Development of high-rate MRPCs for high resolution time-of-flight systems

De elopmen o high- a e MRPCs o high esolu ion ime-o -fligh sys ems Jingbo Wang a , Yi Wang a, n , D. Gonzalez-Diaz a,b,c , Huangshan Chen a , Xingming Fan a , Yuanjing Li a , Jianping Cheng a , Ma cus Kaspa d , Roland Ko e d , Alejand o Laso Ga cia d , Lo ha Naumann d , Daniel S ach d , Ch is ian Wendisch d ,J ¨ o n W¨ us en eld d a Depa men o Enginee ing Physics, Tsinghua Uni e si y, Key Labo a o y o Pa icle & Radia ion Imaging, Minis y o Educa ion, Beijing 100084, China b GSI Helmhol z Cen e o Hea y Ion Resea ch, Da ms ad , Ge many c Labo a o io de Fı ´sica, Nuclea y Al as Ene gı ´as, Uni e sidad de Za agoza, Za agoza, Spain d Fo schungszen um D esden-Rossendo , D esden, Ge many a icle in o A icle his o y: Recei ed 16 July 2012 Recei ed in e ised o m 20 Janua y 2013 Accep ed 20 Feb ua y 2013 A ailable online 6 Ma ch 2013 Keywo ds: MRPC TOF CBM High a e Low- esis i e glass Ra e capabili y abs ac We show how he high cha ged-pa icle flux (1–20 kHz/cm 2 ) expec ed o e he 150 m 2 la ge ime-o - fligh wall o he u u e Comp essed Ba yonic Ma e expe imen (CBM) a FAIR can be ealis ically handled wi h Mul i-gap Resis i e Pla e Chambe s (MRPCs). This c ucial 100- old inc ease o he chambe a e capabili y, as compa ed o ha o s anda d MRPCs p esen ly employed in expe imen s eso ing o sub-100 ps iming, has been achie ed hanks o he de elopmen o a new ype o low- esis i e doped glass. Following he encou aging esul s p e iously ob ained wi h small coun e s, wo ypes o modules (ac i e a ea: 150 cm 2 ) ha e been buil a Tsinghua Uni e si y wi h he new ma e ial. The measu emen s con eyed in his wo k, ob ained wi h a quasi- minimum ionizing elec on beam (g b Z3), p o e hei sui abili y as he building blocks o he p esen had on-iden ifica ion concep o he CBM expe imen . Namely, hey p o ide a ime esolu ion be e han 80 ps and an e ficiency abo e 90% a a pa icle flux well in excess o 20 kHz/cm 2 (up o 35–60 kHz/cm 2 ), being a he co e o a modula concep ha is easily scalable. Recen measu emen s o he elec ical and mechanical p ope ies o his new ma e ial, oge he wi h i s long- e m beha io , a e sho ly summa ized. &2013 Else ie B.V. All igh s ese ed. 1. In oduc ion Mul i-gap Resis i e Pla e Chambe s (MRPCs), fi s de eloped in 1996 [1], a e spa k-p o ec ed gaseous de ec o s wi h good ime esolu ion, high e ficiency and ela i ely low cos . Du ing he las decade, hanks o he p og ess made in iming MRPCs [2–4], la ge-a ea ime-o -fligh sys ems ha e become widesp ead in mode n nuclea and pa icle physics expe imen s, such as ALICE [5–7], STAR [8–10], FOPI [11,12], HADES [13,14], e c. Recen R&D e o s e ealed ha in insic de ec o esolu ions down o 20 ps a e possible unde ce ain condi ions [15]. P esen ly, he main d awback o hese iming MRPCs is hei limi ed capabili y o handle a (minimum ionizing) cha ged pa icle flux exceeding some hund eds o Hz/cm 2 , a ac ha s ems om he usage o con en ional soda-lime floa glass pla es wi h a bulk esis i i y in he ange 10 12 –10 13 O cm. Hence, he insula ion echnique ha endows he coun e wi h i s unique spa k-quenching cha ac e - is ics a e y high uni o m fields is ul ima ely esponsible o he cha ge build-up p ocess ha e ec i ely limi s he ope a ing field a high pa icle flux. The nex gene a ion Comp essed Ba yonic Ma e expe imen (CBM) a he u u e Facili y o An ip o on and Ion Resea ch (FAIR) in Da ms ad , Ge many, is planned o be buil wi h a ime-o -fligh sys em based on MRPCs. The aim is o ob ain good had on iden ifica ion in fixed- a ge hea y ion collisions a p ojec ile ene gies up o E¼25 GeV/A ( o de ails see Re . [17] and e e ences he ein). The challenge is o keep a high e ficiency (abo e 90%) and a good ime o fligh esolu ion (be e han 80 ps) a a pa icle flux up o 20 kHz/cm 2 , which is no accessible o con en ional floa -glass MRPCs ([17], o ins ance). Thus, he ask o finding ma e ials capable o eplacing floa glass o iming applica ions bu p o iding a much less esis i e e u n pa h o he eleased cha ge has been he subjec o in ense R&D e o s in he ecen yea s: no ably, as shown in Re . [16–25], he de elopmen o a ious ypes o semi-conduc i e glass, ce amics as well as he so called wa m glass echnology. The usage o semi-conduc i e glass, in pa icula , da es back o he Pes o coun e ( o an o e iew see Re . [26] and e e ences he ein). Recen ly, a Tsinghua Uni e si y, we ha e managed o de elop a s able p oduc ion line o a p omising new ype o doped glass. Following he encou aging esul s ob ained wi h small-a ea MRPC p o o- ypes based on his ma e ial [24], wo MRPC modules we e Con en s lis s a ailable a SciVe se ScienceDi ec jou nal homepage: www.else ie .com/loca e/nima Nuclea Ins umen s and Me hods in Physics Resea ch A 0168-9002/$ - see on ma e &2013 Else ie B.V. All igh s ese ed. h p://dx.doi.o g/10.1016/j.nima.2013.02.036 n Co esponding au ho . Tel.: þ86 1062771960; ax: þ86 1062782658. E-mail add esses: [email p o ec ed] (J. Wang), [email p o ec ed] (Y. Wang). Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–51 p oduced in o de o sui he cu en design o he CBM–TOF wall. The coun e s ha e been ho oughly cha ac e ized wi h a 30 MeV elec on beam a he Elec on Linac wi h high B illiance and low Emi ance acili y (ELBE) [25,27] a Helmhol z-Zen um D esden- Rossendo (HZDR). The p esen epo is o ganized as ollows: Sec ion 2 desc ibes he mechanical p ope ies o he new low- esis i e glass oge he wi h a sho explana ion o i s elec ical p ope ies; Sec ion 3 p esen s he wo coun e s de eloped o he CBM–TOF wall; Sec ion 4 desc ibes he ELBE acili y and he expe imen al se up; Sec ion 5 p esen s he coun e s pe o mance as a unc ion o high ol age and pa icle flux; and Sec ion 6 is de o ed o a comp e- hensi e compa ison be ween p esen esul s and esul s p e- iously ob ained wi h a had on beam uni o mly i adia ing he chambe . Finally, Sec ion 7 summa izes ou conclusions and ou look. 2. Low- esis i e glass This new glass, de eloped o high cha ged pa icle flux applica- ions, is cha ac e ized by an ohmic beha io and s abili y wi h anspo ed cha ge, p ope ies ha ely on a balanced admix u e o oxides o ansi ion elemen s [24]. I has a black colo and i is opaque o isible ligh , as commonly a ibu ed o glass exhibi ing a o m o elec on conduc i i y. I s final conduc i i y u ns ou o be e y sensi i e o bo h he ini ial chemical composi ion o he aw ma e ial and o he glass mel ing p ocedu e. Di e en composi ions and ela ed p oduc ion p ocedu es ha e been s udied, yielding a unable bulk esis i i y in he ange o 10 10 –10 11 O cm. Fig. 1 shows he measu ed bulk esis i i y as a unc ion o he posi ion along a 30 cm-long glass pla e, o a andomly selec ed sample. The mea- su emen s we e done in a cons an - empe a u e d y box whe e he empe a u e and he humidi y could be con olled. A oom em- pe a u e (25 1C) and a ela i e humidi y o 30%, he esis i i y is seen o change wi hin a mild 30% a ia ion a ound an a e age alue o ¼1.5 10 10 O cm. De ia ions wi h espec o he esul s gi en in Re . [24] s ay wi hin less han a ac o o wo and can be a ibu ed mainly o sample- o-sample a ia ions. The ma e ial exhibi s a highly ohmic beha io up o 1 kV and ollows A henius law as a unc ion o empe a u e, ¼ T 0 10 ðT 0 TÞ= D T [24].S ikingly, i shows one (po en ially usable) o de o magni ude esis ance dec ease e e y 281(i.e., D T¼28 1C), e y close o he beha io p e iously epo ed o floa glass [18,19]. I s dielec ic cons an in he GHz- ange, measu ed bo h wi h capaci i e and ansmission line echniques, shows a alue o e (1 GHz)¼7.5–9.5, jus mode a ely highe han he one measu ed o floa glass ( e (1 GHz)¼670.5). I s loss- angen an d (1 GHz)¼0.035isalsocompa able o heone measu ed o floa glass an d (1 GHz)¼0.02570.05, [28]. A guing along he lines o Re . [28], nei he we expec ed no we obse ed any limi a ion o CBM coun e s coming om hese modes de ia- ions wi h espec o he alues o floa glass, ei he om he poin o iew o signal induc ion, ansmission o c oss- alk. Mo e impo an ly, his new low- esis i e glass shows a la ge s abili y agains elec ical s ess: unde s a ic condi ions i s b eakdown field is abo e 1 kV/mm (a com o able lowe bound) and i s esis i i y inc eases wi hin an accep able ac o o 2 o a densi y o anspo ed cha ge as la ge as 1 C/cm 2 (o e 34 days), a alue close o he expec ed CBM li e- ime [17]. We ha e ecen ly s a ed ageing es s unde dynamic condi ions, i.e., di ec ly on a (ex e nally i adia ed) RPC, and obse ed no dis- ce nible ageing e ec s up o 0.05 C/cm 2 (300 h). This figu e, s ill much lowe han he an icipa ed CBM li e- ime, does al eady exceed he ope a ing li e- imes o so a exis ing iming MRPC walls. De ails ha e been published elsewhe e [29]. As compa ed o he well-es ablished floa glass echnology, 1 howe e , low- esis i e glass p oduc ion is in a less ma u e s age o de elopmen . Simila o he o me , he echnique o mel ing and annealing is e y impo an in o de o p oduce glass wi h high quali y. Bu , con a y o floa glass, he equi ed su ace quali y o he low- esis i e glass can be only achie ed a e a pos -p ocessing polishing p ocess. The oughness o he elec ode in oduces local field a ia ions which may de e io a e he e ficiency and ime esolu ion, [21,30]. I is belie ed ha su ace p ope ies also de e - mine he da k cu en and pho on eedback le els [31]. Since a high su ace quali y is equi ed by a high quali y glass, su ace measu e- men smigh beakeypa o ap o ocoldesigned o hequali y con ol o he glass mass p oduc ion in o de o consis en ly p oduce eliable glass. He e we epo on he su ace oughness o he glass pla es, as de e mined by a Mic oXAM 3D su ace p ofile ha can measu e he su ace p ofile and he oughness o highly polished op ical elemen s and ough su aces, such as s eel, pape , plas ics, and ce amics. A 3-D scan image is shown in Fig. 2. The su ace oughness can be exp essed by a lis o numbe s ollowing di e en defini ions [32]. Th ee o he mos commonly used defini ions a e shown unde an e alua ion scale o abou 857 m m638 m m: a e age oughness (1.06 nm), peak–peak oughness (10.6 nm) and en-poin -heigh oughness (9.42 nm). Fo e e ence, he numbe s o a floa glass sample analyzed h ough he same p ocedu e a e 0.608/8.17/6.22. 3. De elopmen o eal-size modules In he cu en CBM concep ual design, he whole TOF wall is a anged in ou ‘ a e egions’ [17]. In he inne egion o he wall ( egion 1), pad eadou MRPCs based on low- esis i e glass can be e ficien ly used o cope wi h he high pa icle flux (up o 20 kHz/ cm 2 ) a an op imal segmen a ion (4–6 cm 2 /pad), while s ip- eadou MRPCs ep esen a na u al choice o he ou e egion (0.5–8 kHz/cm 2 ) whe e he sys em occupancy and flux a e o de s o lowe magni ude. Recen ly, ou e o is o ealis ically adap he MRPC geome y o sui he concep ual design o he TOF wall, as o Re . [17]. Fig. 1. Bulk esis i i y o he low- esis i e doped glass de eloped a Tsinghua Uni e si y as a unc ion o he posi ion along a andomly selec ed sample ( he dimensions o he pla e a e 20 cm 30 cm 0.11 cm). Tempe a u e was a ound 25 1C. 1 Less o en named soda-lime o window glass echnology. J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–51 41 Fo egion 1, he simula ed pa icle flux anges om 8 kHz/cm 2 o 25 kHz/cm 2 and he pa icle densi y, ob ained o he mos s ingen cen al Au–Au collisions a 25 GeV/A, eaches a max- imum alue o abou 0.01/cm 2 . Hence, a small-pad MRPC (PMRPC), ha could cons i u e he co e o a u u e supe -module, has been de eloped (Fig. 3-up). The coun e has 10 gas gaps and consis s o wo (mi o ed) s acks o pla es made o 0.7 mm- hick low- esis i e glass wi h a esis i i y o abou 2 10 10 O cm. I has 6 2pads o size 2 cm 2 cm each, wi h an in e al be ween pads o 2 mm (occupancy o5%). The gas gap is 0.22 mm, defined by nylon monofilamen s. The high ol age elec odes a e co e ed wi h colloidal g aphi e sp ay, yielding a ypical su ace esis i i y o abou 2 M O /sq. In iew o hei much lowe occupancy a e egions 2, 3 and 4 o he TOF wall a e cu en ly o eseen o be assembled wi h mul i- s ip coun e s. The s uc u e o he s ip- eadou MRPC p oposed he e is depic ed in Fig. 4. As o he pad coun e , i can be conside ed as a supe -imposed s uc u e o wo 5-gap MRPCs mi o ed wi h espec o one o he elec odes, ha allows o easily p o ide he same (nega i e he e) HV o ‘bo h’ chambe s. The posi i e HV is applied o he ou e elec odes. As compa ed o he pad-design, he s ip-coun e has sligh ly la ge gaps o 0.25 mm aimed a pa ially compensa ing o he lowe signal le els cha - ac e ized by he s ip designs. The eadou s ips ha e 24 cm leng h and 2.2 cm wid h. The in e als be ween he s ips a e 3 mm, in which g ounded gua d s ips connec ed wi h ias ha e been pa e ned, o c oss alk minimiza ion. A possible layou based on his ype o modules was desc ibed in some de ail in Re . [33]. In bo h modules, simila ly o he 10-gap coun e in Re . [24], signals induced wi h bo h pola i ies a e sen in di e en ial ashion o he on end elec onics (FEE) used o he STAR TOF p o o ype ‘TOF ’ [34] in RHIC Run 3. This FEE is desc ibed in Re . [35]. The wo king gas is a mix u e o 85% C 2 H 2 F 4 , 5% iso-C 4 H 10 and 10% SF 6 a nea ly a mosphe ic p essu e. 4. Beam es a ELBE 4.1. The ELBE elec on beam The ELBE acili y a Helmhol z-Zen um D esden-Rossendo , D esden, Ge many, consis s o a 250 keV he mionic elec on gun p o iding a pulsed elec on beam. I is ollowed by a wo-s age supe conduc ing linea accele a o (Linac) consis ing o ou ca i ies, each allowing o an ene gy gain o 10 MeV. The accel- e a o p oduces a quasi-con inuous elec on beam wi h a unable mic o-pulse epe i ion a e o 26 MHz/2 Nmp (N mp ¼1, 2, y,8) and Fig. 2. 3-D image o he su ace o he low- esis i e glass: Sa (a e age oughness), Sq ( oo -mean-squa e oughness), Sy (peak-peak oughness), Sz ( en-poin -heigh oughness), Sku (su ace-ku osis oughness) and Ssk (su ace-skewness oughness) a e used o e alua e he su ace oughness. Fig. 3. Pad- eadou MRPC, ea u ing a 10-gap 6 2-pad s uc u e. J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–5142 a co esponding bunch du a ion o less han 5 ps. A desc ip ion o he expe imen al acili ies a ELBE is gi en in Re . [27]. While in ea ly measu emen s wi h iming RPCs a ELBE a sca e ing se up was used [36], ecen ly, a single-elec on mode has been de eloped by a enua ing he beam cu en down o a oampe es, hus allowing de ec o es ing di ec ly in-beam. This mode has been used o demons a e he a e capabili y o ce amic-based MRPCs [37]. The measu emen s p esen ed in his pape we e ob ained wi h an elec on beam wi h a mic o-pulse epe i ion a e o 6.5 MHz and a kine ic ene gy o 30 MeV. A hese ene gies, elec ons ha e an ene gy loss simila o ha o minimum ionizing pa icles [38]. A e y sho bunch du a ion makes i possible o use he adio equency (RF) signal o ELBE as he ime e e ence o TOF measu emen s, hus no s a coun e is equi ed. 4.2. Expe imen al se up The expe imen al se up was al eady used in p e ious s udies [22,37] and a simplified scheme is shown in Fig. 5. All de ec o s we e aligned wi h he help o a s a iona y lase sys em. The MRPC was embedded in a scin illa o chain aimed a selec ing s aigh elec on acks. The MRPC and scin illa o S 5 we e bo h placed on a mo ing pla o m, which allows 50 cm mo emen bo h in he e ical and ho izon al di ec ions ia emo e con ol. Thus, a wo-dimensional posi ion scan was easible du ing i adia ion. A igge condi ion was equi ed in he o m o S 1 ^S 2 ^S 3 ^S 4 ^RF (3 o he used scin illa o s ha e a double-end eadou deno ed by e en and odd numbe s, co espondingly, see Fig. 5). The ime-o -fligh is defined h ough he ex as he ime di e ence be ween he MRPC and RF signals. The RF and RPC disc imina ed signals a e ed in o a ime- o-digi al con e e (TDC, V1290N) wi h a leas -significan bi (LSB) o 24.5 ps, which yielded an o e all esolu ion o 35 ps/channel (measu ed om he ime di e ence be ween a signal spli in o wo di e en chan- nels). Due o he sho 5 ps-bunch we assume he ea e ha he esolu ion o he e e ence ime is domina ed by he TDC esolu- ion alone, ha will be sub ac ed in quad a u e; hus, s ic ly speaking, he RPC ime esolu ions gi en in his wo k ep esen uppe bounds p o ided we ha e no di ec handle on he in insic RF signal ji e . Since op o-couple s we e used o ansmi he signal om he RF ca i ies up o a ansduce ha was placed a nea ly 2 m dis ance om he TDC, we do no expec sizeable e ec s ha migh a ise om loss o signal quali y o he ( e e ence) RF signal du ing ansmission. Following calib a ion in condi ions o d -impulse exci a ion, i was es ima ed ha he signal cha ge was p e-amplified by a ac o o 80 in he on -end elec onics and by a u he ac o o 10 in an auxilia y amplifie module. The amplified cha ge signal was hen ed in o a 25 C/channel cha ge- o-digi al con e e (QDC, CAEN V965) and in eg a ed o e a 50 ns ga e. The da a acquisi ion was ope a ed on a GSI mul i-b anch sys em (MBS) [39]. 4.3. De e mina ion o he pa icle flux and beam p ofile The beam-p ofile was de e mined du ing he expe imen by eplacing he MRPC unde s udy by a 5 mm-diame e finge - scin illa o S 5 ha scanned he elec on beam e ically and ho izon ally in sel - igge mode. A high Gaussian beam o wid h s scin ¼1.1 cm (ho ) 1.5 cm ( e ) was ound unde hese condi- ions (shown as a con inuous line in Fig. 6(a) and (b)) and he single-elec on peak could be clea ly seen on he scin illa o cha ge dis ibu ions. This implies ha he expe imen al si ua ion co esponds o a single-elec on pe bunch ope a ing mode, as desi ed. In iew o he na ow beam and he expec ed la ge mul iple sca e ing o low-ene gy elec ons, we pe o med an independen de e mina ion o he beam-p ofile by using he RPC ime- di e ence om he wo ends o he s ip coun e . Fo ha , he signal p opaga ion eloci y was p e iously ob ained by fi ing he cen oid dis ibu ion a di e en RPC posi ions, yielding a signal p opaga ion eloci y o ¼185 m m/ps. The esul s a e shown in Fig. 6(a), showing a b oade dis ibu ion wi h s x,RPC ¼1.8 cm (ho ) unde RPC sel - igge ed condi ions, while a na owe wid h o 1.4 cm (ho ) was de e mined unde s anda d igge condi ions S 1 ^S 2 ^S 3 ^S 4 ^RF, possibly due o he confining e ec o he scin illa o chain. These wo obse a ions gi e suppo o he ac ha he obse ed s x,RPC ¼1.8 cm wid h has li le con ibu ion coming om he RPC ime esolu ion ( ha was p e iously measu ed o simila coun e sþFEE o be less han s x,RPC ¼0.8 cm). By assuming ha he b oade dis ibu ion obse ed o e he RPC comes om mul iple sca e ing, and ha he b oadening e ec is common o he e ical and ho izon al di ec ions, a ough es ima e o he beam-wid h a Fig. 5. Schema ic iew o he RPC es se up. The beam igge was defined by he coincidence S 1 ^S 2 ^S 3 ^S 4 ^RF. Toge he wi h he MRPC, a small scin illa ion coun e S 5 (finge ) was moun ed a a s ep-mo o d i e o measu e he beam size. Coun e S 9 /S 10 is a 4 cm- hick scin illa o used o moni o ing he single-elec on dis ibu- ion o he bunches. Fig. 4. S ip- eadou MRPC, ea u ing a 10-gap 3-s ip s uc u e. J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–51 43 he RPC posi ion can be made as s RPC ¼1.8 cm (ho ) 2.0 cm ( e ) (o e -es ima ed by less han 10% in each dimension i he fini e RPC esolu ion is aken in o conside a ion). The implica ions o he di e en beam sizes cha ac e ized h ough s scin and s RPC a e subjec o sc u iny in Sec ion 6. As shown, he beam-p ofile u ilized in his expe imen is ai ly na ow, and he e is indeed some his o ic con o e sy on whe he he esul s ob ained unde his non-uni o m si ua ion can be sa ely ex apola ed o he uni o m si ua ion [40,41], ha is usually o highe p ac ical in e es . A na u al in e p e a ion o his disc epancy is based on cha ge di usion o e he esis i e pla es, bu i is no e iden ha i can ha e an impac a he cm- scale [19,21,42], ha is o in e es he e. In he ollowing we will assume ha cha ge di usion is no ele an in he p esen si ua ion and will ollow he analysis p ocedu e ske ched in Re . [33]. Hence, in o de o deal wi h he non-uni o mi y o he beam, we eso o he a e age flux , defined as he expec a ion alue o he local flux ðx,yÞweigh ed wi h i s co esponding p ob- abili y dis ibu ion unc ion, pðx,yÞp ðx,yÞ: ¼ZZ AOI ðx,yÞpðx,yÞdx dy ¼Z x 2 x 2 Z y 2 y 1 2 ps x s y e ððx 2 =2 s 2 x Þþðy 2 =2 s 2 y ÞÞ a 2 ps x s y e ððx 2 =2 s 2 x Þþðy 2 =2 s 2 y ÞÞ dx dy ð1Þ whe e AOI :x 1 oxox 2 y 1 oyoy 2 ( The no maliza ion ac o a is o he o de o 1 o he ypical in eg a ion bounds conside ed he e, o which we e e he ea e as he ‘a ea o in e es (AOI)’. The AOI is chosen depending on he egion conside ed o ob aining a ce ain obse able (e ficiency, esolu ion, e c). Fo he e ficiency es ima e, he AOI is chosen as he whole e ec i e a ea o he coun e ; while o he esolu ion es ima e, he AOI is chosen as he a ea o one eadou cell. The beam a e is ob ained om he fi s scin illa o ups eam S 1 /S 2 ha is known o be la ge enough o p ac ically con ain he en i e beam spo a i s loca ion. 2 Wi hin he ange ha is o in e es in his wo k, he S 1 /S 2 a es showed a linea esponse wi h he luminosi y, be e han 10%. Impo an ly, as demons a ed in Re . [33], i he cha ge di u- sion o e he elec odes can be neglec ed, he a e age flux defined h ough Eq. (1) ep esen s a mos con enien choice: namely, i allows o di ec ly ela e he de ec o pe o mance unde non- uni o m i adia ion (subsc ip n–u) o he one o eseeable in a si ua ion whe e he flux ( ) would be uni o m (subsc ip u) o e he en i e coun e , h ough he equi alence e ðnuÞ ð Þ e ðuÞ ð Þ e ðuÞ ð Þ; s ðnuÞ ð Þ s ðuÞ ð Þ s ðuÞ ð Þ This ac makes he in e p e a ion o he p esen esul s s aigh o wa d and independen om he ac ual beam shape. A sho compila ion o hese esul s can be ound in (Appendix A). Fig. 6. Beam p ofiles ob ained wi h he finge scin illa o S 5 and wi h he s ip MRPC. (a) Along he s ip di ec ion (x), ha is pa allel o he floo : de e mined by S 5 (con inuous line), by he MRPC unde he ex e nal scin illa o igge (do ed/c osses) and by he MRPC i sel in sel - igge (dashed/ci cles). (b) S 5 p ofile pe pendicula (y) o he floo (con inuous) and ex apola ed p ofile a he RPC (dashed), by assuming mul iple sca e ing. (c) 2-D beam p ofile de e mined by S 5 unde an elec on a e o 5.4 kHz. (d) 2-D beam p ofile de e mined by he MRPC a he same coun ing a e and loca ion. 2 Exposu e o films is pe iodically done S 1 /S 2 ups eam o expe imen s a ELBE. The ac ha he beam ‘fi s’ inside S 1 /S 2 was checked in p esen measu e- men s by compa ing he sel - igge a es de e mined om he s ip RPC and S 1 /S 2 , showing an ag eemen be e han 10% a low a es. J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–5144 A his poin , he beam p ofile used o he analysis is he one de e mined wi h he finge scin illa o S 5 , ollowing he choice made in p e ious s udies wi h ce amic-based RPCs [37], and allowing o easie c oss-compa isons. I has o be no ed ha o infini e in eg a- ion bounds (la ge AOI) and pe ec ly Gaussian p ofiles, Eq. (1) yields ¼ 4 ps x s y :ð2Þ In p ac ical e ms, he choice o an ‘e ec i ely i adia ed a ea’ A n ¼4 ps x s y allows ela ing he pa icle a e wi h he a e age flux, unde Eqs. (1) and (2). Di e en choices o he e ec i e a ea na u ally ansla e in o di e en es ima es o he flux. Thus, in he p esen expe imen al si ua ion a sys ema ic unce ain y exis s, o he o de o scin RPC ¼A n RPC A n scin ffi s x,RPC  s y,RPC s x,scin  s y,scin ¼2:3 (1.9 i accoun ing o he fini e RPC esolu ion), sligh ly dependen on he chosen AOI. We will come back o his ele an issue in Sec ion 6. 4.4. Da a analysis The da a analysis has been pe o med ollowing he algo i hms in Re . [24]. The e ficiency is defined as he numbe o RPC e en s wi h alid ime signal in a leas one cell (ei he s ip o pad) and bo h i s wo eadou ends (in he s ip- eadou case) di ided by he numbe o igge ed e en s. In he pa icula case o he pad coun e , since i s a ea does no ully con ain he beam in he e ical di ec ion, an auxilia y hin scin illa o (S 6 ) was used o educe he igge a ea in o de o ob ain he scaling ac o o he e ficiency (1/0.85). Rega ding he ime esolu ion, due o he mul i-cell en i onmen , he e a e se e al ways o ob aining he ime esolu ion by di e en ly weigh ing he ime egis e ed by all fi ed cells (a e age, weigh ed a e age, as es , o men ion some). We chose he e a di ec me hod based on aking he ime om he cell wi h he highes cha ge (o he wo-end a e age ime in he case o he s ip coun e ). In his way we a oid any di e en ial analysis as a unc ion o s ip-mul iplici y, which would equi e o he wise a sepa a e communica ion. We selec hecellbes cen e edwi h espec o hebeam-p ofileanddefine he AOI a i s bounds. So, a ‘cell esolu ion’ is gi en. Ob aining he ime esolu ion o e he whole coun e by applying his p ocedu e o each cell is a edious, albei s aigh o wa d, echnical ma e , ha is o ele ance in he final sys em, bu ha we ha e no included in his analysis, o simplici y. Finally, e y so cu s we e pe o med on he scin illa o cha ges, in o de o selec elec ons wi h simila ene gy deposi s. This has li le o no e ec on he RPC pe o mance. Fig. 7 shows he ypical ime-cha ge slewing co ec ion and he ime dis ibu- ion o he pad MRPC. The a e age pa icle flux is 28(17) kHz/cm 2 as ob ained om he S 5 scin illa o (RPC). 5. Expe imen al esul s 5.1. HV scan MRPCs we e condi ioned unde high ol age o a ew hou s in o de o each a s able, low da k- a e wo king egime. The Fig. 7. Typical T–Qco ec ion and ime dis ibu ion o he pad- eadou coun e (PMRPC). Up: aw T–Qco ela ion and ime dis ibu ion. Bo om: T–Qco ela ion and ime dis ibu ion a e slewing co ec ion. J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–51 45 elec onics h eshold was se a 30 mV ( o d -impulse exci a ion i ansla es o q h ¼30 C). The da k a e o he s ip module (SMRPC) was abou 3.5 Hz/cm 2 a 109 kV/cm, while ha o he pad module (PMRPC) was a ound 10 Hz/cm 2 a 110 kV/cm. Bo h coun e s showed da k cu en s below 0.13 nA/cm 2 . I has o be no ed ha , despi e he app oxima e 10- old inc ease o he da k a e as compa ed o ha o floa glass (1 Hz/cm 2 ) i emains a ac o 1000 below he ypical wo king flux o u u e high- a e applica ions. Fo his glass, as o floa glass, he o igin o he da k a e (in ma ked excess wi h espec o ha expec ed om na u al backg ound adia ion) is poo ly unde s ood a he momen . In o de o find he op imum wo king ol age o he coun e s, he e ficiency and ime esolu ion we e scanned as a unc ion o he applied ol age o a ‘low’ flux o 25kHz=cm 2 .The esul sa e summa ized in Fig. 8. The coun e s exhibi a com o able ope a ing ‘pla eau’, simul aneously p o iding e ficiency abo e 95% and ime esolu ionbelow65pso e asmuchas600V a ia ioninHV.Such a pla eau allows o a con enien (and su ficien , o he in ended applica ion) 10% ole ance agains gas densi y fluc ua ions due o uncon olled en i onmen al condi ions (a ising mainly h ough p essu e a ia ions ha modi y he educed field E/P). The in insic RPC esolu ion can be ob ained a e sub ac ing he con ibu ions o he FEE and TDC as ob ained om pulse measu emen s. A c ude es ima e o he esolu ion down o s RPC 50 ps is ob ained om his app oxima e p ocedu e. Al hough a p ecise posi ion scan is ha dly accessible o a low- ene gy elec on beam, we pe o med a gene ic s udy o he depen- dence o he RPC pe o mance wi h posi ion wi hin ou (limi ed) igge condi ions, ob aining a s a is ical sp ead o 71.573.5 ps a 25 kHz/cm 2 when combining he pe o mance o all 12 pads as measu ed in sepa a e uns (Fig. 9). Pad-e ficiency is mo e sensi i e o beam-de ec o misalignmen and geome ical ine ficiency, howe e he coun e e ficiency was obse ed o be well wi hin a 2% a ia ion when he igge was cen e ed wi h espec o any o he ou cen al pads. We no e ha he e ec i e a ea A n o e which he a e is sp ead (Eq. (2)), ex ends well beyond he pad bounda ies (pad a ea¼ 2.2 cm 2.2 cm¼4.84 cm 2 ,A n ¼4 p (1.1 cm 1.5 cm)¼20 cm 2 )and he e o e he p esen condi ions ep esen , o low a es, a easonable fi s -o de app oxima ion o he coun e pe o mance unde he final uni o m use-case, as a as he e ec o signal loss a ising a he cell bounda ies is conce ned. Coun e non-uni o mi ies, beyond his in insic limi a ion o mul i-cell eadou s ha s ems om he shape o he induc ion p ofile, can be shown o be small h ough he a o emen ioned pad scan and esul s admi a di ec ex apola ion o he s ip coun e , due o he simila a chi ec u e. Las , he pad coun e showed an anomalously high clus e size (a e age numbe o cells fi ed pe p ima y elec on) o 1.4 a ypical wo king fields ( o a ela i e posi ion beam-chambe as shown in Fig. 9). This was aced ( h ough a dedica ed oscillog am s udy) o 5 ns-pe iod oscilla ions in he signal base-line, he e o e ha ing in all likelihood an elec onic o igin (signal ansi ime in he pad is a ound 100 ps). Despi e he clus e size esul ing om his oscilla ing c oss- alk is s ill in e io o compa able o o he p oposals o he CBM expe imen , we in end o minimize i by imp o ing he quali y o he connec ion om he coun e o he elec onics in nea u u e. Elec onics (ASIC-based) and connec- ions will be di e en in he final sys em (see Re . [17] and e e ences he ein). 5.2. Ra e scan In o de o s udy he a e capabili y o he MRPC modules, he coun e s we e es ed as a unc ion o he elec on flux. The a e age flux is defined by Eq. (1), wi h he beam p ofile es ima ed om he finge scin illa o S 5 . Fo e ficiency es ima es we ake as he AOI he whole ac i e egion o he coun e s, while o he ime and cha ge dis ibu ions we es ic he AOI o he cell bes cen e ed wi h espec o he beam p ofile. The e olu ion o he p omp cha ge dis ibu ion o he mos cen al cells can be seen in Fig. 10 o bo h MRPCs. The pad and Fig. 8. E ficiency and ime esolu ion as a unc ion o high ol age (HV). Le : HV scan o he pad- eadou module (PMRPC) a an a e age elec on flux o 1.9 kHz/cm 2 . Righ : HV scan o he s ip- eadou module (SMRPC) a an a e age elec on flux o 5.4 kHz/cm 2 . Fig. 9. Time esolu ion o each pad a 2 ypical pa icle fluxes (5, 25 kHz/cm 2 ). The beam p ofile ( ep esen ed by a shadowed o al) has been chosen o all hese measu emen s o be cen e ed wi h espec o he pad unde s udy (by emo ely mo ing he pla o m on which he RPC was si ing). The AOI used o es ima e he quo ed a e age flux, , is illus a ed wi h dashed lines. Excep o he pad scan (whose nume ical esul s a e shown in figu e), he ac ual ela i e alignmen beam-chambe o all measu emen s o he pad esponse p esen ed in his wo k is he one ske ched in he figu e (and simila ly o he s ip coun e ). The AOI o he flux de e mina ion in he e ficiency s udy is aken o be he ull ac i e size o he de ec o i sel . J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–5146 s ip MRPCs we e ope a ing unde elec ical fields o 109 kV/cm and 103 kV/cm (6.0 kV and 6.45 kV in Fig. 8), espec i ely. Fo he s ip MRPC, he cha ge is ob ained om he sum o i s wo ends. I can be seen ha , wi h he inc ease o he elec on flux, he a e age cha ge o bo h coun e s dec eases and he spec um shi s down o low cha ges as expec ed. Limi ed by he beam ime schedule, we could no ake da a files a high pa icle flux o he s ip coun e ope a ed a an op imal field o 108 kV/cm (6.75 kV in Fig. 8- igh ) and pe o med he s udy a jus 6.45 kV. This depa u e om he cen e o he e ficiency pla eau owa d i s le end (si ing a 6.2 kV o a 5% e ficiency d op; Fig. 8- igh ) made he s ip coun e in his wo k un ealis ically sensi i e o he dynamic ol age d op a ising a high flux ( ha shows an app ox- ima e linea beha io wi h j : see o ins ance Eq. (4) o Re . [17] o a de ailed accoun ). The high-flux beha io o he e ficiency and ime esolu ion o he wo modules a e shown in Figs. 11 and 12. As said, due o he limi ed beam ime, no all combina ions in pa ame e space could be in es iga ed and he da a files o he open symbols in Fig. 11 (a inc eased field) we e absen . Acco dingly, i is impossible o ob ain he ime esolu ion o e ficiency o hese uns. Ne e - heless, he de ec o cu en and a es we e eco ded so we can p o ide an es ima e o he e ficiency h ough i , based on a simple DC model. A na u al consequence o his physical image is ha bo h he e ficiency and cha ge dis ibu ions depend on he e ec i e field in he gap a e co ec ing o he dynamic DC ol age d op a he esis i e pla es. Hence, he e ficiency depends s ongly on he a e age cha ge (i.e., he gain), and he la e can be es ima ed om he cu en . Taking he da a du ing he HV scans as a e e ence, we ex apola e he e ficiency o he gi en o al cha ge in a c ude way. Al hough his p ocedu e canno eplace a di ec measu emen , i is good enough o ob ain a easonable expec a ion. The DC model and he defini ion o he o al cha ge will be desc ibed in he ollowing sec ion. I can be seen in Figs. 11 and 12 ha he maximum ole able pa icle flux app oaches he 100 kHz/cm 2 land-ma k o hese coun e s based on low- esis i e glass, eminiscen o he beha io obse ed o ce amic coun e s unde simila condi ions [37]. F om he indi ec analysis (open symbols), i can be seen ha he s ip coun e unde an elec ic field a ound 108 kV/cm (ci cles) is expec ed o show a highe e ficiency a high pa icle flux, which sugges s ha he coun e pe o mances could s ill be imp o ed by op imizing he wo king ol age. 6. Discussion As desc ibed in Sec ion 4.3, he beam-p ofile was de e mined independen ly h ough a finge -scin illa o (S 5 ) and he MRPC i sel . An ob ious disc epancy be ween he wo es ima es can be obse ed (Fig. 6). By using he simple mul iple sca e ing o mula om Re . [43],weob ainawid ho s x ¼ s y ¼1 cm, o a poin -like 30 MeV elec on beam a S 1 /S 2 , conside ing i s hickness (0.5 cm, PVT) and le e a m (23 cm dis ance be ween S 1 /S 2 and he RPC). This is ai ly consis en wi h he alues measu ed by he finge sicin illa o S 5 placed a he RPC posi ion ( s x ¼1.1 cm, s y ¼1.5 cm). To in e p e he influence o RPC ( s x ¼1.8 cm, s y ¼2.0 cm) wi h simple o mulas is mo e di ficul , since he RPC i sel is ac i e. Taking he o mulas a ace alue, he obse ed sp ead would equi e ha ing an addi ional 1.5 cm Al-equi alen ma e ial budge wi h an inc eased le e a m by a ound 6 cm. These numbe s exceed by some 50% a nai e expec a ion om he known ma e ials and dimensions o he chambe bu can be conside ed easonable, Fig. 10. Cha ge dis ibu ion a di e en flux o e one eadou cell. The field is 103 kV/cm o he s ip coun e and 109 kV/cm o he pad coun e . Fig. 11. Measu ed e ficiencies o di e en uns as a unc ion o he a e age pa icle flux de e mined wi h e e ence scin illa o s. Open symbols a e ob ained h ough an indi ec analysis ha is desc ibed in he ex . J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–51 47 in iew o he limi ed scope o he analy ical o mulas. Since he posi ion esolu ion om he RPC alone is a 0.8 cm le el, i is plausible ha his addi ional sp ead is mainly caused by mul iple sca e ing in he RPC. As p e iously no ed, he e is indeed some con o e sy on whe he he esul s ob ained unde non-uni o m condi ions can be ex apola ed o an uni o m si ua ion, a na u al in e p e a ion o his disc epancy being he di usion o he a alanche cha ge o e he esis i e pla es. The phenomenon is expec ed o occu , al hough a a scale significan ly smalle han he cm-scale ha is o in e es he e. Howe e , e en i cha ge di usion could be neglec ed, he e is s ill some deg ee o sophis ica ion needed in o de o p ope ly in e p e he esul s in ligh o he uni o m si ua ion, and an ope a i e app oach o he p oblem has been ecen ly desc ibed in [33]. In o de o p ope ly add ess he case we spli he discussion in wo pa s. 6.1. Ra e capabili y and locali y I cha ge di usion is o impo ance, i can be expec ed ha he local de ec o beha io will be li le dependen on he local alue o he pa icle flux. As an example, i i is (un ealis ically) assumed ha cha ge di usion o e he su ace o he esis i e pla e is an ins an aneous p ocess, he o al cha ge eleased would uni o mly dis ibu e o e he comple e de ec o , o a ea A. The e ec i e pa icle flux would be ¼ /A, usually much smalle han he plausible es ima e ¼ /A n ( om Eq. (2)) o any p ac ical non- uni o m si ua ion. As a esul , he obse able esponse o he coun e (e ficiency, ime esolu ion) would be uni o m e e y- whe e, and would co espond o he esponse o a pa icle flux a ificially educed by A n /A; a e capabili y would be o e - es ima ed p ecisely by his ac o . Con e sely, i he RPC esponse would show a s ong local beha io , cha ge di usion migh be excluded, a leas quali a i ely. In o de o s udy he locali y o he RPC esponse i is mos con enien o s udy he de ec o gain as a unc ion o i s posi ion, and o see whe he i is ela ed o he alue o he incoming local flux o no . This is s ic ly no possible o an RPC in he p esen si ua ion because: (i) RPC elec onics usually senses he as / p omp cha ge, no he gain and (ii) he e is no ex e nal posi ion- sensi i e de ec o in he p esen se up. An al e na i e app oach is ske ched as ollows. Fi s , he a alanche as cha ge, q as , is ob ained om he QDC (a e calib a ion) as a unc ion o he posi ion along he RPC s ip coun e , de e mined wi h he RPC i sel , a a e e ence elec ical field o 103 kV/cm (Fig. 13-le ). I is indeed possible o app oxi- ma ely ansla e he as cha ge o o al cha ge by assuming ha hey a e unc ionally ela ed in a way ha is independen om he pa icula ol age and pa icle flux. So a calib a ion cu e q as –q o can be ob ained om he low-flux HV-scans o Fig. 8, o which he cu en was moni o ed, by using q o ¼1 2 I ð3Þ This ela ion holds o mi o ed RPCs (wi h up–down symme- y, like he ones s udied he e) wi h being he pa icle a e. Fig. 13- igh shows he o al cha ge (q o ) o he coun e , in esponse o an elec on ha hi s he RPC in a ce ain posi ion, ob ained by his p ocedu e. In his s ep (q as –q o calib a ion) we a e a guing along he lines o he DC model, bu his assump ion is no ins umen al o he e alua ion o he locali y o he RPC esponse ha ollows. The eason o he choice o q o is ha , being ela ed o he gas gain h ough he a e age numbe o p ima y elec ons eleased pe gap m¼q o /n 0 510 6 i eadily allows o in e p e a ion o he final esul based on he con en- ional DC model, as will be shown. Fig. 12. Measu ed ime esolu ions o he cell wi h highes s a is ics (cen e ed wi h espec o he beam) as a unc ion o he a e age pa icle flux de e mined wi h e e ence scin illa o s. Fig. 13. Le : as cha ge as a unc ion o he RPC posi ion o di e en a es. Righ : same figu e, o he o al cha ge a e applying he ex apola ion p ocedu e desc ibed in ex . (We ha e e-s udied ou ea lie da a and confi med ha a numbe o e o s esul ed in a p e ious unde -es ima ion o he a e age cha ge down o 1 pC o ypical wo king condi ions, as gi en in [24]. A figu e q o 2–2.5 pC a he ypical ope a ing poin is mo e adequa e o ou a chi ec u es.) J. Wang e al. / Nuclea Ins umen s and Me hods in Physics Resea ch A 713 (2013) 40–5148