Efficient ion blocking in gaseous detectors and its application to gas-avalanche photomultipliers sensitive in the visible-light range
A. V. Lyashenko, A. Breskin, R. Chechik, J. M. F. Dos Santos, F. D. Amaro, J. F. C. A. Veloso
aa r X i v : . [ phy s i c s . i n s - d e t ] A p r Efficient ion blocking in gaseous detectors and its application to gas-avalanchephotomultipliers sensitive in the visible-light range
A. Lyashenko a , ∗ , A. Breskin a , R. Chechik a J.M.F. dos Santos b , F.D. Amaro b and J.F.C.A. Veloso b , c a Department of Particle Physics, The Weizmann Institute of Science, 76100 Rehovot, Israel b Physics Department, University of Coimbra, 3004-516 Coimbra, Portugal c Physics Department, University of Aveiro, Campus Universit´ario de Santiago, 3810-193 Aveiro, Portugal
Abstract
A novel concept for ion blocking in gas-avalanche detectors was developed, comprising cascaded micro-hole electron multiplierswith patterned electrodes for ion defocusing. This leads to ion blocking at the 10 − level, in DC mode, in operation conditionsadequate for TPCs and for gaseous photomultipliers. The concept was validated in a cascaded visible-sensitive gas avalanchephotomultiplier operating at atmospheric pressure of Ar/CH (95/5) with a bi-alkali photocathode. While in previous works highgain, in excess of 10 , was reached only in a pulse-gated cascaded-GEM gaseous photomultiplier, the present device yielded, forthe first time, similar gain in DC mode. We describe shortly the physical processes involved in the charge transport within gaseousphotomultipliers and the ion blocking method. We present results of ion backflow fraction and of electron multiplication in cascadedpatterned-electrode gaseous photomultiplier with K-Cs-Sb, Na-K-Sb and Cs-Sb visible-sensitive photocathodes, operated in DCmode. Key words: gaseous photomultipliers, ion back-flow, ion feedback, bi-alkali photocathodes
PACS:
1. Introduction
Controlling the back-flow of ions generated in gasavalanches has important consequences on the operationand properties of gaseous detectors. Avalanche ions in-duce space charge effects that limit the gain, counting-ratecapability and localization properties of tracking detec-tors. The ion impact on photocathodes (PC) of gaseousphotomultipliers (GPM) causes their permanent damage[1,2]; more seriously, it induces the emission of secondaryelectrons which, in turn, cause avalanche divergence, dete-rioration of timing and localization information and, mostimportantly, severe gain limits. All these consequences areincluded in the term Ion Feedback effects. GPMs sensitivein the UV range, with CsI PCs, do suffer some ion inducedPC damage [2], but the ion-induced secondary electronemission probability is very low and does not limit their op-eration, even at high gain. There are numerous large-areaCsI-GPMs, presently operating or under construction, in ∗ Corresponding author: tel. +972-8-934-2064, fax +972-8-934-2611,E-mail: [email protected] many particle-physics experiments, e.g. COMPASS [3] andALICE [4] at CERN and PHENIX [5] at BNL.The ion feedback effects are particularly problematic invisible-sensitive GPMs, due to the high electron emissionprobability of bi-alkali and other PCs sensitive in the visiblespectral range. The reader is referred to [6] for an extendeddiscussion on this subject and references to recent worksdealing with methods of ion back-flow reduction.The only method for significantly reducing the ion back-flow fraction (IBF) and reaching high-gain operation invisible-sensitive GPMs has been, so far, their operation in agated mode [1]. Our goal has been to find methods for effi-ciently reducing the IBF to permit the operation of visible-sensitive GPMs in DC mode with single-photon sensitiv-ity. Obviously, large tracking TPCs will also benefit, as lowIBF values would permit their stable DC operation.In the present article, we report on our recent results onblocking of ion back-flow in cascaded micro-hole multiplierscomprising GEMs and other patterned electrodes. We dis-cuss the ion-induced secondary electron emission and pro-vide solutions that permit, for the first time, the operationof visible-sensitive GPMs in DC mode, with gain of 10 . Preprint submitted to Elsevier 20 November 2018 . Requirements for stable operation ofvisible-sensitive GPMs
General consideration on Ion back-flow and Ionfeedback effects.
While cascaded micro-hole multipliers, with their signif-icant optical ”opacity”, efficiently block avalanche-photonfeedback [7], they are less efficient in blocking the back-flow of avalanche ions. The latter, originating from eachavalanche stage in the cascaded multiplier, drift back tothe GPMs’ PC following the device field lines, and a ma-jor fraction of them follow the same paths (in oppositedirection) of the initial photoelectrons and their succes-sive avalanche electrons [6]. When impinging on the PC’ssurface they release secondary electrons; the latter initi-ate secondary avalanches, known as ion-feedback, which,by positive feedback mechanism diverge the proportionalavalanche multiplication into discharge [8]. An example ofion-feedback effect measured in a double-GEM multiplierwith K-Cs-Sb visible-sensitive PC operating in Ar/CH (95/5) at 700 Torr (Fig. 1a), is the deviation of the gain-voltage curve from exponential (Fig. 1b). The measuredgain, G meas , contains contributions from ion feedback andis described by:G meas = G1 − γ + · IBF · ε extr · G (1)where γ + is the ion induced secondary emissionprobability or the ion feedback probability, IBF is theion back-flow fraction namely the fraction of ions, fromall avalanche stages of the multiplier, flowing back to thePC (or to the drift region of a tracking detector), ε extr isthe efficiency of extracting secondary electrons from thePC into the gas and G is the multiplier’s gain withoution feedback. To avoid avalanche divergence into a spark,the above formula should fulfill γ + · IBF · ε extr · G < , required for good single-photonsensitivity in GPMs, implies γ + · IBF · ε extr < − .2.2. Measurement of γ + and ε extr The extraction efficiency ε extr of secondary electrons isthe fraction of electrons emitted from the PC and not scat-tered back (by collisions with gas molecules) into it [9,10]. ε extr depends on the kinetic energy distribution of the elec-trons leaving the PC, which has not yet been measured. Thetheoretical calculations of energy distribution and extrac-tion efficiency ε extr of ion-induced secondary electrons fromPCs are rather complex and differ from those of photon-induced ones [11]. Such calculations are presently underway, in cooperation with T. Dias of Coimbra University,and will be the subject of a future publication.In the absence of any knowledge of ε extr , we have cho-sen to use γ eff + = γ + · ε extr ; the latter can be extractedfrom the experimental GPM’s gain curve and its deviation (a)
200 220 240 260 280 300 320 34010
700 Torr Ar/CH (95/5)E drift =0.5kV/cm T o t a l ga i n V GEM [V]
K-Cs-Sb QE=22%@375nm (b)
Fig. 1. (a) A double-GEM GPM coupled to a semitranspar-ent photocathode; (b) gain-voltage characteristics measuredin this GPM (see conditions in the figure, QE refers to vac-uum) with CsI (dashed) and K-Cs-Sb (open circles) photo-cathodes. The divergence from exponential with K-Cs-Sb isdue to ion feedback. from exponential line. γ eff + is defined as the effective ioninduced secondary emission probability or the effective ionfeedback probability. It was extracted from fitting the ex-perimental gain curve G meas by equation (1) (solid line inFig. 1b). A significant deviation from the exponential gain-voltage characteristic (dashed line in Fig. 1b) is observedwith the bi-alkali PC already at low gain. The IBF and G asa function of GEM voltage were measured in the same de-tector (geometry, gas and voltages), with a CsI PC, ensur-ing no ion feedback. The drift field between the PC and thetop face of the first GEM was kept constant at 0.5kV/cmthroughout the entire measurements. In GPMs it providesabout 60% extraction efficiency of photoelectrons in thisgas [8]. The gain-voltage characteristics, like the one shownin Fig. 1b, were measured for K-Cs-Sb, Na-K-Sb and Cs-SbPCs; they yielded γ eff + values of ∼ · − for all these pho-tocathodes. This study and the results will be discussed inmore detail elsewhere.2.3. Requirements for IBF
Establishing the effective ion feedback value γ eff + withthe visible sensitive PCs under investigation, we can set the2imits on the IBF value needed for stable DC operation ofvisible-sensitive GPMs at a gain of 10 . Requiring γ eff + · IBF < − , and using the estimated value γ eff + = ∼ · − , the IBF value should be < . · − .
3. IBF reduction in cascaded micro-hole multiplierstructures
A straight forward way to reduce the IBF is by loweringthe drift field, since IBF decreases linearly with the driftfield [12]. However, in GPMs the drift field could not be toolow because it controls, the photoelectron extraction intothe gas, ; drift field values of the order of 0.5kV/cm [8] weregenerally applied in our GPMs filled with Ar/CH (95/5).A more detailed report on a comprehensive IBF reduc-tion, study, carried out with a variety of cascaded micro-hole multipliers, can be found elsewhere [6]. All GPM detec-tors investigated had active areas of 30x30mm ; they wereirradiated over a surface of 15mm in diameter, at photonfluxes of about 10 photonssec · cm . The core outcome of this studyis that very low IBF values can be obtained with cascadescombining GEMs and other patterned electrodes derivedfrom the Micro-Hole & Strip Plates (MHSP) [13]. The lattercomprised GEM-like holes and additional patterned stripson one of their faces, and they were operated in differentmodes regarding the voltages and orientation schemes [6]:MHSP, reversed-MHSP (R-MHSP) and flipped-reversed-MHSP (F-R-MHSP). These schemes aim at reducing theIBF by diverting the ions and trapping them on the stripspatterned on the surface. While the MHSP, placed at theend of the cascade, can divert and trap only part of theions generated within its own avalanche stage, the othertwo types of electrodes can divert ions created in succes-sive multiplying elements; therefore their incorporation inthe cascade yielded better results. In a cascade comprisingF-R-MHSP followed by GEM and MHSP (Fig. 2a), IBFvalues as low as 3 · − were recorded [6]; this fulfills ourrequirement for stable DC operation at a gain of 10 withvisible-sensitive PCs (see Fig. 2b). We varied the inter-stripvoltage on the bottom MHSP, to vary the total gain of thedetector.Following the success of the above study, and with ourunderstanding of the operation mechanism of the MHSP-like electrodes [6], a new patterned micro-hole electrodenamed Cobra (Fig. 3) was developed with a geometry thatis expected to improve the ion divergence away from theholes. It has thin anode electrodes surrounding the holesand creating strong electric field inside the holes (requiredfor charge amplification); the more negatively biased cath-ode electrodes cover a large fraction of the area for bet-ter ion-collection as compared to the F-R-MHSP [6]. Theconcept of the Cobra electrode has been recently investi-gated. It was found that when introduced as a first element(with the patterned area pointing towards the photocath-ode), preceding two GEMs in the cascade (Fig. 4a), it dras-tically improved the ion trapping capability. The IBF as a (a) -5 -4 -3 -2 E drift =0.5kV/cmF-R-MHSP/GEM/MHSPE drift =0.2kV/cm Ar/CH (95/5), 760 Torr I B F Total gain (b)
Fig. 2. (a) Scheme of the cascadedF-R-MHSP/GEM/MHSP multiplier coupled to a semi–transparent photocathode; possible avalanche-ion paths areshown. (b) The IBF in correlation with the total gain of thisGPM plotted for drift fields of 0.2 kV/cm (TPC conditions,squares) and 0.5 kV/cm (GPM conditions, circles).Fig. 3. A microscope photographs of one face of a ”Cobra”micro-hole electrode with dimensions given in the figure.The other face is identical to a GEM. function of voltage between electrodes on the top surfaceof Cobra is shown in Fig. 4b. In GPM conditions, with adrift field of 0.5kV/cm, we measured IBF values of 3 · − which is 10,000 times lower than that of cascaded tripleGEMs. In TPC conditions with a drift field of 0.2kV/cmthe same detector configuration provided IBF values as lowas 2 . · − . These are the lowest IBF values ever reached ingaseous detectors. However, while the F-R-MHSP yieldedfull photoelectron collection efficiency into the holes of the3 a) -6 -5 -4 -3 -2 -1 E drift =0.2kV/cm GPMTPC700 Torr Ar/CH (95/5)Flipped-Cobra/2GEME drift =0.5kV/cm I B F Total Gain (b)
Fig. 4. (a) Scheme of cascaded Cobra/2GEM GPM witha semi-transparent photocathode; possible avalanche ionspaths are also shown. (b) The IBF as a function of the totalgain of the Cobra/2GEM cascaded detector for drift fieldsof 0.2 kV/cm (TPC conditions, triangles) and 0.5 kV/cm(GPM conditions, squares). first cascade element, the Cobra, in its present geometry,had a limited electron collection efficiency of about 20%.This can and should be improved by optimizing the geo-metrical parameters.
4. DC operation of a visible-sensitive GPM withmicro-hole multiplier cascades
The operation of a visible-sensitive GPM in DC modewas investigated with a F-R-MHSP/GEM/MHSP cas-caded multiplier, schematically shown in Fig. 2a. A pho-tograph of the experimental detector is shown in Fig. 5.All the multiplier electrodes were mounted between ce-ramic spacers within a UHV vessel [8]. The photocathodewas prepared and characterized in an adjacent vessel ofthe dedicated UHV system, and then transported with anUHV manipulator and placed in a stainless-steel holderabove the detector. Details can be found in [14] and in [8].The K-Cs-Sb PCs had typical quantum efficiency (QE) of30% measured in vacuum at 375 nm. The details of the IBF measurements and results for thismultiplier configuration were reported in [6] both in con-ditions for TPC and for GPM operation (Fig. 2b). Condi-tions for full efficiency of electron collection from the PCwere confirmed and applied in all measurements
Fig. 5. Photograph of the F-R-MHSP/GEM/MHSP detec-tor and the photocathode, mounted in the vacuum chamber.
In Fig. 6 we present gain-voltage characteristics for thecascaded GPM of Fig. 2a with a K-Cs-Sb PC and with a CsIPC, for comparison. The measurements were carried out inAr/CH (95/5) at 700 Torr. The present semitransparentK-Cs-Sb PC had a QE of ∼
27% measured in vacuum at375nm. Its QE value in the gas, with a drift-field of 0.5kV/cm, was estimated to be 16% [8]. The solid and dashedcurves in Fig. 6 represent exponential fits to the data pointsmeasured with K-Cs-Sb and CsI PCs, correspondingly. Inboth cases the GPM could reach a gain of 10 with nodivergence from an exponential gain-voltage characteristic,indicating upon full suppression of ion feed-back effects.
200 220 240 260 28010 K-Cs-Sb (QE~27%@375nm) CsI Exponential fit of Exponential fit of F-R-MHSP/GEM/MHSPE drift =0.5kV/cm700 Torr Ar/CH4 (95/5)UV-LED 375nm T o t a l ga i n V AC2 [V]
Fig. 6. Gain-voltage characteristics of the detector shownin Fig. 2a with a K-Cs-Sb (squares) and CsI (circles) pho-tocathodes. The data was fitted with exponential functions;no divergence from exponential was observed. 700 TorrAr/CH (95/5); E drift =0.5 kV/cm. QE refers to vacuum.
00 250 300 35010 K-Cs-Sb (QE~40%@375nm) CsI Exponential fit of Exponential fit of Flipped-Cobra/2GEME drift =0.5kV/cm700 Torr Ar/CH4 (95/5)UV-LED 375nm T o t a l G a i n V GEM [V]
Fig. 7. Gain-voltage characteristics of the Cobra/2GEMcascaded GPM of Fig. 4a, with K-Cs-Sb (squares) and CsI(circles) photocathodes. The data were fitted with exponen-tial functions. 700 Torr Ar/CH (95/5); E drift =0.5 kV/cm.QE refers to vacuum. The visible-sensitive GPM with a K-Cs-Sb PC coupledto the Cobra multiplier followed by two GEMs (Fig. 4a)was investigated in DC operation mode; the gain-voltageplots are shown in Fig. 7, in comparison with a CsI PC.The measurements were carried out in Ar/CH (95/5) at700 Torr. The semitransparent K-Cs-Sb PC had ∼
40% QEmeasured in vacuum at 375nm. The exponential fits to thedata points measured with K-Cs-Sb and CsI PC are repre-sented by solid and dashed curves, correspondingly. Therewere no feed-back effects as can be seen from the exponen-tial shape of the gain-voltage curve.
5. Conclusions
Ion feedback in cascaded micro-hole gaseous detectorswas studied with a variety of cascade elements and PCs, inconditions of TPC and of GPM operation. The effective ionfeedback probability γ eff + was measured in Ar/CH (95/5)at 700 Torr and found to be 3 · − for Na-K-Sb, K-Cs-Sband Cs-Sb photocathodes. Based on these measurements,the ion backflow fraction (IBF) required for stable DC op-eration of cascaded visible-sensitive gaseous photomultipli-ers (GPM) was estimated to be 3 . · − .Systematic ion blocking investigations with various pat-terned micro-hole cascaded multipliers yielded the requiredIBF values, at gain of 10 . The best results were recordedin a cascaded multiplier of a flipped reversed-bias micro-hole and strip plate flowed by a GEM and by a micro-holeand strip plate (F-R-MHSP/GEM/MHSP). This configu-ration yielded 100 fold lower IBF value than that measuredin cascaded GEMs. This permits reaching stable operationconditions both for TPCs and for visible-sensitive GPMsoperating in DC mode.Even lower IBF values, of 3 · − at a gain of 10 anddrift field 0.5kV/cm, was recorded in a cascade comprisinga novel ”Cobra” micro-hole patterned multiplier, followedby two GEMs. This IBF value is 10,000 times lower thanthat measured in cascaded GEMs. However, the electron collection efficiency of the present ”Cobra” multiplier wasonly 20%, which requires further optimization of its geom-etry.A Visible-sensitive GPM with a F-R-MHSP/GEM/MHSPcascaded multiplier and a K-Cs-Sb photocathode, yielded,for the first time, stable operation at gains of 10 in DCmode with full photoelectron collection efficiency and with-out any noticeable feedback effects. This is a breakthroughin the field of gaseous photomultipliersAcknowledgments This work is partly supported by theIsrael Science Foundation, grant No 402/05, by the MIN-ERVA Foundation and by Project POCTI/FP/63962/2005through FEDER and FCT (Lisbon). A. Breskin is theW.P. Reuther Professor of Research in The Peaceful Useof Atomic Energy.References [1] A. Breskin, et al., Ion-induced effects in GEM and GEM/MHSPgaseous photomultipliers for the UV and the visible spectralrange, Nucl. Instr. and Meth. A 553 (2005) 46 and referencestherein.[2] B. K. Singh, et al., CsBr and CsI UV photocathodes: new resultson quantum efficiency and aging, Nucl. Instr. and Meth. A 454(2000) 364.[3] B. Ketzer, Micropattern gaseous detectors in the COMPASStracker, Nucl. Instr. and Meth. A 494 (2002) 142.[4] F. Piuz, et al., The CsI-based ring imaging detector for theALICE experiment: technical description of a large prototype,Nucl. Instr. and Meth. A 433 (1999) 222.[5] Z. Fraenkel, et al., A hadron blind detector for the PHENIXexperiment at RHIC, Nucl. Instr. and Meth. A 546 (2005) 466.[6] A. Lyashenko, et al., Further progress in ion back-flow reductionwith patterned gaseous hole-multipliers, JINST 2 (2007) P08004and references therein.[7] D. M¨ o rmann, et al., Operation principles and propertiesof the multi-GEM gaseous photomultiplier with reflectivephotocathode, Nucl. Instr. and Meth. A 530 (2004) 258.[8] D. M¨ o rmann, Study of novel gaseous photomultipliers for UVand visible light, PhD thesis, Weizmann Institute of Science,http://jinst.sissa.it/jinst/theses/2005 JINST TH 004.jsp.[9] A. Buzulutskov, et al., The GEM photomultiplier operated withnoble gas mixtures, Nucl. Instr. and Meth. A 443 (2000) 164.[10] J. Escada, et al., A Monte Carlo study of backscattering effectsin the photoelectron emission from CsI into CH and Ar-CH mixtures, JINST 2 (2007) P08001.[11] H. Hagstrum, Theory of Auger neutralization of ions at thesurface of a diamond-type semiconductor, Phys. Rev. 122 (1961)83.[12] A. Bondar, et al., Study of ion feedback in multi-GEMstructures, Nucl. Instr. and Meth. A 496 (2003) 325.[13] J. Veloso, et al., A proposed new microstructure for gas radiationdetectors: The microhole and strip plate, Rev. Sci. Inst. A 71(2000) 2371.[14] M. Balcerzyk, et al., Methods of preparation and performance ofsealed gas photomultipliers for visible light, IEEE Trans. Nucl.Sci. 50 (2003) 847.mixtures, JINST 2 (2007) P08001.[11] H. Hagstrum, Theory of Auger neutralization of ions at thesurface of a diamond-type semiconductor, Phys. Rev. 122 (1961)83.[12] A. Bondar, et al., Study of ion feedback in multi-GEMstructures, Nucl. Instr. and Meth. A 496 (2003) 325.[13] J. Veloso, et al., A proposed new microstructure for gas radiationdetectors: The microhole and strip plate, Rev. Sci. Inst. A 71(2000) 2371.[14] M. Balcerzyk, et al., Methods of preparation and performance ofsealed gas photomultipliers for visible light, IEEE Trans. Nucl.Sci. 50 (2003) 847.