Simulation on the Transparency of Electrons and Ion Back Flow for a Time Projection Chamber based on Staggered Multiple THGEMs
Mengzhi Wu, Qian Liu, Ping Li, Shi Chen, Binlong Wang, Wenhan Shen, Shiping Chen, Yangheng Zheng, Yigang Xie, Jin Li
SSimulation on the Transparency of Electrons and IonBack Flow for a Time Projection Chamber based onStaggered Multiple THGEMs
Mengzhi Wu a , Qian Liu a , Ping Li a , Shi Chen a , Binlong Wang a , WenhanShen a , Shiping Chen a , Yangheng Zheng a , Hongbang Liu b , Yigang Xie c , JinLi c a School of Physics, University of Chinese Academy ofSciences, , Beijing, 100049, , China b School of Physics, Guangxi University, , Nanning, Guangxi, 530004, , China c Institute of High Energy Physics, , Beijing, 100049, , China
Abstract
The IBF and the transparent rate of electrons are two essential indicators ofTPC, which affect the energy resolution and counting rate respectively. Inthis paper, we propose several novel strategies of staggered multi-THGEMto suppress IBF, where the geometry of the first layer THGEM will be opti-mized to increase the electron transparent rate. By Garfield++ simulation,the electron transparency rate can be more than 90% of single THGEMwith a optimized large hole. By simulating these configurations of triple andquadruple THGEM structures, we conclude that the IBF can be reduced to0.2% level in an optimized configuration denoted as ”ACBA”. This strat-egy for staggered THGEM could have potential applications in future TPCprojects.
Keywords: staggered multiple THGEM, ion back flow, electron transparent rate,Garfield++
Email addresses: [email protected] (Mengzhi Wu), [email protected] (Qian Liu)
Preprint submitted to Elsevier February 25, 2021 a r X i v : . [ phy s i c s . i n s - d e t ] F e b . Introduction The time projection chamber (TPC) is an important detector in particlephysics and nuclear physics experiments. Gaseous TPCs are widely usedin nuclear collider experiments like STAR[1] and ALICE[2, 3], and electroncollider experiments like ALEPH[4] retired and CEPC[5] in progress, whileliquid or dual-phase TPCs are used in neutrino experiments NEXT[6], anddark matter experiments XENON-1T[7] and PANDAX[8]. The basic func-tion of TPC is to measure the 3-dimensional track of final-state particles.In a collider experiment, TPC is able to measure charge and momenta ofparticles by a deflection magnetic field. In addition, for a GeV scale exper-iment, TPC can also be used for particle identification (PID) by measuringthe energy loss dE/dx of particles in working gas.Modern gaseous TPC usually use micro pattern gaseous detectors (MPGD)[9]as their readout chamber, in order to adapt the high counting rate MHz / mm in modern colliders. The main technical indicators of a MPGD contains thegain, the transparency of electrons, the ion back flow (IBF) rate and thecounting rate. These factors would influence its performance such as thedetection efficiency, the spacial resolution and the energy resolution.The transparency of electrons is the ratio at which the electrons producedby the primary ionization can pass through the MPGD. This ratio affects theenergy resolution of TPC, so it impacts the measurements on the energy lossdE/dx of particles, which is essential for particle identification(PID) of Pions,Kaons and protons with momenta less than 1 GeV. Hence the transparentrate of electrons is expected to be as large as possible. In reference [10],the authors studied several GEM-like structures by Garfield simulation andoptimized the electron transparency to more than 80%.The ions produced by primary and secondary ionization would flow backto the drift zone of a TPC. These ions will lead the spacial charge effect thatthe drift electric field is distorted which would affect the velocity of driftelectrons and result in the bad spacial resolution of TPC. Therefore, theIBF in a TPC is expected to be suppressed as low as possible. One solutionis to seek new structures of MPGD. For example, the mesh of MicroMegascan absorb a number of ions, so some groups try to use double and triplemeshes structures to reduce ions with IBF 0.01% level[11, 12]. However, it ishard for MicroMegas to achieve large-size manufacturing. For example, thesize MicroMegas in [11, 12] is 2 . × . ∼ H (90/10). Insection 4, we propose several staggered configurations of triple and quadrupleTHGEMs and find out that the IBF can be reduced to 0.01% level in aoptimized configuration.
2. The gas properties
It is important to choose an appropriate working gas to optimize the per-formance of TPC. In laboratory, the binary gas mixtures are commonly usedsuch as Ar-CF [17], Ar-CO [18] Ar-iC H [28] as well as some ternary mix-tures. The drift velocity and diffusion of electrons are quite different in puregases. Generally, the electron drift velocity is supposed to be less affected byelectric field in order to optimize the time and z-coordinate resolution. Also,the diffusion coefficients is expected to be as small as possible to obtain abetter spatial resolution.Thus, in this section, we would simulate the drift velocity and diffusioncoefficients of electrons by Garfield++[19]. We set the temperature as theroom temperature, the pressure as the standard atmospheric pressure, andthe Penning coefficient as 0.57[20]. In the case of Ar-iC H , we simulate the3 igure 1: The drift velocity and transverse diffusion coefficients of electrons in Ar-iC H at different proportions drift velocity and transverse diffusion coefficients of electrons in Ar-iC H at different proportions, shown as Fig. 1.The drift velocity of electrons in TPC is supposed to be stable to improvethe time resolution and z-coordinate spatial resolution and to be as fastas possible to reduce the dead time of the detector, while the transversediffusion coefficients is required to be as small as possible to improve thexy-coordinate resolution, so we choose Ar-iC H at 90:10 and constrain theelectric field between 280 and 650 V/cm where the drift velocity is between4.4 and 4.6 cm/ µ s.Similarly, we simulate the same properties of electrons in Ar-CF , Ar-CH , Ar-CO at different ratio then optimize the proportion of gases andthe electric field, which is summarized as Fig. 2 and Table 1. The resultfor Ar-CH matches well with reference[21]. If a TPC is designed with onemeter drift zone, then the corresponding drift time and xy-coordinate spatial4 igure 2: The drift velocity of electrons in different working gasesTable 1: The appropraite field, drift velocity and transverse diffusion coefficients in differ-ent working gases ratio optimizedfield / V drift velocity/ ( cm/µs ) transverse diffusioncoefficient / ( µm/ √ cm ) transverse diffusioncoefficient under 1 Tesla/ ( µm/ √ cm ) drift time in1 meter TPC/ µ sAr-iC H ∼ ∼ ∼
380 150 ∼
300 21.7 ∼ ∼
550 4.0 ∼ ∼
425 200 ∼
300 22.7 ∼ ∼
550 10.0 ∼ ∼
265 70 ∼
130 9.3 ∼ ∼
220 5.0 ∼ ∼
615 90 ∼
200 17.8 ∼ D T ( B ) D T (0) = (cid:114)
11 + ω τ (1)where ω = q e m e | B | and τ is the mean time between collisions.
3. The transparency rate of electrons
In a gaseous TPC, the energy loss of a charged particle is mainly convertedto primary ionization, so the readout chamber of this TPC is expected tocollect all of the primary electrons to reconstruct precise dE/dx. Thus theratio of primary electrons passing through the multi-THGEMs, especially thetop one, is required to be as large as possible.5enerally speaking, when the electrons of primary ionization reach thefirst THGEM, some may be absorbed by the copper of the upper and lowersurfaces of THGEMs, which could not arrive lower layers of THGEMs toavalanche. So we define the effective transparent rate of electrons as theratio at which the electrons could arrive the second THGEM.The main factor that influence the transparency rate is the optical trans-parency rate that means the ratio of THGEM holes area to total THGEMworking area. Since the THGEM is a regular triangle periodic structure, theoptical transparency can be formulated as η optic = πd √ l (2)where d is the hole diameter and l is the pitch of the top THGEM.In addition, the field lines near THGEM holes have a converging shape,so the motion of electrons near THGEM holes also converges to the holes.Then the transparent rate is generally larger than the optical transparency.Furthermore, the transparency is also influenced by electric field of bothdrift area and transfer area. Thus in this section, we would study the effectof the optical transparency and the electric field on the transparent rate ofelectrons.In this section, we establish a model of a periodic unit of a THGEMby COMSOL. The THGEM geometry is set as the thick 0.4 mm, the pitch1.0 mm and the transfer area height 2 mm, while the THGEM hole diame-ters and electric field in drift and transfer zone are variables to study. Theelectrons of primary ionization is set as a uniform distribution on a xy-planein the drift area. The electric field in the transfer area is fixed as 1500 V/cm, and the driftfield is set according to Table 1, while the hole diameter is a variable between0.4 and 0.8 mm. As for the voltage of the first THGEM, since this THGEMcan only suppress the IBF of lower THGEMs, so the gain of this THGEM issupposed to be low and we require it less than 10. Finally, we set the voltageof this THGEM as 800 V. In one periodic unit, initial electrons is set 10,000.According to the electron motion simulated by Garfield++, we count theproportion of electrons passing through the transfer area, i.e. transparency,and summarize it as Fig. 3. 6 igure 3: The Transparent rate of electrons of THGEM with different hole size
According to Fig. 3, when the aperture is larger than 0.6 mm, the trans-parency would not change significantly with the hole diameter. On the otherhand, the hole size is not supposed to be too large in order to reduce the ionback flow. So it is enough to set the hole diameter as 0.6 mm. In addition,in Fig. 3, it should be remarked that in Ar-iC H (90/10), the transparencycan reach 90%, so we will choose Ar-iC H (90/10) as working gas in furthersimulation.It’s worth remarking that in Ar-CF (93/7), the transparent rate is aslow as 20% level. By checking the end points and end status of electronsin THGEM with hole size 0.6mm in different gases, we summarize the endpoints in Table 2, and the cases of Ar-iC H (90/10) and Ar-CF (93/7) isvisualized as Fig. 4. In Ar-CF (93/7), there are more than 50% electrons thatis attached in the avalanche area. The attachment effect of Ar-CF (93/7) isso serious that the transparent rate is low, thus we don’t choose Ar-CF (93/7)as the working gas in this project.Finally, we did a further simulation on the influence of the initial positionon transparency, shown as Fig. 5. Besides the optical transparency, the electric field also has a great impacton the transparent rate of electrons. The shape of electric lines near theTHGEM hole depends on the ratio of field in drift and avalanche zone, andthat in transfer and avalanche zone, which impacts the efficiency of electronentering and leaving holes respectively.Since the transparent rate is generally insensitive to the exact value of7 a) The ended points of electrons in Ar-iC H (b) The ended points of electrons in Ar-CF Figure 4: The ended points of electronsFigure 5: The transparent rate of electrons at different initial position a) Transparent rate of electrons ofTHGEM under different drift field (b) Transparent rate of electrons ofTHGEM under different transfer fieldFigure 6: The Transparent rate of electrons of THGEM under different drift field andtransfer field Table 2: Electrons ended points Ar-iC H (90/10) Ar-CO (95/5) Ar-CH (90/10) Ar-CF (93/7)upper metal 0% 0.2% 0.05% 2.2%avalanche area 5.25% 15.46% 9.55% 53.82%lower metal 8.23% 8.54% 9.55% 3.48%transfer 86.522% 73.53% 80.69% 21.82%other 0% 2.26% 0.16% 18.67%the electric field strength, we may fix the voltage of THGEM to 800 V, i.e.the avalanche field to 15 kV/cm, and change the electric field in both driftzone and transfer zone. The result can be summarized as Fig. 6.According to Fig. 6, to ensure the transparent rate larger than 90%, thedrift field is supposed to be less than 1/25 of the avalanche field, while thetransfer field should be larger than 1/10 of the avalanche field. In the casewhere the avalanche field is 15 kV/cm, the threshold of the drift and transferfield are 800 V/cm and 1.5 kV/cm respectively.
4. The ion back flow of the multiple THGEM
The ion back flow would cause the space charge effect, i.e. the distortionof the drift field. As a consequence, the drift velocity of electrons wouldbe disturbed, and result in a worse time resolution and z-direction space9esolution, while the counting rate of TPC is also limited. So the IBF rateis an essential technical indicator to research. In early gaseous detectors,Some groups studied a GEM-like structure called MHSP and its improvedconfigurations called R-MHSP and F-R-MHSP, and suppressed the IBF as1 . × − [22, 23, 24]. Another GEM-like structure called Cobra is alsostudied to suppress the IBF [25].MicroMegas is another novel solution to suppress the IBF. The mesh ofMicroMegas can naturally absorb ions, and the IBF can be easily lower 0.1%level [26]. The combination of GEM/THGEM and MicroMegas also shows agood performance on suppression of IBF [5, 27]. More recently, double-meshMicroMegas is proposed [28], and the latest IBF result is 3 ∼ × − atgain > [11, 12]. But the large size and robustness of MicroMegas stillneeds further research.Staggered multi-GEM/THGEMs [15] is also a proposal on the rise. Somenovel configurations are proposed recently [16]. However, all of the pre-vious work only considered staggered strategies of two-layers although theresearchers use triple- or quadruple-THGEM structures. Therefore, in thissection, we would propose several staggered structures of multi-THGEMs,and study the IBF of these configurations.The fundamental staggered THGEM for 2-layers is easy to achieve by”flip flop” or 180 rotation. Then for the triple and quadruple THGEMs,the staggered strategies is the combination of these two structures. Thereare 2 configurations for triple-THGEM and 3 configurations for quadruple-THGEM, shown as Fig. 7 together with their periodic unit. Following thenotation in solid state physics, we use symbols like ”ABA” to represent acertain configuration. To avoid duplication of tags, we use a convention thatthe bottom two THGEMs are always marked as ”A” and ”B”. The size ofTHGEM is chosen: the thickness is 0.4 mm and the pitch is 1.0 mm, whilethe hole size of top THGEM and other THGEM are 0.6 mm and 0.4 mm.In order to perform Monte Carlo simulation of ion back flow, we need toknow the position distribution of ions produced at a THGEM. It’s feasible toassume that the production coordinates are uniform distribution in the holesof a single THGEM, although this algorithm doesn’t contain any informationof avalanche procedures. Thus we simulate the avalanche process in one holeof THGEM, and record the production position distribution of ions, includingthe z-direction and r-direction distribution, shown as Fig. 8. Besides, the gainof single THGEM can be also obtained, shown as Fig. 9, where the absolutegain is defined as the number of the secondary electrons in the avalanche10 a) ”ABA” config-uration (b) ”CBA” config-uration (c) ”BABA” con-figuration (d) ”CABA” con-figuration(e) ”ACBA” con-figurationFigure 7: The upper 2 figures are 2 typical configurations of triple THGEM denotedas ’ABA’ and ’CBA’ structures, and the lower 3 figures are 3 typical configurations ofquadruple THGEM denoted as ’BABA’, ’CABA’ and ’ACBA’ structures, where ’A’ alwaysrepresents the configuration of the bottom THGEM. In all figures, the yellow and bluecircles on the top surface of the bottom THGEM is the projection of holes in THGEM ofupper layers. a) z-coordinate (b) r-coordinateFigure 8: The coordinates distribution of ions production for a single THGEM, where z = 0 is defined as the bottom surface of THGEM, and r = 0 means the center of a holeof THGEM Figure 9: The gain of single THGEM zone, while the effective gain is defined as the number of the electrons thatcan achieve the anode and induce a electronic signal.In order to suppress the IBF by using the upper layers of THGEM, thegain of the upper layers of THGEM should not be too large so that theions are mainly produced in the bottom THGEM. So the ions produced atTHGEM of upper layers are negligible comparing to that at bottom one, andit’s feasible only simulate the ions produced at the bottom THGEM.Commonly-used electronics can generally read out the charge of the 0 . ∼
10 pC magnitude, so the total gain of the multi-THGEM should be 10 ∼ level. On the other hand, the total gain is expected to mainly come from thecontribution of the bottom THGEM so that the upper THGEMs are able12o absorb more ions. Therefore, for triple THGEM structures, we set thegains of each THGEM are 10 , 10 and 10 , while for quadruple THGEMstructures, we would like to set the gains of each THGEM are 10 , 10 , 10 and 10 . According to Fig. 9, the corresponding voltage should be 800 V,1200 V and 1500 V for triple THGEM, while 800 V, 1000 V, 1000 V and 1500 Vfor quadruple THGEM. The hole diameter of top THGEM is set as 0.6 mmand other THGEMs are 0.4 mm. The field of both transfer and inductionzone are set as 2 kV/cm. For a single event, the IBF simulation resultsare summarized as Table 3 for triple THGEM and Table 4 for quadrupleTHGEM. The IBF of more events can be filled in histograms and we usecrystal ball function to fit, an example of ”ACBA” structure is shown as Fig.10. According to the Table 3, we can conclude that the ions absorbed bythe lower two THGEM are similar in both configuration of triple THGEM.The difference is that the top THGEM of ”CBA” configuration can absorbalmost half ions remained of the lower two THGEM, while that of ABAconfiguration can absorb almost none. But the IBF of both configuration islarger 5% level. So we have to consider quadruple THGEM.According to the Table 4, the IBF of configuration ”ACBA” can be re-duced to 0.5%, which is similar to double MicroMegas structure[11, 12]. Asfor the other two configurations, the IBF of both configurations are 3% level,which are too large comparing with the ACBA configuration.Finally, we study the influence of voltage of the bottom THGEM onthe IBF. The voltage of upper layers of THGEMs as 800 V, 1000 V and1000 V, and the simulation result is summarized as the Fig. 11. The IBF canbe reduced as low as 0.2% in the ”ACBA” configuration while the IBF intraditional staggered strategy ”BABA” is 2% level. This means the ”ACBA”structure can reduce the IBF by a factor of 10 − .
5. Conclusion
In this paper, we propose to use large-hole-diameter THGEM to improvethe transparency of electrons and to use staggered multi-THGEM structureto suppress the IBF. The simulation by Garfield++ shows that the trans-parency rate can reach more than 90% in Ar-iC H and the IBF can besuppressed to 0.2% level in ”ACBA” configuration of quadruple THGEM.It’s remarkable that our algorithm to simulate the IBF is that firstlystudy the avalanche procedure in single THGEM to obtain the probability13 able 3: The Ions of each THGEM in triple-THGEM structures Configuration ABA CBAIBF 8.15% 4.17%THGEMtop 0% 3.98%THGEMmid 44.05% 44.15%THGEMbottom 47.80% 47.70%
Table 4: The Ions of each THGEM in quadruple-THGEM structures
Configuration BABA CABA ACBAIBF 3.24% 3.31% 0.46%THGEM1 5.82% 5.03% 1.85%THGEM2 4.69% 4.38% 10.41%THGEM3 27.49% 27.44% 27.47%THGEM4 59.87% 59.84% 59.83%
Figure 10: The IBF histogram of ”ACBA” structure with 800-1000-1000-1500V igure 11: The IBF ratio under different voltages of the bottom THGEM density function(PDF) of the ion production position, then drop ions basedon this PDF. Comparing to uniform distribution of initial ions, this algorithmcontains more information of the avalanche procedure, although it needs morepowerful computing source. Further, if the computing power is even better,one can attempt to do a whole simulation of multi-THGEMs, i.e. one candirectly study the avalanche procedure of the primary electrons and the driftprocedure of ions in multi-THGEM structures.In a large-scale experiment based on TPC, beyond indicators like electrontransparency and IBF, other engineering issues including mechanical stabil-ity, service life and large-size manufacturing should also be considered, andthese are the advantages of THGEM. Further, in recent years, there are somenovel deformations of THGEM such as THCOBRA[29] and M-THGEM[30]which can suppress the IBF effectively. Therefore, it is worth attemptingto use THGEM, THCOBRA and M-THGEM with multi-layer strategies infuture TPC. Acknowledgement
This work was supported by the National Natural Science Foundationof China (Grant Nos. 11575193, U1732266, U1731239, 12027803), Key Re-search Program of Frontier Sciences, CAS, Grant No. QYZDB-SSW-SLH039.