DRAFT

Version V2.3.1

Heavy Photon Search Experiment at Jefferson Laboratory: theINFN-JLAB12 Collaboration involvement

M. Battaglieri, A. Celentano, R. De Vita, S. Fegan, M. Osipenko, M. TaiutiIstituto Nazionale di Fisica Nucleare, Sezione di Genova e

Dipartimento di Fisica dell’ Universita, 16146 Genova, Italy

M. Carpinelli, V. SipalaUniversita e Istituto Nazionale di Fisica Nucleare, 07100 Sassari, Italy

A. D’Angelo, L. Colaneri, A. RizzoIstituto Nazionale di Fisica Nucleare, Sezione di Roma-TorVergata

e Dipartimento di Fisica dell’Universita, Roma, Italy

N. Randazzo, S. Aiello, M. De Napoli, E. Leonora, C. VenturaIstituto Nazionale di Fisica Nucleare, Sezione di Catania, Catania, Italy

D. Calvo, A. FilippiIstituto Nazionale di Fisica Nucleare, Sezione di Torino, Torino, Italy

G. SimiIstituto Nazionale di Fisica Nucleare, Sezione di Padova, Padova, Italy

(Dated: August 8, 2013)

ABSTRACT

This proposal presents the Heavy Photo Search (HPS) experiment planned at

Jefferson Lab for 2014-15 and a possible involvement of the INFN-JLAB12 Collab-

oration. This document is based on the HPS proposal to HEP-DOE and Jefferson

Lab submitted on May 10th, 2013. INFN-JLAB12 groups are expected to mainly

contribute to the electromagnetic calorimeter ECal, the key-element of the HPS trig-

ger system. The expected activities are: the motherboard design, manufacture and

test; the Light Monitoring System (LMS) design and prototyping; the replacement of

the existing APDs with new Large Area APDs. The INFN-JLAB12 commitment is

expected to be both in term of funding (LAAPDs procurement, motherboard man-

ufacturing, LMS prototyping), and manpower (motherboard design and test, LMS

design and prototyping, photo-sensor replacement). New Italian groups (INFN-CA,

INFN-CT and INFN-TO), specifically focused on HPS experiment, are joining the

DRAFT

2

INFN-JLAB12 Collaboration (INFN-GE and INFN-RM2) providing the necessary

manpower to accomplish the proposed activities. The detailed funding request to

the INFN-CSNIII will be placed in July/September 2013.

DRAFT

3

Contents

The HPS experiment 4

The HPS Collaboration 4

Experimental set-up 5

The electromagnetic calorimeter ECal 6

The INFN-JLAB12 commitment 8

Time schedule 12Run plan and beam time request 12HPS schedule 12ECal schedule 13INFN activity schedule 13

Manpower 14HPS manpower 14INFN-JLAB12 manpower and resposabilities 15

Cost and funding 18HPS cost and funding 18INFN-JLAB12 commitment 18Travel 18INFN-JLAB12 Software commitment 21

DRAFT

4

The HPS experiment

The Heavy Photon Search (HPS) is an experiment proposed for Jefferson Laboratory tosearch for new heavy vector boson(s), aka ”heavy photons” or ”dark photons” or ”hiddensector photons”, in the mass range of 20 MeV/c2 to 1000 MeV/c2. Such particles will ariseif there are additional U(1) gauge bosons in nature, and they will couple, albeit weakly, toelectric charge through kinetic mixing. Many BSM theories predict the existence of addi-tional U(1)’s, and recent observations of high energy electrons and positrons in the cosmicrays may be the result of primordial dark matter annihilating into heavy photons. HPSsearches for electro-produced heavy photons using both invariant mass and separated decayvertex signatures using a compact, large acceptance forward spectrometer. The first stageof HPS, the HPS Test Run, was approved by the Jefferson Lab PAC37 on January 14, 2011,after which it was proposed to DOE HEP for funding and approved and funded by Summer2011. The Test Run was built in 2011-2012, and installed and run at JLAB in Spring, 2012.PAC39 reviewed HPS in June, 2012, and on the basis of the successful run, granted it an”A” rating, a commissioning run with electron beams, and approval (C1) to proceed to thefull experiment contingent only on final approval from JLAB management. The full HPSexperiment, is capable of searching for heavy photons over a wide and unchartered regionin parameter space. Schedules at Jefferson Laboratory admit time for HPS to be commis-sioned in Hall B in the Fall of 2014 and to take data beginning in Spring 2015. With finalapproval from JLAB management, and with timely DOE approval and funding, the HPSCollaboration can design, build, test, and commission HPS in time to take advantage ofthese scheduling opportunities, and begin in earnest its search for spectacular new physicsat the intensity frontier.

The HPS Collaboration

The HPS Collaboration whose spokesmen are J.Jaros (SLAC), S.Stepanyan (JLab), andM.Holtrop (University of New Hampshire), includes about sixty people from about twentydifferent institutions from all around the world. In Italy, the HPS experiment is part of theINFN-CSNIII JLAB12 experiment. The participating INFN groups, and related Universi-ties, are: Cagliari, Catania, Genova, Sassari, Roma Tor Vergata. According to the HPSCollaboration Charter and Bylaws to become member of the HPS Collaboration each of theparticipants need to commit him/her-self to contribute to the experiment in instrumentationand/or software. The commitment statement reported in this document is intended to re-spond to this requirement, providing the HPS Collaboration membership to all the authorsnot yet part of the HPS Collaboration. The document will be submitted to the HPS Col-laboration Executive Committee for the necessary approval. Each INFN group will define aPI responsible for informing the HPS executive Committee of any changes within his grouppertaining the membership. The Italian contribution and commitment will be coordinated

DRAFT

5

FIG. 1: The HPS detector components. From left to right: the target, the silicon tracker in the

dipole magnet, the electromagnetic calorimeter ECal and the muon detecor.

with the relevant HPS Working Group.

Experimental set-up

HPS will utilize a setup located at the downstream end, called alcove, of the experimentalHall-B at Jefferson Lab. The setup will be based on a three-magnet chicane, the second dipolemagnet serving as the analyzing magnet for our forward spectrometer. The detector package,shown in Fig. 1, will include a silicon tracker, an electromagnetic calorimeter, and a muondetector. High luminosities are needed to search for heavy photons with small couplingsand masses in the 20 to 1000 MeV range. Utilizing CEBAF essentially continuous dutycycle, the experiment can simultaneously maximize luminosity and minimize backgrounds

DRAFT

6

by employing detectors with fast response and rapid readout. The HPS setup is designedto run with > 200 nA electron beams at energies from 1.1 GeV to 6.6 GeV impinging on atungsten target of up to 0.0025 X0 located 10 cm upstream of the first layer of the tracker.

The HPS tracker consists of six double layer planes, 36 microstrip sensors in total. Placingthe planes of the tracker in close proximity to the target means that the primary beam mustpass directly through the middle of the tracking detector. This has necessitated that thesensors don’t encroach on a dead zone, where multiple Coulomb scattered beam particlesand radiative secondaries are bent into the horizontal plane, the so-called ”wall of flame”.However, since the energy released in the decay of a low mass A′ is small relative to its boost,the opening angle between decay daughters can be quite small. To maximize the acceptancefor low mass A′s, the vertical extent of the dead zone must be minimized and sensors placed asclose as possible to the beam, so our design incorporates precision movers that can bring thesilicon detectors to the required positions. Since interactions of the primary beam with air oreven helium at atmospheric pressure gives rise to low-momentum secondaries that generateunacceptable occupancies, we have chosen to place the entire tracking and vertexing systemin vacuum, in the Hall B pair spectrometer’s magnet vacuum chamber. Silicon microstripsensors are used in the tracker/vertexer because they collect ionization in 10s of nanosecondsand produce pulses as short as 50 − 100 ns. The sensors are read out continuously at 40MHz using the APV25 chip, developed for the CMS experiment at the LHC.

A lead-tungstate electromagnetic calorimeter, with APD readout, is used to trigger theDAQ. Like the tracker system, the electromagnetic calorimeter is split in an upper and abottom part to avoid impinging on the dead zone. The beam and radiative secondaries passbetween the upper and lower ECal modules, which are housed in temperature-controlled en-closures, needed to stabilize the energy calibration. The ECal is the HPS detector componentwhich the INFN-JLAB12 Collaboration wants to contribute to: a more detailed descriptionis reported in the next Section.

The muon detection system will be installed downstream of the ECal, which has absorbedmost of the electromagnetic background produced in the target. The remaining backgroundswill be attenuated by the first absorber layer of the muon system. The muon system willconsist of four double layers of scintillator hodoscopes sandwiched between iron absorbers.Light from scintillator strips will be transported to photo-multiplier tubes via wave-lengthshifting fibers embedded within the strips. As in case of the ECal, the muon system will bedivided into two parts, beam up and beam down. There is a vacuum chamber between thetwo parts to allow the beam and radiated secondaries to be transported in vacuum.

The electromagnetic calorimeter ECal

The Ecal, depicted in Fig. 2, consists of 442 lead-tungstate PbWO4 crystals withavalanche photodiode (APD) readout and amplifiers. Indeed,(PbWO4) crystals with APDreadout are ideally suited to deal with the expected high radiation and high rate environment

DRAFT

7

FIG. 2: Arrangement of Ecal crystals. The two modules are positioned above and below the beam

plane. Each module has 5 layers. There are 46 crystals in each layer, with the exception of the

layers closest to the beam plane in which 9 crystals are removed to allow a larger opening for the

outgoing electron and photon beams.

and they can operate in the fringe field of the HPS magnetic field as well. The lead-tungstatemodules, see Fig. 3, are taken from the Inner Calorimeter (IC) of the JLab CLAS detector,which was built by IPN Orsay (France) and other Hall B collaborators and was used in aseries of high energy electroproduction experiments. Orsay has the responsibility of ECalmechanics and preamplifiers. The Ecal data is digitized in the JLAB FADC250, a 250 MHzflash ADC developed for the JLAB 12 GeV Upgrade program. The full analogue informationfrom the FADCs coupled with the fine spatial information of the calorimeter is available tothe trigger, which uses energy deposition, position, timing, and energy-position correlationsto reduce the trigger rate to a manageable ∼ 30 kHz.

The PbWO4 crystals are 16 cm long and tapered. The cross section of the front face is1.3 × 1.3 cm2, and 1.6 × 1.6 cm2 at the back. Modules in the ECal are arranged in twoformations, as shown in Fig. 2. There are 5 layers in each formation; four layers have 46crystals and one has 37. The ECal is mounted downstream of the analyzing dipole magnetat the distance of about 137 cm from the upstream edge of the magnet. In order to stabilizethe calorimeter’s performance, the crystals, APDs, and amplifiers are enclosed within atemperature controlled environment, held constant at the level of 1 F. The energy resolutionof the system, expected from operational experience with the IC, is σE/E ∼ 4.5%/

√E

DRAFT

8

FIG. 3: The ECal module is composed of a 16 cm long lead-tungstate crystal, Avalanche Photo

Diode, and a amplifier board.

(GeV). As in the IC, PbWO4 modules are connected to a motherboard that provides powerand transmits signals from individual APDs and amplifier boards to the data acquisitionsystem.

The HPS calorimeter was built and used during the HPS test run in April-May of 2012.This was the first time that a readout and trigger system utilising the JLAB FADC250s hadbeen used in a real experiment. While the ECal performance during the test run was sat-isfactory, there are some aspects that can be improved by the INFN-JLAB12 Collaborationcontribution.

The INFN-JLAB12 commitment

1. Matherboards modification - One of the issues faced during the test run was someexcess noise on some channels of the motherboards. Most missing channels were infact switched off because they were very noisy and there was no time to debug them.New motherboards will be designed and built to resolve this issue. One of the optionsunder discussion is to replace long motherboards with shorter ones with power andsignal connectors located on the top (for the top module) and the bottom (for thebottom module) of the thermal enclosure. Figure 4 shows the conceptual design ofthe new HPS motherboard. Each motherboard connects up to 60 APDs. For thewhole ECal readout 6+2 identical motherboards are foreseen. Two of them have areduced number of channels and a different geometry to match the central beam-hole.APDs will be HV-supplied in group of 10 to keep a reasonable gain-matching. HV andsignal connectors will be placed in the upper/bottom part requiring a modification

DRAFT

9

FIG. 4: The new HPS motherboard conceptual design.

of the mechanical enclosure. The Orsay collaborators will take care of the necessarymechanical modifications. The motherboard design takes advantage of the experiencegained by the INFN-JLAB12 groups in building the FT-Cal and the FT-Cal Prototypemotherboards. In particular, specific precautions to reduce noise pick-up by signaltracks and reduce the cross-talk will be taken. These includes: multiple-grounds planesin the PCB and separation of adjacent signal tracks with grounds. The design will bedone by the INFN-GE group supervised by the Electronic group of Orsay that originallydesigned the IC motherboard.

2. Light Monitoring System (LMS) - For an experiment like HPS, where back-grounds must be well understood and need to be strongly suppressed, the trigger biasmust be minimized. In particular, having stable, known, and uniform thresholds atthe trigger readout is necessary in order to avoid bias in the event selection. Suchuniformity and stability can be achieved with the installation of an online LightMonitoring System (LMS). The LMS will introduce short reference light pulses into

DRAFT

10

the front face of the crystals and checking the response it will be possible to identifya reduced light transmission in the crystal (for radiation damage) or a variation inthe APD gain. Such a calibration system has been developed for several experiments(PHOS and CMS at CERN for instance) with various light sources. The system forthe ECal will be similar to the monitoring system developed by INFN-GE for theFT-Cal. In the FT-Cal LMS each crystal is coupled to a blue LED inserted in a plasticsupport glued to the front face. All LEDs are soldered to a PCB (called LED Boardor LB), connected to the LED Drivers Board (DB). Each LED can be independentlyswitched on and off and the light varied to scan the whole dynamic range requested.A Control Board (CB) connected to the DB is equipped with a micro-controller,EPIC-compliant, that can be remotely operated via USB and Ethernet connection.The system has been designed in Genova and a prototype of 16 channel has beenmounted on the FT-Cal prototype and tested on-beam. The system shows excellentperformance in terms of dynamic range (from an equivalent energy deposition of 10MeV to few GeV per crystal), timing (down to ∼200ps) and stability (the ratio oftwo different channels show a stability of few h over a week). We propose to builda similar system for ECal. The CB is identical to the FT-Cal CB while the LB andDB need to be redesigned to match the different geometry keeping the same electriclayout. INFN-JLAB12 will design the whole system, machine the plastic LED holdersthat will be glued to the front of the crystals, while assuming that the final CB, LBand DB production will be funded and produced by a non-italian member of the HPScollaboration.

3. Photosensor replacement - Installing new APDs on the current crystals will sig-nificantly improve the ECal performance. Replacing the old 5 × 5 mm2 HamamatsuS8664-55 APDs with 10 × 10 mm2, Hamamatsu S8664-1010 (LAAPD) will improvetwo critical characteristics of the present calorimeter modules. First, the new APDsfrom Hamamatsu have much better performance than the ones which are currentlyinstalled. In particular, being made by the same wafer, they show an excellent gainuniformity. For a chosen working point of G = 150, a ∆G = ±10% corresponds to asupply bias variation of ∼ 1.5 V. For the current APDs the difference is now in therange of 50V. A reduced variation allows to supply the new LAAPDs in group of 10keeping the gain variation within a minimum value. In this way it will be possibleto achieve much better uniformity in the response of the calorimeter modules, andconsequently better uniformity in trigger thresholds. Secondly, the new APDs havea 4 times larger readout area, ensuring 4 times more light collection and therefore 4times larger signals. This will allow the use of lower gain amplifier modules whichin turn will decrease electronic noise. Tests performed for the FT-Cal, showed thatamplifier boards with a factor 2 lower gains have a noise level < 5 MeV. The energydeposition in the HPS PbWO4 crystals from minimum ionizing cosmic muons passingtransversely to the crystal axis is ∼ 15 MeV. Moving the energy thresholds close to 5

DRAFT

11

FIG. 5: Charge distribution from readout of the HPS calorimeter crystal with Hamamatsu S8664-55

(left) and S8664-1010 (right) APDs, and the new low noise amplifier board. The red line histogram

corresponds to the charge distribution for all triggers comming from the scintillators positioned

above and below the crystal. The black line shows the distribution for hits in the crystal within

100 ns of the trigger signal.

MeV will allow the MIP peak to be clearly distinguished, and will let the calorimeterbe calibrated and monitored with cosmic muons. We performed in Genova the firsttests of the Hamamatsu 10 × 10 mm2 LAAPDs and a new amplifier board on HPScrystals. In Fig. 5, the charge distribution of a single crystal is shown for 5× 5 mm2

(left) and 10× 10 mm2 (right) APDs. A coincidence signal from scintillator telescopespositioned above and below the module provided the trigger. The crystal was posi-tioned horizontally as it would be in HPS, so the cosmic muons would pass throughit perpendicular to the crystal axis. The red line histogram is for all events triggeredby the scintillation telescope and corresponds to the noise. The black line histogramcorresponds to the charge detected within 100 ns. The MIP peak is clearly visible andwell isolated from the noise when the S8664-1010 LAAPD is used. For the S8664-55APD, the MIP signal is also seen, but its charge distribution is under the noise peakand the peak position can not be resolved. MIP calibration, together with the lightmonitoring system, will ensure stable and reliable performance of the ECal and thetrigger system. As a bonus, the lower noise will allow the use of lower thresholds andimprove the calorimeter’s energy resolution. The increased APD size will increase thecapacitance (by almost a factor of 4) requiring a new preamplifier with an adaptedinput impedance and a lower gain. The new preamplifier, whose design was optimisedfor the FT-Cal detector will be built and funded by Orsay Group.

DRAFT

12

Time schedule

Run plan and beam time request

The Jefferson Laboratory PAC39, based on the addendum to the HPS proposal, reportingresults of the test run, graded HPS physics with an ”A”, approved a commissioning run withelectrons, and granted the approval (C1) for the full HPS experiment. (C1 means that theexperiment does not need to go back to PAC for the final approval, leaving to the JLabmanagement the final decision based on the success of the commissioning run).

The total beam time requested in the original proposal submitted to PAC37 is 180 days.Anticipating early running in Hall-B, we propose to conduct HPS in two phases. The firstphase, will consist of 1 week of beam for the beam-line commissioning and beam tuning, 2weeks for detector commissioning, and 4 weeks for data tacking, 2 weeks at 1.1 GeV and2 weeks at 2.2 GeV beam energies. In total, the first phase will use 7 weeks. The secondphase of the HPS, which will use the remaining beam time, can be scheduled later in 2016and beyond, and will continue runs at 2.2 GeV and 6.6 GeV, and possibly other energies. Ifmore beam time is available in 2015, HPS will continue data taking at these energies.

HPS schedule

Our goal is to be ready to install HPS in Hall-B by September 2014, and proceed withcommissioning on beam with the CEBAF early physics beam window opportunity in October2014. The data taking will continue until summer 2015. Meeting this schedule will requireapproval and funding as soon as possible, preferably by August 2013. The total constructionschedule extends over 16 months, assuming the funding will be available from mid-2013. Theschedule includes a contingency of about 20%. Schedules for each of the major subsystemsof the experiment are summarized below.

The conceptual design of the beam-line will be done during 2013. Final Engineering andConstruction will start in Spring 2014 and be completed well before the installation time inOctober 2014, providing substantial float.

The Test Run SVT was shipped back to SLAC by early February 2013 to rework themodules for the first three layers of the HPS and commission the motion control systems.The conceptual design of the Layers 1-2-3 and Layers 4-5-6 have started already in spring2013. The SVT DAQ will formally begin work in the second half 2013. The assembly andintegration test at SLAC are expected in spring 2014 and the SVT will be ready for shippingon July 2014. The SVT will be ready for installation in mid-August 2014. SVT installationin the analyzing magnet vacuum chamber will occur in September, depending on the scheduleof the Hall-B.

DRAFT

13

ECal schedule

The Ecal work will start in the second half 2013 and run through August 2014. The ECalwill be ready for installation by September 2014.Starting in fall 2013 the calorimeter will be disassembled and the crystals kept in a clean roomuntil the LAAPDs will be available on-site. In late summer - fall 2013 the new motherboardswill be designed, produced and tested to be ready before the end of the year. The newpreamplifiers are expected to be ready at the same time. Electric and electronics tests willbe done before the end of the year. The mechanics for the new mounting system will bemanufactured in early spring 2014. During fall 2013 the LED LMS will be designed and theLED holder machined and sent to JLab. The full system should be on-site for spring 2014.The APD replacement is expected for spring/summer 2014 and the final ECal assemblyin August 2014. At the beginning of September 2014 the ECal should be ready for theinstallation in the Hall. During spring 2014, the crystals will be unwrapped, the frontplastic holder replaced with the new LED holders. The crystals will be tested with an LEDsource in order to find an optimal regrouping of the bias voltage. This precaution offers abackup solution if the LAAPDs will not be available onsite at early convenience and the oldAPD kept in place. The APD ungluing operations will only start when the LAAPD willbe available on-site. INFN-JLAB12 will prepare many batteries of tools to speed up theungluing and re-gluing operations. In case the LAAPDs delivering will have some delay andtheir replacement will not be compatible with the ECal re-assembly, the replacement willtake place before the production run will start.

INFN activity schedule

The main INFN contributions to HPS are in the following areas: motherboard design,production and test, light monitoring system design and photo-sensors replacement. Thefirst two activities will start in early fall to be completed before the end of the year. Thereplacement of the APD with the new LAAPD is a more complex task since it requires atight coordination between the funding agency (INFN-CSNIII) and the vendor (Hamamatsu)to be sure that we will keep tightly on schedule. LAAPD funding is expected in October2013 and the order of 500 LAAPDs (442 channels + 58 spares) will be placed as soon aspossible. LAAPDs will be tested and benchmarked to establish the working condition, gainas a function of bias voltage and temperature, upon delivery at JLab. Tests are expected tohappen in April/May 2014. During fall/winter 2013 all tools necessary for APD replacementwill be prepared in advance and provided to JLab. As soon as the new sensor delivery timewill be firmly established, the operations to disassemble (crystal + APD + wrapping +plastic nose) and clean the crystals will start. The glueing and re-assembly procedure willgo in parallel with the un-gluing and cleaning in order to minimize any risk of running outof time. In case the replacement will be partial, the new instrumented crystals will be used

DRAFT

14

on the electron-side of the calorimeter where the background is expected to be more severeand even a partial ECal upgrade will be useful. A final test and characterisation of eachcrystal will be performed at the end of the machining, before mounting back the crystal intothe Ecal. A detailed timetable of the different tasks is reported in Tab I.

TABLE I: INFN-JLAB12 tasks commitment and time-schedule.

Activity Place Time

New motherboards design GE June - September 2013

Iterations with mechanical engineers GE Orsay June - August 2013

New motherboards production and test GE October - December 2013

Light monitoring system design GE September - December 2013

LED holder design GE June - September 2013

LED holder production CT September - December 2013

Tooling for crystal gluing GE October - December 2013

Tooling for crystal assembly test GE October - December 2013

Tooling for LAAPD benchmarking GE January - March 2013

LAAPD procurement October - May 2013-14

LAAPD benchmarking JLab May - June 2014

New LED holder replacement JLab Jan - March 2014

Crystal tests for regrouping JLab Jan - April 2014

APD replacement JLab June - July 2014

Crystal test and benchmarking JLab July - August 2014

ECal assembly JLab August - September 2014

Manpower

HPS manpower

The HPS Collaboration successfully mounted the HPS Test Run experiment on a veryaggressive schedule, and has the personnel needed to realize full HPS.

Beam-line conceptual and mechanical design will be done at JLab with contributions fromthe Engineering staff at SLAC. The Tracker/Vertexer is being designed and engineered bySLAC as well as the SVT DAQ. The Ecal work is being coordinated by the Orsay Group withparticipation of the JLab group. The Ecal Trigger/DAQ work is done in JLAB DAQ group,which supports Hall B activities and with JLAB Fast Electronic group, which has designed

DRAFT

15

the FADC250. JLab will take care of assembling and testing the electronics, programmingthe trigger, and integrating it with the Ecal hardware. Slow control programming is alsobeing done by JLAB personnel. Simulation work is supported by New Hampshire, SLACand JLAB, Data management and storage and computing infrastructure will be overseenby JLAB Computer Center. The HPS collaboration is about 60 strong, so has adequatemanpower for overall installation, commissioning, and data taking.

INFN-JLAB12 manpower and resposabilities

A detailed work-plan for the ECal upgrade has been established with a realistic estimateof the necessary manpower for the different tasks. Here below is reported the detailed list.

1. Motherboard modification - The motherboard design requires 30 days of FTEengineer to define the electrical layout, lines tracing, place components on the PCB andvalidate the project, plus several iterations with the Orsay HPS mechanical engineer todefine the modifications to the enclosure box, since the new design moves the gap wherethe PCB feeds-trough from the lateral sides to the top/bottom sides. After receivingthe motherboard from the vendor, 10 days of FTE engineer are necessary for testingthe electric connections and motherboard functionality using a complete channel (fullcrystals+LAAPD+preamps chain). The cross talk will be tested using the FT-Cal16-channels prototype. The necessary tools to adapt it to the HPS motherboard,including stands and supports, requires 5 days of FTE mechanical engineer and 5 daysof FTE technical collaborator to machine pieces and mount them. The cross talk testsrequires additional 5 days of FTE engineer and 10 days FTE of technical collaborator.This task will be performed in Genova.

2. Light Monitoring System (LMS) - We can adapt the light monitoring systemdeveloped for the FT-Cal, that thanks to its modular design can be widely reused andadapted to the HPS ECal geometry with a minimal effort. The system is split in threeparts: the controller board (CB), the drivers board (DB) and the LED board (LB).The controller does not need any change while the drivers board layout need to beredesigned, as well as the LED PCB. We foresee 30 days of FTE electronic engineer,supported by 5 days FTE of technical collaborator to design and test a prototype ofthe LMS. The LMS design and prototyping will be done in Genova. The new LEDholder placed in front of the crystals need to be redesigned and machined. The designrequires 10 days of FTE of mechanical engineer and will be done in Genova. Themachining of 500 pieces will be done in Catania requiring 20 days of FTE of technicalcollaborator.

3. Photosensor replacement - The procedure will be split in two phases. Phase-1consists in designing and machining the necessary tools for the massive productionand optimizing the procedure on a sample of 30 crystals. This phase will be done in

DRAFT

16

Italy (Genova) where collaborators from other INFN units (Cagliari, Catania, Torinoand Roma-TorVergata) will work together to acquire the necessary skills. This part willbe done well in advance using 30 spare ECal crystals and 30 LAAPDs spares from theFT-Cal, independently on the effective procurement of LAAPDs and with no impacton ECal assembly. Phase-2, consisting in replacing all the ∼450 remaining APDs, willbe done at JLab where the italian personnel will coordinate and collaborate with alocal technical team appointed by JLab. If, for any reason, the LAPPDs will have anunplanned delay incompatible with the HPS run plan, the Phase-2 will be postponedto the first available downtime (more likely in summer 2015).Details of the different tasks are outlined in the following:

• LAAPD benchmarking: the 500 sensors needs to be characterise to measure thegain of each sensor as a function of the applied bias voltage and temperature,G(V, T ). The measurement will be performed using a similar benchmarking sys-tem used to test the FT-Cal LAAPDs. In Phase-1, the FT-Cal facility will beadapted to the HPS requirements (5 days of INFN electric engineer). An INFNelectric engineer and a technical collaborator will be trained for 3 days with asample of 30 LAAPDs. In Phase-2 the two twin systems, that allow to test up to24 APD per day (about 4 hours per each T set-point), will be sent to JLab. Themassive LAAPDs benchmarking requires a total of 25.5 days of FTE technicalcollaborator and 5.5 days of electronic engineer. The operation will be performedat JLab by a JLab technician working for 4 weeks, supervised, for the first twoweeks, by an INFN electric engineer and a technical collaborator.

• ECal disassembling: the calorimeter will be disassembled and the crystals removedfrom the enclosure. The fiber-holder support, sitting in front of the crystal, willbe replaced by the new LED holder. After this operation, the crystals will betested with an LED pulser to allow a better bias voltage grouping in case wewill be forced to remount the old crystal/APD assembly for late delivery of thenew LAAPD. This operation will be performed by a JLab team of two technicalcollaborators and the supervision of one electric engineer for of 40 days FTE and 5days FTE respectively. At the beginning of the operations, an electronic engineerfrom Italy will set up the benchmark bed in two weeks of work.

• Crystal preparation: when the LAAPDs will be on-site or a firm estimate of theirarrival will be possible, the crystal preparation will begin. The PbWO crystalsneed to be carefully de-wrapped from the VM2000 wrapping envelope, placed inalcohol for a whole night to weak or dissolve the glue between the crystal and theAPD, remove the APD and clean the crystal face by acetone. To perform thisoperation two arrays of gas-tight plastic containers, partially filled by alcohol, willbe prepared in advance. For Phase-1, the estimated time necessary for tooling is5 days FTE technical collaborator, and 3 days to test the procedure and train theINFN personnel (3+3 days of FTE technical collaborator). This task will be done

DRAFT

17

in Italy on a sample of 30 crystals, used to test and optimize the whole procedure.In Phase-2, the full job will be performed at JLab, requiring about 29 days of FTEshared between a JLab team of two technical collaborators, overlapping for thefirst 10 days with an italian team of 3, 1 supervisor and 2 technical collaborators.

• LAAPD glueing: having the crystal cleaned the next phase is to glue the LAAPD,and place the old wrapping around. The glueing procedure takes a full night toharden the glue and therefore the job needs to be done in parallel over a set ofmany crystals at time. Two ”L”-shape aluminium+neoprene racks to keep thecrystals in vertical position will be prepared in advance. The time necessary todesign and machine the tools is estimated in 5 days FTE mechanical engineerand 10 days of technical collaborator while the time to optimize the procedure(in Italy) is the same as for the cleaning phase. The same sample of 30 crystalswill be used to practice. The full job on the remaining 450 crystals will beperformed at JLab by a JLab team of 2 technical collaborators working for 2weeks, overlapping for the most part of the time with an italian group of 2 (1supervisor and 2 technical collaborators).

• Crystal assembly test: before placing a crystal into ECal, the full assembly will betested to check, within 20-30% accuracy, if the glueing and wrapping procedurewas correctly performed. About 20 crystals at-time will be placed over a fixed-temperature plate in a T-controlled room and let them wait for a night to reduceany T-related effect (APD-Gain dependence, LED emission, ...). Each crystal willbe tested by connecting one reference LED pulser from one side and a preamplifierto the LAAPD. Each data taking will take 10 minutes to accumulate enoughstatistics. When done, the LED and the preamp will be moved to the next crystal.The data acquisition can be performed with an oscilloscope able to accumulatethe signal amplitudes in a histogram. For Phase-1, we estimate 15 days of FTE todesign and produce the necessary tools (4 identical LED drivers will be producedand tuned), 10 days of FTE mechanical engineer to design the plate and thesupport for the LED pulser and the preamps, 10 days of technical collaborator tomachine it, 5 days to test and optimize the benchmarking procedure in Genova.The training for the visiting INFN personnel will take 3 FTE days of electricengineer and a total of 6 days of technical collaborators. In Phase-2, the 450crystals will be assembled at JLab by the team of 2 JLab technical collaborators in3 weeks, overlapping with a team of 4 (2 supervisors and 2 technical collaborators)from Italy for the first 10 days.For crystals showing a poor light transmission, out of the desired range, thewhole cleaning/glueing/testing procedure will be repeated. We consider a failurefraction of the gluing procedure of about 10%, included in the previous estimate.

• Ecal assembly: the JLab team of two technical collaborators, supervised by anengineer, will mount the refurbished crystals in the ECal box. The operation is

DRAFT

18

expected to take about 2 weeks for a total of 20 days of FTE technical collabora-tors and 5 days of supervisor. No italian personnel is expected to be involved inthis task. Orsay personnel will help JLab technical staff to assemble the ECal.

• ECal final tests the final assembly of the machined crystals in the ECal enclosurewill be done by the 2 JLab technical collaborators for a total of 2 weeks. A Jlabsupervisor and three italian electric engineers will will insure the full overlap.

The detailed work plan is reported in Tables II and III.

Cost and funding

HPS cost and funding

The total cost for HPS is $3.187 M, consisting of $1.692 M capital equipment, $927 Koperations, and $567 K infrastructure. HPS is seeking funding from other sources for theMuon System and upgrades to the Ecal. The College of William&Mary (Williamsburg, VA),together with ODU and Rutgers University have submitted an MRI proposal to NSF for theMuon System, requesting ∼ $280k. IPN ORSAY (France) has submitted a proposal to aFrench funding agency for the various ECal upgrades, including an ECal Light MonitoringSystem ($100k), and other expenses related to ECal fabrication and test.

INFN-JLAB12 commitment

The budget requested to the INFN-CSNIII is reported in Tab. IV. The main cost isrelated to the LAAPD procurement (see the attached quote of 215k euro). The other items(motherboards, LLMSMS prototyping, tooling and shipping) account for a total of 50k euro.The precise sharing among the INFN-JLAB12 groups will be presented to the INFN-CSNIIIin June/September 2013. It worth to be noticed that, since the final destination of themachined crystals is JLab (US), the LAAPD purchase is VAT exempted. The total requestof about 300k euro already takes it into account.

Travel

We estimated about 240 travel days to accomplish the calorimeter upgrade, a total of 12domestic and 16 trips to JLab (see Tabs. II and III for more details about the travel timeestimate). We also foresee about 60 days for detector commissioning and shifts for the firstphysics runs together with 6 trips oversea. Another 20 travel-days and 3 trips are requestedto attend HPS Collaboration meetings. Requests are summarized in Tab. ??.

DRAFT

19

TABLE II: Work plan - Phase-1: activities in Italy (Genova and Catania).

Activity Workers FTE (total) Travel days INFN Units

Motherboards design EE 30 GE

Motherboards tests EE, TC 35 15 GE,TO

EE 15 GE

ME 5 GE

TC 5 5 TO

TC 10 10 TO

LED LMS design/prototyping EE, TC 35 5 GE, TO

EE 30 GE

TC 5 5 TO

LED holders design ME 10 GE

LED holders production TC 20 CT

LAAPD benchmarking tooling EE, TC 11 10 GE,RM2,CT

EE 5 GE

EE 3 5 RM2

TC 3 5 CT

Crystal cleaning tools and proc. optimiz. TC 11 10 GE,CT,CA

TC 5 GE

TC 3 5 CA

TC 3 5 CT

Crystal gluing tools and proc. optimiz. TC 26 10 GE,CT,CA

ME 5 GE

TC 10 CA

TC 5 GE

TC 3 5 CA

TC 3 5 CT

Crystal assembly test tooling and proc. opt. EE, TC 49 15 GE,CT,CA

ME/EE 25 GE

TC 10 CT

TC 5 GE

TC 3 5 CA

TC 3 5 CT

EE 3 5 CT

Total 140 GE

Total 15 25 CT

Total 20 20 TO

Total 9 15 CA

Total 3 5 RM2

Total 187 65

DRAFT

20

TABLE III: Work plan - Phase-2: activities at JLab.

Activity Workers FTE (total) Travel days INFN Units

LAAPD benchmarking (500pcs) EE, TC 31 21 RM2,GE

1xTC 20 JLab

1xEE 5.5 10.5 RM2

1xTC 5.5 10.5 GE

ECal disassembly EE, TC 51.5 10.5 GE

2xTC 40 JLab

1xEE 5 JLab

1xEE 5.5 10.5 GE

Crystal cleaning (450 pcs) EE, TC 31.5 31.5 CT

2xTC 15 JLab

2xTC 11 21 CT

1x EE 5.5 10.5 CT

Crystal gluing (450 pcs) EE, TC 31.5 31.5 CA,CT

2xTC 15 JLab

2xTC 11 21 CA

1xEE 5.5 10.5 CT

Crystal assembly test (450 pcs) EE, TC 47 42 CA,CT,GE

2xTC 25 JLab

1xTC 5.5 10.5 CA

1xTC 5.5 10.5 CT

1xEE 5.5 10.5 CT

1x EE 5.5 10.5 GE

ECal assembly EE, TC 25

2xTC 20 JLab

1xEE 5 JLab

ECal final tests EE, TC 46.5 31.5 CT,GE

2xTC 20 JLab

1xEE 10 JLab

1xEE 5.5 10.5 CT

1xEE 5.5 10.5 GE

1xEE 5.5 10.5 CA

Total 175 JLab

Total 38.5 73.5 CT

Total 22 42 CA

Total 22 42 GE

Total 5.5 10.5 RM2

Total 263 168

DRAFT

21

TABLE IV: INFN-JLAB12 funding contributions.

Components Cost (Euro)

Motherboards design and production 10k

Tools for motherboard testing 3k

LMS design and prototype 5k

LED holders 5k

Tools for crystal assembly 8k

Tools for LAAPD tests 5k

LED system for assembly tests 4k

Shipping 4k

Consumables 6k

LAAPDs (500pcs Hamamatsu S8664-1010 Sel3) 215k

Total + VAT a + 10% contingency 303k

aFor LAAPDs only contingency has been accounted

TABLE V: INFN-JLAB12 travel money.

Motivation Days/number Cost (Euro)

Domestic travels 65 13k

Domestic trips 13 5k

JLab travels: detector construction 168 22k

JLab travels: detector commissioning 20 3k

JLab travels: test run 30 4k

JLab travels: collaboration meetings 30 4k

JLab trips 31 31k

Total + 10% contingency 90k

INFN-JLAB12 Software commitment

The INFN-JLab12 Collaboration is willing to collaborate also in detector monitoring,calibration and data analysis. In particular we can provide an initial contribution in thefollowing area: 80 hours for ECal low-level monitoring; 80 hours ECal reconstruction andCalibration; 80 hours contribution to tracking alignment reconstruction.

DRAFT

22

Offerta n.: I103421Data preventivo: 05/04/13Revisione n.: 2

Ns. riferimento: Mauro BombonatiTel.: 00390293582303Fax: 00390293581741E-mail: bombonati@hamamatsu.it

Codice cliente: HPI000647Vs. Rif.: MARCO BATTAGLIERIRFQ n.:

Dr. MARCO BATTAGLIERIINFN GENOVAVIA DODECANESO, 3316146 GENOVA (GE)ITALY

Pagina 1/1

Pos. Codice prodotto Quantità Prezzo unit. Importo nettoDescrizione [pcs] EUR EUR

1.1 S8664-1010SEL3 500 430,00 215.000,00Si Avalanche Photodiode,SEL:gain +/-10%, M=150 in pairsof ten pcs, laser marked inbackside

Legame cambio:I prezzi riportati in offerta sono riferiti ad un legame cambio di 115 JPY/EUR. Eventuali variazioni della valuta intervenute alladata di fatturazione verranno addebitate/accreditate.

Validità offerta: 6-mag-2013Termini di spedizione: Porto Franco Vs Sede con spese di spedizione in fattura €15,49/shipment (spedizione al CERN

circa 1500€)Termini di pagamento: R.D. 60 GG. D.F.Garanzia: 12 MesiTempi di consegna: Approx. 120 GG DRO (da confermare in prossimità dell'ordine di acquisto)

I prezzi si intendono al netto dell'aliquota IVA.

Questa offerta è soggetta ai Termini e alle Condizioni Generali di Vendita di Hamamatsu Photonics Italia, riportate nel nostrosito: www.hamamatsu.eu/gtc/it

In fede

HAMAMATSU PHOTONICS ITALIA S.R.L. A SOCIO UNICO

Offerta

HAMAMATSU PHOTONICS ITALIA S.R.L. A SOCIO UNICOSede Legale e amministrativa: Coordinate Bancarie: Registrazioni Commerciali:Strada della Moia, 1 int.6 Intesa Sanpaolo Partita IVA IT1024719015920020 Arese (Mi) - Italia BIC: BCITITMM Registro Imprese di Milano n. 312362Tel: +39 02 9358 1733 IBAN: IT02 C030 6932 4411 0000 0001 456 (EUR) Capitale Sociale 728.000 EURFax: +39 02 9358 1741 IBAN: IT60 M030 6932 4411 6100 9347 239 (JPY) REA di Milano n. 1356127Web: www.hamamatsu.it Iscrizione Registro AEE n IT 08030000003931

FIG. 6: Hamamtsu quote