Spatial resolution optimization in a THGEM-based UV photon detector

Original Paper



THick Gas Electron Multiplier (THGEM) is considered in many UV photon detector applications. It has the capability of detecting single photon and imaging with high sensitivity. Operating parameters such as choice of gas mixture, pressure, drift field, drift gap, multiplication voltage, induction field and induction gap play an important role in deciding the spatial resolution of the detector. Detailed simulation study enables to optimize the above-mentioned parameters for a given THGEM-based imaging detector and hence to achieve improved performance for the same.

Materials and methods

Simulation, using ANSYS and Garfield++, starts with the release of primary electrons at random coordinates on the photocathode plane. They are tracked as they pass through the drift gap and THGEM hole till the electron cloud reaches anode plane. Distribution of electron cloud on the anode plane along X and Y axis is plotted in histogram and fitted with Gaussian function to determine spatial resolution. Ar/CO2 (70:30) mixture, which shows higher ETE and lower transverse diffusion, is chosen for this simulation study.


Transverse diffusion has a major impact on both ETE and the spatial resolution. Lower transverse diffusion coefficient is always desired for having better resolution as well as for ETE. It is found from the simulation study that higher gas pressure, lower drift field and induction field, smaller drift and induction gap can provide optimum detection efficiency with the best spatial resolution. The simulation method proposed here can also be extended to X-ray imaging detectors.


Thick gas electron multiplier (THGEM) UV photon detectors Electron transfer efficiency Spatial resolution simulation Garfield++ 


Position-sensitive UV photon detectors play a key role in modern research like detection of particle in high-energy physics, sky observation in astronomy and positron emission tomography in medical diagnosis. Large area and high efficiency are the major requirements in these above-mentioned fields. Thick gas electron multiplier (THGEM), a novel development within the broad family of micropattern gas detectors, is considered in many of these applications [1, 2, 3, 4]. THGEM is a thin copper cladded printed circuit board with mechanically drilled holes and chemically etched rim around the holes. THGEM-based photon detectors assure high gain with single-photon detection sensitivity. The operation mechanism of these detectors is described in detail in [2]. Due to robustness and easy production of THGEM compared to traditional GEM [5], it can be used in many applications wherever spatial resolution in the sub mm range is acceptable. The capability of position sensing with very good efficiency for a single photon makes them very attractive in detection and imaging of UV photons and also recently visible photons [6]. THGEM with reflective (Ref) or semi-transparent (ST) photocathode configuration is considered in many applications such as in RICH detectors [7, 8], gaseous photomultipliers (GPM) [9, 10]. These applications require higher sensitivity with maximum detection efficiency. For maximizing the detection efficiency of the detector, electron transfer efficiency (ETE) should be higher. The ratio of total number of electrons focused into the THGEM hole undergoing multiplication to total number of electrons produced at photocathode surface is termed as ETE. The ETE in a Ref configuration depends on multiplication voltage and hole size [11]. In a ST configuration, ETE depends on geometrical parameters of THGEM like hole diameter, pitch and also drift parameters. Drift parameters can also play role in other performance parameters like spatial resolution in imaging applications. The schematic of THGEM UV imaging detector based on ST photocathode configuration, which is of our interest in this study, is shown in Fig. 1.
Fig. 1

Schematic of THGEM-based UV photon detector operating in ST mode

There are few reported works on spatial resolution measurement and also on its simulations [12, 13, 14, 15, 16, 17]. Most of them deal with traditional GEM but not THGEM. In a GEM/THGEM-based position-sensitive detector, spatial resolution depends on the geometrical parameters like hole diameter, pitch and readout method [12, 13, 14]. Pixelated readout coupled with centre-of-gravity calculation allows in obtaining resolution better than the pixel width. However, the final resolution or position uncertainty is determined by the intrinsic electron diffusion from its creation point. Once the geometrical parameter is fixed, operating parameters play further role in deciding the final resolution. The aim of this work is to study the parameters that affect the spatial resolution. We fixed the THGEM geometrical parameters and simulations carried out to evaluate the dependence of spatial resolution on operating parameters.


Modelling of a THGEM unit cell is done using ANSYS [18], a software tool based on finite element method. Geometrical parameters of THGEM chosen in this study are: insulator thickness 250 µm, hole diameter 200 µm and pitch 450 µm, similar to the THGEM described in earlier work [19]. Drift and induction gaps maintained are 3 and 2 mm, respectively, unless otherwise specified. Electric field values are calculated at various coordinates, and these field map files are imported to Garfield++ [20] for further simulation.

In Garfield++, simulation starts with release of primary photoelectrons at photocathode plane (drift electrode plane) with a given energy (0.1 eV). Initial coordinates of the photoelectrons are chosen randomly within a specified area on drift electrode plane to simulate realistic situation. Each electron is drifted towards THGEM and takes part in avalanche after reaching high field region near the centre of THGEM hole. Figure 2 shows the maximum electric field value (Emax) across the THGEM hole and effective gain values for different ΔVTHGEM. Effective gain values are estimated by estimating the total number of electrons reaching the anode plane. Penning transfer rate of 0.57 is considered to obtain more realistic value of gain [21].Multiplication voltage (ΔVTHGEM) of THGEM was intentionally kept low in the subsequent simulations as simulation with high gain takes huge amount of CPU time. Figure 3a shows drift of a single photoelectron from photocathode plane. Coordinates of each electron are recorded as it arrives at anode plane.
Fig. 2

Maximum electric field (Emax) produced at the centre of the hole and effective gain as a function of ΔVTHGEM for Ar/CO2 (70/30) gas mixture at 760 torr

Fig. 3

a Development of an avalanche, b 2D, c 3D view of electron distribution at anode plane and d histogram with Gaussian fit for Ar/CO2 (70/30) gas mixture at 760 torr, ΔVTHGEM = 1200 V, drift field = 0.67 kV/cm, induction field = 3 kV/cm

Distribution of electron cloud is plotted as histogram along x and y directions. Figure 3b and c presents 2D and 3D view of the electron distribution, respectively. The histogram is fitted with a Gaussian function as shown in Fig. 3d. RMS width of the fitted histogram gives the spatial resolution (sigma in µm) [22].

Results and discussion

Simulations are carried out for various drift and induction parameters which affect electron transport properties such as diffusion and hence spatial resolution. Gas mixture, pressure, drift field, drift gap multiplication voltage, induction field and induction gap are varied to study their effect on detector performance.

Gas mixture

Simulation is carried out for different gas mixtures: Ar/CO2 (70/30), Ar/CO2 (80/20), Ar/CO2 (90/10), Ne/CH4 (90/10), Ne/CF4 (95/5), Ne/CO2 (90/10) and Ar/CH4 (90/10), which are commonly used in UV photon detector. Figure 4a shows variation in spatial resolution for different gas mixtures at a pressure of 760 torr. It is observed that spatial resolution is best for Ar/CO2 (70/30). This difference in spatial resolution can be explained in terms of transverse diffusion coefficient (DT) as calculated by Magboltz and shown in the same Fig. 4a. The spatial resolution improves for gas mixtures having lower DT. In addition to spatial resolution, ETE is also estimated and plotted in Fig. 4b for Ar/CO2 (70/30) and Ar/CH4 (90/10) gas mixtures as ETE decides detection efficiency of the ST-based detector system. Higher ETE enhances the probability of electrons, reaching the THGEM holes. Ar/CO2 (70/30) shows higher ETE in comparison with other gas mixtures due to its lower value of DT [23]. ETE variation with drift field is influenced by the transverse diffusion as discussed in the later “Drift field (ED)” section. Gas gain is another factor while choosing the gas mixtures. Gain is higher for relatively lower multiplying voltages in neon-based gas mixtures. However, higher gain can be obtained with single THGEM with Ar/CO2 (70/30) gas mixture at higher operating voltages as discussed in [7], which reported a gain of above 5 × 104. Hence, the work is focussed mainly on spatial resolution in Ar/CO2 (70/30), which is of interest in UV imaging applications.
Fig. 4

a Effect of gas mixture on spatial resolution and b ETE for different gas mixtures as a function of drift field (ED)

Gas pressure

The effect of gas pressure on spatial resolution is shown in Fig. 5. It is observed that the spatial resolution improves with increased pressure. DT plotted with pressure in the same figure shows that the reduction in DT with increased gas pressure enhances spatial resolution. However, higher pressure lowers the gas gain of a detector. Thus, gas pressure must be chosen to satisfy both the gain and spatial resolution requirements depending on the application.
Fig. 5

Spatial resolution and transverse diffusion coefficient (DT) as a function of gas pressure

Drift field (E D)

An important parameter of UV photon detector operating in ST mode is drift field applied between photocathode and THGEM top electrode. Drift field has strong influence on electron diffusion. Hence, it affects both spatial resolution and ETE as shown in Fig. 6a and b, respectively. A plot of DT is also presented in Fig. 6b. Drift field value of 0.7 kV/cm showing maximum ETE and minimum diffusion is optimum for best spatial resolution and this changes as the field is increased or decreased. Drift field value in this range has also the advantage of reduced ion-feedback and hence ageing of CsI material [24], which is an important aspect in UV detector [7, 25].
Fig. 6

a Spatial resolution (in µm) with drift field and b ETE as a function of drift field

Drift gap

Drift gap is another parameter that influences spatial resolution of UV photon detector based on ST photocathode. In this simulation, the drift gap is varied from 0.1 to 12 mm keeping drift field constant. 3D histograms for different drift gaps are plotted in Fig. 7a to see the spreading at anode plane. The electrons spread out more at 10-mm drift gap compared to 0.5- or 5-mm drift gap. Width of the electron cloud (σ) at the anode plane increases with the drift gap according to the relation \( \sigma \propto \sqrt d \) [26], where d is the drift distance. Figure 7b shows the fitted data obtained from the simulation. As the drift gap increases, the width of the electron cloud also increases, thus degrading the spatial resolution. However, drift gap below 1 mm shows poor ETE due to less focusing ability as described in previous work [19].
Fig. 7

a 3D histograms for 0.5-, 5- and 10-mm drift gaps and b effect of drift gap on spatial resolution

Multiplication voltage (ΔV THGEM)

THGEM multiplication voltage (ΔVTHGEM) decides the overall gain of a detector. Choice of voltage is mainly based on the required gain for operation. ΔVTHGEM is varied from 800 to 1200 V. Effective gain of 10 to 104 is obtained in this voltage range as shown in Fig. 2. Spatial resolution estimated by simulations for this range is presented in Fig. 8a. We can observe that the sigma value starts increasing after 1000 V though the increment is marginal. This can be explained from electric field configuration of THGEM. As ΔVTHGEM increases, maximum electric field (Emax) at the centre of the hole increases. In addition, threshold field, minimum electric field required for multiplication, extends outside the hole with increasing ΔVTHGEM. Threshold field for Ar/CO2 (70/30) is about 12 kV/cm at 760 torr. Figure 8b shows the electric field distribution through the THGEM hole for different ΔVTHGEM. It is seen that for ΔVTHGEM of 800 V, the threshold electric field extends around 20 µm towards both the induction and drift region from the hole. Field extension increases linearly with ΔVTHGEM. For ΔVTHGEM = 1200 V, field extends up to 70 µm from the hole in both sides. The effect of field extension is noticeable after ΔVTHGEM = 1000 V. Spreading of electrons increases at higher ΔVTHGEM due to avalanche confinement spilling outside the THGEM hole and results degradation in spatial resolution.
Fig. 8

a Spatial resolution for different multiplication voltages, b electric field distribution as a function of ΔVTHGEM

Induction field (E I)

Induction field is responsible for efficient collection of electrons at anode plane. Higher the induction field better is the collection. Simulations are carried out from 1 to 7 kV/cm to investigate the effect of induction field on spatial resolution, and the results are presented in Fig. 9a. DT is also plotted in the same figure. Spatial resolution follows the behaviour of DT at low induction field values. This is similar to the results seen for drift field above 1 kV/cm. Above 5 kV/cm, spatial resolution shows an improvement with increasing induction field though diffusion is almost constant. To understand this behaviour, drift lines are simulated by Runge–Kutta–Fehlberg (RKF) integrator (Garfield::DriftLineRKF) [20] for three different induction fields (1, 3 and 6 kV/cm) and are shown in Fig. 9b. It can be seen that at higher induction field electron drift lines show reduced spread in the induction region with increasing induction field. This effect contributes to the improvement of the spatial resolution above 5 kV/cm. However, a very high induction field may extend the avalanche out of the hole and lead to early discharges during operation.
Fig. 9

a Spatial resolution with induction field, b drift lines simulated in Runge–Kutta for 1, 3 and 6 kV/cm induction fields

Induction gap

Figure 10 shows the variation of spatial resolution with induction gap keeping induction field constant. Here, spreading increases with the increase in induction gap similar to drift gap. So it is always advantageous to keep anode plane closer to the THGEM to minimize spreading of electrons at anode plane. The optimum gap is to be chosen for a reliable operation with high voltages involved.
Fig. 10

Effect of induction gap on spatial resolution


The simulation tools are used to understand various operating parameters that affect the THGEM performance as UV imaging detector. RMS spread of electron cloud (sigma), representation of spatial resolution, is studied in detail. Optimization is required to have best possible spatial resolution with maximum detection efficiency for any high sensitivity application. The effect of operating parameters such as gas mixture, pressure, drift field, drift gap, multiplication voltage, induction field and induction gap on spatial resolution have been simulated. ETE data are compared with spatial resolution as ETE is very important factor in maximizing the detection efficiency in a THGEM-based photon detector with semi-transparent photocathode configuration.

Considering higher ETE and lower transverse diffusion coefficient, Ar/CO2 (70/30) gas mixture is chosen for the present simulation studies. Transverse diffusion coefficient has major impact on both ETE and the spatial resolution. Lower transverse diffusion coefficient is always desired for having better resolution as well as for ETE. Therefore, higher gas pressure, lower drift field, higher induction field, smaller drift and induction gap with sufficient gain may provide optimum detection efficiency of a single photon with best spatial resolution. Effects of multiplication voltage and induction field on spatial resolution are not significantly high, and these voltages can be chosen as per operating gain and collection efficiency requirements, respectively. The method proposed here, can be extended to THGEM-based X-ray detectors also. In X-ray detectors, primary electron cloud, which is energy dependent, has to be considered for simulations instead of single electron.


  1. 1.
    F. Sauli, Nucl. Instrum. Methods Phys. Res. A 477(1), 1–7 (2002)ADSCrossRefGoogle Scholar
  2. 2.
    R. Chechik et al., Nucl. Instrum. Methods Phys. Res. A 535(1), 303–308 (2004)ADSCrossRefGoogle Scholar
  3. 3.
    A.F. Buzulutskov, Instrum. Exper. Tech. 50(3), 287–310 (2007)CrossRefGoogle Scholar
  4. 4.
    M. Cortesi et al., J. Instrum. 8(10), C10009 (2013)CrossRefGoogle Scholar
  5. 5.
    F. Sauli, Nucl. Instrum. Methods Phys. Res. A 386(2–3), 531–534 (1997)ADSCrossRefGoogle Scholar
  6. 6.
    F. Tessarotto, Evolution and recent developments of the gaseous photon detectors technologies. In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment (2017).
  7. 7.
    R. Chechik et al., Nucl. Instrum. Methods Phys. Res. A 553(1), 35–40 (2005)ADSCrossRefGoogle Scholar
  8. 8.
    M. Alexeev et al., Nucl. Instrum. Methods Phys. Res. A 732, 264–268 (2013)ADSCrossRefGoogle Scholar
  9. 9.
    R. Chechik et al., Nucl. Instrum. Methods Phys. Res. A 595(1), 116–127 (2008)ADSCrossRefGoogle Scholar
  10. 10.
    A. Di Mauro, Nucl. Instrum. Methods Phys. Res. A 766, 126–132 (2014)ADSCrossRefGoogle Scholar
  11. 11.
    C. Shalem et al., Nucl. Instrum. Methods Phys. Res. A 558(2), 475–489 (2006)ADSCrossRefGoogle Scholar
  12. 12.
    J.M. Maia et al., Nucl. Instrum. Methods Phys. Res. A 580(1), 373–376 (2007)ADSMathSciNetCrossRefGoogle Scholar
  13. 13.
    A. Bamberger et al., Nucl. Instrum. Methods Phys. Res. A 572(1), 157–159 (2007)ADSCrossRefGoogle Scholar
  14. 14.
    M. Cortesi et al., J. Instrum. 2(09), P09002 (2007)CrossRefGoogle Scholar
  15. 15.
    V.N. Kudryavtsev et al., Nucl. Instrum. Methods Phys. Res. A 845, 289–292 (2017)ADSCrossRefGoogle Scholar
  16. 16.
    Y. Lan-Lan et al., arXiv preprint arXiv:1306.0270 (2013)
  17. 17.
    H. Zhao et al., Radiat. Detect. Technol Methods 1(1), 6 (2017)CrossRefGoogle Scholar
  18. 18.
  19. 19.
    G. Baishali et al., Nucl. Instrum. Methods Phys. Res. A 729, 51–57 (2013)ADSCrossRefGoogle Scholar
  20. 20.
  21. 21.
    C.D.R. Azevedo et al., J. Instrum. 11(08), P08018 (2016)CrossRefGoogle Scholar
  22. 22.
    L. Moleri et al., On the localization properties of an RPWELL gas-avalanche detector. J Instrum 12(10), P10017 (2017)CrossRefGoogle Scholar
  23. 23.
    Y. Assran, A. Sharma, arXiv preprint arXiv:1110.6761 (2011)
  24. 24.
    A. Breskin, Nucl. Instrum. Methods Phys. Res. A 367(1-3), 326–331 (1995)ADSCrossRefGoogle Scholar
  25. 25.
    A. Breskin et al., Nucl. Instrum. Methods Phys. Res. A 478(1–2), 225–229 (2002)ADSCrossRefGoogle Scholar
  26. 26.
    C. Shalem, M.Sc. thesis, Weizmann Institute of Science.

Copyright information

© Institute of High Energy Physics, Chinese Academy of Sciences; Nuclear Electronics and Nuclear Detection Society and Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  • P. Ray
    • 1
  • G. Baishali
    • 2
  • V. Radhakrishna
    • 3
  • K. Rajanna
    • 1
  1. 1.Department of Instrumentation and Applied PhysicsIndian Institute of ScienceBangaloreIndia
  2. 2.Department of PhysicsDayananda Sagar UniversityBangaloreIndia
  3. 3.Space Astronomy GroupISRO Satellite CentreBangaloreIndia

Personalised recommendations