Hamamatsu R3600-02 used in the KamLand experiment, the peak QE is about 20\% and the peak photon detection efficiency (PDE) is only about 16\% since the photoelectron collection efficiency is 80\% as stated by Hamamatsu for the R3600-02. The 16\% peak PDE is far lower than the desired 35\% PDE for JUNO. Another factor that also needs to be considered is the effective photocathode area. The R3600-02 PMT has an outside diameter of 508~mm and its effective photocathode diameter is only 460~mm. This factor puts a limit on the capability for JUNO to maximize its photocathode coverage. Given these factors, it became apparent that we need to start our own large PMT R\&D program mainly to improve the photon detection efficiency. Note that new 20 inch PMT (R12860 Under development) samples with super bialkali photocathode from Hamamatsu have shown great improvement for the PDE. Our PMT R\&D was based on a concept proposed by IHEP (Beijing) in early 2009. The conceptual schematic is shown in Fig.~\ref{fig:Fig.5.3}. The concept calls for a spherical PMT with top hemisphere, as shown in the figure, used as transmission photocathode and its bottom hemisphere used as reflective photocathode which converts the photons that passed through the transmission photocathode. It is hoped that by optimizing the thickness of the photocathode in the top and bottom halves, the QE of such a PMT can be significantly improved compared to the conventional design that has only the top transmission photocathode. This new design requires two back-to-back compact electron multipliers, since conventional dynode chains would be too bulky. The proposed electron multipliers are two back-to-back pairs of microchannel plates (MCPs), so that photoelectrons from the top and bottom photocathodes are focused on the two sets of MCPs facing left and right, as shown in Fig.~\ref{fig:Fig.5.3}. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.3.png} \caption[The conceptual drawing of the high efficiency spherical PMT with two sets of MCP electron multipliers]{\label{fig:Fig.5.3} The conceptual drawing of the high efficient spherical PMT with two sets of MCP electron multipliers } \end{center} \end{figure} As a starting point, a contract for a 5 inch MCP-PMT prototype was signed with the Nanjing Electronic Devices Institute, which was the only place in China that had made MCP PMTs. The progress was slow at the beginning and after about two years, working prototypes were successfully made. While working with the Nanjing Electronic Devices Institute, it became apparent that we needed to expand our efforts in order to broaden our technology bases to solve many complex and difficult technical issues. The Large Area MCP-PMT Development Collaboration was formally established in 2011. Coordinated by IHEP, the collaboration consists of the CNNC Beijing Nuclear Instrument Factory, the Xi'An Institute of Optics and Precision Mechanics (XIOPM) of the CAS, Shanxi Baoguang Vacuum Electric Device Co., Ltd., China North Industries Group Corporation (NORINCO GROUP), Nanjing University, and the PMT Division of the CNNC Beijing Nuclear Instrument Institute (formally No. 261 Factory). Great progress has been made since the start of the collaboration almost three years ago. \subsection{Large Area High Efficiency MCP-PMT Project} Many critical components and technologies need to be developed for the success of the Large Area High Efficiency MCP-PMT Project. First, the raw materials and methods to produce large glass bulbs with high strength, high optical quality and extremely low radioactivity must be identified. Among many requirements, the glass bulbs must be able to work in pure water under high pressure. The technology for vacuum sealing between the glass and the Kovar metal leads must be developed. Technologies required for making large area high QE photocathodes are also needed. The alkali metals and Antimony sources and thin film deposition technologies necessary for photocathode fabrication also require development work. Fabricating MCP pairs as electron multipliers with low noise and gain exceeding 10$^{7}$ is very complex and requires a lot of development work. Other key technologies include electro-optic, mechanic and electric designs for MCP-PMTs and the testing of such PMTs. At the time when the collaboration was started, most of the required technologies did not exist in China. After three years of collective work, 8 inch and 20 inch MCP-PMT prototypes have been successfully made and evaluated. \subsubsection{Basic Concept of Large Area MCP-PMT} 1) Effect of combining the transmission and reflection photocathodes Conventional wisdom says that the QE of a reflective photocathode, that can be made thicker, has much higher QE than that of a transmission photocathode. Coating the lower half hemisphere of a spherical PMT bulb to convert the light that has passed through the transmission photocathode on the top hemisphere into photoelectrons can in principle improve the QE of the device. Scientists at XIOPM have done theoretical analysis and computer simulation, showing that by optimizing the thicknesses of transmission and reflection photocathodes, the QE of the spherical PMT can indeed be improved. Such improvement has also been seen in prototypes of MCP-PMTs. 2) Electro-optical Simulation The shapes of photocathodes of the focusing electrodes required for effective photo-electron collections, and of the other structures inside the PMT bulb, require careful electro-optical calculations. Commercial computer software specially written for this purpose are extensively used by XIOPM and NORINCO to optimize the MCP-PMT design both for 8 inch and 20 inch MCP-PMTs. An example of the computer simulation results is given in Fig.~\ref{fig:Fig.5.4}. The simulation includes the photocathode, MCP assembly, its support structure and metal leads, focusing electrodes and the Earth's magnetic field. Electric potentials for various electrodes were optimized. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.4.png} \caption[MCP-PMT electro-optic simulation with Earth magnetic field]{\label{fig:Fig.5.4} MCP-PMT electro-optic simulation with Earth magnetic field } \end{center} \end{figure} 3) MCP computer simulation XIOPM has performed computer simulations for photo-electron collection and amplification in the microchannels of MCPs. Electrode arrangement, gap size between two MCP plates, size of microchannels and inclination angles were studied. A diagram of such simulation is given in Fig.~\ref{fig:Fig.5.5}. The MCP assembly has several advantages such as compactness insensitivity of the electron amplification to the Earth's magnetic field, high dynamic range, etc. Additional concerns, including the photoelectron collection efficiency and lifetime, etc. still need to be addressed. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.5.png} \caption[Simulation for MCP functionality]{\label{fig:Fig.5.5} Simulation for MCP functionality } \end{center} \end{figure} \subsubsection{MCP and MCP Assembly} NORINCO has decades of extensive experience in MCP production and development. The company has organized a special group to develop the MCP and MCP assembly required for the MCP-PMT project. Large capital and manpower investments have been made. MCP parameters and processing procedures have been adjusted. High gain MCP assemblies with gain exceeding 10$^{7}$ for a pair of MCPs have been developed after the design of the MCP assembly was optimized. This level of gain is critical for single photoelectron detection. \subsubsection{Glass Tube and Raw Materials} Our 8 inch MCP-PMTs use the traditional electro-vacuum glass, whose coefficient of thermal expansion (CTE) matches the CTE of Kovar, which is used for vacuum feedthrough pins. The vacuum seal is easy to make. However, the electro-vacuum glass is difficult to blow into 20 inch glass tubes due to its poor mechanical properties above the softening point. Also there are other issues that make the electro-vacuum glass unsuitable for 20 inch PMT tubes even if it can be used to make 20 inch glass tubes. Its mechanical strength is not good enough to withstanding the high hydraulic pressure, while oxides of alkaline metal in the glass can slow leach out and affect the optical transparency of the glass. Therefore, the Pyrex type of borosilicate glass that has the required mechanical strength and chemical stability is preferred for making 20 inch glass tubes. Unfortunately, the CTE of Pyrex glass does not match the CTE of Kovar so a section of glass tube that joins the Perex to electro-vacuum glass needs to be made by using glasses with different CTEs. Techniques and glass materials must be developed for making this multiple transition glass tube section. The aforementioned technical difficulties have been solved by the collaborative effort of our collaboration and industrial partners. The resulting 20 inch glass bulb has a good appearance and excellent quality. Its mechanical accuracy and thickness uniformity are also excellent. Some completed glass bulbs, including the Pyrex sphere and neck, multiple transient section and Kovar seal, as shown in Fig.~\ref{fig:Fig.5.6}, have been tested in a pressure test set-up with pressure up to 10 bars. Since the neutrino event rate in JUNO is extremely low, radioactivity of the PMTs for JUNO must be tightly controlled. The radiative decay of radioactive materials in PMT can potentially be a major source of background for a low energy neutrino experiment. The radioactive elements $^{238}$U, $^{232}$TH, and $^{40}$K in the glass are the major contributors of PMT radiation backgrounds. The glass of the Hamamatsu 8 inch R5912 used in Daya bay detector has rather high radioactivity level, which is acceptable for a relatively high rate experiment like Daya bay, but far exceeds the level that can be tolerated by JUNO. To control the radioactive level in the PMT glass tubes, care must be taken to select the raw materials and also to control the glass melting process preventing the radioactive materials contained in the crucibles to be dissolved into the molten glass. The main component of Pyrex glass is SiO2. We have identified sources of high purity quartz sand in China and indeed the measured $^{238}$U, $^{232}$TH, and $^{40}$K levels are satisfactory for JUNO. Low radioactivity raw materials for other ingredients of the Pyrex glass have also been identified. Costs of high purity raw materials are significantly higher than those of materials commonly used to make Pyrex glassware. However, the cost of raw materials for glass, even the high grade materials used, are still an insignificant part of the total PMT production cost. Although we believe we know how to do it, so far low radioactivity glass tubes have not yet been manufactured. The 20 inch glass tubes for our prototype PMTs have been made with regular materials and using regular procedures in the factory. This is because even just to make a test run for low radioactivity glass requires the glass factory to stop their regular production for an extended period to purge their ovens, and the test would cost a significant amount of money. We will do a low radioactive test production when we are ready to make a small batch of 20 inch PMTs. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.6.png} \caption[Vacuum sealed 20inch glass bulb on a CMM for mechanical accuracy measurements]{\label{fig:Fig.5.6} Vacuum sealed 20inch glass bulb on a CMM for mechanical accuracy measurements } \end{center} \end{figure} \subsubsection{Photocathode} The Photocathode is a critical component of our PMT R\&D project. Typically to make bialkali photocathode, a very thin Antimony layer is first deposited on the inner surface of an evacuated glass tube, then layers of the alkali metals Potassium and Cesium are deposited on top. Under high temperature a semiconductor layer of K$_{2}$CsSb photocathode is formed. Both XIOPM and NORINCO, on the basis of their previous experiences, have tried to make bialkali photocathode for 8 inch MCP-PMTs successfully. Recently NORINCO has also made a 20 inch photocathode with good results as shown in Fig.~\ref{fig:sec2QE}. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/QE.jpg} \caption[QE]{\label{fig:sec2QE} QE } \end{center} \end{figure} \section{Organization of the MCP-PMT R\&D Project} \subsection{Introduction} The required effort to develop a brand new product such as 20 inch MCP-PMT is very technically demanding, and the R\&D process is complex and time consuming both for skilled technicians and scientists. The cost of the R\&D, including the basic equipment and the investment for developing the various technologies involved, is quite high. IHEP (Beijing) would be unable to take on this task by itself. The approach we have taken is to organize a concerted effort by factories that have experience of producing PMTs and MCPs, as well as by research institutions and university groups that have strong theoretical and practical background in opto-electronic devices. This approach led to creating a collaboration able to exploit effectively the expertise of each individual member institution toward the final goal of designing and fabricating 20 inch PMTs that meet the requirements for JUNO. Unlike other electronics devices, PMTs often cannot be tested at every production stage; instead meaningful tests can only be made after all the production steps are completed and the tube is evacuated to high vacuum level. We must do as much as possible of the testing and Q\&A work before the mass PMT fabrication is started. Therefore, we divide our R\&D project into the following categories, as shown in Fig.~\ref{fig:Fig.5.7}. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.7.png} \caption[Flow chart of the large area MCP-PMT R\&D process]{\label{fig:Fig.5.7} Flow chats of the large area MCP-PMT R\&D process} \end{center} \end{figure} The development process includes many different aspects that must proceed in parallel as listed below. 1. Alkali sources: Develop and produce standardized alkali sources that ensure high QE and stable photocathode. 2. Glass: Develop 20 inch high quality and low radioactivity glass bulbs. 3. MCP: Develop and produce low noise, high gain MCPs and optimize the performance of MCP assemblies 4. Electronics: Optimize the signal propagation from the anode to readout electronics 5. R\&D to fabricate large high vacuum equipment for high throughput PMT production and testing . 6. Vacuum seal: Techniques for hot Indium seal and glass seal 7. Electro-optic: Design and performance simulation 8. Performance tests: Platform for 20 inch PMT testing that can be used in the PMT factory and other places. \subsection{Capability of the Collaboration} \subsubsection{IHEP (Beijing)} The Institute of High Energy Physics, Chinese Academy of Sciences, is the initiator and coordinator of the Large Area MCP-PMT Project. IHEP has extensive experience in nuclear and particle instrumentation, vacuum technology, PMT testing, and PMT applications in large particle physics experiments. The Large Area MCP-PMT Project is organized within the Time-of-Flight detector group, which is a subgroup of the State Key Laboratory of Particle Detection and Electronics at IHEP. Funded partly by the Major Research Instrumentation Initiatives of CAS, the IHEP group has built a PMT Performance Testing Lab. Many PMT parameters such as the single photoelectron (SPE) spectrum, absolute QE and photocathode sensitivity, dark current and dark pulse rate, linearity and gain curves, etc. can be measured quickly and accurately. Figure~\ref{fig:Fig.5.8} shows graphically the functionality of the PMT testing lab. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.8.png} \caption[ Organization of the existing PMT test lab at IHEP]{\label{fig:Fig.5.8} Organization of the existing PMT test lab at IHEP} \end{center} \end{figure} The PMT Performance Lab is equipped with a 10000 class clean room, a dark room with electromagnetic shielding, an optic testing system based on LEDs, a variable frequency laser, light flux calibration systems and associated optic components. Advance testing instrumentation, including pA meters and digital oscilloscopes has been installed. With a VME based automated PMT input surface scanning system and testing system for photo gain curves, the absolute photocathode sensitivity (accurate to 1~$\mu$A/lm), QE (relative uncertainty < 2\%) and single photon transit time spread TTS can be measured in the frequency range 230~nm - 1000~nm (accurate to 50~ps). The dark current can be measured to pA level in the range of 1~pA - 10~nA. This test system is completely automated and the HV, laser wavelength, data recording and data analysis can be controlled by computers. Moreover, a VME based automated single photoelectron (SPE) spectrum recording system has been built. The SPE testing usually requires the PMT gain to be set to gain > 10$^{7}$, but with a preamplifier the SPE spectrum can be acquired also at a reduced gain > 10$^5$. Fig.~\ref{fig:Fig.5.9} shows measurement results by such a system. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.9.png} \caption[PMT measurement results at IHEP]{\label{fig:Fig.5.9}PMT measurement results at IHEP} \end{center} \end{figure} In addition, IHEP has recently assembled a small group to perform fundamental studies on MCP optimization to prepare for the MCP-PMT production and to study optic thin films. An ultra low radioactivity analysis system based on a Canberra High-purity Germanium (HPGe) Detector with low background shielding and BGO veto counter has been built at IHEP. This system is used to study the radioactivity of PMT glass and raw material samples. \subsubsection{XIOPM} The XiAn Institute of Optics and Precision Mechanics (XIOPM) of CAS has extensive experience in research and fabrication of various opto-electronic devices and related technologies. With significant government funding, the State Key Laboratory of Transient Optics, Photonics and Optoelectronics Laboratory and the Photoelectric Measurement and Control Technology Research Department at XIOPM have recently upgraded laboratories with new clean rooms and new equipment, including three research platforms. 1. Platform for opto-electronic devices: Design software for opto-electronic devices, high precision assembly table, plasma surface cleaning, vacuum seal and welding machines, vacuum leak detection systems, photocathode activation station, photocathode testing station, and others. This platform provides basic tools for high sensitivity, high quality opto-electronic devices R\&D and fabrication. 2. Platform for high speed electronic fabrication and testing: Software for designing radiofrequency circuits, circuit fabrication facility. 3. Platform for opto-electronics testing: static and dynamic opto-electronics testing system, image intensify and fast DAQ system. The 8 inch MCP-PMT design and prototype fabrication were started at XIOPM in 2011, followed by the design and fabrication of the equipment needed for the 20 inch MCP-PMTs after some success in their 8 inch MCP-PMT program. In this process, they have performed extensive opto-electronics simulation, theoretical analysis for high QE photocathode, photocathode fabrication and measurements, MCP functionality studies and optimization. Part of their work is shown in Fig.~\ref{fig:Fig.5.10}. The 20 inch MCP-PMT fabrication equipment, both the non-photocathode transfer and photocathode transfer types, is now almost ready at XIOPM and they are essentially ready to produce some 20 inch MCP-PMT prototypes. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.10.png} \caption[MCP-PMT fabrication process at XIOPM]{\label{fig:Fig.5.10}MCP-PMT fabrication process at XIOPM} \end{center} \end{figure} \subsubsection{NORINCO Group} China North Industries Group Corporation (NORINCO GROUP) is one of five producers of image intensifier tubes based on MCPs. NORINCO has teams of engineers to design and fabricate special equipment for fabricating and testing vacuum opto-electronic devices. In particular it has more than 30 years experiences in R\&D, design, fabrication and testing for various photocathodes, MCP and MCP assemblies. Their original technology came from PHOTONIS in the Netherlands. After many years refinement, they are now able to routinely realize multialkali photocathodes in small image tubes with sensitivity exceeding 900~$\mu$A/lm, which is higher than that reached by any other manufacturer in the world. For the Large Area MCP-PMT R\&D Project, their expertise in bialkali photocathode is assisted by the expertise of No. 741 PMT factory. For the R\&D of Large Area MCP-PMT project, NORINCO GROUP has organized a team of experienced engineers, made available workshops, lab spaces and equipment, and invested a large amount of capital. Three of their platforms, some new and some existing with additional equipment, can be used for various MCP-PMT R\&D tasks. 1. MCP Research Platform: Glass ovens, glass fiber pulling machines, MCP slicing machines, automated etching machines, electrode deposition stations, MCP testing station. 2. Platform for low level opto-electronic devices: Software for opto-electronics design and simulation, instrument for high sensitivity photocurrent measurements, photocathode production station, including plasma cleaning, monitoring during photocathode fabrication, thin film deposition stations for high QE photocathodes, photcathode sensitivity testing station, light transmission and reflection testing station, HV power supply design and fabrication and testing, etc. 3. Platform for Large Area PMT R\&D and prototyping: Glass lathe, large annealing oven, RF welding machines, large ultrasound cleaning stations, clean baking oven, point welding machines, thin film deposition stations, 20 inch PMT fabrication station using non-vacuum transfer technology, 20 inch PMT fabrication station using vacuum transfer technology (shown in Fig.~\ref{fig:Fig.5.11}), photocathode activation monitoring station, PMT dynamic testing system, PMT static testing system, opto-electronics design software, etc. These platforms provide a solid foundation for large area MCP-PMT design, R\&D, prototype testing and industrialization. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.11.png} \caption[ Equipment for fabricating 20 inch MCP-PMTs developed by NORINCO using photocathode transfer technology]{\label{fig:Fig.5.11} Equipment for fabricating 20 inch MCP-PMTs developed by NORINCO using the photocathode transfer technology} \end{center} \end{figure} \section{R\&D Results and Plan} \subsection{Status of MCP-PMT R\&D} The main achievements up to June of 2014 made by the Large Area MCP-PMT Collaboration are summarized in Table~\ref{Achievementsandstatus}. \begin{table}[!hbp] \centering \caption{Achievements and status\label{Achievementsandstatus}} \begin{tabular}{|c|c|} \hline \hline Items & Status \\ \hline 8 inch MCP-PMT prototypes & Completed \\ \hline PMT testing system & 8 inch system completed at IHEP and XIOPM \\ & 20 inch system near completion at NORINCO \\ \hline 20 inch fabrication equipment & Near completion at XIOPM; Completion at NORINCO \\ \hline 20 inch glass bulb & Completed and met expectations \\ \hline 20 inch Prototypes & Three produced at NORINCO \\ \hline \end{tabular} \end{table} \subsubsection{8 inch MCP-PMT Prototype R\&D} The Large Area MCP-PMT collaboration had its effort focused on the 8 inch MCP-PMT prototypes in the first two years in order to develop the necessary technologies for the 20 inch MPC-PMT. The opto-electronic design and simulation, photocathode fabrication techniques, design of the MCP assembly, MCP scrubbing, electrode design, etc. have gradually become mature and many difficult problems have been solved. The highest QE achieved in 8 inch MCP-PMT slightly exceeds 30\%. The required gain 10$^{7}$ has been achieved and the dark current can be as low as 10~nA. The Peak/Valley (P/V) ratios of SPE spectra for prototype tubes have been in the range of 1.6 to 2.5, with one prototype tube reaching 3.8. QE and SPE spectra of the 8 inch MCP-PMT made at NORINCO and tested at IHEP are given in Fig.~\ref{fig:Fig.5.12}. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.12.png} \caption[QE values and SPE spectra of the 8 inch MCP-PMT]{\label{fig:Fig.5.12}QE values and SPE spectra of the 8 inch MCP-PMT} \end{center} \end{figure} \subsubsection{20 inch Glass Bulb and Transition Section} As discussed previously, we use the so-called Beijing hard glass (Chinese standard CG-17), that is equivalent to the Pyrex borosilicate glass, to make the 20 inch glass bulbs. The CTE of the CG-17 glass is very low at 0.0000033/\SI{}{\degreeCelsius}. It has very high mechanical strength and can withstand various strong acids, alkalis and pure water. Glass bulbs as large as 508 mm are hand-blown by skilled glass technicians. Very few people have such skills and experience in the country. The collaboration was able to identify a glass factory that was very cooperative, which successfully made more than 50 sample bulbs with very good quality. In particular, the achieved mechanical tolerance (< 0.6 mm) is outstanding. We have tried three methods to make the glass transition section match the low CTE of the Pyrex glass with the slightly higher CTE of the glass used to make the vacuum seal with the Kovar pins and the end flanges. After being sealed and annealed, the evacuated glass bulb was placed in a test container filled with water at 1~MPa pressure (equivalent to 100 meters of water) and observed for 24 hours without any problem. Photos of a 20 inch glass bulb, its visual defect inspection and transition sections are shown in Fig.~\ref{fig:Fig.5.13}. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.13.png} \caption[20 inch glass bulb, its defect inspection and transition sections]{\label{fig:Fig.5.13}20 inch glass bulb, its defect inspection and transition sections} \end{center} \end{figure} 20 inch MCP-PMT prototypes were successfully made at the Nanjing factory of the NORINCO in June, 2014 by using the photocathode vacuum transfer equipment. After increasing the pumping capacity of the fabrication station, a QE level of 26\% has been reached. More detailed testing is in progress and we expect there will be further improvements when more prototypes will be made. In addition, prototypes of 20 inch MCP-PMT will be fabricated by using the non-vacuum transfer station and comparison will be made of the 20 inch prototypes made by both technologies. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.14.png} \caption[ 20 inch MCP-PMT Prototype]{\label{fig:Fig.5.14} 20 inch MCP-PMT Prototype} \end{center} \end{figure} \subsubsection{Improving the Testing System} The testing system for large MCP-PMTs with spherical photocathodes are different in many aspects from the system used for conventional PMTs. IHEP, XIOPM and NORINCO are all in the process to improve their current testing system. The existing testing system at IHEP is shown in Fig.~\ref{fig:Fig.5.15}. Photocathode sensitivity, QE, dark current, SPE spectra, gain, etc. can be measured. The automated scanning station can accommodate 8 inch PMTs. Test results using this system agree well with parameters provided by Hamamatsu and Photonis for their standard PMTs. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{PMT/figures/Fig.5.15.png} \caption[ PMT test system at IHEP]{\label{fig:Fig.5.15} PMT test system at IHEP} \end{center} \end{figure} \subsection{Status of Commercial PMTs} \subsubsection{Hamamatsu} In recent years, some PMT manufacturers have been developing so called SBA photocathode and try to use such photocathode in large size PMTs. Hamamatsu has made a new SBA 20 inch R3600-02 with Venetian Blind type dynode. Five such tubes have been recently tested. \cite{Nishimura2014Jan}. Hamamatsu is also developing a 20 inch PMT (R12860) with box and line type dynodes, expecting 93\% photoelectron collection efficiency compared to about 80\% for the R3600-02. The TTS will also be improved with the box and line type dynodes. The QE of the Hamamatsu 20 inch SBA is expected to be about 30\% (Private communication with Hamamatsu) at the sensitivity peak. The sensitivity peak of the current SBA photocathode is located at 390~nm, which matches the requirement of the HyperK water Cherenkov detector, but slight mismatchs the scintillation light, which is mainly emitted at 390~nm to 450~nm. Based on what we know now, we can expect that the PDE of Hamamatsu R12860 may reach 25\% for scintillation photons. Another problem of using theR12860 for JUNO is that the size of its effective photocathode is somewhat smaller than the 508 mm outside diameter , which would affect the global useful coverage. Hamamatsu has promised to deliver samples of the R12860 to IHEP in June 2014. \subsubsection{HZC Photonics} HZC Photonics in Hainan Province, China acquired PMT production line and related technologies, including the SBA technology, from PHOTONIS, France in 2011. HZC Photonics started their PMT production in late 2013 and now have a 12 inch PMT XP1807 available. Its stated photocathode sensitivity 60~\SI{}{\uA} /lm is rather low. Their preliminary product, the XP53B20 3 inch circular tube, has a very high peak photocathode sensitivity, listed as 160~\SI{}{\uA} /lm, but with rather low gain of $ 6.25 \times10^{5}$ . HZC Photonics have promised IHEP to produce samples of 20 inch SBA PMT for JUNO soon. \section{Risk Analysis} \subsection{Risk due to Low Photon Detection Efficiency} PMTs are critical components for JUNO. The current MCP-PMT R\&D plan and backup options carry certain risks that need to be considered. In order to achieve the required energy resolution of less than 3\% at 1 MeV, which is expected to be dominated by photon statistics, high efficiency liquid scintillator with a long attenuation length must be used to detect enough photons when neutrinos interact in the liquid scintillator. In addition, the fraction of the detector surface that is covered by PMT photocathodes must be as large as possible and the PDE of PMTs must be as high as possible. For a 508 mm diameter spherical PMT, assuming glass wall thickness of 4~mm, the maximum diameter of the photocathode would be 500~mm, or 97\% of the area of the PMT is actually covered by photocathode. Space needed for anti-implosion covers and gaps among PMTs futher reduce the maximum possible photocathode coverage. We also must make the QE of PMTs as large as possible. As stated earlier, the PDE of PMTs equals the QE of the photocathode multiplied by the photoelectron collection efficiency (PCE). The goal of the peak photocathode QE for the JUNO PMTs is 38\%, which is a very demanding goal. In Fig.~\ref{fig:Fig.5.16} (taken from Hamamatsu PMT handbook), the solid black curve is the typical cathode radiance sensitivity in unit of mA/W. The typical QE curve can be calculated from it. The computed QE as a function of wavelength is given as the dashed curve in Fig.~\ref{fig:Fig.5.16} and the scintillation light spectrum is plotted as the red solid curve (linear scale) with double peaks at 400~nm and 425~nm. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.16.png} \caption[Spectral response of common bialkali photocathode and the emission spectrum of liquid scintillator (red curve) ]{\label{fig:Fig.5.16}Spectral response of common bialkali photocathode and the emission spectrum of liquid scintillator (red curve) } \end{center} \end{figure} It can be seen from Fig.~\ref{fig:Fig.5.17} that the typical QE of regular bialkali photocathode is 20 - 25\% within the wavelength range 350~nm to 450~nm. The value of the peak QE typically is no more than 30\% in commercial PMTs. In recent years, Hamamatsu has developed SBA and UBA photocathodes with higher QE. The QEs as functions of wavelengths of these new photocathodes are given in Fig.~\ref{fig:Fig.5.17}. We will not discuss the UBA photocathode since the UBA is only available in small flat faced PMTs made by Hamamatsu. The peak QE of the SBA photocathode shown in Fig.~\ref{fig:Fig.5.16} is about 35\% and the peak sensitivity occurs at ~390~nm. There are two important issues to be considered. It can be seen that the spectral response curve of the Hamamatsu SBA is rather narrow and there is a slight mismatch between the SBA spectrum response peak at 390 nm and the scintillation light emission peak at ~ 425~nm. In Fig.~\ref{fig:Fig.5.17}, the black curve that is shifted slightly toward longer wavelengths indicates the desired spectrum response, which matches the spectrum of emitted scintillation light better. The second issue is that according to our conversation with Hamamatsu, the typical peak sensitivity they expect for the future 20 inch SBA PMT is 30\%, not the 35\% given in their published documents. The average QE for scintillation light is expected to be less than 30\%. With the stated 93\% photoelectron collection effiiency of the new SBA R12860, we speculate that its average PDE may be less than 25\% due to the spectral response mismatch discussed above. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.17.png} \caption[QEs vs. wavelengths for conventional bialkali, SBA and UBA photocathodes]{\label{fig:Fig.5.17}QEs vs. wavelengths for conventional bialkali, SBA and UBA photocathodes} \end{center} \end{figure} The specified peak QE for JUNO PMTs as listed in Table~\ref{tab:PMTsrequiredbyJUNO} is 38\% and we hope the average PCE of JUNO PMTs can be higher than 30\% in the wavelength range 400~nm to 450~nm. Therefore one of the critical risks is whether or not we can achieve this. Our approaches to limit this risk are the following. 1. Transmission + Reflection Photocathode Use the spherical photocathode, combining the power of transmission and reflection photocathodes. We hope the targeted peak QE 38\% and average QE > 30\% within 400 - 450~nm can be achieved by optimizing the photocathode design and fabrication process. 2. Transmission photocathode + reflection film Coat a layer of thin film with high reflectivity in the bottom half of the sphere to reflect photons that have penetrated the transmission photocathode back and convert these photons into photoelectrons. \subsection{Risk Related to MCP Aging} The operational lifetime of JUNO is 20 years and the working lifetime of the PMTs must also be greater than 20 years. It is expected that as long as the vacuum in the MCP-PMTs can be maintained, their lifetime will mainly be determined by the gain reduction due to MCP aging. It is expected that in the completely dark environment of JUNO, accumulated charges will mainly be caused by the photocathode thermal noise and the dark current of MCPs. At gain 10$^{7}$ and dark counting rate 10~kHz (equivalent to dark current 16~nA), the accumulated charge due to photocathode thermal emissions will be ~10~C for 20 years operation. The dark counting rate of a PMT is typically measured by counting dark pulses with a threshold of 0.25~pe. The true dark current, including the MCP dark current and dark current from other sources, can be much higher than the 16~nA used in the above estimations. Some of this additional dark current should be considered when the MCP aging effect is evaluated. At this level of charge accumulation, MCP aging can become a problem for MCP-PMT operation. The charge accumulation level of the MCPs developed by NORINCO can reach 30~C. In addition, PMT anode assembly typically has a gain of 3 - 5 and the charge accumulation is reduced by the same factor if such an anode design is used. This implies that the accumulated charge that the MCP-PMT can be operated with can reach the level of 100~C. In order to control the dark pulse rate and dark current due to MCPs, we are testing two different methods to fabricate the MCP-PMTs. The photocathode vacuum transfer method has separate vacuum chambers for photocathode fabrication and MCP processing and can in principle lower the risk of MCP contamination by alkali materials, thus achieving lower dark current. In the past few years, MCP manufacturers and research labs have developed a new type of MCP that use the ALD technique to deposit a thin secondary electron emission layer on top of a resistive layer on the wall of the microchannels of the MCPs. It has been shown that the lifetime of ALD coated MCPs can be an order of magnitude higher than the conventional MCPs. IHEP has started a R\&D program to develop MCPs using ALD technology. \subsection{Risk due to PMT Radiation Background} \subsubsection{Radiation Background Requirements} The PMT radiation backgrounds are dominated by the radioactive elements, most importantly $^{238}$U, $^{232}$Th and $^{40}$K in the glass bulb. In the current 20 inch spherical MCP-PMT design, the weight of the glass bulb is ~9~kg. According to the current detector physics simulation, the upper limits for $^{238}$U, $^{232}$Th and $^{40}$K contents in the PMT glass are 12.1 ppb, 26.1 ppb and 1.5 ppb, respectively. For comparison, the contents of these elements in Hamamatsu R5912 8 inch PMT glass are 153 ppb, 335 ppb and 18.5 ppb, while they are 540.3 ppb, 568.3 ppb and 75.2 ppb in the 20 inch R3600-02 PMTs. The radioactivity levels of Hamamatsu PMTs are many times higher than that allowed by JUNO. In a sample of regular CG-17 borosilicate glass, the type of glass to be used to fabricate the 20inch bulbs for MCP-PMTs, $^{238}$U, $^{232}$Th and $^{40}$K contents are measured to be $349.1 \pm 18.7$ ppb, $851.2 \pm 72.8$ ppb and $26\pm3.$1 ppb, which are 29, 32 and 17.5 times higher than the limits for JUNO PMTs. Clearly new glass materials with lower radioactive contents must be developed. In order to meet the required extra low radiation background requirements, we must use ultra-pure raw materials for making the glass and also control the possible contamination occurring in the material handling and glass melting process. Chemical composition of the GG-17 glass and of raw materials are listed in Table~\ref{tab:PMTglass}. \begin{table}[!hbp] \centering \caption{Chemical composition of the GG-17 glass and raw materials\label{tab:PMTglass}} \begin{tabular}{|c|c|c|c|c|c|} \hline \hline Items & Status \\ \hline Compostion & SiO2 & B2O3 & Na2O & Al2O3 & NaCl \\ \hline Content (\%) & 80.0 & 13.0 & 4.0 & 2.5 & 0.5 \\ \hline Raw materials & Quartz sand & Boric acid & Na2CO3 & AlOH & Salt \\ \hline \end{tabular} \end{table} We have acquired a large number of raw material samples from different manufacturers in different regions of China and measured contents of the three radioactive elements in these samples carefully. We have identified suppliers who can provide the necessary raw materials with $^{238}$U, $^{232}$Th and $^{40}$K contents that are low enough to meet JUNO PMT requirements. Obviously these high purity materials must be handled properly to prevent any contamination. We have also investigated the possibilities for contamination during the glass melting process. The most worrisome concern is that at high temperature radioactive elements can be introduced from crucibles and other debris into the molten glass in the oven. A Platinum crucible can mostly eliminate this problem, but it would be a very expensive solution. We believe corundum crucibles commonly used in Pyrex glass factories should be good enough based on the radiation background test for AlOH. So far, for the 50 or so 20 inch glass bulbs that have been made, regular raw materials were used and we know their radiation levels are very high. We have not attempted to make low radiation background PMT bulbs yet. \subsection{Risk due to Nonuniformity of MCP PCE} The photon detection efficiency (PDE) of PMTs can be expressed as $PDE=QE*PCE$, where QE is the photocathode quantum efficiency and PCE is the photoelectron collection efficiency of the electron multiplier. In our case the electron multiplier is the MCP assembly. With a spherical glass bulb and the photocathode vacuum transfer approach, we believe it is possible to fabricated completely uniform photocathode since we can place the evaporation alkali sources at the center of the sphere. To make the photoelectron collection completely uniform in our current MCP-PMT design is more difficult, since photoelectrons generated by a spherical photocathode arrive at the flat MCP input surface from different angles and they may have different efficiencies of collection and amplification. We are following several possible approaches to find remedies for this potential problem. 1. Optimize the focusing design to limit the incident angular spread of photoelectrons. 2. Optimize the MCP assembly design to add a metal mesh in front of MCP and coating the MCP with MgO or Al2O3 thin film, which has a high secondary electron emission coefficient, or use other schemes to improve the MCP photoelectron collection. Fig.~\ref{fig:Fig.5.18} shows the metal mesh and MCP coating schemes. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{PMT/figures/Fig.5.18.png} \caption[The MCP coating scheme]{\label{fig:Fig.5.18} The MCP coating scheme} \end{center} \end{figure} 3. Further increase the gain saturation level of the second MCP such that the output pulse heights are less dependent on the numbers of secondary electrons produced by the photoelectron striking the first MCP. \section{Schedule} The global schedule from prototyping to mass production is shown in Table~\ref{the schedule of the mass production}. \begin{table}[!hbp] \centering \caption{the schedule of the mass production\label{the schedule of the mass production}} \begin{tabular}{|c|c|c|} \hline \hline Year & Tasks & Parameters \\ \hline 2013 & 8inch prototype & QE $\ge$ 25\%, P/V $\ge$ 1.5 \\ \hline 2014 & 20 inch prototype & QE $\ge$ 30\%, P/V $\ge$ 2.0,\\ & & other parameters meet specs \\ \hline 2015 & Engineering design & QE $\ge$ 30\%, P/V $\ge$ 2.0, \\ & & other parameters meet specs and Yield $\ge$ 80\% \\ \hline 2016 & Preproduction & 1,000 20inch PMTs \\ \hline 2017 & Mass production & 5,000 20inch PMTs \\ \hline 2018 & Mass production & 5,000 20inch PMTs \\ \hline 2019 & Mass production & 5,000 20inch PMTs \\ \hline \end{tabular} \end{table} \chapter{PMT} \label{ch:PMT} \input{PMT/CH5.1.PMTDesigngoalsandspecifications.tex} \input{PMT/CH5.2.MCP-PMTRandDforJUNO.tex} \input{PMT/CH5.3.OrganizationofMCP-PMTRandDproject.tex} \input{PMT/CH5.4.RandDResultsandPlans.tex} \input{PMT/CH5.5.Riskanalysis.tex} \input{PMT/CH5.6.ScheduleandCost.tex} \chapter{Readout electronics and trigger} \label{ch:ReadoutElectronicsAndTrigger} The JUNO readout electronics system will have to cope with the signals of 17,000 PMTs of the central detector as well as 1,500 PMTs installed in the surrounding water pool. The average number of photoelectrons being registered by an individual PMT will reach from below one for low-energy events up to several thousands in case of showering muons or muon bundles. In both extreme cases, the number of photoelectrons has to be counted and their arrival time profiles have to be determined. A dedicated trigger system will be necessary to perform a pre-selection of correlated PMT hits caused by physical events from a sea of random dark noise hits. Based on the selected arrival time and pe pattern, software algorithms will reconstruct energy, type and position of the event vertex. Thus, the main task of the readout electronics is to receive and digitize the analog signals from the PMTs and transmit all the relevant information to storage without a significant loss of data quality. The basic concepts considered in the design of the system as well as a list of specifications are given in Sec.~\ref{sec:roe:specs}. The PMTs will be submerged in the water pool surrounding the acrylic sphere that contains the LS. Based on this layout, three different approaches for the realization of the read-out electronics are discussed: A mostly "dry scheme" for which individual cables carry the analog signals out of the water pool to an external electronics hut containing front-ends and digitization units. A "wet scheme" for which digitization happens at the PMT bases and the digital signals are bundled in submerged central underwater units before being passed on to an external online computer-farm. And finally, an intermediate readout scheme where a considerable part of the electronics is submerged inside the water pool but still accessible for replacements. The three schemes are presented in Secs.~\ref{sec:roe:dry}$-$\ref{sec:roe:wet}. Sec.~\ref{sec:roe:dev} will provide an overview of further R\&D efforts to be undertaken. A first time line for the realization of the system and the most important aspects to be dealt with is laid down in Sec.~\ref{sec:roe:rea}. Possible risks are summarized in Sec.~\ref{sec:roe:risk}. \section{Design considerations and specifications} \label{sec:roe:specs} The concept for the design of the electronics is driven by the following goals, which are listed here in order of their priority: \begin{enumerate} \item {\bf Energy reconstruction:} Optimization of the energy measurement of $\bar\nu_e$, especially at the lowest energies. A basic limitation on the energy resolution arises from the statistics of detected photoelectrons (pe). This limit must not be worsened significantly by the effect of electronics. At a photoelectron yield of 1100 pe per MeV, the achievable relative resolution is $\sim$3\,\% at 1\,MeV. \item {\bf Reconstruction of the photon arrival time pattern:} A precise measurement of the arrival times of the photons is required to reconstruct the position of the events. This is particularly important for the IBD events induced by $\bar\nu_e$'s because (a) the energy response of the detector will be position-dependent and (b) the formation of spatial coincidences between prompt positron and delayed neutron events will support background suppression. Moreover, track reconstruction for cosmic muons and especially for muon bundles is very sensitive to timing and will be an important tool for the veto of cosmogenic backgrounds. \item {\bf Low dead time and dynamic acquisition rate:} During usual data taking conditions, all neutrino events from reactor, solar or geological origin should be acquired without in-efficiency, i.e. with a minimum or zero dead time. In addition, readout electronics have to be able to cope with short episodes of extremely high trigger rate that might be caused by the neutrino burst from a nearby galactic Supernova. Dependent on the distance, thousands of neutrino events are expected within a period of 10 seconds. Therefore, the system has to be able to bear a dramatical increase in its acquisition rate for short periods without significant data loss or dead time. \end{enumerate} \subsection{Energy reconstruction} For the low-energy events, the deposited energy can be almost directly inferred from the overall amount of photons or pe detected by the PMTs. However, there are several ways of performing this integral measurement of collected pe (see below). Additional complications arise for high energy events, where the accuracy of determining the number of pe detected for a given PMT will depend on the dynamic range of the readout electronics. \subsubsection{Photoelectron counting} There are two basic methods to determine the visible energy created by an interaction in the detector. One can either integrate the charge created by the PMTs or one can try to identify individual photons and count them. The reconstruction is complicated by the spread in the arrival time of the photons, which is caused both by the fluorescence time of the LS, photon time-of-flight and photon scattering which all we contribute to delays in the order of several hundred nanoseconds\footnote{Note that once the vertex position is known, the delay from the time-of-flight can be approximately corrected offline. Based on this, most of the light will be concentrated towards the start of the event.}. An additional distortion of the time pattern will be introduced in case of cosmic muons where the light is generated along the extended particle track, causing a delay by the time-of-flight of the muon ($\leq$120\,ns). The trigger system will have to take into account this spread in arrival times when defining a threshold condition. \medskip\\ \noindent{\bf Charge Integration} is technically the simpler solution. During a certain time window the signal of the PMTs is integrated and the integrals are summed up into a global charge. This global charge is proportional to the visible energy of the event. For an online measurement the time window has to be in the order of 300\,ns from the arrival of the first photon at a given PMT. If time-of-flight effects are taken into account, e.g. for an off-center event close to the surface of the acrylic sphere or a cosmic muon, the acquisition gate has to be extended to 500\,ns. The length of the time window must be optimized between two aspects. With longer time windows more and more of the scintillation light is included, but the contribution from the electronics noise increases. Due to the electronics noise the potential of the baseline will vary around zero. Integrating such a baseline even without a signal will not give exactly zero. The contribution from the baseline noise will scatter around zero with a width of the distribution that increases with the square root of the integration time. There are two subversions of the charge integration. One can either sum all the PMTs or just those PMTs above a certain threshold (e.g.~0.25\,pe). Due to fluctuations in the amplification process of the PMTs (in the dynode structures of a conventional PMT or in the MCP) the charge produced by a single photoelectron varies. The variation is usually classified by the so-called peak-to-valley ratio of the PMT. Due to these fluctuations it can happen that the signal of a single pe is not visible above the baseline noise level. But the information is not completely lost. In the absence of electronics noise the sum of all PMTs will give a better estimate of the visible energy than just the sum of those PMTs above a certain threshold. For a real system one has to optimize the threshold. It largely depends on the noise of the system and on the average number of pe per PMT for the relevant events . For JUNO it is not yet clear what the best value is. For Daya Bay it turned out to be 0.25\,pe. Note that this charge integration threshold is not necessarily the same as the trigger threshold applied for self-triggering of a channel. \medskip\\ {\bf Photon Counting} is usually a self-triggered method. One analyzes the wave-form of each PMT individually in an attempt to identify the arrival of individual photons. Then the number of these photons is counted over at time window comparable to the above. The total number is proportional to the visible energy. The advantage of this method is that the number of photons is more directly related to the energy than the total charge. Some photoelectrons produce higher charges and some lower. This introduces an additional fluctuation into the charge measurement which is unrelated to the visible energy. With photon counting this additional fluctuation is eliminated. With the method of photon counting a threshold for each PMT is unavoidable. Photo electrons with very low charge due to fluctuations in the amplification process will be lost. This degrades the resolution by an amount that strongly depends on the peak-to-valley ratio of the PMTs. Whether photon counting is superior or inferior to charge integration depends on the details of the detector and the electronics. For JUNO the answer is not yet known. A detailed simulation is necessary to produce it. For Borexino photon counting turned out to be the preferred method at low energies and also DoubleChooz is using a method of approximate photon counting. For Daya Bay the electronics does not allow to do photon counting. To extract the number of photons from the waveform of a PMT, algorithms of different levels of sophistication are possible. In case of a software trigger, a very simple algorithm that counts the number of transitions of the photon pulse(s) across a threshold is probably sufficient. A more sophisticated algorithm will identify the charge of a pulse and its shape. Effectively at very low occupancies any pulse consistent with the charge of a single pe will be counted as one photon. At very high occupancies, the shape of the pulse will not contain any significant information. Therefore the number of photons will be derived from the integrated charge. The two basic methods become identical at very high occupancies. At intermediate occupancies, when the pulses of a few photons partially overlap in time, shape information like satellite peaks in the wave-form combined with the integrated charge will give the best answer \subsubsection{Dynamic range} The electronic system must be able to cope with a large range of signal heights (and also durations). For low-energy neutrino events, the average occupancy is below 1\,pe per PMT. Therefore the electronics must have good resolution for single pe's. Cosmic muon events are forming the high-energy end of the the dynamic range. These events can reach hit occupancies of 4000 pe for individual PMTs, and this has been defined as the maximum occupancy the electronics will have to cover. However, even larger signals are expected in case the muons arrive in bundles or create hadronic showers inside the target volume. \subsection{Photon timing} \label{sec:roe:timing} For all event types, the arrival time of the first photons at the individual PMTs will be the primary information available for the reconstruction of the vertex position (in case of neutrino events) or the particle track (in case of muons). This will be of particular importance for the energy reconstruction of antineutrino events because the effective pe yield will be position-dependent. A precise reconstruction of the vertex position will allow for low systematic uncertainties when applying correction terms. \subsubsection{Reconstruction of point-like events} Reconstruction algorithms will start from the hypothesis of a simultaneous emission of all photons from a single event vertex inside the detector. Then, the assumed vertex position will be adjusted to fit the observed photon arrival time pattern when considering the time-of-flight of the photons. The assumption of simultaneous emission is usually well fulfilled for the first photons detected for an event. More sophisticated algorithms will take into account the time delays introduced by the finite fluorescence times, photon scattering in the LS and the transit time spread of the PMTs by using appropriate probability density functions. Moreover, also the density of photon hits distributed over the surface of the detector sphere bears information on the vertex position. What is more, the precision of the photon timing will be decisive for the practicability of pulse shape discrimination. Heavy particles like $\alpha$'s or neutrons (proton recoils) create more light than electrons in the slow components of the fluorescence profile. Based on the information of the vertex position, individual hit times can be corrected for the time-of-flight of the photons. The resulting sum pulse of all detected pe corresponds to the original scintillation profile of the event (plus impact of light scattering). Profiles can be characterized for their long-lived decay component, allowing to quite clearly distinguish electrons from $\alpha$-particles and proton recoils and with lower efficiency even positrons from electrons. The quality of the separation will largely depend on the gate length over which the photoelectrons are acquired. The time window should cover a significant portion of the slow fluorescence component which is of the order of several 100\,ns. In this respect, a decentralized trigger schemes featuring a variable gate start time and gate width for each channel seems preferable. In this case, acquisition will stop only when the acquired pulse becomes indistinguishable from dark noise (see below). \subsubsection{Reconstruction of cosmic muons} Reconstruction of cosmic muon tracks is more complex. For single, minimum-ionizing muon tracks, the emitted light front will closely resemble a Cherenkov cone in the forward-running direction\footnote{This shape is due to the relative speeds of muon (speed of light in vacuum) and photons (speed of light in the medium) which results in the same geometrical condition as in case of the Cherenkov effect.}, plus a superposition of spherical back\-ward-running light fronts. Such events are reconstructed in a similar fashion as point-like events, relying on a fit of the muon time-of-flight plus the photon time-of-flight for a given track hypothesis to the arrival time pattern of first photons observed at the surface of PMT photocathodes. Also charge information from the many hundred photons per channel may be included as the collected number of photoelectron at a given PMT will depend on the distance from the muon track. However, showering muons as well as muon bundles create much more complex event topologies that lead in effect to a superposition of several light fronts arriving at the photocathodes with short time-delays. First attempts at the reconstruction of such events suggest that the full time profile of photon arrival times have to be recorded for each individual PMT in order to be able to discern secondary maxima or at least substructures that are created by the delayed arrival of subsequent light fronts. While a variety of algorithms are currently investigated, none has reached a state from which hard requirements on the necessary photon timing information can be formulated. Note that especially for these events, the ability for a separate detection of the true Cherenkov cone might provide valuable information for disentangling the contributions of individual particles. However, it seems unlikely that conventional or MCP-PMTs will reach the necessary time resolution. \subsubsection{Limitations in timing} While accurate timing information is for sure of great value, there are intrinsic limitations to the accuracy that can be reached in JUNO that are inherent to the LS detection technique. The first one arises from the finite time width of photon emission from the scintillator that is described by the fluorescence time profile. The fast component that is most relevant for timing questions features a decay time in the order of a few ns. This leads to a smearing of the timing information especially in the regime of low-photon statistics, i.e. at low energies. The second and more severe constraint arises from the transient time spread (tts) of the PMTs. The tts characterizes the variation of the time between the arrival of the photon and the creation of of the electrical pulse. As the final PMT is not yet available, the precise tts profile is not known. But the ($1\sigma$) uncertainty introduced is expected to be in the range of 3$-$4\,ns. Taking both inaccuracies into account, MC simulations of scintillation signals have been performed to obtain realistic PMT signals that were then "digitized" with different sampling rates. These studies indicate that a time resolution of 1\,ns resp.~a sampling rate of 1\,GS/s will be sufficient to record the signals without loss of information. \subsection{Triggering schemes} A further important consideration is the triggering scheme: It may either be global, i.e. based on the information of all channels, or de-centralized, i.e. based on self-triggering channels that send all their digitized signals (including all dark noise photoelectron) to a computing farm. In the latter case, a (online) software trigger is used to identify physical events. Both options are discussed more closely in the following. \medskip\\ In case of a {\bf Global Trigger}, the information of all PMTs is combined in order to form a global triggering decision that starts the readout of all PMT signals. The information from a limited local group of PMTs will not be sufficient to form the trigger decision. Information from all of the PMTs is required. Monte Carlo studies have shown that a simple trigger logic based on the total number of active PMTs (PMTs with at least one hit) within a 300\,ns window is sufficient to suppress the background from random coincidences of dark noise and to guarantee 100\,\% efficiency for the detection of $\bar\nu_e$ events. Threshold would be set to 500 hits ($\sim0.5\,$MeV) which is well above the average number of 225 dark noise hits within the same time span\footnote{This number is based on a trigger coincidence window of 300\,ns and 17,000 PMTs featuring a dark noise rate of 50\,kHz each.}. Such a trigger can easily be realized in the more conventional scheme of the external readout electronics (sec.~\ref{sec:roe:dry}) that relies on the collection of the analog signals in a physically confined space. In this case, the delay introduced by the decision for triggering reaching the digitization units is fairly small. However, a global trigger is much harder to apply for a de-centralized digitization as it is propagated by the underwater electronics scheme (sec.~\ref{sec:roe:wet}). Hit information from the spread-out PMTs would have to be transmitted either by individual cables/fibers or by combining the information from a group of PMTs underwater into one cable/fiber per group. The trigger decision would then be sent back along the same network to the PMTs to start the readout. During the latency of the trigger (several ms), the wave forms of the PMTs would have to be buffered either on the PMTs or in the underwater units. An important aspect regarding energy reconstruction that potentially favors a global trigger is the fact that this is the only configuration that will allow the recording of waveforms for channels in which the PMT signals are below an individual triggering threshold, i.e. the detection of low-charge photoelectron. On the other hand, the existence of a global trigger threshold at $\sim$0.5\,MeV poses a special challenge for data acquisition during a Supernova. Besides a large number of $\bar\nu_e$ events, a substantial amount of the resulting neutrino interactions in JUNO will be from elastic scattering of protons. Due to reaction kinematics and quenching, the visible light will be fairly low and necessitate a temporal lowering of the threshold below 500\,pe. While this does not pose a problem from point of view of radioactive background (there might be an interference with dark noise, though), the only information available to make the system aware of the on-going neutrino burst is the sudden increase in event rate. It will take a certain number of triggers to realize that neutrinos from a nearby supernova are arriving. Then the threshold will have to be lowered in retrospective. This is not impossible with an appropriate amount of buffers, but it does impose a significant complication of the trigger scheme. \medskip\\ \noindent{\bf Decentralized triggering.} In principle this "untriggered" scheme is much simpler. The signal from each PMT is continuously monitored. If it exceeds the acquisition threshold for a single photoelectron (SPE), i.e. probably around 0.25 of the single photoelectron charge, the signal is digitized and transmitted to the online farm. On the online farm the signals from the individual PMTs are combined into events and a decision is formed whether to store the event or to drop it. It is mainly a question of data reduction and storage capacity to decide which fraction of events or even individual PMT waveforms to store. The advantage is that much more information is available for the decision on storage of the event. Time-of-flight information can be taken into account to associate true photon hits inside a 300\,ns window to the emission by a single vertex. This potent mean for the suppression of background from random dark noise coincidences will allow to lower the threshold below 0.5\,MeV during normal data taking. For the underwater readout electronics, this scheme bears the considerable advantage of reducing the required cabling. There is no need for a network to send trigger information to the central trigger and to return the trigger decision. The disadvantage is a higher load on the data links to the farm and on the required computing power. In this scheme not only the waveforms of signal but also background events have to be transferred to the farm in order to form a trigger decision. The difference in the data transmission rates between the two trigger schemes is somewhat reduced by the length of the time windows required for the readout. For the global trigger scheme it must cover the full events ($\geq300$\,ns) for all PMTs while in the untriggered scheme the gate width can be adjusted to the number of photoelectrons detected, e.g. 32\,ns per single photoelectron pulse. If for a given event multiple photoelectrons are detected by a single PMT, several internal triggers will occur and several such windows will be transferred. Also the timing between the PMTs is simpler in the untriggered scheme. There is no need to synchronize the clocks between the PMTs. $t_0$-corrections can be applied in the farm. A White Rabbit system as proposed for the global triggering scheme is not necessary. Only the synchronization of the FADC frequencies has to be ensured, which will be achieved by locking the internal oscillators by a centrally distributed clock. Last, but not least one should mention that a software algorithm for data reduction on a farm is much more flexible to react to unforeseen complications or opportunities than a hardware-based trigger. \subsection{List of specifications} Based on the above discussion, the following list of specifications has been compiled: \begin{itemize} \item {\bf Waveform sampling} should be available over the whole energy range with a sampling rate of 1\,GS/s. \item {\bf Photoelectron resolution:} In the signal range from 1$-$100\,pe, the charge resolution should increase linearly from 0.1$-$1\,pe.\\ In the background range of 100$-$4000\,pe, the charge resolution should be 1\,\%.\\ Signal range and background range should overlap. \item {\bf Dynamic range} should reach from 1 to 4000\,pe per channel. \item {\bf Arrival time resolution}, e.g. by fitting of the signal leading edge, should be $\sigma_t\approx100ps$. \item {\bf Noise level} should be below 0.1\,pe for single photoelectron detection. \item {\bf Maximum acquisition rate} should be on the order of $10^3-10^4$ triggers per second. \item Acquisition should be low in or without any dead time. \end{itemize} \noindent Beyond these performance requirement, it has to be ensured that the system can be reliably and continuously operated for at least 10 years with no or minimal access to the parts of the readout system submerged inside the water pool. It is also important to facilitate the installation by minimize the cabling between the PMT matrix and the outside electronics. \section{External readout scheme} \label{sec:roe:dry} The philosophy of the external read-out scheme foresees to place as large a part of the electronics as possible in a dedicated electronics area or hut outside the detector. In particular, the analog-to-digital converters are to be located outside the water pool. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{ReadoutElectronicsAndTrigger/figures/OutwaterScheme.jpg} \caption[Scheme for external readout electronics]{\label{fig:out water scheme}Schematic drawing of the external readout scheme.} \end{center} \end{figure} \subsection{System layout} The general layout is shown in Fig.~\ref{fig:out water scheme}: The electronics attached directly to the PMTs is kept to a minimum. At the base of the PMT, the signal is decoupled from high voltage and increased by a pre-amplifier. From there, the analog signal travels along a 100\,m long cable to the input at the external front-end electronics (FEE). These will be arrayed in crates of the ATCA chassis specification. Each crate will hold ten FEE modules and two trigger modules. Each FEE module will provide for 32 channels backend amplification, signal distribution and processing, analog-to-digital conversion and data processing, as well as signal processing for the trigger system. The FEE modules will send the digitized data to the trigger modules via the backplane of the ATCA crate. These will create the trigger information that will be sent to the central trigger crate by fiber links. In case of a positive trigger decision, the trigger signal will be sent back to each signal processing crate and then fanned out to each FEE module. Each FEE module buffers the data and sends the necessary data asynchronously to the DAQ. Trigger and slow control system will follow the White Rabbit protocol to transfer the clock and other service signals to each crate. The clock signal will be distributed together with the DAQ and slow control through an optical fiber network. This has several advantages: The simple structure of the external scheme allows for an easy implementation. A global trigger can be easily formed. Maybe most importantly, broken components of the FEE and trigger logic are easy to access for repairs or replacement, and an upgrade of the system is easy to realize during the 10 years or more of detector operation time. The main disadvantage is the transmission of the analog signals from each individual PMT over the 100\,m long cables. While signal attenuation can be compensated by placing the preamplifier at the PMT, some shaping of the leading edge of the analog pulses will be unavoidable. In addition, the sheer amount of underwater cables complicates the installation and is a severe cost driver. \subsection{Signal amplification and digitization} In the external electronics scheme, pre-amplification will be performed directly at the PMTs and a second amplification stage is located directly after the input of the FEE modules. The bandwidth of the amplification circuit has to be chosen appropriately to minimize information loss by shaping. From there, as displayed in Fig.~\ref{fig:Amplification}, the signal will be divided into two paths: One is fed to a discriminator in order to provide a trigger flag, the other is sent to an FADC for charge measurement. Due to the large dynamic range of the PMT signals (1\,pe to $\geq$4,000\,pe), the use of a single FADC for the digitization seems hard to achieve while meeting the resolution requirements for low pulses. Therefore, the amplification will be divided into two or three paths with different amplification gains \cite{Qiuju}. In each path, signal amplitude will be adjusted to match the input range of the FADCs. For most of the fast, high-resolution FADCs \cite{AD9234} the maximum range of the input voltage is only 0.5$-$1\,V. Furthermore, minimization of the input electronics noise is quite important to keep up the specified voltage resolution. This can be achieved both by design and by filtering. The FADCs will be operated in differential mode to increase the performance. The foreseen sampling rate is 1\,GS/s, the required voltage resolution 14 bit. There are two possibilities: Either to purchase commercial chips or to design a customized ASIC. In the first case, the ADS5409 from TI and AD9680 from Analog Device would fulfill the requirements. If relying on a self-designed ASIC, the most practicable approach is a multi-channel time-interleaved architecture which will feature a low power consumption compared to a realization based on a single channel. For each channel, a 12\,bit, 500\,MS/s pipeline quantizer (or a 12\,bit, 250\,MS/s SAR quantizer) will be used for meeting a good balance between power, speed and accuracy. Offsets and gain mismatches between channels can be calibrated offline. A differential clock input will be used to control all internal conversion cycles. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/amplification.png} \caption[Scheme for signal amplification and digitization]{\label{fig:Amplification}Scheme for signal amplification and digitization.} \end{center} \end{figure} \subsection{Data processing and transmission} In a first step the FEE system will process the data and provide some simple information to the trigger system. Only when a trigger decision is formed and a corresponding signal is sent back to the crates, the full data (including waveforms) of all channels will be read out, bundled to a combined output in units of modules and crates and then transferred to the DAQ. There are 10 FEE modules plus 2 trigger modules in each crate. Data is transmitted between the modules via the backplane BUS system of the crates. Communication is done by a dual-star structure. Each FEE module can receive the trigger signal from either of the two trigger modules via a pair of high-speed serial connections. The digitized data is transferred via ethernet connection to a switchboard and from there to the DAQ. A block diagram of the system is shown in Fig.~\ref{fig:DataProcessingAbove}. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{ReadoutElectronicsAndTrigger/figures/DataProcessingAbove.jpg} \caption[Data processing and transmission in the external scheme]{\label{fig:DataProcessingAbove}Block Diagram of the data processing and transmission of the external electronics scheme} \end{center} \end{figure} \subsection{Trigger system} The design goals for the trigger system are the following: \begin{itemize} \item The trigger efficiency for IBD events induced by reactor $\bar\nu_e$ should be high, i.e.~reach 99\,\% detection efficiency for an energy deposit of $\sim$0.5\,MeV, so well below the minimum energy deposit of a prompt positron. \item The system should be able to suppress detector-related backgrounds as PMT dark noise coincidences and surface radioactivity to reduce the overall event rate. \item The latency introduced for the processing of 17,000 PMTs should not exceed a few $\mu$s. \end{itemize} \subsubsection{Trigger rates and threshold} The expected trigger rates generated by neutrino signals and backgrounds were studied by simulations. Signal sources considered were IBD signals, residual radioactive background events in the LS, cosmic-ray muons, and the dark noise of PMTs. We assumed that the buffer was thick enough to shield the gamma-ray background from the PMTs and support structures. The details on signal and background assumed are summarized in Tab.~\ref{tab:trigger}: Radioactive background levels in the LS were estimated conservatively based on the conditions in the Daya Bay detectors \cite{Qiuju}. However, radiopurity requirements for JUNO are much stricter. If i.e.~levels from KamLAND \cite{K.Eguchi} are used, the rates will be lower by several orders of magnitude. Muon background was estimated with 1,500\,mwe of rock overburden, taking into account a coarse image of the three-dimensional surface topology. Muon-induced spallation isotopes (i.e.~$^9$Li etc.) were not considered. \begin{table}[htbp] \begin{center} \begin{tabular}{llc} \hline Signal or background & Calculation standard & Event rate before trigger \\ \hline IBD & 80\,/day & 80\,/day \\ Background in LS & ${}^{238}$U: 2.0$\cdot$10$^{-5}$\,ppb & 69\,Hz \\ & ${}^{232}$Th: 4.0$\cdot$10$^{-5}$\,ppb & 32\,Hz \\ Muons & 1,500\,mwe & 3\,Hz \\ PMT dark noise & 50\,kHz/PMT & 3.3\,MHz \\ \hline \end{tabular} \caption[Input for trigger rate MC]{\label{tab:trigger}Input for trigger rate simulations} \end{center} \end{table} For the parameters characterizing the detector, we assumed an LS light yield of 11,000 photons per MeV, an attenuation length of 20\,m, a PMT coverage of 78\,\%, and a quantum efficiency of 35\,\% \cite{Dayabay}. In the simplest approach, the trigger decision will be based on the hit multiplicity inside a coincidence window of 300\,ns (see above),~i.e. on the number of simultaneously occurring PMT hits. Fig.~\ref{fig:trigger coincidence} shows the event rate as a function of this number of coinciding PMT hits. Based on these spectra, it seems that a simple hit multiplicity trigger will return both sufficiently low trigger rates and high trigger efficiency for IBD events if the trigger threshold is placed at $\sim$500 coincident hits. This coincidence trigger scheme is technically easy to implement and will create only low latencies. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/TriggerCoincidence.jpg} \caption[Event spectrum for coinciding PMT hits]{\label{fig:trigger coincidence} Event rates as a function of the number of PMT hits coinciding in a 300\,ns trigger window.} \end{center} \end{figure} \subsubsection{Trigger setup} Each FEE will handle 32 PMT signals. The hit flags of ten FEEs inside a crate will be bundled by one Level 1 Trigger Collection Board (L1TCB). In turn the information of ten L1TCB will be fanned to one level-2 board (L2TCB). The Trigger Board (TB) itself will receive data from six L2TCBs. This will be sufficient to read out the trigger flags of 17,000 channels. The L1TCBs and L2TCBs will also work as a fan-out for re-transmitting the trigger decision from the TB to individual FEEs. \subsection{Clock system} The clock system for JUNO has two major functions. It will provide: \begin{itemize} \item a standard reference frequency to all FEE modules to support the waveform sampling, TDC measurement and process logic; \item an absolute timestamp for each event to correlate the events of the central and veto detectors and for comparison with other experiments. \end{itemize} Based on the experience gained in the Daya Bay experiment and following the latest developments, a distributed clock network based on the "White Rabbit" standard will be applied for JUNO. The "White Rabbit" (WR) system is a technology originally developed by CERN and GSI for the use at accelerators. It is based on IEEE1588 (PTP) including two further enhancements: The precise knowledge of the link delay and a possibility of clock synchronization over the physical layer with Synchronous Ethernet (SyncE). Applying WR in JUNO will offer the following advantages: \begin{itemize} \item The low-jitter recovery clock from SyncE provides a reference frequency for the electronics. \item Sub-ns clock phase alignment can be achieved among all nodes by phase measurement and phase compensation. \item The precise time protocol (PTP) synchronizes the timestamp among all nodes. \item An external Rubidium oscillator and GPS can be used for frequency and UTC reference. \end{itemize} The system will synchronize all electronics crates. Each crate will contain a custom-designed clock module to recover the reference frequency and the absolute timestamp. This will then be broadcasted to all FEE and trigger modules in the crate via the dedicated high-speed clock traces on the back-plane for ATCA crates and via front-panel fan-out cables for non-ATCA crates. \subsection{Power supply and electronics crates} Conventionally, high voltage for the PMTs would be generated by a dedicated HV supply outside the water pool. However, the preferred solution is the generation of the HV by a Cockroft-Walton type power supply either attached to the PMT base or in underwater units. In this case, these generators have to be fed with low or medium voltage cables from outside the tank. Solutions with voltages of 5\,V or 24\,V are conceivable, the former requiring a larger cable diameter to transfer the larger current without loss. In order to avoid the power loss by using a linear power supply, a DC/DC module should be used for high efficiency. In order to reduce the ripple, the power supply should be independent. The FEE and trigger modules will be designed according to the specifications of the ATCA standard. Fig.~\ref{fig:ATCA Crate} shows the ATCA crate that will house the modules. It provides a large bandwidth for digital transmission which allows to pass trigger information between FEE and trigger modules via the bus system of the crate. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/ATCACrate0.JPG} \caption[ATCA crate]{\label{fig:ATCA Crate}The ATCA crate} \end{center} \end{figure} \section{Intermediate readout scheme} \label{sec:roe:int} The primary disadvantage of the external electronics scheme (Sec.~\ref{sec:roe:dry}) is the length of the coaxial cables needed for the transmission of the analog PMT signals. This will lead to a shaping of the leading edge of the signals, deteriorating the time resolution. To avoid this problem, the intermediate scheme shown in Fig.~\ref{fig:under water scheme} places not only the pre-amplifier but most of the FE electronics including digitization in a water-tight box inside the tank, considerably shortening the signal transmission length for the analog signals. Each of these 500 submerged processing units will supply an array of 32 PMTs. From the box, only few cables carrying the digitized signals as well as trigger and clock information will run to the external electronics. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{ReadoutElectronicsAndTrigger/figures/UnderwaterScheme.jpg} \caption[Intermediate scheme for readout electronics]{\label{fig:under water scheme}Schematic drawing of the intermediate, partially submerged readout scheme.} \end{center} \end{figure} \subsection{Signal processing} In this partially submerged configuration, the analog PMT signal will be transmitted only over a short cable to the amplification stage included in the underwater processing units (cf.~figure\ref{fig:Amplification}). Therefore, the signal shape will not change significantly. Otherwise, the design of the FEE boards inside the processing units will be mostly analogous to the external scheme. The digitized signals will be bundled for all channels connected to one box and passed by an optical fibers to external modules collecting the acquired data. From there, data will be passed on via ethernet. The block diagram is shown in figure\ref{fig:DataProcessingUnder}. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/DataProcessingUnder.jpg} \caption[Readout Electronics]{\label{fig:DataProcessingUnder} Block Diagram of the data processing and transmission of the under-water scheme} \end{center} \end{figure} \subsection{Trigger and clock systems} To simplify the design and maintenance of the trigger system, the trigger modules will be located with the other external electronics. The trigger flags from the FEE discriminators as well as the trigger signals can in principle be transmitted on the same optical fibers as the data. It is possible to transmit the same optical cable the data. However, separate cables might be used for the trigger system in order to increase the flexibility. As for the external scheme, a White Rabbit system can be applied for synchronization of the FEEs in the watertight boxes, providing both a reference frequency for FADCs and an absolute timestamp. In case the optical fibers required by White Rabbit are not compatible to oil, a dedicated clock and data recovery (CDR) link could be designed to provide the reference frequency. The physical calibration system could be exploited to determine the absolute timestamps of events relative to calibration pulses fed in from outside for which GPS time would be known. \subsection{Submerged processing units} An important aspect for the long-term stability of the system is the possibility to replace the part of the electronics submerged in the water, i.e.~the processing units. The structure of a readout strand including one processing unit is shown in Figure\ref{fig:underwater changeable system}. The enclosing box will be 30\,cm high, 40\,cm wide and 60\,cm long. There will be mounting holes and the possibility to remove the top lid. Fig.~\ref{fig:watertight case appearance} shows a three-dimensional representation, Fig.~\ref{fig:watertight case} displays the side panels: The left panel of each watertight case is connected to coaxial-cables by 32 connectors, while the right panel is connected to a composite cable including an optical fibre for signal transmission and a power cable for HV generation inside the underwater box. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{ReadoutElectronicsAndTrigger/figures/UnderwaterChangeableSystem.jpg} \caption[Representation of the intermediate scheme]{\label{fig:underwater changeable system}Structural representation of the readout chain in the intermediate scheme.} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/WatertightCaseAppearance.jpg} \caption[Readout Electronics]{\label{fig:watertight case appearance}The diagram of the watertight case appearance} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/WatertightCase.jpg} \caption[Side panels of submerged processing units]{\label{fig:watertight case}Diagrams of the side panels of the submerged processing units} \end{center} \end{figure} The further components shown in Fig.~\ref{fig:underwater changeable system} have to fulfill a number of requirements: \begin{itemize} \item Cable $Aa$ should be coaxial, 10$-$20 meters long, and with a characteristic resistance of 50\,$\Omega$. \item Connector $A$ is watertight. \item Connector $a$ is watertight and has to be pluggable under water. \item Cable $Bb$ is a composite cable of 100\,m length containing at least 4 optical fibres and 4 power wires (24-48V, 10-20A) usable for underwater environments. \item Connector $B$ is watertight and does not have to be pluggable under water. \item No special requirements are set for connector $b$. \item Parts $B$, $Bb$ and $b$ can be replaced by non-watertight cables covered by a corrugated pipe. \end{itemize} \section{Underwater readout scheme} \label{sec:roe:wet} The underwater concept places almost the entire readout electronics inside the water pool, as close as possible to the PMTs themselves. Front-end electronics (FEE) and HV generation are done at the PMT base, all data transfer is done based on digitized data and bundled in central underwater units (CUU). Thus, the underwater scheme minimizes the deteriorating effects of analog signal transmission over long cables and the overall amount of cables needed. On the other hand, it seems unlikely that broken FEE channels can be replaced during operation, while CUUs probably can. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/Scheme2.png} \caption[Readout Electronics]{\label{fig:Scheme 2}Schematic drawing of the underwater readout scheme.} \end{center} \end{figure} \subsection{Overview} The layout of the fully-submerged readout scheme is sketched in Fig.~\ref{fig:Scheme 2}. At the end of each PMT there is a small water-tight housing that contains the front-end electronics (FEE) with all the essential electronics of the system. It contains the base of the PMT and a module that generates the HV from a low-voltage AC input. And it contains a PC-board with a system on a chip (SoC) that includes the digitization of the signal, the discriminator to detect photoelectrons, and other functionalities. The FEE is connected to the online computer farm through data links. It transmits the data to the farm and receives commands from it. It also receives a clock signal from a central clock unit and the power for the PCB and the HV from corresponding power supplies. In order to reduce the cabling the PMTs are grouped together under water. We estimate that up to 128 PMTs might be grouped together. In this case the full detector would consist of 128 groups. The grouping of the PMTs takes place in central underwater units (CUU) mounted between the backend of the PMTs. To reduce the length of the cables from the PMTs to the CUUs the groups will be formed from nearby PMTs. The cables from the PMTs to the CUUs will have a unique length somewhere between 5 and 10 meters. They contain a by-directional data link, the clock signal, the low voltage power for the PCB (probably 3V), a 24 V AC power for the generation of the HV and a ground potential. On the PMT end the cables will be glued into the housing of the FEE. On the end of the CUU there will be a water-tight connector to allow for the mounting. From the CUU there will be a by-directional data link to the farm (probably two for redundancy) and power cables and the connection to the clock on top of the detector. \subsubsection{The `intelligent' PMT} With the FEE the PMT is turned into an `intelligent' PMT that can be connected directly to a computer. No further electronics is needed. Besides the module that generates the HV and contains also the voltage divider to derive all potentials there is a PCB in the FEE with an ASIC. It contains the SoC with the following functionalities: \begin{itemize} \item Amplification of the signal. \item Shaping of the signal, should it be necessary. \item Regulation of the baseline to zero. \item Detection of single photon signals through a leading-edge discriminator. \item Generation of a time-stamp for each readout window. \item Digitization of the input signal with an FADC. \item If necessary, a TDC to measure the precise start of each pulse (relative to the time-stamp of the readout window. \item Feature extraction, i.e. determination of the number of photons contributing to a pulse and their arrival-time. \item Self-calibration, i.e.~adjustment of the charge integral of single photons to a predefined value through a regulation of the HV and the amplification of the input stage. \item Creation of test pulses and injection into the input of the chip and into the PMT. \end{itemize} These functionalities are provided for each of the two MCPs of one PMT individually (2 channels). The following additional functionalities are common to both channels: \begin{itemize} \item Creation of an optical test pulse through an LED on the PCB that illuminates the photo cathode. \item Creation of an internal clock signal with the appropriate frequency locked to the input of the external clock. \item Internal functionality tests. \item Data exchange with the CUU. \item Limited data buffer. \end{itemize} A sketch of the SoC components is displayed in Fig.~\ref{fig:SoC}. Its functionalities are described in more detail in Sec.~\ref{sec:roe:ipmt}. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{ReadoutElectronicsAndTrigger/figures/SoC.png} \caption[Readout Electronics]{\label{fig:SoC}The SoC scheme} \end{center} \end{figure} \subsubsection{Central Underwater Units} The CUU functions as the connection between the PMTs of the assorted array to the infrastructure on top of the detector. It receives the data from the PMTs. It combines the data and sends it by a single optical link to the computer farm. It must have a sufficient data buffer to compensate fluctuations in the data rate. At the moment it is envisioned to use a commercial PC-board for the CUU. For the data-links to the PMTs the USB-2 standard might be appropriate. Commercial USB-routers could combine the data from the PMTs into a single data-stream which is then sent to the farm through an optical 10\,GBit/s link. If this works out, no development is necessary for the hardware of the CUU. The CPU power of the CUU could be used for monitoring, pre-sorting of the data, data reduction, or other purposes. The most critical parts of the CUU are the many connections attached to the CUU in an under-water environment (at least one connector per PMT plus the up-links). The distribution of the power and the clock to the PMTs will also be done inside the housing of the CUU, so that only one combined cable with connector is necessary for each PMT. For the uplink this is less critical. One could use a power cable separate from the fiber(s). A substantial amount of monitoring will be necessary to control the detector. The CUUs are also the proper place where the data can be injected into the data stream. Sensors for temperatures, voltages, scintillator flows, can be mounted in or near the CUUs and connected to dedicated inputs of the PC. Monitoring data on the PMTs will be collected by the SoC in the FEE and transmitted through the data link to the CUU. An interesting alternative to the routing of the data as described above is a passive routing of many data streams through one fiber. Such systems transmit each data stream with light of a different color. They are commercially available. They are potentially more reliable than a CUU based on a PC in a water-tight housing. In this case we would need a connector on the PMT to connect its fiber and a separate power cable to connect to a power distribution box under water. \subsubsection{The Online Computer-Farm} The online computer farm will receive the data through 128 optical links with 10\,GBit/s capacity. The data will arrive asynchronous. First is has to be sorted into proper time order. Then events are created as time-slices with a minimum activity of the PMTs. These proto-events are subjected to a reconstruction algorithm and from there the decision is derived whether to keep the event or to drop it (or to keep a reduced amount of data). After another sorting in time the accepted events are permanently stored. More details need to be worked out by the DAQ group. \subsubsection{Expected data rates and buffer sizes} As PMT pulses are converted early into digitized data, the amount of data to be transferred via fibers and data links as well as local buffers for temporary storage are important characteristics in this scheme. An estimate of the data volume to be processed starts with the amount of digital data caused by the detection of a single photoelectron (SPE). We assume a sampling rate of 500 MHz with one out of three 7-bit FADCs. With a typical rise time of 5\,ns and a fall time not much longer a time window of 32\,ns per spe is quite conservative. With the MCP-PMTs we will always digitize both channels at the same time. The data volume per spe is summarized in the table\ref{tab:Data Volume}: A single waveform is recorded with 144\,bits. For two channels and with the auxiliary information this adds up to 360 bits per photoelectron. For the potential TDC we have assumed 8\,bits with 125\,ps per unit to cover the full length of the readout window. \begin{table}[htbp] \begin{center} \begin{tabular}{lcl} \hline {\bf Data volume of a single waveform}\\ sampling frequency & 500\,MHz \\ readout window & 32\,ns \\ number of time-slices per readout & 16 \\ number of bits & 7 \\ range bits & 2 \\ number of channels & 2 \\ \hline Intermediate sum & 288\,bit \\ \hline {\bf Additional information} \\ address (PMT-ID within group, channel) & 8 \\ time stamp & 40 \\ control flags & 8 \\ TDC (2 x 8 bit) & 16 \\ \hline {\bf Total sum} & 360 bit \\ \hline \end{tabular} \caption{\label{tab:Data Volume}Data volume needed per single photoelectron. \label{Data volume per single photoelectron.}} \end{center} \end{table} Based on the spe volume, the data rate to be transmitted between a PMT and the CUU can be derived. It is completely dominated by the dark rate of the PMT consisting of single photons. We assume a conservative 50\,kHz for the dark rate, which corresponds to 18.2\,Mbit/s. Any data volume of physics events is negligible compared to the dark rate, even if the physics events contain much more than single photons. For comparison the maximum data rate of a USB-2 link is 280\,Mbit/s. In the event of a supernova the data rate will probably exceed the capacity of the link to the CUU. Therefore a buffer is needed that allows a temporary storage of the data on the FEE itself. A nearby ($\sim$3\,kpc) supernova will create 60,000 neutrino events in the detector within $\sim$10 seconds. The event energy ranges up to 50~100\,MeV and will create relatively large pulses in almost all of the PMTs. Therefore, we assume that the average sampling period per neutrino event corresponds to five 32\,ns readout-windows for each PMT. This produces a data volume of $60,000\times5\times360\,{\rm bit} = 13.7$\,MByte for each FEE channel. On top of this there is the data volume from the dark noise which adds up to 22.8\,MByte in 10 seconds. The total sum is 36.4\,MByte. This is the minimum size of the buffer on each FEE. Further along the readout chain, the uplink from the CUU to the computer farm has to be considered. A maximum of 128 PMTs will be combined into one group read out by a single CUU. For smaller numbers the requirements are less severe. To obtain rates and data volumes, the requirements for of a single PMT have to be multiplied by 128, corresponding to a data rate of 2.3\,Gbit/s for the link. A link capacity of 10\,Gbit/s which is available for standard PCs today will be sufficient. The overall buffer storage needed for a supernova event amounts to 4.7\,GByte. The data can potentially be stored in the RAM of the CUU. \subsection{The `intelligent' PMT} \label{sec:roe:ipmt} This section describes the functionalities of the single elements contributing to the underwater readout scheme. \subsubsection{High Voltage} The high voltage (HV) will be generated directly on the PMT from a low-voltage input. A single HV potential will be generated for each PMT from which all voltages required for the photo cathode, the field shaping electrodes, and the two MCPs are derived through a voltage divider. The final PMTs will have their cathodes on ground potential. Consequently the output of the signal will be on positive HV. Fig.~\ref{fig:MCP base} shows the corresponding circuit proposed by the Dubna group. The HV will be derived with a Cockcroft-Walton multiplier, also known as a Greinacher cascade, displayed in Fig.~\ref{fig:Greinacher}. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{ReadoutElectronicsAndTrigger/figures/MCPbase.jpg} \caption[Voltage divider for the MCP-PMT base]{\label{fig:MCP base}Circuit for the voltage divider at the MCP-PMT base} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/Greinacher.png} \caption[Cockroft-Walton multiplier]{\label{fig:Greinacher}The Cockroft-Walton multiplier for HV generation.} \end{center} \end{figure} The Dubna group has already developed a working module for the 8'' MCP-PMTs, which is displayed in Fig.~\ref{fig:MCP-PMT prototype}. The input is an AC voltage of 24V. For the final module the input voltage might be higher (up to 200V). The value still needs to be decided. The cascade is embedded in a HV-module that allows regulation of the output voltage. The module already includes the voltage divider and the capacitor necessary to decouple the signal from the HV. It also has the ability to monitor and limit some of the parameters, i.e.~the output current. The current prototype is controlled through a USB interface. The final version will be controlled by the SoC on the PMT. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/MCPprototype.jpg} \caption[Prototype for HV module]{\label{fig:MCP-PMT prototype}Prototype of the HV module for the 8'' MCP-PMT prototype} \end{center} \end{figure} \subsubsection{Clock System} It is essential that all PMTs are running with the same frequency. This frequency will be provided by a central clock in the cavern above the detector. The clock signal is distributed to all PMTs. For most physics topics knowledge of the absolute time is not required, with the exception of the signals from a nearby supernova. A GPS-system that allows absolute timing should be considered for this case. The GPS receiver must be mounted on the surface, probably in one of the entrance areas. The cables or fibers that connect a PMT to the central clock generate a time delay between the clock in the PMT and the central clock. These are called time-offsets $t_0$. For the reconstruction of events in the detector these offsets must be calibrated on the level of 100\,ps. If the system runs with a global trigger, a rough calibration is necessary already during running (maybe on the level of 1\,ns). For the untriggered system online-calibration is not required. The values of the $t_0$'s can be measured from calibration runs with a light-pulser (LED) in the center of the detector and from the data itself. The electronics must ensure stable $t_0$'s over longer running periods to reduce the calibration runs to a reasonable frequency. A White Rabbit system has been proposed to calibrate the offsets between the PMTs and the central clock . Such a system seems well suited for a triggered system. For the untriggered system such an effort is not necessary. Offsets ($t_0$'s) from the previous calibration run will be more than sufficient. In the event of a nearby supernova all offsets (including the offset between the GPS receiver and the central clock) must be calibrated in absolute terms (maybe on the level of 1\,ns). The calibration can be done before or after the supernova event. A calibration after the event would save the effort, if no supernova appears, but a plan must be available in any case. The simplest method to distribute the clock to the PMTs is through individual cables. Distribution of the clock signal could use the same infrastructure as the CUUs, i.e.~a single cable transmitting the signal from the surface to the CUU that than spreads it out to the whole group of PMTs. Alternatively one could add the clock signal to the power cables or distribute it through the data links. In principle a wireless distribution is also possible. The central clock will produce a reference frequency from which the PMTs derive their frequency through a phase lock loop (PLL). Most likely the reference frequency will be lower than the PMTs frequency. To give an example: For a 500 MHz digitization on the PMTs one could distribute a 62.5 MHz reference frequency and divide the clock cycles by a factor 8 on the PMTs. \subsubsection{Baseline Adjustment} The output of the PMTs will be on the potential of a positive high voltage. A capacitor will be used to decouple the signal from the HV. This capacitor will be integrated into the HV module. The potential behind the decoupling capacitor will be referenced to ground. For the determination of the charge integral of signal pulses a precise knowledge of this potential (i.e.~the signal baseline) is essential. One possible approach uses the FADC to measure the baseline directly prior to the pulse. The digitization of the input is started already before the arrival of the pulse. It is important not to under-sample the baseline. It must be determined with the same precision as the pulse itself. The baseline must be sampled for the as long as the pulse itself, doubling the data output. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/BaselineAdjustment.png} \caption[Circuit for baseline adjustment]{\label{fig:Baseline adjustment}Circuit for the baseline adjustment} \end{center} \end{figure} Here we propose to use a different approach. We want to measure the baseline continuously and adjust it to ground potential through a feedback loop (Fig.~\ref{fig:Baseline adjustment}). The concept is very simple. A discriminator compares the input against ground potential with the frequency of the clock. Due to noise on the input the input potential will fluctuate between positive and negative potentials. On average, time slices with positive ($+$'s) and negative ($-$'s) voltage bits will be balanced. The discriminator will produce as many $+$'s as $-$'s. If the input potential changes, the balance between $+$'s and $-$'s will be distorted. For example an excess of $+$'s will indicate a deviation of the input potential to positive values. A simple logic will determine the difference of the number of time slices with $+$'s and $-$'s. The result will be used to adjust the potential in one of the input amplifiers until the balance is restored. Two special situations must be considered to avoid a bias in the baseline: \begin{itemize} \item If a true pulse arrives the feedback loop will detect it as a positive deviation of the baseline and counter-correct the deviation. To avoid this bias the feedback loop has to be stopped if an input pulse is detected. \item The return of the signal to the baseline after a pulse will take some time. If a second pulse arrives before the signal has returned to zero, its charge integral will be measured with a too large value. This effect can be corrected offline. All photon pulses will be read out. Time and charge of the previous pulse can be determined independently. An algorithm on the online-farm can either correct for the previous pulse, or store the previous pulse for later correction, or do both. \end{itemize} There is, however, one special case, where this will not work. If the pulseheight of the previous pulse is below detection threshold, the correction cannot be applied. But one should keep in mind, that close pre-pulses will be rare, the undetected fraction of pre-pulses even rarer, and the bias will be very small, especially for pulses below threshold. A number of parameters of the baseline adjustment need to be optimized. The difference of $+$'s and $-$'s outputs of the discriminator needs to be averaged over a certain amount of time slices. The larger the number, the more precise the adjustment becomes, but it also gets slower. This parameter should be kept flexible to adapt to the situation in the real experiment. Another important parameter is the step-size with which the baseline is adjusted. The overall concept of the intelligent PMT will include a random trigger. That is, at adjustable time intervals the electronics will digitize a readout window and store the output. These readout slices should be empty and can be used to check the performance of the baseline adjustment. \subsubsection{Internal Trigger} A simple level discriminator will be used to trigger the readout of a single PMT. It will continuously monitor the output of the PMT. The threshold must be programmable. It will be set directly above the noise, as low as acceptable in terms of rate of noise triggers. The threshold can be adjusted individually for each PMT. We envisage a threshold at a level of 0.3 photoelectrons. In the case of the MCP-PMTs there will be separate triggers for each MCP. Most likely the electronics will be configured in such a way that a signal above threshold from one of the MCPs will trigger the readout of both channels. The readout window will be adjusted to the length of a pulse from a single photoelectron, probably around 32\,ns. In the case of high energy events such as cosmic muons or atmospheric neutrinos several or even many photons will be detected by a PMT. The output signal will extend over more than 32\,ns. If the photons are separated in time, they will be read out as individual triggers. If not, the full signal will be covered by reading several consecutive 32\,ns windows. The trigger logic monitors the threshold discriminators. If at a certain time during the readout (i.e. in the middle of the 32\,ns window) the signal has not returned to below threshold, the logic will continue the readout for another 32\,ns. The readout will be sustained until the signal finally falls below threshold. No information is lost. \subsubsection{Flash-ADC} A FADC will digitize the input signal. For the MCP-PMTs two separate FADCs will be needed for the two MCPs. The required dynamic range of amplitude has been estimated to 1000\,pe for a whole 20'' MCP-PMT, or 500\,pe per MCP. The prototypes of the MCP-PMTs show an electronic noise level which is in the order of 10\,\% of the amplitude of a single photo-electron. Therefore a 10\,\% resolution on the amplitude of a single pe will be appropriate. The corresponding resolution on the spe charge integral will be better, on the level of a few percent. If the amplitude of a single pe corresponds to 10 ADC units, a dynamic range of 5,000 to 10,000 ADC units will be needed, corresponding to 13 bits. It is technically not possible to derive 13 bits from a single ADC. Multiple ADCs with staggered ranges are required. We propose to use 3 ADCs with 7 bits each, featuring the following scheme: \begin{center} \begin{tabular}{ccc} \hline amplitude & range & resolution \\ \hline $\times1$ & 0$-$13\,pe & 0.1\,pe \\ $\times1/8$ & 0$-$100\,pe & 0.8\,pe \\ $\times1/64$ & 0$-$800\,pe & 6.4\,pe \\ \hline \end{tabular} \end{center} \subsubsection{Time Measurements} For the reconstruction of vertices or tracks it is necessary to measure the time of arrival of the photons on the PMTs. A resolution of 1\,ns is required. The arrival time can be extracted from the waveform recorded by the FADC. With the proposed 500-MHz FADC (2\,ns time slices) it is possible to extract the starting time with 1\,ns resolution. More precise timing information is encoded in the digital pulse of the level discriminator of the internal trigger. The discriminator triggers the first photon of the pulse. This is the most valuable information for the reconstruction. A TDC could be included to measure the time of the trigger relative to the preceding clock pulse. For example to get a 500\,ps resolution a 2-bit TDC would be sufficient. \subsubsection{Self-Calibration} The ASIC will contain a mechanism to calibrate its PMT to a predefined charge integral of a single photoelectron (SPE). The basic scheme is shown in Fig.~\ref{fig:Self-calibration}. Whenever the input signal crosses the trigger threshold for a single photoelectron the flash-ADC will digitize the input and write the data into an internal memory. The logic in the digital part of the chip will analyze the pulses within a single photoelectron window and determine the average charge integral, $\langle q_{\rm pe}\rangle$. The $\langle q_{\rm pe}\rangle$ of each channel is compared to a default charge value, and the HV applied to the PMT adjusted until the predefined value is reached. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/Selfcalibration.png} \caption[Self-calibration scheme]{\label{fig:Self-calibration}Schematic drawing of the self-calibration feedback loop} \end{center} \end{figure} The presence of two MCPs with different gains inside one PMT adds a complication because both MCPs are powered from a common high voltage. The algorithm on the chip will compare the $\langle q_{\rm pe}\rangle$ of both MCPs to the predefined value. It will first adjust the gain of the input amplifiers until the $\langle q_{\rm pe}\rangle$ of the two MCPs agree with each other. Then it will adjust the high voltage to match the $\langle q_{\rm pe}\rangle$ to the predefined value. \subsubsection{Internal test devices} The circuitry will incorporate extensive possibilities for testing. This is important to guarantee the functionality of the system and to understand how to react in case of problems. The details still need to be worked out. Two test installations that will be integrated for sure are: \begin{itemize} \item A test pulser within the SoC: The pulser can be used to inject charge into the input of the chip. It could either be an analogue circuit which produces pulses similar to the real pe's or a digital pulse generator. \item A light pulser: A LED will be mounted in the electronics in a position where the light illuminates the transmission cathode from the backside (through the glass neck of the PMT). \end{itemize} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/PMTMechanics.png} \caption[Mechanical realization of the intelligent PMT ]{\label{fig:PMT Mechanics}Mechanical realization of the intelligent PMT design.} \end{center} \end{figure} \subsubsection{Mechanics} The scheme for mounting the FEE directly to the base of the PMT is shown in Fig.~\ref{fig:PMT Mechanics}: The HV module plugs directly onto the pins that connect to the electrodes in the PMT. It includes the wire(s) that carries the output signal. A capacitor is needed to decouple the signal from the high voltage. This capacitor will be included in the high voltage module. On top of the high voltage module there will be a PCB that carries the ASIC with the SoC and some external components (voltage regulator for the low voltage, etc.). The cable to the central underwater units is connected to this PCB. The whole electronics is inside a watertight metal cap that connects to the mechanical fixture of the PMT. The details on how to seal the housing still need to be worked out. For the default scenario, where data is routed through the central underwater units (CUU) there is a single cable from each PMT to the CUU. It contains power, clock and the data line. It is firmly attached to the electronics housing on the PMT without a connector. The cable length will be between 5 and 10 meters. It has a watertight connector that connects to the CUU. For redundancy there might be two such cables per PMT connecting to two different CUUs. In case passive data links will be used, an optical fiber will be connected to each PMT. The driver circuit and the laser will be included on the PCB. There must be a watertight optical connector on the fiber where it connects to the PMT. For each PMT there will be a separate cable for the power that connects the PMT to the power distribution box. \subsubsection{Power} Each PMT needs a low voltage DC-power for the SoC and a low voltage AC-power for the high voltage module. The exact values have not yet been defined. For the time being we assume a +3V DC-voltage for the SoC and a +24V AC-voltage for the high voltage module. The voltages are generated in power supplies in the cavern on top of the detector. They are accessible and can be replaced in case of failures. From the power supplies, cables will transmit the power to distribution boxes inside the water pool. Each PMT is connected to one of the boxes with a cable and a water tight connector on the power distribution box. The cable contains the DC and the AC voltage, the ground potential and maybe the clock. In case the scheme with active data links is adopted, the power distribution will be integrated into the CUUs. In the case of passive data links there will be distribution boxes just for the power. One power distribution box or CUU will supply power to a group of 128 PMTs. This system must be protected against failure of the whole group. If a cable to one of the PMTs breaks, a single PMT will be lost which is not catastrophic. However, a short in one PMT could take out the whole group. Therefore a serial resistor RP has to be introduced in the PMT contacting scheme shown in Fig.~\ref{fig:MCP base}. These resistors must be large enough to limit the current through a PMT with a short, at the same time avoiding a too large power consumption in the resistor if the voltage on the PMT changes. For example, if 3\,V is required for the SoC, a 6\,V power supply could be used. We choose a value for RP that creates a 1\,V voltage drop. A voltage regulator located on the PCB at the PMT base will reduce the remaining 5\,V to 3\,V. The protection resistor increases the power consumption by 20\,\%. which will still be acceptable. In case of a short the power of this channel goes up by a factor of 5. This is still negligible compared to the power of the whole group. Another critical item is the cable from the power supply to the distribution box. It might break or shorten out. We propose to have two identical cables, but only one is connected to the power supply. Inside the distribution box there are switches on the power. Without power the switches are closed. Let's assume cable A is connected to the power supply. If we turn on the power, it will open the switch on cable B and disconnect it completely from the system. If cable A fails, it will be disconnected from the power supply. This closes the switch to cable B. Cable B can now be connected. It will open the switch on cable A and disconnect it from the system, bringing the system is back into operation. \section{Development plan} \label{sec:roe:dev} \subsection{Characterization of PMT response} To get a better understanding of the dynamic range and voltage span required, the PMT response and in particular the pulse height and linearity will be studied as a function of the incident number of photo-electrons. For this, a laboratory test stand is currently realized. As the MCP-PMTs are not yet available, characteristics will be first determined for the 20'' PMT produced by Hamamatsu. For covering the whole range specified, waveforms containing 1$-$4000 pe have to be recorded. To match this requirement, the tests will be performed following this procedure: \begin{enumerate} \item Record {\bf single photoelectron pulses} at a gain level of $~$10$^7$. The right gain is assured by integrating the single photoelectron waveform (cf.~Fig.~\ref{fig:waveform for SPE}), subtracting the baseline pedestal and dividing by the elementary charge. An additional factor of 2 has to be taken into account as a double-end termination is used in the PMT base. \item {\bf Ramp up of LED intensity} until 4000\,pe are reached, while continuously recording the waveforms for different LED output and thus multiple photoelectron levels. As before, the equivalent charge will be calculated based on waveform integration. \end{enumerate} At low hit multiplicities, a discret cascade of broadening peaks will be observed that corresponds to defined numbers of photoelectron. At higher LED intensities, the charge spectrum will become a Gaussian distribution. The number of incident photoelectrons can be calculated based on the gain. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/SPEWaveform.jpg} \caption[Readout Electronics]{\label{fig:waveform for SPE}The waveform for SPE} \end{center} \end{figure} \subsection{Signal amplification} A careful study of the amplification requirements will be performed for the three readout schemes, including pre-amplification, gain division, and ASIC design. The relevant design specifications are the dynamic range of 1$-$4000\,pe and an equivalent input noise RMS of less than 0.1\,pe. The following aspects should be investigated: \begin{itemize} \item {\bf Impedance matching for long cable.} Reduce the reflection between transmission lines and other devices, considering the effect of parasitics and couplings of cables and transmission lines. \item {\bf Gain division.} The dynamic range is divided into three overlapping FADC acquisition ranges, 1$-$67\,pe, 1$-$529\, and 1$-$4000\,pe. As the original amplification by the PMT dynode chain or MCP will depend on the PMT chosen in the final design, so will the amplification stages. \item {\bf Amplifier selection.} Commercial amplifiers have to be selected for large dynamic range and low noise. Signal clipping and overload protection will be studied carefully for MCP signal. \item {\bf Amplifier for FADC driver.} Since the PMT signal amplitude is large, the amplification factor will be small (less than 2), which makes the noise of ADC driver important to the resolution of the total system. Low noise of this amplifier is thus mandatory. Other performance parameters such as settling time have to be studied. \item {\bf Discriminator selection}. A discrimination level of less than 1/4 pe is required. \item {\bf ASIC design.} Amplifiers and discrimators have to be integrated in an ASIC holding 16 or 32 channels. \end{itemize} \subsection{Cables and connectors} \subsubsection{External readout scheme} The external scheme will employ 20,000 coaxial cables of 80$-$100\, in length and compatible with the underwater environment. Due to the great length, special attention should be given to low signal delay and attenuation to ensure a high quality for the transmission of the analog signals. A selection of coaxial cables from the have been surveyed Tianjin 609 Cable Co. Ltd.. Samples of 100\,m length have been tested for attenuation. Results are listed in Table\ref{tab:Coaxial Cable}. While the Type III cable clearly shows the best performance in terms of signal dampening, it is also the most expensive. \begin{table}[htbp] \begin{center} \begin{tabular}{lccc} \hline Parameter & Type I & Type II & Type III \\ \hline Product Model & SYVF -50-3-1 & 609C5021A & 609C5019A \\ Outer Diameter & 10.5\,mm & 3.5\,mm & 4.5\,mm \\ outer sleeve & ordinary sleeve & FEP & FEP \\ Attenuation @ 100\,MHz & 17.7\,dB & 15.9\,dB & 12.3\,dB \\ Attenuation @ 200\,MHz & 24.8\,dB & 23.4\,dB & 17.7\,dB \\ Price (RMB/m) & 10 & 14 & 18 \\ \hline \end{tabular} \caption{\label{tab:Coaxial Cable} Survey and attenuation results for three samples of 100\,m long coaxial cables from Tianjin 609 Cable co.} \end{center} \end{table} The sleeves have also been tested for compatibility with the LS. The effects on the attenuation spectra of LS samples that have been left for some time in contact with the cables are shown in Fig.~\ref{fig:compatibility of the sleeves and LS}. The resulting changes for cables of Type II and III are less severe than for Type I. \begin{figure}[htb] \begin{center} \includegraphics[width=0.9\textwidth]{ReadoutElectronicsAndTrigger/figures/CompatibilityOfSleevesAndLS.jpg} \caption[Compatibility tests of cables and LS]{\label{fig:compatibility of the sleeves and LS}Effect of the exposure of LS to the cable sleeves on the LS attenuation spectrum.} \end{center} \end{figure} \subsubsection{Intermediate readout scheme} The requirements for the cabling of the intermediate readout scheme are somewhat different. As the FEE are moved into the submerged processing units, watertight boxes as well as underwater connectors are needed. In detail, this scheme needs: \begin{itemize} \item About 20,000 coaxial cables of 10$-$20\,m length for the signal transmission from PMT to the processing units. Cables should be covered by a sleeve from fluoroplastics. \item About 20,000 connectors that can be plugged under water. Fig.~\ref{fig:pluggable connector} displays suitable products by the Shanghai Rock-firm Co.~Ltd. \item About 500 watertight cases for the central processing units \item About 500 composite cables for data transmission via optical fibers and low-voltage power supply \end{itemize} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/PluggableConnector.jpg} \caption[Underwater connectors]{\label{fig:pluggable connector}Under-water connectors from Shanghai Rock co.} \end{center} \end{figure} The feasibility of this scheme is currently evaluated at IHEP. Contacts have been established to a variety of providers for watertight connectors and cases, including Beijing the Great Wall Electronic Equipment Co., Ltd. China aviation optical-electrical technology CO., Ltd. Shanghai Rock-firm CO., Ltd. No.23 Research Institute of China Electronics Technology Group Corporation and others. While watertight cables can in any case be provided, the sleeve material and the fluorination treatment of the rubbers are still under investigation. This is an important aspect as fluorination will protect the rubber against the corrosive strength of pure water, allowing for a life time of 10$-$20 years. Further important questions are the impermeability of the watertight cases, the characteristics of the signal transmission and last but not least the additional cost introduced by this scheme. \subsection{ADC selection and circuit design} A further important step is the selection of the FADC chip. Several FADC chips have been included in a preliminary survey. Tab.~\ref{tab:ADC Select} summarizes the main characteristics for a variety of commercial and self-developed chips. Circuit designs are currently studied to evaluate the FADC performance and its compatibility. \begin{table}[hbtp] \begin{tabular}{l|lcccc} \hline & & Sampling & No.~of & Resolution & \\ Type & Model & rate [GS/s] & channels & bits & ENOB \\ \hline FADC & ADC10D100D & 2/1 & 2 & 10 & 9.1 \\ & ADC12D1000 & 2/1 & 2 & 12 & 9.6 \\ & EV10AQ190A & 1.25 & 4 & 10 & 8.6 \\ & ADS5409 & 0.9 & 2 & 12 & 9.8 \\ & ADS5407 & 0.5 & 2 & 12 & 10.3 \\ & Stefan's (ASIC) & 1 & 1 & N/A & N/A\\ & Fule's (ASIC) & 1 & 1 & 12 & >10\\ \hline Sampling & DRS4 + AD9252 & 1-5 & 8 & 11 & 8(?) \\ + ADC & Weiw's (ASIC) & 1 & 1 & N/A & N/A\\ \hline \end{tabular} \caption[Overview of FADC chips]{\label{tab:ADC Select}Overview of selected (Flash-)ADC chips} \end{table} \subsection{Design of waveform digitization ASIC } To fulfill the requirements of JUNO concerning timing and pulse height resolution, the acquisition and digitization of PMT waveforms at a sampling rate of 1\,GS/s and an acquisition range of 12\,bit is mandatory. However, commercial high-speed FADCs meeting these specifications are currently banned from import to mainland China. While possibilities exist to obtain suitable chips by transfer via a Sino-American collaboration, the resulting price will be undoubtedly high. In order to reduce the cost per electronics channel, design efforts for implementing a new form of ADC on an ASIC or for a self-made high-speed FADC have been started. \subsubsection{Design of an ASIC for high-speed waveform sampling} The bottleneck of the design of the high speed-waveform sampler is the huge data throughput. It is very difficult to store a huge data volume within a very short time as it would be encountered during a supernova. However, as long as the event rate is limited, continuous storage will not be necessary as most of the acquired date is either baseline or background signals. The waveform-sampling chip will rely on the online selection and caching of the analog data. The chip will first sample the waveform at high speed, and save the resulting waveform information in an analog memory. When there are no physics triggers, the sampling process will keep running, thus refreshing the cached analog data and overwriting the old one; when a trigger is issued, the waveform stored in the analog cache will be digitized immediately. The post-processing logic therefore only has to deal with waveforms of the length of the analog buffer. This method allows for both high-speed sampling and caching. Moreover, the conventional approach of simultaneous analog-to-digital conversion and readout are thus separated. The analog memory first acquires the pulse, then the ADC is converting the stored waveform and sends only a short piece of waveform to the backend system for storage. Therefore, the data processing load for the backend systems is greatly reduced. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/waveformASIC.png} \caption[Readout Electronics]{\label{fig:Waveform ASIC}The architecture of one channel of the waveform sampling ASIC} \end{center} \end{figure} The architecture of the waveform sampling chip is given in Fig.~\ref{fig:Waveform ASIC}. The high speed sampling and analog memory is realized by a large switched-capacitor array (SCA). When no trigger comes, the SCA will keep sampling in a cycling style, and the voltage values held on the capacitors are consequently refreshed; when a trigger is issued, the held voltage will be compared with the common ramp signal of the chip by an exclusive comparator for each capacitor. Meanwhile, a counter will monitor the time duration from the start of the common ramp signal until stop signals are issued by the comparators. The stored analog information is then converted to digital output in a single step. The individual start-to-stop time counts will be stored in the corresponding latches. When the the ramp signal ends, the latched data will be transmitted by the serializer. In order to be able to cope with the expected dark rate per channel of 50\,kHZ, each ASIC channel will include 256 storage cells. \subsubsection{Design of a high-speed FADC} A further solution is a self-designed high-speed FADC. The idea is based on a more conventional architecture, for which a high-speed ADC will take charge of sampling the waveform and digitization in a single step. All the converted digital data will then be transmitted to the backend system for data buffering, trigger selection processes and so on. Due to the import prohibitions, a self-designed FADC put into mass production will greatly decrease the cost of the system. The ADC will be based on a hybrid architecture combining a pipeline A/D convertor and flash type. The signal will be sampled through an input buffer and a sample-and-hold circuit, and then be pre-converted by a four-stage pipeline conversion. A flash conversion stage will finally digitize the signal. In this way, the stress of high sampling speed and high voltage resolution are distributed to multiple stages. Also power dissipation per area will be reduced. The chip is expected to achieve asampling rate of 1\,GS/s and a resolution of 12\,bit, with a guaranteed effective number of 10\,bits. \subsection{Data processing and transmission} \subsubsection{External readout scheme: ATCA modules} The FEE modules used in this electronics scheme will be based on the ATCA standard. The structure block diagram is shown in figure\ref{fig:ATCA Structure}: The modules will be based on a daughter-mother board structure. The function of the daughter board is to receive the PMT signals and to perform the digitization. The mother board is a standard ATCA module. As the main unit of data transmission and processing, it will aggregate the data from all daughter boards, transfer the clock and trigger signals via the high-speed backplane bus, transfer the final data to the DAQ via a Gigabit ethernet connection and take care of system control and configuration. The FPGA of the module will perform the digital signal processing and information extraction. \begin{figure}[htb] \begin{center} \includegraphics[width=0.8\textwidth]{ReadoutElectronicsAndTrigger/figures/ATCAStructure.jpg} \caption[FEE block diagram for ATCA modules]{\label{fig:ATCA Structure}Standard ATCA module structure block diagram} \end{center} \end{figure} \subsubsection{Intermediate readout scheme} {\bf Submerged processing unit.} In the partially submerged scheme, the signal processing unit will still be based on a daughter-mother board structure. The FEE daughter board will be mostly identical to the external case. The mother board will package the data and transfer it to the external electronics via an optical fiber. For reliable data transmission, some redudancy will be introduced for the mother board:two FPGAs will serve as backup for each other. When one of the FPGAs is not working properly, the system will switch to the other one. A corresponding diagram is shown in Fig.~\ref{fig:UnderWaterSignalProcess}. \medskip\\ \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{ReadoutElectronicsAndTrigger/figures/UnderWaterSignalProcessing.jpg} \caption[Block diagram of processing unit]{\label{fig:UnderWaterSignalProcess} Diagram of the submerged processing unit} \end{center} \end{figure} \noindent {\bf Signal collection unit.} The data sent by the underwater processing units via optical fibers will be received by a collecting ACTA module based on a daughter-mother board structure. The motherboard is an ATCA standard module, while the daughter board completes some simple function like the collection of the optical data and the transmission to the mother board. \subsection{Design of trigger and clock} The White Rabbit technology has been used extensively in the LHAASO experiments. A compact FMC-form White Rabbit node (cute-wr) has been developed which can be easily integrated as a network mezzanine card on other electronics system. It will provide a uniform 125\, MHz clock signal, a Pulse-Per-Second signal and an encoded UTC timestamp signal. First tests with a prototype system show a 100\,ps accuracy and 21\,ps precision among several cute-wr cards. \begin{figure}[htb] \begin{center} \includegraphics[width=\textwidth]{ReadoutElectronicsAndTrigger/figures/WhiteRabbit.png} \caption[White Rabbit system]{\label{fig:WhiteRabbit}Compact FMC-form White Rabbit nodes and the measurement of PPS synchronization} \end{center} \end{figure} \subsection{Design for system reliability} \subsubsection {Specifications for reliability} The reliability requirement for the electronic system is a failure rate of less than 0.1\,\% per year of operational time for each of the 16,500 detector channels. The design life time of the system is 20 years. At the desired failure rate, it should be able to run for at least 10 years without need of maintenance action. In the intermediate and underwater schemes, most of the electronics will be put into sealed boxes at a maximum water depth of 40\,m. The temperature of the water will be in the range of +10 to +20 Centigrade. The ambient radiation levels will be normal. \subsubsection {Reliability engineering} As an important aspect of product assurance management, reliability engineering will cover the entire electronics design and development for JUNO. This will include: \begin{itemize} \item {\bf Management of components.} The used parts will extremely affect the quality and reliability of the final product. There will be a management plan for the components, and a list of preferred components to provide guidance in their selection and quality control. \item {\bf Reliability design,} including reliability definition, system reliability analysis, reliability allocation, reliability design and prediction \item {\bf Analysis of failure modes and effects}, including a fault tree analysis. \item Analysis of {\bf sneaks} \item Analysis of {\bf ASIC reliability} \item Assurance of {\bf firmware reliability}. \end{itemize} For achievement of the high quality and reliability requirements for the electronics, all steps of production and installation of the JUNO electronics (except the reliability design in R\&D procedure) should be controlled by the overall Product Assurance (PA) management plan. This will include \begin{itemize} \item Risk management \item Safety management \item Control of critical items \item Procurement, verification, screening, handling and storage of components \item Control of techniques and production procedure \item Verification and Environmental tests. \item Software assurance, etc. \end{itemize} There will be a PA organization supervised by PA manager to establish and implement the management plan. \section{Manufacture and assembly} \label{sec:roe:rea} \subsection{Production} Production of $\sim$20,000 channels will be a demanding task that must be subject to a strict quality management and continuous supervision. The final design will be chosen after a phase of elaboration and prototyping of the three design concepts that have been introduced above and a careful evaluation of their respective merits and disadvantages. Once the design is fixed, production will start in several steps: \begin{itemize} \item {\bf Ordering of electronic components}, including cables and crates. Each supplier and batch will be strictly controlled to ensure the reliability. \item {\bf PCB manufactory and assembly} will be carried out by professional manufacturers. A close personal contact with the companies will be established to allow for quality supervision and sampling at the production site. \item {\bf Aging tests} with the components that should be performed well in advance of the start of JUNO data taking. Components failing the tests can be repaired and re-introduced after the performance has been tested. \end{itemize} \subsection{Aging and long-term tests} According to typical reliability curves for electronic devices, the failure rate will be relatively high in the beginning while forming a plateau of constant low failure rate after. Therefore, all critical components should be installed and running for some time before the experiment starts, allowing to pick out and repair malfunctioning components. Aging tests will be performed by exposing electronic parts for a period of the order of one week to an environment of increased temperature. If 1000 channels can be tested simultaneously, the whole process will require about 20 weeks. In order to optimize the conditions of the aging tests, pre-studies risking the destruction of some of the components should be conducted. Before installation, the system needs to undergo a long-term test run at a dedicated test stand at a laboratory that allows for debugging, identification of failing components and further aging control. This thorough test of at least part of the system should last at least for 3$-$4 months, and should be performed with real PMTs connected. \subsection{Installation} The readout system will be installed along with the central detector and PMTs. The foreseen time span for installation is 6$-8$ months. It will start from the bottom and end at the top cap. The external testing stand should be completed before installation to allow to perform the tests during the installation phase. After completing the installation, a debugging phase to test the stability of the whole system is foreseen to ensure optimal working conditions at the start of the experiment. \subsection{Reliability} All steps require rigorous quality and reliability management and control, from ordering the components to the finalization of the installation. This will ensure the system stability, reduce losses in the production process, and minimize the consumption. \section{Risk analysis} \label{sec:roe:risk} \subsection{Three read-out schemes} Some risk obviously exists for such a 20,000-channel high-speed sampling system, in which all of the PMTs are located inside the water pool and therefore not immediately accessible to repairs. The "external readout scheme" (Sec.~\ref{sec:roe:dry}) will place most of the electronics outside the water pool. This will greatly simplify the replacement of broken modules but potentially reduces the measurement quality and increases the costs. In the "underwater readout scheme" (Sec.~\ref{sec:roe:wet}), almost all the critical components of the electronics will be placed underwater and are possible without reach for replacement (especially the FEE). This assures optimum measurement performance, but will increase the risk of failure unless special attention is given to an increased redundancy of vital system components. In case of the "intermediate readout scheme" (Sec.~\ref{sec:roe:int}), a large part fraction of the electronics is located inside the water. All vital electronics are concentrated inside the central processing units that are designed to be replaceable. Compared to the other schemes, both performance and repairability are optimized but costs and design difficulty are increased. The final design will be chosen after all three designs have been properly elaborated and prototypes have been tested. \subsection{Design risks} Design risks comprise mainly four aspects: \begin{itemize} \item Quality of signal measurement \item Reliability \item Cost control \item Design integrity \end{itemize} The biggest risk is the signal measurements, which followed by the risk of reliability and design integrity. The reliability is associated with cost control. Risks in design integrity will be mostly caused by two aspects: Insufficient communication of the groups designing different parts of the system and a lack of understanding of the system itself. These risks can be mitigated by a good collaboration in between the participating groups. \subsection{Production risks} Production risks are mainly arising in the procurement of electronic components, PCB manufactory and assembly, aging tests and cost control. A very strict quality control system will be established. \subsection {Installation risks} Risk in installation will also be minimized by a strict management and quality control, including three aspects: installation quality, time and safety. \section{Schedule} \begin{itemize} \item[2013] Start of design of the readout schemes. \item[2014] R\&D for key components. \item[2015] Test of a prototype system of 200 channels for the "external readout scheme" at Daya Bay. Finalization of a 32-channel demonstrator for the "intermediate readout scheme". Review of all three schemes and choice of the final scheme. \item[2016] Prototype system for the final scheme, about 200 channels \item[2017] Mass production \item[2018] Mass production, testing ans start of installation \item[2019] Installation and testing \end{itemize} \chapter{ScheduleAndBudget} \label{ch:ScheduleAndBudget} \chapter{Veto Detector} \label{ch:VetoDetector} \section{Experiment Requirements} The main goal of JUNO is to determine the neutrino mass hierarchy, which is one of the most important unsolved problems related to neutrinos. The neutrinos are detected via the inverse beta decay by measuring the correlated positron and neutron signals. With a careful design of the detector, the neutrino spectrum can be measured precisely, and we are looking for the distortion of the spectrum for the sign of neutrino mass ordering. Compared to the small number of signal events (60/day), the number of background events is still very high due to the large volume of detector (20~kton). The cosmic ray muon induced backgrounds are the main backgrounds and they are hard to remove. The cosmic ray generated backgrounds are: 1. $^9$Li/$^8$He background from muon spallation and muon shower particles. 2. Fast neutron background in the detector from muon induced high energy neutrons. The cosmic ray induced backgrounds also effect the study of the diffuse supernova neutrino flux. In order to reduce the experimental backgrounds, the neutrino detector must be placed in deep underground and a veto system is used to tag muons. The muons should be detected with high efficiency for the purpose of background reduction. This chapter is mainly about research and design of the veto system. Due to the strict requirements on background suppression, larger overburden of rocks on top of the detector is needed to reduce the cosmic ray muon flux. The experiment is located at a site of about 53~km equal distance to the Yangjiang and Taishan nuclear power plants. The height of the mountains is 270 meters and the experimental hall is located underground at the depth of 460 meters. Therefore, there is about 700-meter rock on top of the experimental hall. Muon rate is estimated at about 0.003~Hz/$m^{2}$ and the average muon energy at about 214~GeV from simulation. The cosmic ray muon flux is reduced by $\sim$60,000 times compared to that at the ground surface. The remaining energetic cosmic ray muons can still produce a large number of neutrons in the rocks and other detector materials surrounding the central detector. These neutrons can produce fast neutron background in the central detector which mimics the inverse beta decay signal. This kind of background cannot be ignored. In order to shield the neutrons and the natural radioactivities from the surrounding rocks, at least 2 meters of water surrounding the central detector is needed. The water is effective for shielding against neutrons and gamma (Fig.~\ref{gamma_neutron_shield}). And when being instrumented with photomultiplier tubes (PMTs), the water pool can serve as a water Cherenkov detector to tag muons. The Daya Bay experiment shows that the veto system utilizing water Cherenkov detector can be very successful \cite{DYB:muon2014}. From simulation, muons with relatively long track in the detector can be detected with very high efficiency, and the undetected muons are mainly of short track lengths, which would induce less background because they are relatively away from central detector. Based on the Daya Bay experimental results (fast neutron background about $\sim$0.2\%), if the water shield thickness is at least 2.5 meters in JUNO, fast neutron background to signal ratio is $\sim$0.3\%, after taking into consideration the large detector volume and geometry effects. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/Gamma_neutron_attenu_water.jpg} \caption[Thickness of water vs gamma and neutron attenuation]{Left: Gamma attenuation vs thickness of water shield; Right: Neutron attenuation vs thickness of water shield.} \label{gamma_neutron_shield} \end{center} \end{figure} Another background is produced from $^9$Li/$^8$He resulting from muon spallation in the scintillator and muon shower particles. The beta-n decay of $^9$Li/$^8$He would mimic inverse beta events. This kind of background can not be reduced from increasing the thickness of the water shield. The number of $^9$Li/$^8$He background events is estimated to be $\sim$80/day. Their existence would greatly reduce JUNO's capability to determine the neutrino mass hierarchy. To reduced the $^9$Li/$^8$He background, we need precise muon track information. The $^9$Li/$^8$He background is reduced by excluding a certain cylindrical region along the muon track within a certain period of time after the muon had passed through the detector. Therefore, the $^9$Li/$^8$He background reduction depends on precise muon track reconstruction. We will build a tracking detector on top of the water Cherenkov detector to help muon tagging as well as track reconstruction. The top tracker can tag muons and reconstruct muon tracks, independent of the central detector and water Cherenkov detector. The tracking results from the top tracker can be extrapolated to the central detector. Therefore, it can provide an independent measurement of the $^9$Li/$^8$He background. The muons which pass through the water pool but not the central detector can also produce high energy particles (pions, gamma, neutrons, etc.), which could migrate into the central detector and produce $^9$Li/$^8$He there. A simple simulation shows that this kind of background which cannot be detected by the central detector is $\sim$1.2/day. A top tracker covering a large area can help directly measure such background. Through optimization of the top detector area and its layout, the detector placed at the top with a long strip structure provides better results. This configuration can identify up to 1/4 of this kind of background, which would be helpful for the study of such background. It can also be used to measure the rock neutron related background. Also the top tracker and water Cherenkov detector can help identify multi-muon events for the central detector to reduce the $^9$Li/$^8$He background. \section{Detector Design} The JUNO veto system is shown in Fig.~\ref{vetodet}. The system consists of the following components: 1. Water Cherenkov detector system, a pool filled with purified water and instrumented with PMTs. When energetic muons pass through the water, they can produce Cherenkov light. The Cherenkov photons can be detected by PMTs. The arrangements of the PMTs as well as the number of PMTs needed in the water pool is currently under study to facilitate the high efficiency detection of muons as well as good muon track reconstruction. One of the options is to have the water pool surface and central detector outer surface covered with reflective Tyvek to increase the light detection by PMTs without using a large number of PMTs. 2. Water circulation system. There will be 20,000-30,000 tons of water in the pool depending on the different central detector designs. This system will include a water production system on the ground and a purification/circulation system underground in the experimental hall. 3. Top tracker system. It can provide independent muon information to help muon tagging and track reconstruction. The OPERA top tracker are going to be transported and installed in JUNO. 4. Geomagnetic field shielding system. Though small, the Earth's magnetic field can affect the performance of PMTs. Either compensation coils or magnetic shields will be used to reduce the effect on PMTs. 5. Mechanical system. The system includes top tracker support structure, water pool PMT support structure as well as a light and air tight cover for the water pool. \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{VetoDetector/Figures/VetoDetector.png} \caption[Veto systems of JUNO]{Veto systems of JUNO.} \label{vetodet} \end{center} \end{figure} \subsection{The Central Detector as a Muon Tracker} The large liquid scintillator central detector itself is one of the most important muon trackers in the veto system. The muon rate for the central detector is about 3~Hz due to 700-meter rock overburden above. A minimum-ionizing muon deposits about 2~MeV cm$^{2}$/g. On average, a muon will pass through 23~m of LAB based liquid scintillator (density of about 0.85~g/cm$^{3}$) inside the central detector, resulting in about 3.9~GeV energy deposition. A muon generates Cherenkov photons and scintillation photons along its track in the detector. The Cherenkov photons are emitted in a cone with an opening angle $\theta = \cos^{-1}(1/n)$ relative to the muon track trajectory, where $n$ is the index of refraction for liquid scintillator $\approx 1.45$. Although the scintillation photons are emitted isotropically, the path for the fastest photon arriving at any PMT follows the same angle as that of the Chrenkov light angle $\theta$ when the velocity of the muon is nearly the speed of light. Based on the earliest hit time from each PMT, it is straightforward to reconstruct the muon track. Such a method, called the ``Fastest-Light'' muon reconstruction, has been proved to be very successful in reconstructing more than 99\% of the non-showering muons in the KamLAND experiment. Based a preliminary muon tracking study on the simulated JUNO muon tracks, the bias of this method on the zenith angle is less than 5$^{\circ}$ and the distance to the detector center bias is less than 20~cm. Although the current study has not considered many real detector effects, such as the resolution of PMT transition time, given the much better energy resolution and comparable or better PMT transition time resolution for JUNO detector compared with KamLAND, the fitting quality is expected to be further improved. However, the fitting quality of this algorithm becomes worse for the corner clipping muons, which are close to the edge of the detector. And more importantly, this reconstruciton method is also not appropriate for multiple muons or showering muons inside the detector, which is estimated to be about 10-20\% of all muons. Since the muon rate for the JUNO central detector is about 10 times higher than that of the KamLAND detector, it is not acceptable to veto the whole detector for a few milliseconds, which will significant reduce the detector live time. In order to reduce the $^{9}$Li/$^{8}$He background introduced by those muons, we have to rely on the other veto system, such as the water Cherekov detector and the top tracker to track those muons. \subsection{Water Cherenkov Detector} PMTs are placed in the water pool to tag cosmic ray muons by detecting water Cherenkov light that is produced. The number of photons produced is proportional to the muon track length in the water. Due to the limited thickness of water as well as the presence of the central detector, the majority of the muon tracks going through the water pool are not very long. Therefore, a large photo-coverage in the water pool is desired to detect muons with high efficiency. For the purpose of tagging muon, normally the pool surface would be covered with high reflectivity Tyvek film to help collect the photons without using a large number of PMTs. Based on the experience of the Daya Bay experiment and with Geant4 for simulation, muon detection efficiency is about 98\% in (Fig.~\ref{junoEffi}) by using 1,600 8-inch PMTs with at least 2.5 meter thickness of water in the pool, if we can keep the noise level as low as that of the Daya Bay experiment. The long track muons can be detected with extremely high efficiency, as shown in Fig.~\ref{iwsEffi}. Since the central detector size is larger than that of the Daya Bay detector, the large surface area of stainless steel tank might have a big effect on light transmission. We carried out a simulation and found that there is no major impact on detector performance if the surface of the steel tank is covered with reflective Tyvek. \begin{figure}[htb] \begin{center} \includegraphics[width=6cm]{VetoDetector/Figures/JUNO_vetoEffi.png}~ \caption[Muon detection efficiency vs Number of PMT threshold]{Muon detection efficiency vs threshold in number of PMT (1,600 PMTs, with a number of PMTs threshold at 20, the efficiency is larger than 97\%).} \label{junoEffi} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/IWS-effi.png} \caption[Muon efficiency vs time]{Detection efficiency (vs time) for long track muons going through antineutrino detector of Daya Bay. } \label{iwsEffi} \end{center} \end{figure} The challenge for the JUNO water Cherenkov detector is that we not only need to tag the muon with high efficiency but also want to reconstruct the tracks of those muons. Ideally full photo-coverage on the water pool surface would be the best choice. Yet the budget constraint limit the number of PMTs available for the water pool. Considering the directional nature of the Cherenkov light and that all muons are coming from outside, it would be natural to have PMTs installed on the central detector. Only few clipping muons at certain particular incident angles would not have direct photons projected on the central detector if not utilizing reflective water pool surface. Some of these clipping muons would be detected by the top tracker. Most of these clipping muons are further away from central detector ($>$ 2 m) which would less likely to produce cosmogenic backgrounds in the central detector. The possibilities of different configurations of the central detector as well as the uncertainty to utilize the outer surface of the central detector, require careful study on PMT arrangements, placement as well as the number of PMTs in the water pool. Large photo coverage enhances the muon detection but also drives the cost high. Therefore, it needs study of utilizing larger size PMTs or wavelength shifting plates around the PMTs as done in the Super-Kamiokande experiment \cite{SK:detector2003} and find the optimal solution. Since the muon track reconstruction is important for $^9$Li/$^8$He background reduction, we would consider to increase the water pool muon track reconstruction capability for the purpose of background reduction. For JUNO, the shower muons will induce non-negligible dead time if the central detector can't reconstruct the shower muon tracks due to detector saturation. If the water Cherenkov detector can help reconstruct those muon tracks, it would reduce the central detector dead time. Reflective surfaces can enhance the Cherenkov photon detection, and therefore help tag muons. But the reflection would smear out the position information. We can use more PMTs to build optically segmented water Cherenkov detectors to improve muon reconstruction resolution. But these segmented detectors could normally provide only one point along the muon track, therefore it might need to have multiple layers of these kind of detectors installed to reconstruct the muon tracks, which would pose engineering as well as installation challenges. In the mean time, developing better algorithms for track reconstruction utilizing both the PMT charge and timing information would be useful. For this purpose, a better water pool PMT calibration system is essential to not only calibrate the PMT gains but also the PMT timing. Now, all these options are under study. The simplest option would be to have a single water Cherenkov detector by installing PMTs on the surface of the water pool. There are other options, such as having multiple layers of PMTs or modular PMT boxes installed. Since there is not much space in the water pool and there will be structures for the central detector in the pool, therefore from the engineering point of view, a simple water Cherenkov detector is desired. A simulation was performed to study the muon reconstruction capability of the simple water Cherenkov detector. PMTs are uniformly distributed on the surface of the water pool as well as the outer surface of the central detector (Figure~\ref{fig:watertank_config1}). Of the 2,000 PMTs, 200 PMTs are on the top facing toward the outside of the pool, 200 PMTs are at the bottom and 800 PMTs are on the barrel all facing inward. Another 800 PMTs are on the outer surface of the central detector. This gives roughly 2.5~m distance between PMTs. All the surfaces are covered by Tyvek with reflectivity of 95\%. Figure~\ref{fig:watertank_config2} shows various muon tracks with points on different kinds of surface where the PMTs are installed \begin{figure}[htpb] \centering \includegraphics[width=0.36\textwidth]{VetoDetector/Figures/watertank_config1.png} \caption{Simple water Cherenkov detector option. Uniformed distribution of PMTs. PMTs at the top of water pool are facing outward. PMTs at the bottom and on the barrel are facing inward.} \label{fig:watertank_config1} \end{figure} \begin{figure}[htpb] \centering \includegraphics[width=0.42\textwidth]{VetoDetector/Figures/watertank_config2.png} \caption{Various muon tracks with reconstructed points at pool surface or central detector. } \label{fig:watertank_config2} \end{figure} \begin{figure}[htpb] \centering \includegraphics[height=2.in, width=0.25\textwidth]{VetoDetector/Figures/watertank_CDresolX.png} \includegraphics[height=2.in, width=0.25\textwidth]{VetoDetector/Figures/watertank_CDresolY.png} \includegraphics[height=2.in, width=0.25\textwidth]{VetoDetector/Figures/watertank_CDresolZ.png} \caption{Resolution of the reconstructed points on the central detector} \label{fig:watertank_CDresol} \end{figure} \begin{table}[htdp] \caption{Resolution of reconstructed points on the muon tracks.} \begin{center} \begin{tabular}{|c|c|c|c|} \hline & X & Y & Z \\ \hline \multicolumn{4}{|c|}{Points on the top of the pool}\\ \hline Resolution & $\sim$1.5 m & $\sim$1.5 m & \\ \hline \hline \multicolumn{4}{|c|}{Points on the bottom of the pool}\\ \hline Resolution & $\sim$0.8 m & $\sim$0.75 m & \\ \hline \hline \multicolumn{4}{|c|}{Points on the barrel of the pool}\\ \hline Resolution & $\sim$0.3 m & $\sim$0.4 m & $\sim$0.8 m \\ \hline \hline \multicolumn{4}{|c|}{Points on the surface of CD}\\ \hline Resolution & $\sim$0.4 m & $\sim$0.4 m & $\sim$0.4 m \\ \hline \end{tabular} \end{center} \label{tab:watertank_muonpoints} \end{table} The muon tracks are represented by the interceptive points on either the top, barrel, bottom of the pool or on the central detector. The points are obtained using charge center method. The reconstructed points on the barrel and bottom are the exit points of the muon track. And the reconstructed points on the top and central detector are the entry points. Table \ref{tab:watertank_muonpoints} shows the resolution of the reconstructed points at various location. The reconstructed points on the top are poor due to thin water the muons going through and thus having fewer photons for PMTs. The reconstructed points on the barrel are good for X and Y and average with a bias of $\sim$0.5~m for Z, this could be the results of all the tracks are going downward. The reconstructed points are very good on the central detector for all coordinates. It might be possible to use the reconstructed points from top tracker for muons passing through the top of the pool. It is essential to have good reconstructed muon track in central detector. Though the central detector can be excellent muon tracker when the muon path is close to the center of the central detector, the muon track reconstruction gets worse when the muon is close to edge of the central detector. Therefore, the presence of good reconstructed points from the water Cherenkov detector would definitely help and cross check with the central detector muon track reconstruction. Optimizing the number of PMTs, the PMT arrangement as well as taking into account the PMT hit timing information in reconstruction definitely warrant further study. Of all the muons, about 10\% are multiple muons. For double muons, it was estimated that the central detector could distinguish the muons with distance greater than 4~m. Water Cherenkov detector with good resolution of muon entry points on the central detector has the potential to be able to separate multiple muon tracks by looking for clusters of hit PMTs. This would be helpful to reduce the dead time caused by multiple muons. \subsubsection{Tyvek Reflector Film} As proved in the Daya Bay experiment, reflective surface is effective in tagging muons without larger number of PMTs. Tyvek film as a reflective material has the merits of high reflectivity and stable performance, therefore it is widely used in water Cherenkov detectors. The Daya Bay experiment uses Tyvek film as to make reflective surface in the water Cherenkov detector. In JUNO, the Tyvek reflective film could be used to cover the inner wall of the pool and the outer surface of the central detector. The Cherenkov photons produced in the pool can be detected by the PMTs after multiple reflections to improve the detection efficiency. Figure~\ref{MakeTyvek} shows the process of making reflective Tyvek. Tyvek pieces with a width of 1 meter and short lengths can be welded together to form a larger piece of reflector for large detector installation. The performance of various Tyvek reflective films is shown in Fig.~\ref{TvyekReflect}. \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/Tyvek_made.png} \caption[Production of the Tyvek reflector]{The process of making the tyvek reflector.} \label{MakeTyvek} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/Tyvek_reflectivity.png} \caption[Reflectivity of different types of Tyvek]{Reflectivity of different type Tyvek films .} \label{TvyekReflect} \end{center} \end{figure} The reflectivity of the Tyvek depends on the thickness of Tyvek. The Daya Bay experiment uses two layers of thick Tyvek (1082 D) with PE between them to form a multi-layer Tyvek. The reflectivity of the multilayer Tyvek is shown as the red points in Fig.~\ref{TvyekReflect}. The reflectivity is >95\% when the wavelength > 300~nm. \subsubsection{Calibration system} \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/LED_diffuse_ball.png} \caption[LED]{LED and diffuse ball for PMT calibration.} \label{LED} \end{center} \end{figure} The gain and the timing calibration of the PMTs will be monitored by a LED system. No radioactive sources will be required. As Fig.~\ref{LED} shows, we will use LED or diffuse ball by LED for the PMT calibration. LED flashing is triggered by pluse generator. The diffuse balls could be put at differnt positions of water pool. So every PMT can receive the photons from diffuse ball. PMTs gain and timing calibration could be done once a week. \subsection{Top Tracker} The JUNO cosmic muon tracker will help enormously to evaluate the contamination of the cosmogenic background to the signal. The OPERA detector has to be dismounted soon and the OPERA TT ~\cite{Adam:2007ex}($\sim4.5$M EUR) will become available by mid--2016. It will be placed on top of the JUNO water Cherenkov detector to be used as a cosmic muon tracker. \subsubsection{OPERA Target Tracker} The TT is a plastic scintillating detector which had several critical roles in OPERA: it was used to trigger the neutrino events, to identify the brick in which the neutrino interaction took place and to reconstruct muons therefore reducing the Charm background. Its performances well met the expectations: only limited aging was observed over the 2007 - 2012 data taking period, and the detector understanding was well demonstrated by showing very good data/MC agreement in particular on the muon identification and energy reconstruction~\cite{Agafonova:2011zz}. The TT is composed of 62 walls each with a sensitive area of 6.7$\times$6.7~m$^2$. Each wall is formed by four vertical ($x$) and four horizontal ($y$) modules (Fig.~\ref{wall_schematic}). The TT module is composed of 64 scintillating strips, 6.7~m length and 26.4~mm wide. Each strip is read on both sides by a Hamamatsu 64-channel multi-anode PMT. The total surface which could be covered by the 62 x-y walls is 2783~m$^2$. All TT walls in OPERA are hanged by the top part of the detector and are thus in vertical position (Fig.~\ref{hanging}). In the case of JUNO, all TT modules have to be placed in horizontal position, in which case more supportive mechanical structure is needed. \begin{figure}[hbt] \begin{minipage}[b]{.45\linewidth} \centering \includegraphics[width=7cm]{VetoDetector/Figures/wall_schematic.pdf} \caption{\small Schematic view of a plastic scintillator strip wall.} \label{wall_schematic} \end{minipage} \hspace{1.cm} \begin{minipage}[b]{.45\linewidth} \centering \includegraphics[width=7cm]{VetoDetector/Figures/hanging.pdf} \caption{\small Target Tracker walls hanging in between two brick walls inside the OPERA detector.} \label{hanging} \end{minipage} \end{figure} The particle detection principle used by the TT is depicted by Fig.~\ref{principle}. The scintillator strips have been produced by extrusion, with a $TiO_2$ co-extruded reflective and diffusing coating for better light collection. A long groove running on the whole length and at the center of the scintillating strips, houses the wavelength shifting (WLS) fiber which is glued inside the groove using a high transparency glue. This technology is very reliable due to the robustness of its components. Delicate elements, like electronics and PMTs are located outside the sensitive area where they are accessible (Fig.~\ref{endcap_schematic}). \begin{figure}[hbt] \begin{minipage}{.45\linewidth} \centering \includegraphics[width=7cm]{VetoDetector/Figures/wls_en.pdf} \caption{\small Particle detection principle in a scintillating strip.}\label{principle} \end{minipage} \hspace{1.cm} \begin{minipage}{.45\linewidth} \centering \includegraphics[width=7cm]{VetoDetector/Figures/endcap_schematic.pdf} \caption{\small Schematic view of an end--cap of a scintillator strip module.}\label{endcap_schematic} \end{minipage} \end{figure} Figure~\ref{ttelectronics} presents details about the end-caps of the TT modules hosting all electronics, front end and acquisition. These electronics are composed of: \begin{itemize} \item a front end card located just behind the multi-anode Hamamatsu PMT, hosting the two OPERA ROC chips (32 channels each, BiCMOS 0.8 microns)~\cite{Lucotte:2004mi}, \item an acquisition card (DAQ)~\cite{Marteau:2009ct}, also hosting an ADC for charge digitization, \item a light injection card located on the DAQ card, \item two LEDs able to inject light at the level of the WLS fibers near the PMT and driven by the light injection card, this system is used to regularly calibrate the detector (PMT gain, stability etc.), \item an ISEG High Voltage module located on the DAQ card. \end{itemize} \begin{figure}[htb] \centering \includegraphics[width=12cm]{VetoDetector/Figures/electronics.pdf} \caption{TT electronics including DAQ card. All elements in red will be replaced.} \label{ttelectronics} \end{figure} The TT electronics record the triggered channels and their charge, thanks to the OPERA-ROC chip. A schematic view of this chip is given by Fig.~\ref{operaroc}. Each channel has a low noise variable gain preamplifier that feeds both a trigger and a charge measurement arms. The adjustable gain allows an equalization of all PMT channel gains which can vary from channel to channel by a factor 3. The auto-trigger (lower part) includes a fast shaper followed by a comparator. The trigger decision is provided by the logical ``OR" of all 32 comparator outputs, with a threshold set externally. A mask register allows disabling externally any malfunctioning channel. The charge measurement arm (upper part) consists of a slow shaper followed by a Track \& Hold buffer. Upon a trigger decision, charges are stored in 2~pF capacitors and the 32 channels outputs are readout sequentially at a 5~MHz frequency, in a period of 6.4~$\mu$s. All charges are digitized by an external ADC (12--bit AD9220) placed on the DAQ cards. \begin{figure}[htb] \centering \includegraphics[width=12cm]{VetoDetector/Figures/opera_roc.pdf} \caption{Architecture of a single channel of the OPERA--ROC chip.} \label{operaroc} \end{figure} {\bf TT dismounting in Gran Sasso}\ The TT dismounting in Gran Sasso underground laboratory will start in summer 2015 (first OPERA Super Module) and will end in Spring 2016 (second OPERA Super Module). The OPERA detector dismounting and cost sharing among the funding agencies are defined in a special MoU. The cost of the TT dismounting up to its storage area is part of this MoU and thus it is not considered in the IN2P3 JUNO requests. All TT modules will be stored in Gran Sasso in 10 containers before sending them to China. The shipping of the containers will be done when storage halls are ready near the JUNO underground laboratory. This is expected to take place in 2016. In all the cases the TT will not be mounted on top of JUNO detector before 2019. This implies that the TT containers will be stored somewhere for about three years. The best place, in order to avoid big temperature variations and scintillator aging, is the Gran Sasso underground laboratory. Negotiations are engaged with LNGS on this possibility. If this is not possible, the TT will be temporarily stored in a hall in the surface LNGS laboratory waiting to be shipped to China. {\bf TT in JUNO}\ This muon tracker, called now Top Tracker (again TT), will be needed in JUNO in order to well study the cosmogenic background production. The most dangerous background is induced by cosmic muons generating $^9$Li and $^8$He unstable elements, and fast neutrons, which could fake an IBD interaction inside the central detector. Figure~\ref{ttnoise} presents schematically the most important noise configurations. The first two, (a) and (b), mainly concern $^9$Li and $^8$He production directly in the central detector (a) and in the veto water pool (b) while the last one (c) concerns the neutron production in the surrounding rock. \begin{figure}[htb] \centering \includegraphics[width=\textwidth]{VetoDetector/Figures/Cosmogenic_background.png} \caption{Configurations considered for the induced background.} \label{ttnoise} \end{figure} The surface on the top of the JUNO detector is of the order of $40\times 40$~m$^2$. The total surface that the TT could cover depends on the number of superimposed $x-y$ layers (composed by consecutive TT walls). In any case it will never be able to cover the entire surface, half of cosmic muons crossing the JUNO detector can pass by the sides. The number of layers will depend on several parameters: \begin{itemize} \item the minimum statistics needed to well measure the cosmogenic background ($^9$Li and $^8$He production), \item the muon tracking accuracy needed up to the bottom part of the central detector (of the order of the one induced by the multiple scattering), \item the noise rate reduction using coincidences (affordable by the acquisition system), \item the rate reduction of fake tracks. \end{itemize} First studies show that 3 to 4 layers will be needed that means that the total surface on top of JUNO liquid scintillator detector will be 690~m$^2$ to 920~m$^2$, respectively. The noise rate in JUNO underground laboratory is expected to be significantly higher than that in Gran Sasso. During OPERA operation the noise rate for 1/3 p.e. threshold was of the order of 10~Hz. This low rate was mainly due to the low PMT dark current (2.5~Hz) and the fact that the TT walls were shielded by the OPERA lead/emulsion bricks. Before the insertion of the bricks the noise rate was of the order of 25~Hz/channel. In JUNO this rate is expected to be significantly higher for several reasons: \begin{itemize} \item The JUNO overburden is lower (2,000~m.w.e.) than for Gran Sasso (4,200~m.w.e.) implying more cosmic muons crossing the detector. \item There will be no shielding between the TT walls. \item The radioactivity of the rock, measured by the Daya Bay experiment 200~km away, is expected to be significantly higher than the one observed in the Gran Sasso underground laboratory. \item There will be no concrete on the walls surrounding the JUNO underground laboratory (contrary to the Gran Sasso underground laboratory), concrete that could absorb part of the $\gamma$'s emitted by the rock. \end{itemize} In the case of 4-layers cover, the configurations of Fig.~\ref{coverage} have been considered. The first configuration has the advantage compared to the others of covering the maximum surface of the central detector region. The second one is more representative of all detector configurations and noise measurements could be extrapolated to the total detector top surface. The third configuration is best for rock muon production estimation but is not very representative of the rest of the detector. Finally, as baseline it has been defined the second configuration. In all the cases, the TT has to well cover the chimney region (central region not covered by the veto water Cherenkov) from where radioactive sources will be introduced to calibrate the detector. \begin{figure}[htb] \centering \includegraphics[width=\textwidth]{VetoDetector/Figures/TT_coverage.pdf} \caption{Configurations considered for the Top Tracker coverage.} \label{coverage} \end{figure} {\bf Noise rate estimations}\ Table~\ref{noise} presents the rock noise rate for one $x$ layer on the top, for one $y$ layer just below and for the case of one $x-y$ coincidence of these two layers ($x-y$ correlated coincidence). For the $x-y$ correlated coincidences the $x$ and $y$ hits must come from the same radioactivity event while for the $x-y$ accidental coincidences come from different events occurring during 200~ns time--window. The slight noise reduction from the $x$ layer to the $y$ one is due to the fact that the top layer shields a bit the layer above. A strong noise rate reduction is observed when an $x-y$ coincidence is required. For all the cases a threshold of one and two photoelectrons is considered. The results are shown considering no concrete on the JUNO cavern walls and with 10~cm concrete. The presence of concrete would reduce the radioactivity noise by about a factor of 4. \begin{table}[htdp] \caption{Noise rate for several TT conditions. (a) is under 1 p.e. threshold and (b) is under 2 p.e. threshold. (1)no concrete; (2)10~cm concrete.} \begin{center} \begin{tabular}{|c|c|c|c|c|}\hline Config. & \multicolumn{2}{c|}{Rate (Hz/m$^2$),(a)} & \multicolumn{2}{c|}{Rate (Hz/m$^2$),(b)} \\ \cline{2-5} & (1) & (2) & (1) & (2) \\ \hline $1-x$ layer (top) & 6800 & 1780 & 2700 & 710 \\ \hline $1-y$ layer (bottom) & 6500 & 1780 & 1200 & 310 \\ \hline $x-y$ layer (correlated) & 540 & 150 & 29 & 7 \\ \hline $x-y$ layer (accidentals) & 400 & 29 & 29 & 2 \\ \hline \end{tabular} \end{center} \label{noise} \end{table} The noise rate per channel is not the only parameter to take care during the design of the new TT configuration. The number of $x-y$ layers will mainly be determined by the number of ``fake'' muons (virtual tracks due to many hits in a time--window of 200~ns) produced by the environmental radioactivity. Table~\ref{fake} presents the ``fake'' muons' rate estimation considering two and three TT $x-y$ layers. For the case of the two layers, the distance between is considered to be 2~m, while for the case of three layers the third layer is inserted at the middle between the two considered layers. It can be observed that the rate is considerably reduced for the case of three layers compared to the one of only two layers. From this table the case of two layers can be excluded. The rate being of the order of few Hz and less for the case of three layers, the case of four layers initially considered probably will not be necessary. A fourth layer would provide some redundancy in case of inefficiencies due to scintillator aging, dead time or dead zones. It has also to be noted that each ``fake'' muon will bias the cosmogenic $^9$Li and $^8$He production and thus this rate has to be very low. \begin{table}[htdp] \caption{Rate of ``fake'' tracks for several TT conditions.(a) is under 1 p.e. threshold and (b) is under 2 p.e. threshold. (1)no concrete; (2)10~cm concrete.} \begin{center} \begin{tabular}{|c|c|c|c|c|}\hline Config. & \multicolumn{2}{c|}{Fake muon rate (Hz), (a)} & \multicolumn{2}{c|}{Fake muon rate (Hz), (b)} \\ \cline{2-5} & (1) & (2) & (1) & (2) \\ \hline 2 TT $x-y$ layers & 340,000 & 12,500 & 1300 & 31 \\ \hline 3 TT $x-y$ layers & 2.6 & 0.02 & $6\times 10^{-4}$ & $2\times 10^{-6}$ \\ \hline \end{tabular} \end{center} \label{fake} \end{table} In order to study the background induced by $^9$Li, $^8$He and fast neutrons, enough statistics has to be collected using the TT. To estimate the background rate induced by these three source for the three configurations of Fig.~\ref{ttnoise}, the following parameters have been considered: \begin{itemize} \item $E_\mu$: the muon energy, \item $L_\mu$: the average track length of muons in the liquid scintillator (case a), \item $R_\mu$: the muon rate in the liquid scintillator (case a), \item $F_{nCap}$: the neutron capture ratio (1 for JUNO, case a), \item $N$: the muon number (case b), \item $L_{att}$: the attenuation length of $^9$Li/$^8$He in water pool (0.5~m, case b), \item $S_{wp}$: surface of water pool (case c). \end{itemize} The following formulas give the background rate of the three considered configurations: \begin{itemize} \item ${R_{Li,He}} \propto E_\mu ^{0.74} \cdot {L_\mu } \cdot {R_\mu } \cdot {f_{nCap}}$ for case (a), \item ${R_{Li,He}} \propto \frac{{\sum\limits_{i = 1}^N {E_{{\mu ^i}}^{0.74}(\sum\limits_{j = 1}^M {L_{{\mu ^i}}^j{e^{ - \frac{{{d_j}}}{{{L_{att}}}}}}} )} {f_{nCap}}}}{N} \cdot {R_{wp}}$ for case (b), \item $R_n^{rock} \propto E_{{\mu ^i}}^{0.74} \cdot {R_\mu } \cdot {S_{wp}} \cdot {f_{nCap}}$ for case (c) \end{itemize} Table~\ref{casea} presents the expected background rate from $^9$Li and $^8$He for all three TT coverage configurations of Fig.~\ref{coverage}. From these results, the TT configuration (4XY, ``O'') can be excluded because of the poor rate observed of the $^9$Li and $^8$He induced background. It has to be noted that in absence of the TT the $^9$Li and $^8$He rate per day is of the order of 90 to be compared to $\sim$50~IBD interactions induced by nuclear reactors. Obviously, this background has to be reduced or at least its rate must be well known. \begin{table}[htdp] \caption{Noise rate induced by $^9$Li and $^8$He for case (a) of Fig.~\ref{ttnoise}.} \begin{center} \begin{tabular}{|c|c|c|c|}\hline & $L_\mu$ (m) & $R_\mu$ (Hz) & $^9$Li/$^8$He rate/day \\ \hline all muons & 22.5 & 3.5 & 90 \\ TT (4XY, Mid) & 23.5 & 0.94 & 27 \\ TT (4XY, Rtg) & 23.4 & 0.80 & 23 \\ TT (4XY, ``O") & 21.8 & 0.30 & 9 \\ \hline \end{tabular} \end{center} \label{casea} \end{table} Table~\ref{caseb} presents the expected background rate from $^9$Li and $^8$He for all three TT coverage configurations of Fig.~\ref{coverage} where the muons only cross the water pool without passing through the central detector. As already said, the TT configuration (4XY, Rtg) is the most representative of all parts of the detector and, not seen a significant difference between this configuration and the configuration (4XY, Mid), the configuration (4XY, Rtg) has been considered as baseline. \begin{table}[htdp] \caption{Noise rate induced by $^9$Li and $^8$He for case (b) of Fig.~\ref{ttnoise}.} \begin{center} \begin{tabular}{|c|c|c|c|}\hline & $^9$Li/$^8$He rate/day & Back./signal (\%) \\ \hline all muons & 1.2 & 3.0 \\ TT (4XY, Mid) & 0.35 & 0.9 \\ TT (4XY, Rtg) & 0.3 & 0.7 \\ TT (4XY, ``O") & 0.16 & 0.4 \\ \hline \end{tabular} \end{center} \label{caseb} \end{table} Table~\ref{casec} presents the expected background rate from fast neutrons obtained by extrapolations for Daya Bay observations for case (c) of Fig.~\ref{ttnoise} and for the baseline configuration of TT (4XY, Rtg) of Fig.~\ref{coverage}. It is expected to have a fast neutron induced background rate of the order of 0.13 per day which represents 0.3\% of the expected signal. \begin{table}[htdp] \caption{Expected noise rate from fast neutrons for case (c) of Fig.~\ref{ttnoise} and for the configuration of TT (4XY, Rtg) of Fig.~\ref{coverage}.} \begin{center} \begin{tabular}{|c|c|c|c|}\hline & Daya Bay & Far detector & JUNO \\ \hline $E_\mu$ (GeV) & 57 & 137 & 215 \\ $R_\mu$ (Hz) & 1.2 & 0.055 & 0.003 \\ $S_{wp}$ (m$^2$) & 724 & 1032 & 3740 \\ $f_{nCap}$ (\%) & 45 & 45 & 100 \\ Rock neutron bkg. rate/day & 1.7 & 0.22 & 0.13 (B/S$\sim$0.3\%)\\ \hline \end{tabular} \end{center} \label{casec} \end{table} The noise measured with the help of the TT will be extrapolated to the whole detector and will be introduced in the simulations in order to well estimated the related systematic errors. {\bf Modifications to the Target Tracker}\ The present OPERA TT acquisition system can afford up to 20--25~Hz trigger rate per channel. For higher rates (as those expected in JUNO) this system would have severe limitations inducing huge dead time and thus significantly reducing the TT efficiency. Even if the $x-y$ coincidences already reduce the noise rate (see above), the acquisition system specifically developed for OPERA needs will not be able to be used for the TT installed in JUNO cavern. The high noise rate expected in JUNO obliges the replacement of the acquisition system (DAQ) by a new one. By the same way, the front end electronics based on OPERA-ROC~\cite{Lucotte:2004mi} chip will be replaced. Indeed, this chip is now obsolete and not enough spares are available. The new generation chip MAROC3~\cite{Blin:2010tsa} (AMS SiGe 0.35~$\mu$m technology, developed by IN2P3 Omega laboratory) disposing of more functionalities can replace the OPERA-ROC chip. On OPERA-ROC the word registering the triggered channels was not functioning properly while on MAROC3 this has been corrected. This gives the possibility to only read the triggered channels and not all of them as it was done in OPERA TT, thereby gaining in readout speed. The MAROC3 chip has a similar architecture (Fig.~\ref{maroc}) than the OPERA--ROC (fast part for triggering and slow one for charge measurement). The difference is that in MAROC3 the ADC is already integrated in the chip while for OPERA--ROC the ADC was externally implemented on the DAQ cards. The MAROC3 chip has two ``OR''s, one could be used for the timestamp while the second one could be used for fast coincidences with other sensors. This would allow to do fast coincidences with at least $x$ and $y$ strips to reduce the DAQ trigger rate. For all coincidences among TT sensors a dedicate electronics card has to be developed receiving the ``OR'' of all sensors and making coincidences. The timestamp can be provided by a special clock card to be developed. OPERA institutes participating now in JUNO have expressed a strong interest to develop the DAQ cards which will be located just behind the multi--anode PMTs, as done in OPERA. The front end card where the MAROC3 chip will be located could be produced by the same institutes, which would also have the responsibility to check and characterize all MAROC3 chips produced for this application (992 plus spares). These institutes could also develop the clock card and the coincidence card. The drawings for the installation of the TT in JUNO will be produced by IN2P3 and IHEP--Beijing. It is already agreed with IHEP that the TT supporting structure will be financed and produced by Chinese institutes. This structure, very different from the one used in OPERA due to the fact that in JUNO all modules will be in horizontal position instead of vertical position in OPERA, has to be drawn with care and in a close collaboration between IN2P3 and IHEP. A special care has also to be taken in order to leave accessibility to all electronics of all TT modules. \begin{figure}[htb] \centering \includegraphics[width=\textwidth]{VetoDetector/Figures/MAROC_architecture.pdf} \caption{Architecture of the MAROC3 chip.} \label{maroc} \end{figure} \section{Water System} For the JUNO water Cherenkov detector, it requires ultrapure water for high muon detection efficiency. There will be different kinds of materials submerged in the water, including stainless steel, Tyvek, PMT glass, cables, etc. The complexity of the underground environment makes it difficult to seal the pool completely and almost impossible to passively keep the water quality good for a long period of time. Therefore, it is necessary to build a reliable ultrapure water production, purification and circulation system. Another important function of the water system is to keep the overall detector temperature stable. The stability of the temperature of the central detector is critical for the entire experiment. JUNO is located in southern China, and the temperature of the surrounding rocks can reach 32 degrees Celsius. We intend to maintain the temperature of the central detector at 20 degrees Celsius. Water is a good conductor of heat. Lowering the temperature of the water flowing into the pool is a natural choice to offset the heat from the surrounding rocks. The Daya Bay experiment has 1,300 tons of water in the near experimental hall water pool and 2,000 tons of water in the far experimental hall water pool. The water production rate is 8~tons/hour and the purification rate is 5$\sim$8 tons/hour. It can circulate one volume of water in two weeks. The water Cherenkov detector's muon detection efficiency is high to 99.9\% and its long-term performance is stable. The designed water purification system satisfies the experiment requirements~\cite{Wilhelmi:2014rwa}. \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/WaterSystemFlowRate.png} \caption[Flow rate vs one cycle time]{Flow rate vs one cycle time.} \label{flowRate} \end{center} \end{figure} The JUNO water Cherenkov detector will have 20,000 - 30,000 tons of ultrapure water. Based on the experience of Daya Bay experiment, circulating one volume of water in two weeks requires a flow rate of about 80~tons/hour at JUNO. With this flow rate, we can keep good water quality as well as maintain pool water at stable temperature. Fig.~\ref{flowRate} shows the number of days to circulate one volume of water as function of the flow rates of the proposed water system for the two candidate designs of the central detector. The red line indicates the flow rate at 80~tons/hour. It would take about 16 to 20 days to circulate one volume of water and this would satisfy the experiment requirements. In the proposed design, we divide the water system into two parts. One part is for the production of ultrapure water and the other part is for the circulation/purification of water. For the water production system, one part of the system would be on the surface ground and the other part of the system would be underground. The reason for this design is that there is a strict requirement of waste water rate to be less than 5~tons/hour at 500 meters underground. Therefore, we have the part of the water production system which generates the most waste water on the surface ground. \subsection{Water Production on Surface} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/WaterProductionGround.png} \caption[General view of water production system on ground]{General view of water production system on surface ground.} \label{WaterProdGround} \end{center} \end{figure} As Fig.~\ref{WaterProdGround} shows, water flows through the disk filter first, which can block solid sediments and other large particles from flowing into the raw tank before goes through the ultrafiltration hosts. The ultrafiltration hosts can filter out small particles of insoluble material in the water before it goes into ultrafiltration tank. The water then passes through a reverse osmosis (RO) system, which can remove most of the ions in the water. Finally the water flows into the primary pure water tank. After the RO process, a third of the original water is produced as waste water with high ion concentration at a rate of 15~tons/hour. The waste water is of no harm to the environment and can be directly drained. \subsection{Water Production Underground} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/WaterProductionUnderGround.png} \caption[General view of water production system underground]{General view of water production system underground.} \label{WaterProdUnderGround} \end{center} \end{figure} From primary pure water tank on the surface ground, the water is pumped through pipes at about 85~tons/hour going down a height of 500 meters into the underground raw water tank. From there the water will go through another RO system as well as other water treatment systems, like reduction of Total Organic Carbons (TOC) and Electrodeionization (EDI) systems. The resulted ultrapure water with resistance greater than 16~MOhm flows into experiment hall at a rate of 80~tons/hour. \subsection{Water Circulation/Purification Underground} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/CirculationUnderground.png} \caption[General view of water circulation/purification system on ground]{General view of water circulation/purification system on ground.} \label{WaterCircuUnderGround} \end{center} \end{figure} We would start the circulation process after the pool is fully filled with water. The water will be drawn drained from the pool at 80~tons/hour. The water will first go through the heat/cooling equipment . Then the water will flow through the TOC system (to remove organic carbon), the polishing mixed bed (to remove ions in the water to improve water quality), UV sterilization system, membrane filters, the degassing apparatus (to remove oxygen and other gases in water), and finally a cooling device (to control the water temperature). In the end, the purified water will return to the pool. For the whole water production, purification and circulation system, we will use a water control system to automate the running of the system. The people on duty will only need to replace filters and other supplies on a regular basis. \section{Geomagnetic Field Shielding System} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/jiangmen_EarthMagnet.png} \caption[Earth magnetic in Jiangmen]{Earth magentic in Jiangmen.} \label{JUNO_filed} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/PMT_CEvsMagnet.png} \caption[PMT collection efficiency vs the external magnetic field]{PMT collection efficiency vs the external magnetic field.} \label{CEvsField} \end{center} \end{figure} At the experimental site, the horizontal component of the geomagnetic field is about 40~$\mu$T and vertical component is about 24~$\mu$T. Since JUNO will use 20 inch PMTs as the main device for central detector, the geomagnetic field will have a big effect on large size PMT. It will deteriorate the whole detector performance. Figure~\ref{CEvsField} shows the relationship between the PMT collection efficiency and the external magnetic field intensity. We need to establish an effective magnetic shielding system to ensure good performance of central detector. \begin{figure}[htb] \begin{center} \includegraphics[width=8cm]{VetoDetector/Figures/OneGroupCoils.png} \caption[One group earth magnetic shield system]{One group earth magnetic shield system.} \label{OneGroupCoil} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/MagDis_onegroup.PNG} \caption[Magnetic field distribution with one group system]{Magnetic field distribution with one group system.} \label{MagDis} \end{center} \end{figure} We intend to ensure that a spherical region with a radius of 37.5~m would be well shielded from geomagnetic field. We plan to use coils with current flowing through them to compensate the geomagnetic field. The residual field intensity could be reduced to be below 10~$\mu$T. The baseline design is of one coil system. We can perform accurate measurement of the geomagnetic field to determine its main direction (mapping the geomagnetic field in the pool) and then do compensation along its main direction. As Fig.~\ref{OneGroupCoil} shows, this system is easy to control. Figure~\ref{MagDis} is the magnetic field distribution in the sphere from one coil system. It can have a good shielding. The uniformity within the 37.5~m volume is 1.8\%. The PMTs in central detector will not be effected by the Earth's magnetic field after compensation. Careful selection of non-magnetized materials used in the detector for parts such as the supporting structure for PMTs as well as stainless steel for the detector, etc. is also important. They could cause an additional local magnetic field in additional to the Earth's field, which could affect the performance of the PMTs. The compensation coils would be installed on the central detector which means the PMTs of the water Cherenkov detector would still be subject to the Earth's magnetic field plus some additional field generated from the coils. Therefore, it needs to have the pool PMTs shielded against magnetic field with passive magnetic shield like the one used in the Daya Bay experiment \cite{DeVore:2014nim}. \section{Mechanical Structures and Installation} \subsection{Support Structure over the Pool} The tracking detectors over the pool are divided into two groups, each consisting of several layers of detectors stacked together. One group is placed on the steel bridge, while the other is put three meters above the first group. Hence a two-story bridge is needed, as shown in Fig.~\ref{SupportStr}. The parameters of the two-story bridge are as follows: bridge width is between 17~m and 20~m and the distance between two layers of the bridge is 3 meters. The support structure has three functions: (1) supporting the tracking detectors; (2) supporting the calibration systems of the central detector; (3) possibly supporting the electronics rooms. The diameter of pool is 42 meters and the supports of bridge are located at both ends of the bridge. Considering the height limit of the underground space and the structure stiffness, a box structure is adopted for the bridge. Two columns of tracking detectors, whose width is about 15~m, are placed on the bridge. Considering the passageway and the width of the electronics room (about 3 meters), the bridge width is between 17~m and 20~m. The load of the bridge includes the weight of the central detector cable and electronics rooms (about 100 tons), weight of the tracking detectors (about 70 tons), weight of the electronics room (20 tons) and weight of the supports of the tracking detectors. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/SupportStructure.png} \caption[The top support structure]{The top support structure.} \label{SupportStr} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{VetoDetector/Figures/PMT.png} \caption[PMT support structure of DayaBay]{PMT support structure of DayaBay.} \label{PMTstr} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/cover.png} \caption[The cover design of Daya Bay]{The cover design of Daya Bay. a) gastight zipper b) Th connection between pool cover and pool dege} \label{DYBcover} \end{center} \end{figure} \begin{figure}[htb] \begin{center} \includegraphics[width=10cm]{VetoDetector/Figures/wire.png} \caption[The wire design option for cover installation of Daya Bay experiment]{The wire design option for cover installation of Daya Bay experiment.} \label{cover1} \end{center} \end{figure} \subsection{Support Structure for PMTs} The support structure of PMT adopts the design for the Daya Bay veto PMTs, as shown in Fig.~\ref{PMTstr}. The PMT support structure is then attached to a steel-frame fixed on the waterpool wall. The design of the steel-frame structure will be similar to that for the central detector. The tyvek film will also be fixed to the steel frame. \subsection{Sealing of the Pool} A black rubber cloth served as the pool cover in the Daya Bay experiment. The pool cover is connected to the pool edge through an air-tight zipper which shields the pool from the external air and light, as shown in Fig.~\ref{DYBcover}. With the seal zipper, the pool is easily accessible for routine maintenance, such as replacing the liquidometer. We will adopt the same technique for JUNO's pool seal. As the JUNO pool is very large with a 40-meter span, it is difficult to unfold the black rubber cloth. Therefore, we propose two solutions: the first is the tight-wire scheme used in the Daya Bay experiment; the second is a slide-guide scheme. The two schemes are similiar. For the slide-guide scheme shown in Fig.~\ref{DYBcover}, the cloth is unfolded along the guide, instead of the tight-wire in the tight-wire scheme shown in Fig.\ref{cover1}. For comparison, the tight-wire scheme is more difficult to operate. In order to prevent the cloth from touching the ground, we will adjust the height of the wire rope pillar. The adjustment process is shown in Fig.\ref{cover1}. Firstly, wire is put on the higher rope pillar and pulled tight; secondly, unfold the black rubber cloth and fix the edge of the cloth; finally, put the tight-wire on the lower pillar slowly. \begin{figure}[htb] \begin{center} \includegraphics[width=12cm]{VetoDetector/Figures/DetView.png} \caption[Lead rail for rubberized fabric cover option installation]{Lead rail for rubberized fabric cover option installation.} \label{Detview} \end{center} \end{figure} The slide-guide scheme costs more, but is more easily implemented. The main structure contains the main beams, columns, slide guides and pulleys. A couple of slide guides is fixed on the main beams under bridge. One end of the main beam is supported on the bridge, and the other end is connected to the pillar fixed on the ground. Each of the two schemes have advantages and disadvantages, and the final design choice is still under discussion. \subsection{Lining and Thermal Insulation Layer of the Pool} The cylindrical pool with 42-meter diameter and height will be filled with 20,000-30,000 tons of ultra-pure water. Since the laboratory is located 400 meters underground, the pressure of underground water is high. A powerful lining is needed to prevent water from the surrounding rock seeping into the pool. We propose to adopt high-density polyethylene (HDPE) material as the waterproof lining between rock and pool. HDPE has good corrosion resistance, high strength, good ductility, and does not pollute ultrapure water. The design can be similar to that for landfill, but the lining for the JUNO water pool contains many through holes and further studies are needed. The operating temperature of the JUNO detector system is 20 degrees, while the temperature near the laboratory rock is around 32 degrees, significantly higher than the operating temperature. The granites surrounding the pool have poor permeability, which has certain heat preservation effect. The concrete wall of the pool is a poor heat conductor with a heat transfer coefficient of only 1.63 W/(mK). According to an independent heat transfer calculation and the example of calculation for the Daya Bay experiment, the pool thermal insulation may not require additional insulating layer. However, to reduce operating costs, it is proposed to add an insulating layer. \subsection{Installation} The detector installation includes the assembling of PMT and bracket, and the fixation of PMT bracket. The assembling of PMT and bracket is shown in Fig.~\ref{PMTstr}. To make full use of the embedded parts in concrete pool wall, the PMT bracket is considered to be fixed on the embedded parts. In addition, special tools and operation management are needed to ensure safe and smooth detector installation. \section{Prototype} \subsection{ Geomagnetic Shielding Prototype } Through numerical calculation, we provide an option for the geomagnetic shielding system for JUNO. A prototype of coil system is required for validation and material property study. After the prototype study, we will have a more accurate estimate for the shielding effect and can then optimize the design of the geomagnetic shielding system. \section{Detector Construction and Transportation} The veto system mainly consists of the water Cherenkov detector and the tracking detector. For the water Cherenkov detector, the most important tasks are PMT test and transportation, water purification system construction and transportation, and support structure manufacturing and transportation. The veto PMT test and transportation will be done together with the central detector PMTs. Water system manufacturing and transportation and support structure construction await the final design decisions before the bidding and other processes can be started. The option for the top trackers also needs to be finalized before the detailed plans for the constructions and transporting can be made. \section{Risk and Safety Issues} The large size of the JUNO detector limits the available space in the experimental hall. For the veto system, it is important to carefully coordinate the installations of various detectors and hardwares. If the assembly coordination is not optimal, it could influence the whole veto systems installation. To avoid this risk, a careful coordination and strict schedule should be performed for various assembly tasks. Any potential conflicts in the installation schedule should be discussed and resolved among the detector coordinators to ensure a smooth installation process. Most of the safty issues occur during the detector installation stage. The safety procedure must be strictly followed when the large JUNO detector is installed. \section{Schedule} The plan for the period 2014-2020 is as follows: \begin{itemize} \item 2014: Setup the following two kinds of veto detector prototypes and take data with cosmic muons: 1) liquid scintillator + fiber prototype; 2) a multilayer resistive plate chamber (RPC). \item 2015: Finalize the top tracker option. \item 2016: Determine the pool liner design and begin to prepare lining installation; Complete the design of water pool insulating layer design and start installation preparation; Determine geomagnetic shielding design and start the installation preparation. \item 2017: Start to install water pool insulating layer; Start to install water pool lining; Start to install the geomagnetic shielding system. \item 2018: Complete the production of water Cherenkov detector PMT support structure and top track detector supporting structure; Start the installation of the top track detector and water Cherenkov detector PMT installation. \item 2019: Complete water Cherenkov detector installation and top track detector installation; Start pump ultra pure water into water pool. \item 2020: Complete pump process and start veto system commissioning. \end{itemize}

{'timestamp': '2015-09-29T02:17:39', 'yymm': '1508', 'arxiv_id': '1508.07166', 'language': 'en', 'url': 'https://arxiv.org/abs/1508.07166'}

high_school_physics

Controlled Substance Sale and Distribution | Legal Services for Metro Houston Understand your legal rights if you are arrested on charges of controlled substance sale and distribution. Connect with the Law Office of Michael Lamson, Houston, experts in defending those charged with drug trafficking charges. Drug offenses in Texas, especially the more serious circumstances of selling and distributing controlled substances, carry harsh penalties, including possible prison time and large fines. Obtain expert legal advice on your case. Call us today to review the specifics surrounding your arrest and charges for controlled substance sale and distribution. Effective Defense for Cases Involving Controlled Substance Sale and Distribution Michael Lamson Attorney at Law, Houston, has provided legal assistance to defendants accused of drug violations, including controlled substance sale and distribution, since 1977. Call our office today for a free, no-obligation consultation on your controlled substance sale and distribution case. Our expert legal team will review your case and explain your best options for legal defense to seek the best-possible outcome in court. Besides possession, prosecutors will attempt to present evidence of an intent to sell and distribute controlled substances, including: Large amounts of the particular controlled substance Possession of large amounts of cash Packaging of the controlled substance in bags or other containers Testimony and surveillance photos and videos of people interacting with you on the street in your car or in your home. Measuring devices such as scales Phone calls or texts communicating intent to sell or distribute controlled substances Effective defenses against unlawful arrest for selling and distributing controlled substances include lack of a legal search or warrant, possession of a small amount of the drug, inability to prove intent and lack of physical evidence of drug sales or movement of the drug. Call the Law Offices of Michael Lamson to review the specifics of your case involving the sale and distribution of a controlled substance. Be fully prepared to effectively defend yourself against a drug trafficking charge. Call the legal experts at Michael Lamson Attorney at Law for assistance in controlled substance sale and distribution cases. Serving Metro Houston with Legal Help for Drug Trafficking Cases The Texas law office of Michael Lamson provides expert legal services for controlled substance sale and distribution cases for the following communities and jurisdictions: Experienced Legal Experts in Controlled Substance Sale and Distribution Cases Get a highly experienced attorney specializing in controlled substance sale and distribution cases on your side. Call the Texas law office of Michael Lamson, Houston, for help if you or a family member has been arrested for selling or distributing a controlled substance. Get help today with a no-cost, no-obligation case review. Call us today!

{"pred_label": "__label__wiki", "pred_label_prob": 0.6303417086601257, "wiki_prob": 0.6303417086601257, "source": "cc/2022-05/en_middle_0024.json.gz/line1515572"}

high_school_physics

Home » Out-of-place artefacts » Mysterious objects » The Wonderstone spheres Published 7 May 2007 | By Keith Fitzpatrick-Matthews One of the Wonderstone (Ottosdal or Klerksdorp) spheres In the Wonderstone Mines at Ottosdal (West Transvaal, South Africa), hundreds of metallic spheres have been found over several decades in pyrophillite (or wonderstone) deposits of Precambrian date, some 2.8 billion years old. Pyrophillite is a metamorphosed sedimentary mineral mined for use in masonry buildings, as an absorbent, as a filler, as an ingredient in industrial compounds and for a variety of other uses. Two types of sphere are alleged to have been found in the pyrophillite; one is a bluish metal with white inclusions, while the other has a metallic coating around a spongy material. Only examples of the second type seem to have been examined by geologists, who have found that they consist of pyrites and goethite. At least one of these spheres has three parallel fine grooves running around its centre. Lab technicians were said to be at a loss to explain how they could have been formed by any known natural process. The first report citing these spheres has been traced to The Weekly World News, 27 July 1982. This is a publication similar to the British Daily Sport or the American National Inquirer, neither of which is noted for the quality of its investigative journalism. The spheres are much softer than has been claimed, but at least some of the goethite nodules have a parallel groove, although only one has ever been identified with three such grooves. However, there is a serious objection to regarding them as artefactual, since they formed in a metamorphic deposit and could not have existed before the metamorphosis of the deposit took place. If the grooves are artificial, they can only have been carved after the sphere was removed from the pyrophillite. In other words, the spheres are natural, although a question hangs over the grooves, which are either natural and immensely old, or artificial and modern. You can find more dismantlings of the claims here, by Tim McGuinness and here, by Paul V Heinrich. 6 Responses to The Wonderstone spheres Marilyn Reid on 26 June 2012 at 11:46 am Where I come from in Scotland, we have many, many spherical carved stones made from many, many different types of stone. Hundreds have been found, some hidden, but lots have turned up in ploughed fields, rivers etc. They are carved in a variety of shapes but one of them is repeated and the most common one found, with six boss shapes to it. People link them with the Platonic solids but they shouldn’r really as they are far older than that – by a couple of thousand years. They are probably over 5 thousand years old and are found mostly in an area of Scotland where many symbol stones are found, although a few have been found further afield. I was wondering if you knew anything about them? I held a couple of them in my hands the other week and would love to gather more information about them, or anything new, Marilyn Reid Paul Winterbourne on 12 September 2012 at 12:19 pm You should contact Exeter University Archaeology department. One of my friends is doing a docoiral thesis on the stone carved balls from Scotland. We did an undergraduate archaeology degree. He kept up the good work but alas I went on to work in IT! Yourname on 18 November 2012 at 9:14 pm The Scottish stones are just sling stones that have in many cases been personalized so that people can reclaim them or know whose stone took down the prey animal. Humans are not dumb and have known of slings for as long as we have been using animal skins and ligaments. Pingback: 5 to 1 Billion Year | Pearltrees Pingback: Spheres | Pearltrees James Eel on 7 November 2015 at 7:23 am I think it’s quite clear that either Cthulhu or Satan manufactured these. Nobody else has taken credit, and they’re the only two guys around that are that old. Q.E.D., really.

{"pred_label": "__label__cc", "pred_label_prob": 0.7475358843803406, "wiki_prob": 0.2524641156196594, "source": "cc/2020-05/en_head_0061.json.gz/line515703"}

high_school_physics

Police officer, public works employee nabbed in massive drug sting 32 charged in massive drug sting Gallery: 32 charged in massive drug sting By Alex Napoliello | NJ Advance Media for NJ.com A police officer and a public works employee were among 32 people arrested in a sting that netted large amounts of cocaine, heroin, painkillers, ecstasy and more than $125,000 in cash, authorities announced Wednesday. The investigation, dubbed "Operation Tidal Wave," began in October 2014 and honed in on drug-trafficking organizations in Middletown and Beachwood, State Police said in a news release. As a result, police seized $400,000 worth of drugs in the massive sting. At the helm of the "Scattaglia Network" in Middletown, police said, was Ed Scattaglia Jr. The 39-year-old used his home to meet with individuals who would then distribute cocaine throughout Monmouth and Ocean counties, police said. Authorities said they discovered during an investigation that Scattaglia received multiple kilograms of cocaine from the Englewood home of Francisco Romero, 38. Romero used his house on John Street to store "bulk" quantities of cocaine, police said. On June 2, 2014, State Police detectives and the Englewood Police Department searched Romero's residence and seized one pound of cocaine, $75,000 in cash, an Acura TSX and other assets valued at $50,000, police said. On that same day, police said, they searched Scattaglia's home and found $15,000 in cash, drug paraphernalia, marijuana and a BMW 325XI. In all, approximately one kilogram of cocaine, valued at $40,000, was seized from the "Scattaglia Network," police said. The second trafficking operation, the "Beachwood Network," was responsible for distributing cocaine, heroin, marijuana and pharmaceutical pills in Beachwood Township and the surrounding areas, police said. When police raided the network's heroin mill in Seaside Heights on April 9, they seized more than 15 grams of raw heroin, more than 275 decks of packaged heroin, at least an ounce of cocaine and other paraphernalia used to prepare and package heroin, police said. The total value of the drugs seized was more than $10,000, police said. Detectives also arrested Barry "Buzzy" McCombs at the mill, located on Blaine Avenue, police said. A search on April 20 of a second heroin mill, located in Elizabeth, also netted 1.5 kilograms of cocaine with a street value of more than $75,000, police said. Andres Genao and Jose Limardo were arrested at the Elizabeth heroin mill, police said. Also arrested as a result of "Operation Tidal Wave" was former Camden County Police Officer Matthew Van Houten, police said. Houten was charged with possession of cocaine while he was employed as a police officer. An Ocean County Road Department employee, Albert Logan, was charged with hindering apprehension and loitering to purchase a controlled dangerous substance. Acting Attorney General John J. Hoffman said Monmouth and Ocean counties have been hit hard by the "devastating epidemic" of opioid addiction rampant in New Jersey and around the country. "Through operations such as this one, we are taking aim at the criminals who are maliciously fueling the epidemic and profiting from it," Hoffman said. Rick Fuentes, superintendent of the State Police, heralded the Trafficking Central Unit and the other agencies that assisted in the investigation. "These (drug) organizations poison our communities with drugs and the violence that goes hand-in-hand with distribution," Fuentes said. Elie Honig, director of the Divison of Criminal Justice, echoed that sentiment. "By partnering across all levels of law enforcement," Honig said, "we have shut down a significant pipeline through which heroin, oxycodone and other deadly and addictive drugs were entering these Shore communities." The New Jersey State Police Trafficking Central Unit and the U.S. Drug Enforcement Admistration-Atlantic City Office were the lead agencies in the investigation. Assistance was provided by the Division of Criminal Justice, Monmouth County Prosecutor's Office, Middletown Township Police Department and the Jersey City Police Department. A total of 32 individuals were arrested and charged with various drug charges, police said. Scattaglia Network Edmund A. Scattaglia Jr., 39, of Middletown: Distribution of cocaine, conspiracy to distribute cocaine, possession of cocaine, distribution of drug paraphernalia. Francisco J Romero, 37, of Englewood: Distribution of Cocaine, conspiracy to distribute cocaine and possession of cocaine. Rick J. Hiemenz, 33, of Toms River: Distribution of cocaine, conspiracy to distribute cocaine and possession of cocaine. Tabraz Mohammed, 34, of Red Bank: Distribution of cocaine, conspiracy to distribute cocaine and possession of cocaine. Mark Ugaro, 35, of Middletown: Conspiracy to distribute cocaine. Scott M. Leander, 55, of Keyport: Conspiracy to distribute cocaine. Michael G. Dietz, 51, of Toms River: Conspiracy to distribute cocaine. Michael J Ko, 37, of Toms River: Conspiracy to distribute cocaine. Armando Calderon, 42, of Toms River: Loitering to purchase cocaine. James J McCann, 54, Ortley Beach: Loitering to purchase cocaine. William J. Bowes III, 36, of Toms River: Loitering to purchase cocaine. Augusto Deleonardis, 44, of Staten Island, N.Y.: Loitering to purchase cocaine. Shawn P Cotter, Jr., 28, of Middletown: Possession of marijuana and possession of drug paraphernalia. Brian J. Rosenberg, 30, of Red Bank: Contempt of court. Qasim H. Mir, 35, of Edison: Contempt of court. Beachwood Network Jose M. Limardo, 22, of Newark: Distribution of cocaine, conspiracy to distribute cocaine, maintaining a manufacturing facility for cocaine, possession of cocaine and distribution of drug paraphernalia. Andres M. Genao, 22, of Newark: Distribution of cocaine, conspiracy to distribute cocaine, maintaining a manufacturing facility for cocaine, possession of cocaine, distribution of drug paraphernalia. Clarence D Grant, Jr., 31, of Beachwood: Leader of narcotics trafficking network, distribution of cocaine, conspiracy to distribute cocaine, possession of cocaine, distribution of prescription pills and conspiracy to commit burglary. Barry J. McCombs, 40, of Seaside Heights: Distribution of heroin, possession of heroin, distribution of heroin within 1,000 feet of school property, distribution of drug paraphernalia and possession of marijuana. Jesse Gorman, 34, of Toms River: Distribution of cocaine, conspiracy to distribute cocaine and possession of cocaine. Joseph D. Powers, 28, of Toms River: Conspiracy to distribute a controlled dangerous substance and conspiracy to commit burglary. Bobby Hardy, 43, of Beachwood: Conspiracy to distribute a controlled dangerous substance. Dennis F. Gargano Jr., 42, of Beachwood: Conspiracy to distribute a controlled dangerous substance and conspiracy to commit burglary. Patrick J. Cavanaugh Jr., 37, of Bricktown: Conspiracy to distribute a controlled dangerous substance, conspiracy to commit burglary and hindering apprehension. Albert E. Logan, 28, of Beachwood: Hindering apprehension and loitering to purchase a controlled dangerous substance. Matthew R. Van Houten, 28, of Pennsauken: Possession of cocaine. Possession of Cocaine - 3rd degree Frank Caruso, 29, of Toms River: Possession of cocaine. Angela Robert, 33, of Toms River: Possession of a controlled dangerous substance. William T. Hiering III, 31, of Seaside Heights: Loitering to purchase a controlled dangerous substance. Anthony Alexander, 30, of Toms River: Contempt of court. Armando Freitas, 27, of Island Heights: Distribution of Marijuana, possession of Marijuana, possession of prescription pills (Xanax) and possession of steroids. Stefania Dockum, 25, of Island Heights: Distribution of marijuana, possession of marijuana, possession of prescription pills (Xanax) and possession of steroids. Alex Napoliello may be reached at [email protected]. Follow him on Twitter @alexnapoNJ. Find NJ.com on Facebook.

{"pred_label": "__label__wiki", "pred_label_prob": 0.9169088006019592, "wiki_prob": 0.9169088006019592, "source": "cc/2020-05/en_middle_0062.json.gz/line527550"}

high_school_physics

In chemistry, a Zintl phase is a product of a reaction between a group 1 (alkali metal) or group 2 (alkaline earth metal) and main group metal or metalloid (from groups 13, 14, 15, or 16). It is characterized by intermediate metallic/ionic bonding. Zintl phases are a subgroup of brittle, high-melting intermetallic compounds that are diamagnetic or exhibit temperature-independent paramagnetism and are poor conductors or semiconductors. This type of solid is named after German chemist Eduard Zintl who investigated them in the 1930s. The term "Zintl Phases" was first used by Laves in 1941. In his early studies, Zintl noted that there was an atomic volume contraction upon the formation of these products and realized that this could indicate cation formation. He suggested that the structures of these phases were ionic, with complete electron transfer from the more electropositive metal to the more electronegative main group element. The structure of the anion within the phase is then considered on the basis of the resulting electronic state. These ideas are further developed in the Zintl-Klemm-Busmann concept, where the polyanion structure should be similar to that of the isovalent element. Further, the anionic sublattice can be isolated as polyanions (Zintl ions) in solution and are the basis of a rich subfield of main group inorganic chemistry. History A "Zintl Phase" was first observed in 1891 by M. Joannis, who noted an unexpected green colored solution after dissolving lead and sodium in liquid ammonia, indicating the formation of a new product. It was not until many years later, in 1930, that the stoichiometry of the new product was identified as Na4Pb94- by titrations performed by Zintl et al.; and it was not until 1970 that the structure was confirmed by crystallization with ethylenediamene (en) by Kummer. In the intervening years and in the years since, many other reaction mixtures of metals were explored to provide a great number of examples of this type of system. There are hundreds of both compounds composed of group 14 elements and group 15 elements, plus dozens of others beyond those groups, all spanning a variety of different geometries. Corbett has contributed improvements to the crystallization of Zintl ions by demonstrating the use of chelating ligands, such as cyrptands, as cation sequestering agents. More recently, Zintl phase and ion reactivity in more complex systems, with organic ligands or transition metals, have been investigated, as well as their use in practical applications, such as for catalytic purposes or in materials science. Zintl Phases Zintl phases are intermetallic compounds that have a pronounced ionic bonding character. They are made up of a polyanionic substructure and group 1 or 2 counter ions, and their structure can be understood by a formal electron transfer from the electropositive element to the more electronegative element in their composition. Thus, the valence electron concentration (VEC) of the anionic element is increased, and it formally moves to the right in its row of the periodic table. Generally the anion does not reach an octet, so to reach that closed shell configuration, bonds are formed. The structure can be explained by the 8-N rule (replacing the number of valence electrons, N, by VEC), making it comparable to an isovalent element. The formed polyanionic substructures can be chains (two-dimensional), rings, and other two-or three-dimensional networks or molecule-like entities. The Zintl line is a hypothetical boundary drawn between groups 13 and 14. It separates the columns based on the tendency for group 13 elements to form metals when reacted with electropositive group 1 or 2 elements and for group 14 and above to form ionic solids. The 'typical salts' formed in these reactions become more metallic as the main group element becomes heavier. Synthesis Zintl phases can be prepared in regular solid state reactions, usually performed under an inert atmosphere or in a molten salt solution. Typical solid state methods include direct reduction of corresponding oxides in solution phase reactions in liquid ammonia or mercury. The product can be purified in some cases via zone refining, though often careful annealing will result in large single crystals of a desired phase. Characterization Many of the usual methods are useful for determining physical and structural properties of Zintl phases. Some Zintl phases can be decomposed into a Zintl ion—the polyanion that composes the anionic substructure of the phase—and counter ion, which can be studied as described below. The heat of formation of these phases can be evaluated. Often their magnitude is comparable to those of salt formation, providing evidence for the ionic character of these phases. Density measurements indicate a contraction of the product compared to reactants, similarly indicating ionic bonding within the phase. X-ray spectroscopy gives additional information about the oxidation state of the elements, and correspondingly the nature of their bonding. Conductivity and magnetization measurements can also be taken. Finally, the structure of a Zintl phase or ion is most reliably confirmed via X-ray crystallography. Examples An illustrative example: There are two types of Zintl ions in K12Si17; 2x (pseudo P4, or according to Wade's rules, 12 = 2n + 4 skeletal-electrons corresponding to a nido-form of a trigonal-bipyramid) and 1x (according to Wade's rules, 22 = 2n + 4 skeletal-electrons corresponding to a nido-form of a bicapped square antiprism) Examples from Müller's 1973 review paper with known structures are listed in the table below. Exceptions There are examples of a new class of compounds that, on the basis of their chemical formulae, would appear to be Zintl phases, e.g., K8In11, which is metallic and paramagnetic. Molecular orbital calculations have shown that the anion is (In11)7− and that the extra electron is distributed over the cations and, possibly, the anion antibonding orbitals. Another exception is the metallic InBi. InBi fulfills the Zintl phase requisite of element-element bonds but not the requisite of the polyanionic structure fitting a normal valence compound, i.e., the Bi–Bi polyanionic structure does not correspond to a normal valence structure such as the diamond Tl− in NaTl. Zintl Ions Zinl phases that contain molecule-like polyanions will often separate into its constituent anions and cations in liquid ammonia, ethylenediamene, crown ethers, or cryptand solutions. Therefore, they are referred to as Zintl ions. The term 'clusters' is also used to emphasize them as groups with homonuclear bonding. The structures can be described by Wade's rules and occupy an area of transition between localized covalent bonds and delocalized skeletal bonding. Beyond the "aesthetic simplicity and beauty of their structures" and distinctive electronic properties, Zintl ions are also of interest in synthesis because of their unique and unpredictable behavior in solution. The largest subcategory of Zintl ions is homoatomic clusters of group 14 or 15 elements. Some examples are listed below. Many examples similarly exist for heteroatomic clusters where the polyanion is composed of greater than one main group element. Some examples are listed below. Zintl ions are also capable of reacting with ligands and transition metals, and further 'heteroatomic examples are discussed below (intermetalloid clusters). In some solvents, atoms exchange can occur between heteroatomic clusters. Additionally, it is notable that fewer large cluster examples exist. Examples Homoatomic clusters [Si4]4- [Si5]2- [Si9]2- [Si9]4- [Ge4]4- [Ge5]2- [Ge9]3- [Sn4]4- [Sn5]2- [Pb4]4- [Pb9]4- [P4]2- [P7]3- [P11]3- [As6]4- [As7]3- [Sb8]8- [Sb11]3- [Bi4]2- [Bi7]3- [Bi11]3- Heteroatomic clusters AsP3 [Ge2Sn2]4- [Sn2Bi2]2- [Sn3Bi3]5- [Pb2Sb2]2- Sn5Sb3 [InBi3]2- Bi14Ge4 [GaBi3]2- [In4Bi5]3- [TlSn8]3- [TlSn9]3- [Sb@In8Sb12]3- [Sb@In8Sb12]5- Synthesis Zintl ions are typically prepared through one of two methods. The first is a direct reduction route performed at low temperature. In this method, dry ammonia is condensed over a mixture of the two (or more) metals under inert atmosphere. The reaction initially produces solvated electrons in ammonia that reduce the more electronegative element over the course of the reaction. This reaction can be monitored by a color change from blue (solvated electrons) to the color of the Zintl phase. The second is method, performed at higher temperatures, is to dissolve a Zintl phase in liquid ammonia or other polar aprotic solvent like ethylenediamine (on rare occasions DMF or pyridine is used). Some Zintl ions, such as Si and Ge based ions, can only be prepared via this indirect method because they cannot be reduced at low temperatures. Characterization The structure of Zintl ions can be confirmed through x-ray crystallography.  Corbett has also improved the crystallization of Zintl ions by demonstrating the use of chelating ligands such as cryptands, as cation sequestering agents. Many of the main group elements have NMR active nuclei, thus NMR experiments are also valuable for gaining structural and electronic information; they can reveal information about the flexibility of clusters.  For example, differently charged species can be present in solution because the polyanions are highly reduced and may be oxidized by solvent molecules. NMR experiments have shown a low barrier to change and thus similar energies for different states. NMR is also useful for gaining information about the coupling between individual atoms of the polyanion and with the counter-ion, a coordinated transition metal, or ligand.  Nucleus independent chemical shifts can also be an indicator for 3D aromaticity, which causes magnetic shielding at special points. Additionally, EPR can be used to measure paramagnetic in relevant clusters, of which there are a number of examples of the [E9]3- type, among others. Reactivity As highly reduced species in solution, Zintl ions offer many and often unexpected, reaction possibilities, and their discrete nature positions them as potentially important starting materials in inorganic synthesis. In solution, individual Zintl ions can react with each other to form oligomers and polymers. In fact, anions with high nuclearity can be viewed as oxidative coupling products of monomers. After oxidation, the clusters may sometimes persist as radicals that can be used as precursors in other reactions. Zintl ions can oxidize without the presence of specific oxidizing agents through solvent molecules or impurities, for example in the presence of cryptand, which is often used to aid crystallization. Zintl ion clusters can be functionalized with a variety of ligands in a similar reaction to their oligomerization. As such, functionalization competes with those reactions and both can be observed to occur. Organic groups, for example phenyl, TMS, and bromomethane, form exo bonds to the electronegative main group atoms. These ligands can also stabilize high nuclearity clusters, in particular heteroatomic examples. Similarly in solids, Zintl phases can incorporate hydrogen. Such Zintl phase hydrides can be either formed by direct synthesis of the elements or element hydrides in a hydrogen atmosphere or by a hydrogenation reaction of a pristine Zintl phase. Since hydrogen has a comparable electronegativity as the post-transition metal it is incorporated as part of the polyanionic spatial structure. There are two structural motifs present. A monatomic hydride can be formed occupying an interstitial site that is coordinated by cations exclusively (interstitial hydride) or it can bind covalently to the polyanion (polyanionic hydride). The Zintl ion itself can also act as a ligand in transition metal complexes. This reactivity is usually seen in clusters composed of greater than 9 atoms, and it is more common for group 15 clusters. A change in geometry often accompanies complexation; however zero electrons are contributed from the metal to the complex, so the electron count with respect to Wade's rules does not change. In some cases the transition metal will cap the face of the cluster. Another mode of reaction is the formation of endohedral complexes where the metal is encapsulated inside the cluster. These types of complexes lend themselves to comparison with the solid state structure of the corresponding Zintl phase. These reactions tend to be unpredictable and highly dependent on temperature, among other reaction conditions. Examples Group 14 anions functionalized with organic groups: [Ge9Mes]3- , [Ge9(CHCHCH2NH2)2]2- , [(CH2CH)Ge9Ge9(CHCH2)]4- , [Ge9(CHCHCHCH)Ge9]6- , [(CH2CH)Ge9(CH)4Ge9(CHCH2)]4- Silated anions: Ge9Hyp3Tl , [Ge9Hyp3]- Intermetalloid deltahedral clusters: [Co@Sn9]4− , [Ni@Pb10]2− , [Au@Pb12]3− , [Mn@Pb12]3− , [Rh3@Sn24]5- Exo coordinated transition metal complexes: [(ŋ2-Sn9)Hg(ŋ2-Sn9)]6− , [Ge5Ni2(CO)3]2-, [Sn8TiCp]3-, [(tol)NbSn6Nb(tol)]2- [Ni5Sb17]4- (Ni4Sb4 ring inside Sb13 bowl) Electronic Structure and Bonding Wade's rules The geometry and bonding of a Zintl ion cannot be easily described by classical two electron two center bonding theories; however the geometries Zintl ions can be well described by Wade’s rules of boranes. Wade’s rules offer an alternative model for the relationship between geometry and electron count in delocalized electron deficient systems.  The rules were developed to predict the geometries of boranes from the number of electrons and can be applied to these polyanions by replacing the BH unit with a lone pair. Some unique clusters of Ge occur in non-deltahedral shapes that cannot be described by Wade’s rules.  The rules also become more convoluted in intermetallic clusters with transition metals and consideration needs to be taken for the location of the additional electrons. Zintl-Kelmm-Busmann Concept The Zintl-Klemm-Busmann concept describes how in an anionic cluster, the atoms arrange in typical geometries found for the element to the right of it on the periodic table.  So “the anionic lattice is isometric with elemental lattices having the same number of valence electrons.” In this formulation, the average charge on each atom of the cluster can be calculated by: where na is number of anion atoms and VEC is the valence electron concentration per anion atom, then: . The number of bonds per anion predicts structure based on isoelectronic neighbor. This rule is also referred to as the 8 - N rule and can also be written as: . Not all phases follow the Zintl-Klemm-Busmann concept, particularly when there is a high content of either the electronegative or electropositive element.  There are still other examples where this does not apply. Electronic Theory Wade's rules are successful in describing the geometry of the anionic sublattice of Zintl phases and of Zintl ions but not the electronic structure. Other 'spherical shell models' with spherical harmonic wave functions for molecular orbitals—analogous to atomic orbitals—that describe the clusters as pseduo elements. The Jellium model uses a spherical potential from the nuclei to give orbitals with global nodal properties. Again, this formulates the cluster as a 'super atom' with an electron configuration comparable to a single atom. The model is best applied to spherically symmetric systems, and two examples for which it works well are the icosahedral Al13- and [Sn@Cu12@Sn20]12- clusters. DFT or ab initio molecular orbital calculations similarly treat the clusters with atomic, and correspondingly label them S, P, D etc. These closed shell configurations have prompted some investigation of 3D aromaticity.  This concept was first suggested for fullerenes and corresponds to a 2(N+1)2 rule in the spherical shell model. An indicator of this phenomenon is a negative Nucleus Independent Chemical Shift (NICS) values of the center of the cluster or of certain additional high symmetry points. Use in catalysis and materials science Some Zintl ions show the ability to activate small molecules. One example from Dehnen and coworkers is the capture of O2 by the intermetallic cluster [Bi9{Ru(cod)}2]3-. Another ruthenium intertermetallic cluster, [Ru@Sn9]6-, was used as a precursor to selectively disperse the CO2 hydrogenation catalyst Ru-SnOx onto CeO2, resulting in nearly 100% CO selectivity for methanation. In materials science, Ge94- has been used as a source of Ge in lithium ion batteries, where is can be deposited in a microporous layer of alpha-Ge.  The discrete nature of Zintl ions opens the possibility for the bottom up synthesis of nanostructured semiconductors and the surface modification of solids. The oxidation and polymerization of Zintl ions may also be a source of new materials.  For example, polymerization of Ge clusters was used to create guest free germanium clathrate, in other words a particular, pure Ge. See also Plumbide Stannide References Additional resources Video of preparation of K4Ge9 (subscription required) Inorganic compounds Intermetallics Cluster chemistry

{'title': 'Zintl phase', 'url': 'https://en.wikipedia.org/wiki/Zintl%20phase', 'language': 'en', 'timestamp': '20230320'}

high_school_physics

Revision as of 05:09, 3 April 2014 by Z3418837 (talk | contribs) (→‎Lab3) 1 Add a sub-heading. 3 Individual Assessments 3.1 Lab 1 4 Lab 2 5 Lab3 5.1 1.Select 4 reference papers related to your selected topic sub-section. Read these papers and write a brief description of their findings and relevance to the selected topic sub-section. The reference along with your description should then be pasted on both your group discussion page and your own personal page. 5.1.1 Article 1 -Charge as a Selection Criterion for Translocation through the Nuclear Pore Complex. 5.1.2 Article 2 - Role of Molecular Charge in Nucleocytoplasmic Transport. 5.1.3 Article 3- Higher Nucleoporin-Importin β Affinity at the Nuclear Basket Increases Nucleocytoplasmic Import. 5.1.4 Article 4 - Assembly of Nsp1 Nucleoporins Provides Insight into Nuclear Pore Complex Gating. This is a wonderful picture of two cells. <pubmed>24603758</pubmed> This is about Prokaryotes.[1] <pubmed limit=5>Prokaryote</pubmed> Individual Assessments Chloroplast present in Eukaryotes and not in Prokaryotes [1] ↑ <pubmed>17710148</pubmed>| PLoS One. Copyright: © 2007 Henderson et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 1. On your own student page upload an image with the reference using the Pubmed formatting shown in the practical class tutorial last week. Plant Nucleus [1] Copyright © 2012 Gao et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. 2. Identify a recent research article (not review) that uses either confocal microscopy or super-resolution microscopy as one of the study's techniques. Explain briefly (1 paragraph) how the microscopy technique specifically contributed to the article's findings. In Vivo Determination of Organellar pH Using a Universal Wavelength-Based Confocal Microscopy Approach. The purpose of the study was to measure the intracellular pH of organelles in situ using the physics behind confocal microscopy. This was achieved by extracting the colonies ofSaccharomyces cerevisiae which were associated at various developmental stages. Confocal fluorescence microscopy was then used to monitor the variation in intracellular pH by targeting fusion proteins to subcellular compartments and then quantifying the fluctuations in the emission wavelengths obtained. The use of confocal microscopy imaging enabled researchers to distinguish between the various fluorescent fractions that were present in single cells. It also provided enhanced readings of emissions by the decreased photobleaching and amplified signal-to-noise ratios which detected a wide range of fusion proteins and intracellular pKa levels leading to precise quantifications of organellar pH. Ultimately, the use of confocal microscopy imaging has advanced to such a high level that determining intracellular level of pH is simple and efficient compared to the previous methods of microscopy and spectrofluorometry used. Confocal microscopy also promises much greater result as research proceeds. 1.Select 4 reference papers related to your selected topic sub-section. Read these papers and write a brief description of their findings and relevance to the selected topic sub-section. The reference along with your description should then be pasted on both your group discussion page and your own personal page. Article 1 -Charge as a Selection Criterion for Translocation through the Nuclear Pore Complex. Nuclear pore complexes (NPCs) are porous structures that selectively control the components that pass through the nucleus to the cytoplasm and vice versa. A common understanding on the selective aspect of NPCs comes from the various diameter sizes that NPCs can occupy, however scientists have been focusing on another governing principle of translocation that they believe is related to charge. It is known that proteins that undergo rapid translocation are highly negatively charged, whereas proteins that are blocked from this translocation process are positively charged. Analysis has shown that, proteins that occupy the inside of the pore channel are net positively charged and transport receptors are negatively charged. This indicates that translocation rates within NPS are dependent on the electrostatic interactions between transport receptors and NPC due to an immense gain in energy. Overall, the investigation suggests that the negative charge plays a significant role in determining which components pass through the NPC.[1] Article 2 - Role of Molecular Charge in Nucleocytoplasmic Transport. Similarly to article one, this study focuses on the role of charge in the selectivity of Nucleocytoplasmic transport. It highlights the significance of phenylalanine-glycine (FG) nucleoporins (Nups) in creating a selective barrier with net positive charges in the nuclear pore complex. It is known that the positive charge accounts for the passive diffusion of small molecules and transport-receptor facilitated diffusion of cargo molecules. However, it was recently hypothesized that a negative surface charge plays a significant role in determining which components pass through the NPC. This study aims to unpack how these charge interactions may impact transport kinetics and spatial transport routes for both passive diffusion and facilitated translocation. By using high-speed fluorescence microscopy, scientists were able to determine that the electrostatic interactions between the negatively charged surface receptors and positively charged FG-Nups although increased the likelihood of NPC binding, did not reveal the nuclear transport mode or spatial transport routes. Instead, Nucleocytoplasmic transport was found to be dependent on molecular size, signal and surface charge and not one or the other.[2] Article 3- Higher Nucleoporin-Importin β Affinity at the Nuclear Basket Increases Nucleocytoplasmic Import. For many years scientists have debated on whether the presence of an affinity gradient in NPCs for the import receptor Importin β increased or decreased Nucleocytoplasmic Import. However the answer to this postulation remained obscure. This study aimed to understand how this affinity gradient may have enhanced Nucleocytoplasmic Import by using agent based modelling (ABM) that looked at the association between rate constants and molecular binding. By employing different values of the affinity gradient, they have found that the rate of transport had increased by 10% compared to the pores lacking an affinity gradient. They also found that this effect was maximised at 200 µM for Importin β. Overall, this study highlighted the significance of Importin β Affinity in Increasing Nucleocytoplasmic Import rate[3] Article 4 - Assembly of Nsp1 Nucleoporins Provides Insight into Nuclear Pore Complex Gating. Our current understanding suggests that the central transport channel which is a part of the NPC is formed by repeating subunits of FG-nups and serves as the selective barrier for incoming material. By analysing the spatial arrangements of nups, scientists can understand how NPC serves as a filter for macromolecules and allows for diffusion of small molecules below 40kDA. In this study, scientists used molecular dynamics to model the possible variations of Nsp1 (Nsp1-FG). They have discovered brush-like structures that consist of bundles that have been cross-linked to various nups. They have also found that transport factors are tightly associated with multiple FGS in cross-linking zones and also dissociate the bundles to widen the pores and allow molecules to enter. Overall, the model of the nuclear pore complex gating shows that the periphery of the NPC central channel consists of a few brushes with many cross-linked bundles due to the tethering of nups. However, the central regions shows a sieve-like structure of bundles with repeated cross-links as tethering of nups is less prominent.[4]

{"pred_label": "__label__cc", "pred_label_prob": 0.6649181246757507, "wiki_prob": 0.33508187532424927, "source": "cc/2020-05/en_middle_0095.json.gz/line626555"}

high_school_physics

A vortex filament cannot end in a fluid; it must extend to the boundaries of the fluid or form a closed path. In addition to his theorems in fluid dynamics, Helmholtz’s fundamental theorem of vector calculus states that any sufficiently smooth, rapidly decaying vector field in three dimensions can be resolved into the sum of an irrotational (curl-free, longitudinal component) vector field and a solenoidal (divergence-free, transverse component) vector field; this is known as the Helmholtz decomposition. In the mathematics of group theory, we use orthogonal O(n) and the special orthogonal SO(n) groupings to classify vortexes. Imagine using a computer drawing program like Microsoft Paint. It has five options under the category rotate. They are rotate 90, 180, 270, flip vertical, flip horizontal. Some drawings look the same after a “flip” as they do after a “rotate 180”. But some drawings do not look the same. This is the difference between O(n) groupings and SO(n) groupings. O(3) – Before getting into specific vortexes, think about symmetry. In three dimensions, the O(3) group is all rotations about the origin and all reflections along an axis through the origin. In the O(3) group, particles can have an up or down direction, but reflections must still be in the same O(3) group. These same particles in the O(3) group cannot have a left or right direction, as this would be reversed in a mirror. A particle with only transverse (vertical) spin is a member of an O(3) group and will be the same particle after being reflected in a mirror. SO(3) – The SO(3) is the group of all rotations about an axis while relections are reversed in a mirror. For the SO(3) group, the fact that the reflection of something in an SO(3) group is not in that same group ensures that when you look at a particle from an SO(3) group, you can tell which division it belongs to. That is the basis for the anti-particle being distinct from the regular particle. Spinning particles with two components of spin form an SO(3) group. SO(3) group members are classified as either right-handed or left-handed using the right hand rule. You can see from the pictures, that when you bring together one of these SO(3) particles with a mirror image of itself, the spins will conflict with each other no matter what the angle. They will destroy each other. On 4 July 2012, important experimental results were announced at CERN (Conseil Européen pour la Recherche Nucléaire). The discovery was the Higgs Boson with mass of approximately 125 GeV. This amazing event highlights the direct relationship between mathematics and physics, including the predictive power that mathematics gives to physics. The Higgs was predicted solely through the mathematical modelling of the standard model. To match the mathematical properties of the Higgs, the Higgs Boson is modelled as the most basic vortex of swirling energy. The Higgs has only transverse components of spin and is unstable with a lifetime around 1.6×10‾²² seconds. The Higgs Boson is its own anti-particle, with only a transverse component of spin. The mathematics behind a Higgs Boson decay, tells us that a 125 GeV Higgs, has a 23.3% chance of decaying into a W+/W- particle combination and a 2.9% chance of decaying into a pair of Z bosons. The W bosons have transverse and longitudinal components of spin and are unstable with a lifetime around 3 x 10‾²⁵ seconds. The Z boson has only a longitudinal component of spin. The W+ and W- bosons are anti-particles to each other. If you follow the longitudinal spin of the W+ with your fingers, your thumb must point up making the W+ a right-handed particle. If you follow the longitudinal spin of the W- with your fingers, your thumb must point down making the W- a left-handed particle. The Z boson is its own anti-particle and has longitudinal spin with no transverse component of spin. Looking closely at the model, try to visualize the decay of an unstable Higgs. The most likely event, is the breakup of the Higgs into a W+ and a W-. Also, try to visualize the Higgs breaking up into two Z particles, one turning left and one turning right. These are the decays that are seen in particle accelerators like CERN. It is important to note what happens when these particles interact with other matter. Both the W and Z bosons, being spin 1 particles, change the direction of spin of other particles they hit. A spin up electron (+1/2) will become a down electron (-1/2). The W bosons having a +1 or -1 charge, will change the charge of particles they hit. A +1/3 charge quark will be changed to a -2/3 charged quark if hit by a W- boson. The mathematical modelling of the electron began in 1922. German physicists Otto Stern and Walther Gerlach shot electrons through a magnetic field and found that half the electrons went up a bit, half the electrons went down a bit and none stayed straight. This clearly demonstrates the principle of spin and that any model of an electron has to have an up and a down state. To explain this and many other features of electrons and photons, the physicist Paul Dirac (ca. 1902–1984) represented particles in very new ways. He derived the “Dirac Electron” from earlier wave equations. Wave equations had been around for a long time. The French physicist Jean D’Alembert back in 1750, was the first to study wave equations to model vibrating strings in music. Dirac modelled the electron mathematically as two interacting wave equations. He demonstrated the concept of mass as trapped energy. The model allowed Dirac to derive the mass of the electron and predict the existence of anti-matter. Dirac’s one equation for a massive particle can be rewritten as two equations for two interacting massless particles, where the Dirac spinor ψ, with its 4 complex components, can be represented, as a pair of 2-spinors αA and βA’ … with a ‘coupling constant’ M describing the strength of the ‘interaction’ between the two. To visually model the mathematical model of the electron, we get what looks much like a Higgs Boson trapped around a Z boson, but acts very different to both. An electron contains a tiny mass, a strong electrical charge, and a large (relative to other particles) magnetic moment. In an electron, a strong electrical current has trapped an orthogonal magnetic field. One spin around the electrical loop is only 1/2 a spin around the magnetic field. A double spin of 720° is required to return the electron to its original orientation. An electron is the common decay element of a W- boson. The W- boson is modelled with circulating magnetic and electrical field lines running orthogonal to each other, rotating around each other as they both circulate around the center. The electron breaks this apart. The electrical field lines now spin round in a loop, with the magnetic field lines spinning as a torus orthogonal to the electrical lines. The electron is a lepton and is very stable. The electron has two heavier unstable cousins, with the same charge, spin, and similar magnetic moments. The Muon with a mass of 106 MeV and the Tau with a mass of 1,777 MeV are modelled with additional windings. Keep in mind, an electron interacting with a W+ Boson, will be flipped into a positron (anti-electron). The boson (with its much higher mass) will reorient the handedness of the red spin compared to the blue spin, turning a right handed particle into a left handed particle. The Z Boson has a chance of turning an up electron into a down electron. Finally, the amount of force required to flip an electron from an up state to a down state is called the magnetic moment. The magnetic moment of the electron is not quite exactly −1.0 μʙ. There exists a very real (measureable) possibility that as part of the flipping process, a photon will be emitted and reabsorbed by the electron. The mathematical modelling of the magnetic moment of the electron must be adjusted or “renormalized” to take this and other similar interactions into account. One thing about energy is that it moves very fast. In one second it is 186,000 miles away. A photon is a packet of energy. In 1807, Thomas Young published his double-slit experiment proving the wave nature of light. James Maxwell (ca. 1831–1879), a Scottish mathematical physicist, wrote “A Treatise on Electricity and Magnetism” in 1873 and formalized the mathematical basis of the electromagnetic wave. Quantum Mechanics forces you to visualize the photon as an oscillating plane wave moving along the Maxwell’s electromagnetic wave. Both the electric and magnetic fields are oscillating positive to negative, in a coordinated manner, orthogonal to each other. The two fields get big, then small, then reverse big, then small, then big, and so on, continuously as the photon moves through space. This action of getting big, then small, then reverse big, then small is one cycle that takes a certain amount of time. This action also acts as a harmonic oscillator and can store energy. The amount of energy in a specific photon is inversely proportional to the amount of time it takes for the photon to grow big then small. Large energy photons grow large (but not very large) and contract very fast as if they are on a very tight spring. Low energy photons take much longer to grow large, but expand to a much larger size, as if they were on a very loose spring. Based on the connectivity of space itself, the relationship between cycle time and energy is the Planck constant. In 1900, Max Planck, laid out the equations to show a single orange colored photon containing 1.97 eVolts of energy will travel 630 nanometers in one full expansion/contraction cycle. The same 1.97 eVolt photon will take 2 femtoseconds to complete one full cycle. The location of the photon along this expansion/contraction cycle is called the “phase” of the photon. A photon in its “large phase” is much more likely to be reflected then a photon in its “small phase”. In a similar manner, a photon in its “small phase” is much more likely to be absorbed by an electron than a photon in its “large phase”. One photon in its “large phase” in the same location with another photon in a “large negative phase”, will not look like a photon at all as one interferes with the other and they cancel each other out. Photons are said to be “coherent” if the phase of their expansion and contraction cycles match. Polarization can be a direction like up or down, left or right. Polarization can also mean spin in a clockwise or counter-clockwise direction. The axis of the electric field may be rotating or spinning in time and we would call the photon spin left or spin right. If the axis appears to stay vertical, we do not know for sure that the axis is vertical. We only know for sure that it has the highest probability of being measured in a vertical direction and zero probability of being measured in a horizontal direction. The mathematics of polarization starts with the representation of a photon as a vector containing complex numbers (the Jones Vector). To represent this visually, the real part of the variable is where the polarization is right now. The imaginary part represents where the polarization angle is going. Remember, these angles only represent the amplitude of probability. In other words, these angles represent the most likely angle for the polarization to be found. The mathematics captures a particle with a spin axis capable of moving around over time. In addition, the mathematics captures the expansion or contraction of the particle overall. Using this mathematical representation of the probability of the polarization of a photon highlights the direct and important relationship between mathematics and physics. Matrix multiplication of a vector is now equivalent to an action on the photon. In the real world of physics, the orientation of the photon polarization angle can be manipulated and changed by passing it through a wave plate. In the modelling world of mathematics, multiplication of the Jones vector (representing the photon) by a matrix (used to represent the wave plate), provides the calculation to see what the probability of orientation of the photon is in the real world. In 1931 Isidor Rabi predicted that the Stern–Gerlach experiment (used to determine the electron was a spin 1/2 particle) could be modified to confirm the properties of protons and neutrons in the atomic nucleus. Rabi built a molecular beam apparatus and employed a weak magnetic field to detect the nuclear spin of sodium. When the experiment was conducted, four beamlets were found, indicating that the nuclear spin of sodium was 3⁄2. To model his findings, Rabi used the work of Joseph Larmor. Joseph Larmor (ca. 1857–1942) was a physicist and mathematician who made substantial contributions to the view of matter as spinning vortexes. Where mass is involved, Larmor laid the groundwork for mathematically modelling the concepts of spin, precession and coupling for nucleons. Coupling the up or down energy states of multiple spin 1/2 particles allows for mathematical modelling of more complex particle spin. One spin 1/2 particle will go up or down. Two spin 1/2 particles coupled together (spin 1) can go in three different directions. Three spin 1/2 particles coupled together (spin 3/2) will form 4 beamlets, just like the Rabi experiment. With the development of high speed colliders the model becomes far more complex. An electron modelled as a special unitary SU(2) gauge symmetry group, only deals with two properties, magnetic up or down and spin left handed or right handed. The combinations of properties generate only four different particles. With hadrons, there is a literal “particle zoo” with a multitude hadrons discovered. In 1964, The quark model was proposed by physicists Murray Gell-Mann and George Zweig using special unitary SU(3) flavor symmetry. SU(3) symmetry has three properties, often labeled red plus or minus, green plus or minus and blue plus or minus. SU(3) symmetry gives you eight different ways of binding quarks together through gluons. To visually present the SU(3) group, we use three properties: specifically magnetic up or down; spin left handed or right handed; and, flow in or out. Close your eyes for a moment and use your imagination. Imagine a large empty sphere, maybe like a big beach ball. Looking closer, notice the surface of the beach ball is actually made up of several layers of different beach balls. Not only that, the layers are able to spin independent of each other with one beach ball spinning one way and the beach ball just inside it spinning in a different direction! Finally, notice the surface of any one beach ball. It’s spinning in two directions. Put your finger on a point of the sphere and track it. You find your finger not only rotating around the sphere, but also, up and down directly through the north and south poles. The poles are spinning very fast and different layers can “couple” together at the poles. OK, time to open your eyes. Quarks are large empty 3D spinning spherical shells with both longitudinal and transverse components of spin. Quarks colored blue symbolize a +2/3 charge and red symbolizes a -1/3 charge. When layered, these spheres make up Mesons and Baryons. Mesons have two layers of quark shells, Baryons have three layers. Protons and Neutrons (both three layered Baryons) are the final, and most noticeable of the layered quarks. Quark shells are able to stack or layer together with the help of gluons. Gluons act both as “lubricant” between quark layers and as special “taps” that allow energy to flow in or out of a particular region. An important attribute of spin on a quark is what happens at the poles. If you follow the spin of an individual point on the surface of a shell modeled this way, you will notice it goes directly through both the north and the south pole of the quark on every turn. This means a lot of paths are directly converging and diverging at the poles. To accommodate this we assume the converging lines are coming in either above or below the outgoing diverging lines. Also, the left or right handedness of the poles tells us which way the magnetic field is pointing for both regular and anti-quarks. The properties of the gluon are called its color charge. There are three color charges, specifically spin (clockwise or counter clockwise) magnetic direction (up or down) and flow (in or out), making 8 unique gluons. The “Color charge” rules are detailed in quantum chromodynamics (QCD), and are a basis for the interaction of quarks and gluons to form more complex elements. Visual imaging of the mathematical modelling of elementary particles of the Standard Model significantly improves the usefulness of these models. Different visual images within the same mathematical framework can co-exist until some aspect of the imaging is shown to be inconsistent with the real world. These images must then be corrected, removed or placed within a specific scope. The massive particles built from a variety of swirling energy vortex styles exhibit the same properties seen in particle colliders. Massless particles expanding and contracting as they zip through space do the same unusual things waves do. These images provide a basis to build an understanding of much deeper concepts and help foster predictions that will lead to a deeper understanding of the world around us. When one accepts one theory and rejects another which is equally consistent with the phenomenon in question, it is clear that one has thereby blundered out of any sort of proper physics and fallen into mythology.

{'timestamp': '2019-04-23T13:50:15Z', 'url': 'http://animatedphysics.com/insights/particles-of-the-standard-model/', 'language': 'en', 'source': 'c4'}