MPGD-based photon detectors for the upgrade of COMPASS RICH-1 and beyond
J. Agarwala, M. Alexeev, C.D.R. Azevedo, F. Bradamante, A. Bressan, M. Buchele, C. Chatterjee, M. Chiosso, A. Cicuttin, P. Ciliberti, M.L. Crespo, S. Dalla Torre, S. Dasgupta, O. Denisov, M. Finger, M. Finger Jr, H. Fischer, L. García Ordóñez, M. Gregori, G. Hamar, F. Herrmann, S. Levorato, A. Martin, G. Menon, D. Panzieri, G. Sbrizzai, S. Schopferer, M. Slunecka, M. Sulc, F. Tessarotto, J.F.C.A. Veloso, Y.X. Zhao
PPrepared for submission to JINST
International conference on Instrumentation for Colliding Beam Physics INSTR (cid:48) T h
February, 2020 to 28
T h
February, 2020Budkar Institute of Nuclear Physics, Novosibirsk, Russia
MPGD-based photon detectors for the upgrade ofCOMPASS RICH-1 and beyond
J. Agarwala , b , M. Alexeev , C.D.R. Azevedo , F. Bradamante , A. Bressan , M. Büchele ,C. Chatterjee , M. Chiosso , A. Cicuttin , P. Ciliberti , M.L. Crespo S. Dalla Torre ,S. Dasgupta , a O. Denisov , M. Finger , M. Finger Jr , H. Fischer ,L. GarcÃŋa OrdÃşÃśez , , , M. Gregori , G. Hamar , c , F. Herrmann , S. Levorato ,A. Martin , G. Menon , D. Panzieri , G. Sbrizzai , S. Schopferer , M. Slunecka , M. Sulc ,F. Tessarotto , J.F.C.A. Veloso , Y.X. Zhao , d Abdus Salam ICTP and INFN Trieste, Trieste, Italy Charles University, Prague, Czech Republic and JINR, Dubna, Russia Engineering Dep. of Trieste University, Trieste, Italy. INFN Torino, Torino, Italy INFN Trieste, Trieste, Italy Technical University of Liberec, Liberec, Czech Republic Universität Freiburg, Freiburg, Germany I3N, Department of Physiccs, University of Aveiro, Aveiro, Portugal University of East Piemonte, Alessandria and INFN Torino, Torino, Italy University of Torino and INFN Torino, Torino, Italy University of Trieste and INFN Trieste, Trieste, Italy a Corresponding author b Present address: University of Pavia and INFN Pavia, Pavia, Italy c Present address: Wigner Research Centre for Physics, Budapest, Hungary d Present address: Institute of Modern Physics, Chinese Academy of Sciences, Lanzhou, 730000, China
E-mail: [email protected] a r X i v : . [ phy s i c s . i n s - d e t ] J un bstract: COMPASS is a fixed target experiment at CERN SPS aimed to study hadron structure andspectroscopy. Hadron identification in the momentum range between 3 and 55 GeV / c is provided bya large gaseous Ring Imaging Cherenkov Counter, RICH-1. To cope with the challenges imposedby the new physics program of COMPASS, RICH-1 has been upgraded by replacing four MWPC-based photon detectors with newly developed MPGD-based photon detectors. The architecture ofthe novel detectors is a hybrid combination of two layers of THGEMs and a Micromegas. The topof the first THGEM is coated with CsI acting as a reflective photo-cathode. The anode is segmentedin pads capacitively coupled to the APV-25 based readout. The new hybrid detectors have beencommissioned during the 2016 COMPASS data taking and stably operated during the 2017 run.In this paper design, construction, operation and performance aspects of the novel photondetectors for COMPASS RICH-1 are discussed.Keywords: RICH, PID, MPGD, Photon Detectors, THGEMs, resistive MicromegasArXiv ePrint: 1234.56789 ontents RICH-1 [1] is a large gaseous Ring Image Cherenkov (RICH) Counter providing Particle Identifi-cation (PID) for hadrons within the momentum range from 3 to 55
GeV / c for the COMPASS [2]Experiment at CERN SPS. It consists of a 3 m long C F gaseous radiator, where Cherenkov pho-tons are generated when ultra-relativistic charged particles cross it; VUV spherical mirrors coveringa surface of 21 m reflect the photons and focus them onto a 5.5 m photo-detection surface sensitiveto single photons (Fig1-A). Three photo-detection technologies are used in RICH-1: Multi WireProportional Chambers (MWPCs) with CsI photo-cathodes, Multi Anode Photo-Multipliers Tubes(MAPMTs) and novel Micro Pattern Gaseous Detector (MPGD) based Photon Detectors (PDs)(Fig.1-B). The novel PDs are the focus of this article.RICH-1 was designed and built between 1996-2000, commissioned in 2001-2002 and is inoperation since 2002. The whole photo-detection surface was originally equipped with 16 MWPCswith CsI photo-cathodes of ∼ ×
600 mm active area. In-spite of their good performance,MWPCs have limitations in terms of time resolution, maximum effective gain ( ∼ ), timeresponse( ∼ µ s), rate capability and ageing of the CsI photo-cathodes. In 2006, four centralchambers were replaced with detectors consisting of MAPMTs coupled to individual fused silicalens telescopes to cope, thanks to their excellent time resolution, with the high particle rates ofthe central region. In parallel, an extensive R&D program [3] aimed to develop MPGD-basedlarge area PDs established a novel hybrid technology combining Micromegas [4] and THick GasElectron Multipliers (THGEMs) [5]. In 2016 COMPASS RICH-1 was upgraded by replacing fourof the remaining 12 MWPCs with new detectors based on the novel hybrid MPGD technology [6].The new detectors have been successfully commissioned and operated during the 2016 and 2017COMPASS data taking periods. – 1 – B Figure 1 . A. The Cherenkov photon propagation and focusing: the principle is illustrated by a schematicside view. B. Photon detector arrangement (not to scale), where the different detection technologies in useare indicated.
Figure 2 . Schematic of the hybrid MPGD based PD (not to scale).
The basic structure of the hybrid module (Fig.2) consists of thre multiplication layers: twolayers of THGEMs and a final Micromegas one. The architecture is completeded by two planes– 2 –f wires. UV light sensitivity is obtained via a thin (300 nm) CsI film deposited on the top of thefirst THGEM electrode which acts as a reflective photo-cathode, an approach studied also by otherresearch groups [7].The detectors are operated with Ar : CH =
50 : 50 gas mixture. The Drift wire plane isinstalled at 4 mm from the CsI coated THGEM and is biased to a suitable voltage in order tomaximize the extraction and collection efficiency of the converted photo-electrons [8]. The otherwire plane guarantees the correct closure of the drift field lines and is positioned 4.5 mm away fromthe quartz window which separates the radiator gas volume from the photon detector gas volume.The photo-electron generated by the conversion of Cherenkov photon from the CsI surfaceis guided into one of the first THGEM holes where a first avalanche process takes place due tothe electric field generated by the biasing voltage applied between the top and bottom THGEMelectrodes. The electron cloud is then driven by the electric field across the 3 mm transfer regionto the second THGEM, where thanks to complete misalignment of the holes with respect to thefirst THGEM, the charge is distributed, typically, among three holes and undergoes a secondmultiplication process.
Figure 3 . A: Exploded view of one single readout pad structure. B: The schematic of the circuit diagramof the resistive/capacitive anode concept. C and D: Metallographic section of the PCB: the detail of thethrough-via contacting the external pad through the hole of the buried pad [6] .Finally the charge is guided by the electric field across the 5 mm gap to the bulk Micromegaswhere the last multiplication occurs. The spread of the charge over a large surface enhance theMicromegas electrical stability at higher gains. The Micromegas mesh, which is the only non-segmented electrode is kept at ground potential while the anode, segmented in square pads of7.5 × mm (with 0.5 mm inter-pad gaps) is biased at positive voltage (Fig.3-A and -B). The– 3 –icromegas PCBs are based on the capacitive/resistive concept: the signal generated on the anodepad is capacitively transfered to a second pad, parallel to the first one and burried inside the PCB.The two pads are separed by a fiberglass layer 70 µ m thick. Each anode pad is powered throughindividual resistors. Therefore, these detectors represent the first implementation in an experimentof the concept of resistrive Micromegas; this first implementation is by discrete elements, namelythe resistors in series with the individual pads .The high voltage connection to the anode pads is by vias passing through holes of the readoutpads (Fig.3-A and -C). Special attention was paid in order to obtaining a very flat surface of theanode pad via the careful engineering of the connections (Fig.3-D).The typical voltage applied to the multiplication stages are 1270 V across THGEM1, 1250 Vacross THGEM2, and 620 V to bias the MM. The drift field above the first THGEM is 500 V/cm,the transfer field between the two THGEMs is 1000 V/cm and the field between the second THGEMand the MM micromesh is 1000 V/cm. The effective gain-values for the three multiplication layersare around 12, 10 and 120; these values include the electron transfer efficiency.The intrinsic ion blocking capabilities of the Micromegas as well as the arrangements of theTHGEM geometry and electric fields configuration grant an ion back flow to the photo-cathodesurface lower or equal to 3% [9]. Concerning the figures of ion backflow rates, this hybridarchitecture has results more effective than the use of triple staggered THGEMs [10]. Figure 4 . Top: Cad drawing of a 287 × mm THGEM with 12 sectors; Bottom left: Zoom of a cornerof the CAD drawing showing the increased hole diameter of the border holes; Bottom right: Picture of thecorner for a THGEM.
All THGEMs have the same geometrical parameters: thickness of 470 µ m, total length of 581mm and width of 287 mm (Fig.4). The holes are arranged in a hexagonal matrix formation having– 4 –iameter of 400 µ m and pitch of 800 µ m. The holes are produced by mechanical drilling and haveno rim. To obtain a more symmetric field line configuration near the edges of the THGEM, theholes located along the external borders have 500 µ m diameter. This solution provides an improvedelectrical stability of the whole system. The electrodes on the two faces of each THGEM aresegmented in 12 sectors separated by 0.7 mm clearance area. The biasing voltage is individuallyprovided to each sector.The THGEMs are produced following a dedicated protocol, which is one of the results of aneight-year long dedicated R&D. The protocol includes raw material selection, THGEM production,quality assessment, characterization, CsI coating, storage and installation.Achieving an effective gain uniformity over large areas is challenging due to the poor thicknesstolerance of the raw PCB sheets available from producers. A setup based on MITUTOYO EUROCA776 coordinate measuring machine operated in a climatized room was used to map the localthickness values in order to select adequate raw material. In total 50 foils were measured from which100 THGEMs (2 THGEMs/foils) could be produced. Foils with a maximum thickness variationof 15 µ m peak-to-peak were selected for the industrial production including transfer of the maskimage, etching and drilling of the holes. 60% of the foils have been accepted and 60 THGEMs havebeen industrially produced. Thanks to the raw material selection the gain uniformity obtained is atthe at the ∼
7% level over the THGEM active surface of ∼ . m .A post production treatment was applied at INFN Trieste to the industrially produce THGEMs:it consists in polishing the raw THGEM with pumice powder and cleaning by high pressurizedwater, ultrasonic bathing with high pH detergent solution ( pH ∼ C for 24 hours [11]. Then, a measurement of the discharge rate wasperformed using an automated test setup: in an Ar : CO = 70 : 30 gas mixture the bias voltage ofthe THGEM was increased in 10 V steps and the number of sparks (events with more than 50 nAcurrent) was measured for 30 minutes until the bias voltage was increased again. The THGEMs witha discharge rates of less than 1/hour at 1200 V were validated. Effective gain uniformity study wasperformed using a dedicated test setup consisting of a mini-X X-Ray generator and APV-25[12]based SRS[13] read-out system. The best pieces were selected for the upgrade.The Micromegas were produced by bulk technology [14] at CERN EP/DT/EF/MPT workshopover the pad segmented multilayer PCBs. The 600 × mm PDs were built by mounting two300 × mm modules side by side in the same frame. The Micromegas PCBs have been gluedto the supporting frame making use of a volumetric dispenser coupled to a computer numericalcontrol machine. The Micromegas multipliers have been assembled in a clean room.A special box to transport validated THGEMs under controlled atmosphere was used before andafter their Au-Ni coating. The deposition of the solid photo-converter for the hybrid photo-cathodeswas performed at the CERN Thin film Laboratory following the procedure described in [15]. Thephoto-cathodes (THGEMs with CsI coating on one side) were mounted inside a dedicated glove-box. The PDs were then installed on COMPASS RICH-1 and equipped with front-end electronics,low voltages, high voltages and cooling services. – 5 – igure 5 . Amplitude spectrum from one of the hybrid PDs; each entry is the response of an anode pad withamplitude above threshold; a. No beam (random trigger); b. With beam (physics trigger). The new hybrid detectors were commissioned during the COMPASS data taking period from Mayto October 2016. The observed average equivalent electron noise is ∼ e − . A zero suppressionprocedure with a 3 σ threshold cut and a common mode reduction are applied for the standarddata taking. After ensuring accurate timing, the amplitude spectra for noise and signals have beencollected. With no beam and random trigger, the noise part of the amplitude spectrum is observed(Fig.5-a). With physics triggers the amplitude spectrum shows the noise part, a prominent singlephoton exponential part and a tail due to charged particle signals (Fig.5-b).Important gain variation with environmental conditions (pressure and temperature) are expectedin a multilayer gaseous detector. During the R&D phase, these effects had been studied in order tointroduce an automated high voltage correction system to compensate for the effect and to stabilizethe PD gain response [16]. During the data taking, the performance, in terms of voltage and gainstability of the PDs have been studied. In Fig.6-A. the average spark rate per day for each of thehybrid detectors is provided for the COMPASS 2017 data taking period. A typical discharge rate offew events per day is seen. No sizable voltage drop is observed when a discharge occurs, confirmingthe validity of the design of the HV distribution system [16]. In Fig.6-B. a typical signal amplitudespectrum has been shown. The estimated gain is around 13.5k. In Fig.6-C. the measured PDgas pressure and temperature curves are shown. The effective gain stability can be extracted fromFig.6-D, where the values of the effective gains extracted from the data, two measurements per day,at the time when the minimum and maximum temperature is expected, are presented. The gain isstable at the 6% level.Typical Cherenkov rings are presented in Fig.7-left. The centres of the expected ring patternsis obtained from the reconstructed particle trajectories; the particle momentum and the expectedCherenkov angle in the pion hypothesis are also reported. No image elaboration or backgroundsubtraction is applied. The analysis shows that the residual angular resolution of those PDs are ∼ N θ ch = p . sin θ ch + p .θ ch (4.1)– 6 – B C
Figure 6 . A. Average discharge rate per day per detector over COMPASS 2017 data taking period. B. Atypical amplitude spectrum for single photons from one of the hybrid detectors with bias voltages: 1250 Vfor THGEM1, 1200V for THGEM2, 630 V for the Micromegas and with 1 kV . cm − transfer fields betweenamplification stages. The estimated effective gain is 13511 . ± . where N θ ch is the average number of detected photons per Cherenkov ring for a particularCherenkov angle θ ch , p and p are the fit parameters for the Cherenkov photon part accordingto the Frank and Tumm distribution and for the background part of the data respectively, assumedlinear as suggested by geometrical considerations. The fit estimation indicates 12.9 photons perring at Cherenkov angle equals to 55.2 mrad; the signal part is 10 . ± . . ± . The THGEM-MM hybrid PD technology could be used for other applications, for instance fora RICH dedicated to PID of high momentum particles at experiments at the future EIC collider[17], where higher space resolution of the PD would be needed because of the more compressedgeometry that imposes a reduced lever arm.A prototype [18] similar to the COMPASS PDs has been designed and built: it has an activearea of 10 ×
10 cm , an anode segmented in 1024 square pads having 3.5 mm pitch. The prototype isfully modular, with front-end electronics and almost all services contained in the 10 ×
10 cm activearea: detectors covering larger areas could be designed by multiple replica of the basic modulerepresented by the prototype. – 7 – B Figure 7 . Left: Images of hit pattern in the novel photon detectors. The center of the expected ring patternsis obtained from the reconstructed particle trajectories; the particle momentum and the expected Cherenkovangle in the pion hypothesis are also reported. No image elaboration or background subtraction is applied.Right, A: Distribution of the difference between the Cherenkov angle calculated from the reconstructedparticle momentum and the Cherenkov angle provided by single detected photoelectrons; a sample ofidentified pions is used. Right, B: Average number of detected single photons per Cherenkov rings versusthe Cherenkov angle.
The internal structure of the modular minipad prototype presented in Fig.8 and it reproducesthe basic scheme of the hybrid MPGD implemented for COMPASS RICH-1.The detector components have been prepared adopting procedures and protocols similar tothose used for COMPASS hybrid PDs. The prototype has been tested at INFN Trieste laboratoriesand, in October - November 2018, in a test beam at CERN SPS. The beam particles, alternatively π and µ , traversed on a truncated cone fused silica Cherenkov radiator aligned with the center ofthe pad plane. A shutter, remotely controlled, sitting between the radiator and the detector, madepossible to collected data including Cherenkov photons from the radiator or excluding them.Cherenkov rings has been observed for different CH rich gas mixtures as shown in Fig.9. Theanalysis of the test beam data is ongoing. Novel MPGD-based hybrid detectors of single photons, matching the present status of the art in theMPGD technology, have been designed, engineered, built, tested and operated in a RICH detectorfor the first time. The main achievements are: – 8 – igure 8 . Exploded view of Mini-PAD hybrid schematic from CAD drawing
Figure 9 . Observed Cherenkov rings in Ar : CH CH gas. First column:data collected with the shutter open. Central column: data collected with the shutter closed. Rightcolumn: distributions obtained subtracting those obtained with shutter open and shutter closed after properlynomalization to the same number of triggers. – 9 – The four new PDs have been stably operated during 2017 COMPASS data taking periodmatching the COMPASS RICH requirements.• It is the first time that single photon detection is performed in an experiment by MPGD PDs.• It is the first application in an experiment making use of THGEMS and resistive Micromegas.• The typical gain, well above 10 , is the highest at which MPGD have been operated inexperiments.• The characterization shows the expected angular resolution and around 10-12 detected pho-tons per ring at Cherenkov angle saturation.• The hybrid Micromegas and THGEM photon-detection technology has proven to be success-ful and robust.• There are promising perspective for the increase of the spacial resolution using this technology.A new R&D has already been started for the future needs of the EIC experiments with modifieddesign parameters to achieve a higher space resolution. The activity is partially supported by the H2020 project AIDA2020 GA no. 654168.The authors are member of the COMPASS Collaboration and part of them are members ofthe RD51 Collaboration: they are grateful to both Collaborations for the effective support and theprecious encouragements.The Portuguese collaboration was partially supported by CERN/ FIS-PAR/0022/2019 through FCT (Lis-bon). [1] E. Albrecht, et al ., Status and characterisation of COMPASS RICH-1, Nucl. Instrum. Meth.
A 553 (2005) 215; P. Abbon et al ., Read-out electronics for fast photon detection with COMPASS RICH-1,Nucl. Instrum. Meth.
A 587 (2008) 371; P. Abbon et al ., Design and construction of the fast photondetection system for COMPASS RICH-1, Nucl.Instrum. Meth.
A 616 (2010) 21; P. Abbon et al .,Particle identification with COMPASS RICH-1, Nucl. Instr. and Meth.
A 631 (2011) 26.[2] P. Abbon, et al., The Compass Experiment at CERN , Nucl. Instrum. Meth.
A 577 (2007) 455;Ph. Abbon et al., The COMPASS setup for physics with hadron beams Nucl. Instrum. Meth.
A 779 (2015) 69;[3] M. Alexeev et al ., The quest for a third generation of gaseous photon detectors for Cherenkov imagingcounters Nucl. Instrum. and Meth.
A 610 (2009) 174; M. Alexeev et al ., Long term experience withCsI photocathodes in gas photon detectors JINST (2014) P01006 ; M. Alexeev et al ., Measurementof the evolution of the CsI quantum efficiency in gas photon detectors Nucl. Instrum. and Meth.
A 766 (2014) 199; references therein. – 10 –
4] Y. Giomataris et al ., MICROMEGAS: a high-granularity position-sensitive gaseous detector for highparticle-flux environments, Nucl. Instrum. Meth.
A 376 (1996) 29.[5] L. Periale et al ., Detection of the primary scintillation light from dense Ar, Kr and Xe with novelphotosensitive gaseous detectors Nucl. Instr. and Meth.
A 478 (2002) 377; P. Jeanneret, Timeprojection chambers and detection of neutrions PhD thesis, Neuchatel University, 2001; P. S. Barbeau et al ., Toward coherent neutrino detection using low-background micropattern gas detectors IEEE
NS-50 (2003) ; R. Chechik et al ., Thick GEM-like hole multipliers: properties and possibleapplications Nucl. Instrum. and Meth.
A 535 (2004) 303; R. Checkik et al ., Thick GEM-likemultipliersâĂŤa simple solution for large area UV-RICH detectors Nucl. Instrum. and Meth.
A 553 (2005) 35.[6] J. Agarwala et al ., The MPGD-based photon detectors for the upgrade of COMPASS RICH-1 andbeyond, Nucl. Instrum. and Meth.
A 936 (2019) 416; J. Agarwala et al., The hybrid MPGD-basedphoton detectors of COMPASS RICH-1, Nucl. Instrum. and Meth.
A 952 (2020) 161832.[7] C. D. R. Azevedo et al ., Towards THGEM UV-photon detectors for RICH: on single-photon detectionefficiency in Ne/CH4 and Ne/CF4 JINST (2010) P01002 .[8] G. Hamar et al ., Direct measurements of the properties of Thick-GEM reflective photocathodes,https://arxiv.org/abs/1910.05257, submitted to Nucl. Instr. and Meth. A .[9] F. Tessarotto, Status and perspectives of gaseous photon detectors Nucl. Instrum. and Meth. A 876 (2017) 225[10] M. Alexeeev et al ., Ion backflow in thick GEM-based detectors of single photons JINST (2013) P01021 .[11] M. Alexeev et al ., Status and progress of the novel photon detectors based on THGEM and hybridMPGD architectures Nucl. Instrum. and Meth.
A 766 (2014) 133.[12] M. J. French, et al ., Design and results from the APV25, a deep sub-micron CMOS front-end chip forthe CMS tracker Nucl. Instrum. and Meth.
A 466 (2001) 359.[13] S. Martoiu et al ., article is Development of the scalable readout system for micro-pattern gas detectorsand other applications JINST (2013) C03015 .[14] I. Giomataris et al ., Micromegas in a bulk, Nucl. Instrum. and Meth.
A 560 (2006) 405.[15] A. Braem et al ., Technology of photocathode production Nucl. Instrum. and Meth.
A 502 (2003) 205.[16] J. Agarwala et al ., The high voltage system with pressure and temperature corrections for the novelMPGD-based photon detectors of COMPASS RICH-1, Nucl. Instrum. and Meth.
A 942 (2019)162378.[17] A. Accardi et al ., Electron-Ion Collider: The next QCD frontier - Understanding the glue that binds usall Eur. Phys. J A52 (2016) .[18] J. Agarwala et al ., Study of MicroPattern Gaseous detectors with novel nanodiamond basedphotocathodes for single photon detection in EIC RICH Nucl. Instrum. and Meth.
A 952 (2020)161967.(2020)161967.