Study on the TU gas for the GEM-TPC detector

cPC(HEP&NP),2009,33(4):289-291Chinese Physics CNo.4,Apr:,2009Study on the tu gas for the GEM-TPC detectorQI Hui-Rong(祁辉荣)1) LI Yu-Lan(李玉兰)LIJn(李金GAO Yuan-Ning(高原宁)2 LI Yuan-Jng(李元景)11( Deptartment of Engineering Physics, Tsinghua University, Beijing 100084, China)2(Institute of High Energy Physics, CAS, Beijing 100049, China)Abstract In this paper several different working gas mixtures for GEM-TPC were evaluated based on aGarfield simulation. Among them, Ar: CH: CF4=90: 7-3(named herein tu gas) was selected for a detailedstudy because of its better performance. Some performances of drift velocity transverse diffusion, spatialresolution and the effective number of electrons in various electric fields were obtained. The performance of aGEM-TPC prototype working in the tu gas was studied and compared with that in Ar: CHa=90: 10( P10 gas)Key words GEM detector, TPC (time projection chamber), spatial resolution, transverse diffusionPACS 29.40. Gx1 Introductiondiffusion. In the study the following gases weresimulated: TU gas(Ar: CH,: CF4=90: 73), Iso gasGEM-TPC (time projection chamber)is a new (Ar: CH iCH10=94: 3: 3 ), P10(Ar: CH=90: 10), TDRtype gas tracking detector with good position resolu-(Ar: CO2 CH4=93: 2: 5)and P5(Ar: CH4=95: 5)tion and track reconstruction , 2. Its main advantageis a novel readout detector--GEM. At Tsinghua Uni. 2 Simulationversity it has been studied extensively as a promising 2.1 Transverse diffusioncandidate for the ILC 3I central tracking detector andalso studied as a main detector for the inner-targetThe transverse diffusion of electrons degrades theexperimental spectrometer in the CSr(cooling stor-age ring) of Institute of the Modern Physics, CASIn our laboratory, a GEM-TPC prototype was suc-cessfully developed and its performance was studiedusing cosmic rayGarfield was written by Professor Rob Veenhofbased on the monte carlo method and finite ele-ment analysis at CERN5. An interface to the Mag- 5P10boltz program was provided for the electron transportproperties in nearly arbitrary gas mixtures. Com-bined with Maxwell software it can simulate the de-tails of two-dimensional and three-dimensional driftchambers and optimize their designs. The drift ve-locity, attachment coefficient and diffusion cocan be simulated in high precision using aenough collision number. A collision number setE/CV/cm)of 100x 96000 leads to a statistical error of 0. 13%Fig 1. Transverse diffusion constant as a funcin the drift velocity and of 4.5% in the transversetion of drift field for different gas mixtures.中国煤化工Received 7 July 2008, Revised 22 August 2008CNMHGSupported by CAS/SAFEA International Partnership Program for Chear1)E-mail:qihr@tsinghua.edu.cn@2009 Chinese Physical Society and the Institute of High Energy Physics of the Chinese Academy of Sciences and the Instituteof Modern Physics of the Chinese Academy of Sciences and IOP Publishing LtdChinese Physics C(HEP NPspatial resolution of the r-o plane(or r-y in our TPO一 hodoscopeprototype). The relation between transverse difusionconstant Cp and Drift is shown in Fig. 1. From thereone can see that both the tu gas and the Iso gashave smaller Cp than all the other gases in questionWhen E is near the saturation field Emax, Cp reachesFig 3. Diagram of the system setupits minimum valueIn the test of the drift velocity, a hodoscope2.2 Drift velocitywith its small size in the z direction was placed at fiveIn certain gases and electric fields, the electrons test points(their positions were determined based onare moving without a preferred direction on the basis the mechanical design)along the whole drift distanceof random fluctuations. The drift velocity is deter- The drift time was obtained by taking the peak timemined by the functionof a Gaussion fit to the raw data( see Fig. 4).Thedrift velocity was calculated by a linear fit of the rela(1 tion between the drift distance and drift time. In thetests for the tPc performance studies another ho-Here, E is the electric field, P is the gas pres- oscope trigger system covering the whole drift volsure and ve is the drift velocity. When the velocityume was usedreaches the saturation velocity(nearly maximum)the electron drift in the gas depends slightly on EFig. 2 shows the velocity from the garfield simulationfor these gases for different electric fields. Orne canee that the saturation velocity of TU is the largestamong all these gases, and its saturation field is about250 V/cm, which makes it a promising candidate forTPC systems, especially for systems which16182022242628high drift velocity to implement a high event rateime05ns(×103)ig. 4. The histogram of drift time(0.75 ns,E7654324 Test results4.1 Drift velocityThe drift velocities for two different edrift valuesrere measured and listed in Table 1. They are veryclose to the simulated resultsE/(V/cm)Table 1. values of drift velocities from simula-Fig. 2. Drift velocity of the electron as a func-tion and measurement.tion of the drift field for different gas mixtures.Drift/(V/cm)drift velocity/(em/us)3 Test system setup180220048.1士03The GEM-TPC prototype includes the following4.2 Transverse diffusion constantmain parts: a p300 mm gas chamber with a 500 mmdrift distance, a triple gem readout detector with anThe transverse diffusion constant Cn can be deactive area of 100 mmx 100 mm, 312 readout padsterminedfollowing relationeach with a size of 1.6 mm(a direction)x100 mm(y di-中国煤化工rection), a vme based daQ system, a cosmic-ray ho-Zdoscope trigger system and a GEM readout mounted Here.NMHGpad response at aon the center of the end of the chamber. The Tu drift distance of Z, and aPR(O)=aPR(Z)le=0, whichgas was supplied as a pre-mixed gas. Fig. 3 shows a is caused by the finite width of the readout pad andawingof thethe electron diffusion in the GeM detectorQI Hui-Rong et al: Study on the tu gas for the GEM-TPC detectorFigure 5 provides the measured Cp of the TU and 4.4 T-resolutionP10 gases for the diffusion drift field in a magneticfield of 0 Tesla(all detectors and chamber mounted inThe T-resolution o. is determined as the r msthe magnetic tunnel). Here o is the azimuthal angle value of the distribution of the residues between mea-and 0 is the polar angle, both measured with respect surement points and the fitting trackto the y-direction. The diffusion constant of the TUFigure 7 gives the resolution of our GEM-TPCgas is obviously smaller than the one of the P10 gas. prototype working in TU and P10 gases in a 0TeslaHowever, all the values measured are smaller than magenetic field with VGEM=370 V. It is obvious thatthose predicted by Garfieldthe TPC working in a tu gas has a better resolutioneven at drift distances(> 300 mm)where the spatialC,ArCH,CF49070T,◆°,6l0°resolution is very poor for the P10 gasmulation2.0TUTPC,P10,0T,◆<°VGEM370 V, Edif"13001201401601802002206420Telectric field of dnf/(V/cm)VGEM370 V, Edit" 180.2 V/cmsured transverse diffusion constantat different drift fields for tU gas and P10 gas.8°<:,,土43N。m··50100150200250300350400450500Ne is the effective number of electrons on theift distance/mmreadout pad. At relative large drift distances, theelation of the spatial resolution and Net in a TPcFig. 7.esolution at different drift distancessystem width a MPD(micro pattern gas detector,for tu gas and P10 gasincluding GEM)is 02=02+(Cp/v/ 2Z. So, Nef isan important performance parameter in a GEM-TPC5 ConclusionFigure 6 shows Ne for the TU and P10 gases fordifferent drift fields with VGEM=370 V. It is obviousBecause of the small diffusion constant and fastthat the value of N reaches its maximum at a drift drift velocity, CF4 was added to the Ar and CH4field close to its saturation value. In our test, Nef for mixture as a working gas for the GEM-TPC, espe-the P10 gas has reached the maximum, while for the cially for applications where a large drift velocity isTU gas this is has not yet been the case because the needed. Based on a Garfield simulation the tU gasmaximum available drift field at this moment is still has been selected for a detailed experimental studylower than its emaxTest results confirmed that the GEM-TPC prototypecan achieve a better performance working in Tu gas80TUTPC,ACH,CF49073叮◆2,rather than in P10 gas, which is a common gas in60}·Twc,ACH10.02traditional TPCsWe are grateful to Professor Fabio Sauli from100120140160180200220CERN, Professor Takeshi Matsuda, Prof. Keisukeelectric field of drift/(V/cm)Fujii from Kek, for their help. We also thank colFig. 6. Measured effective number of electronsleagues from IHEP for adapting MDC DAQ to ourat different drift fields for TU gas and P10 gasprototype systemReferences4 LI Yu-Lan, CAO Liang-Jun, QI Hui-Rong et al. CPC(HEP&1 Sauli F Nucl. Instrum. Methods A. 1997. 386: 531-535中国煤化工2 Blum W, Rolandi L Particle Detection with Drift Cham- 6 LICN SCience Symposirs. Berlin: Springer- Verlag Second printing, 1994, 120-w:.EE:2452471507 Kobayashi M. Nucl. Instrum. Methods A, 2006, 562: 136-3http://www.linearcollider.org/