- Email: [email protected]
Development of a glass GEM
Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima
Development of a glass GEM Hiroyuki Takahashi a,n, Yuki Mitsuya a, Takeshi Fujiwara b, Takashi Fushie c a b c
Department of Nuclear Engineering and Management, The University of Tokyo, Japan Nuclear Professional School, The University of Tokyo, Japan Hoya Corporation, Japan
art ic l e i nf o
a b s t r a c t
Article history: Received 5 June 2012 Received in revised form 25 April 2013 Accepted 29 April 2013 Available online 16 May 2013
Gas electron multipliers (GEMs) apply the concept of gas amplification inside many tiny holes, realizing robust and high-gain proportional counters. However, the polyimide substrate of GEMs prevents them from being used in sealed detector applications. We have fabricated and tested glass GEMs (G-GEMs) with substrates made of photosensitive glass material from the Hoya Corporation. We fabricated G-GEMs with several different hole diameters and thicknesses and successfully operated test G-GEMs with a 100 100 mm2 effective area. The uniformity of our G-GEMs was good, and the energy resolution for 5.9 keV X-rays was 18.8% under uniform irradiation of the entire effective area. A gas gain by the G-GEMs of up to 6700 was confirmed with a gas mixture of Ar (70%)+CH4 (30%). X-ray imaging using the charge division readout method was demonstrated. & 2013 Elsevier B.V. All rights reserved.
Keywords: Proportional counter X-ray GEM MPGD Position-sensitive detector
1. Introduction Gaseous radiation detectors are used in various applications. Since the gas electron multiplier (GEM) was introduced [1], its multi-stage high-gain capability and robustness have attracted many researchers, and many efforts have been made to develop the fabrication methods [2–4]. Although the double GEM or triple GEM structure can achieve a very high total gas gain, a higher gas gain in a single GEM is preferable to prevent unwanted discharge and reduce the number of stages for simplicity. Therefore, thick alternative substrates such as printed circuit boards have been investigated [5]. In this study, we developed new glass GEMs (GGEMs) with substrates made of a photo-etchable glass called PEG3 from the Hoya Corporation. The fabrication process is shown in Fig. 1. The photo-etchable glass used for the substrate in this study is cerium–silver-doped lithium–aluminum-silicate glass. Under illumination with UV light, cerium is oxidized and releases one electron. In the first annealing step, the released electrons reduce the positively charged doped metal ions (Au, Ag, Cu). The neutral metal atoms have high mobility, and they move to form metal clusters within the substrate. In the second annealing step, at a higher temperature, these metal clusters act as seed crystals for crystallizing the glass around them. The crystal can then be easily
n Correspondence to: Department of Nuclear Engineering and Management, Faculty of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan. Tel.: +81 3 5841 7007; fax: +81 3 3818 3455. E-mail address: [email protected] (H. Takahashi).
0168-9002/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nima.2013.04.089
etched with hydrogen fluoride to form holes. Finally, a Cr/Cu layer is sputtered as an electrode for the G-GEM. The use of glass material produces a self-standing, outgas-free substrate. The outgassing problem is particularly important for sealed counters such as those in neutron detectors and gas photomultipliers. Another favorable feature of photosensitive glass is its compatibility with photolithography techniques. Thus, we can obtain a very fine electrode pattern on the glass material. In fact, we have made patterns that form guard rings to suppress the effective capacitance around the holes; however, in the present study, we focus on a normal structure. Table 1 summarizes the characteristics of the glass substrates. We selected PEG3 because it shows a rather low volume resistivity, which is favorable for removing the surface charge. Another potential advantage of this material is its opaqueness; therefore, we could realize a see-through structure if indium tin oxide (ITO) electrodes are used with this glass. 2. Design and fabrication of G-GEMs Because the fabrication process of the photosensitive G-GEM substrate relies entirely on the photolithography technique, we can define a very fine structure for the holes. However, we selected a hole diameter of 120–170 μm and a hole pitch of 360 μm as our standard dimensions considering the yield and requirements for neutron detection applications. To prevent metal spilling into holes, we filled the holes with resist before applying the metal layers. After coating the substrate with metal, we removed the resist.
2
H. Takahashi et al. / Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
Fig. 1. Fabrication process of G-GEMs. Table 1 Charactersitics of substrate materials.
Thermal conductivity (W/mK) Young's modulus (Gpa) Relative permittivity Volume resistivity (Ω cm)
PEG3
PEG3C
Polyimide
0.795 79.7 6.28 8.5 1012
2.72 90.3 5.26 4.5 1014
0.3 18.6 3.55 1018
The effective area of our G-GEM is set to 100 100 mm2. However, the maximum size of the glass is currently limited by the size of the ingots of the glass material, which we hope can be extended to 300 300 mm2. The yield of G-GEM fabrication has been continuously improved; it is currently about 80–90%. Fig. 2 shows a photograph of the G-GEM, and Table 2 lists the characteristics of the five G-GEMs that we fabricated and tested. The resistances are measured across the G-GEM substrates and greatly affected by the thicknesses of G-GEMs. For example resistances of G-GEM 1 and G-GEM 2 are almost identical because their thickness is same. However, the thickness of G-GEM3 is smaller than that of G-GEM 1, therefore, G-GEM3 provides lower resistance compared with G-GEM1. We tested the abovementioned G-GEMs with Ar (70%)+CH4 (30%) at atmospheric pressure in a gas flow mode. Fig. 3 shows the experimental setup. Our custom-designed CMOS ASIC preamps [equivalent noise charge: 880 electrons, full width at halfmaximum (FWHM)] were used. Pulse signals were obtained from a single planar anode as a function of the voltage across the GEM layer. The pulse height was measured using a general-purpose shaping amplifier and multi-channel analyzer (MPA3 from FastComtec). An 55 Fe source was used for the gas gain measurement. The drift field was set to 42 V/mm, and the induction field was set to 250 V/mm for G-GEMs 1–3 and 500 V/mm for G-GEMs 4 and 5. Fig. 4 shows the measured gas gains for these G-GEMs. Although G-GEMs 3 and 4 had the same dimensions, G-GEM 4 showed a higher gas gain. The highest gas gain of 6700 was achieved by G-GEM 5. This is attributed to the lower voltage across the G-GEM substrate needed to realize a higher electric field thanks to its thickness. Generally speaking, lower
Gas Gain
10000
Fig. 2. Photograph of a G-GEM.
1000
Table 2 Five different G-GEMs. Number Metal
#1 #2 #3 #4 #5
Copper Copper Copper Chromium Chromium
Thickness (μm)
Hole size Resistance between the (μm) front and the rear surfaces (MΩ)
700 700 580 580 420
170 140 120 120 120
280 315 180 60 50
100
1300
1500
1700 1900 2100 Applied Voltage (V)
No.2
No.1 No.4
No.3 No.5
Fig. 4. Gas gain and applied voltage.
5mm
2mm
Fig. 3. Experimental setup.
2300
H. Takahashi et al. / Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
1900V
1950V 2000V
Fig. 5. Energy spectra for an 55Fe source obtained with G-GEM 1 (uniform irradiation of entire effective area, induction field: 250 V/mm).
2050V
2100V
2150V 2190V
Fig. 6. Energy spectra for an 55Fe source obtained with G-GEM 2 (uniform irradiation of entire effective area, induction field: 250 V/mm).
3
voltage operation is favorable. If the GEM shows a sufficient gas gain at a lower applied voltage, then, the electric field around the edge of the GEM holes can be lower, which is in turn favorable for highergain operation. Assuming the same maximum tolerable voltage across the G-GEM substrate, a smaller voltage across the substrate is generally preferable. If gas amplification starts at a smaller voltage, we have considerable room to increase the voltage and achieve a higher gas gain. The gas gain of G-GEM 2 was the lowest among the tested prototypes. Because its hole size is small, and the thickness of the substrate and the aspect ratio are large; some charge collection loss might have degraded the gain. Figs. 5–7 show the energy spectra of G-GEMs 1, 2, and 4, correspondingly. These spectra were obtained with uniform irradiation from an 55Fe source. Assuming the same process tolerence, the relative tolerance of a structure depends on the size of the structure itself. Therefore, a larger hole size yields better tolerance, which in turn contributes to smaller relative variation in the hole sizes. G-GEM 1, which has the largest hole diameter, exhibited the best FWHM energy resolution, 18.8%. This is considered to be attributable to the good uniformity of the hole size over the entire effective area. For demonstration purposes, we took X-ray transmission images of a 0.5-mm-thick copper plate with holes using a charge division method. We placed a thin double-layer ITO-coated glass plate behind the G-GEM and obtained an X-ray image (see Fig. 8) by irradiating the detector with the 55Fe source. The measurement time was about 30 min owing to the slow data transfer rate of our measurement system, which uses a general-purpose 4-ch waveform digitizer (Agilent L4534A) to obtain the waveform data. The gas gain was 3000 in this measurement. We successfully obtained an X-ray image with our G-GEM.
3. Conclusions 1800V
1800V 1900V
Fig. 7. Energy spectra for an 55Fe source obtained with G-GEM 4 (uniform irradiation of entire effective area, induction field: 250 V/mm).
We designed and fabricated G-GEMs and tested their basic performance. The solid substrate of a G-GEM was easy to handle even without any support frame for tested plates. We successfully operated prototype detectors with effective areas of 100 100 mm2. The highest gas gain of 6700 was achieved by G-GEM 5. The best energy resolution of G-GEM1 for an 55Fe source was 18.8% under uniform irradiation. Note that these two results were obtained from two G-GEMs with very different materials and geometries. G-GEMs based on a photo-etchable glass technique have promise as micropattern gaseous detectors that yield high gain and good energy resolution with a single plate. A demonstration X-ray image was obtained successfully.
Fig. 8. Optical and X-ray transmission images “UT”.
4
H. Takahashi et al. / Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
References [1] F. Sauli, Nuclear Instruments and Methods A386 (1997) 531. [2] B. Ketzer, et al., Nuclear Instruments and Methods A535 (2004) 314.
[3] A. Sharma, Nuclear Instruments and Methods A666 (2012) 98. [4] H. Ohshita, S. Uno, T. Otomo, et al., Nuclear Instruments and Methods A623 (1) (2010) 126. [5] R. Chechik, A. Breskin, C. Shalem, D. Mörmann, Nuclear Instruments and Methods A535 (2004) 303.
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research A journal homepage: www.elsevier.com/locate/nima
Development of a glass GEM Hiroyuki Takahashi a,n, Yuki Mitsuya a, Takeshi Fujiwara b, Takashi Fushie c a b c
Department of Nuclear Engineering and Management, The University of Tokyo, Japan Nuclear Professional School, The University of Tokyo, Japan Hoya Corporation, Japan
art ic l e i nf o
a b s t r a c t
Article history: Received 5 June 2012 Received in revised form 25 April 2013 Accepted 29 April 2013 Available online 16 May 2013
Gas electron multipliers (GEMs) apply the concept of gas amplification inside many tiny holes, realizing robust and high-gain proportional counters. However, the polyimide substrate of GEMs prevents them from being used in sealed detector applications. We have fabricated and tested glass GEMs (G-GEMs) with substrates made of photosensitive glass material from the Hoya Corporation. We fabricated G-GEMs with several different hole diameters and thicknesses and successfully operated test G-GEMs with a 100 100 mm2 effective area. The uniformity of our G-GEMs was good, and the energy resolution for 5.9 keV X-rays was 18.8% under uniform irradiation of the entire effective area. A gas gain by the G-GEMs of up to 6700 was confirmed with a gas mixture of Ar (70%)+CH4 (30%). X-ray imaging using the charge division readout method was demonstrated. & 2013 Elsevier B.V. All rights reserved.
Keywords: Proportional counter X-ray GEM MPGD Position-sensitive detector
1. Introduction Gaseous radiation detectors are used in various applications. Since the gas electron multiplier (GEM) was introduced [1], its multi-stage high-gain capability and robustness have attracted many researchers, and many efforts have been made to develop the fabrication methods [2–4]. Although the double GEM or triple GEM structure can achieve a very high total gas gain, a higher gas gain in a single GEM is preferable to prevent unwanted discharge and reduce the number of stages for simplicity. Therefore, thick alternative substrates such as printed circuit boards have been investigated [5]. In this study, we developed new glass GEMs (GGEMs) with substrates made of a photo-etchable glass called PEG3 from the Hoya Corporation. The fabrication process is shown in Fig. 1. The photo-etchable glass used for the substrate in this study is cerium–silver-doped lithium–aluminum-silicate glass. Under illumination with UV light, cerium is oxidized and releases one electron. In the first annealing step, the released electrons reduce the positively charged doped metal ions (Au, Ag, Cu). The neutral metal atoms have high mobility, and they move to form metal clusters within the substrate. In the second annealing step, at a higher temperature, these metal clusters act as seed crystals for crystallizing the glass around them. The crystal can then be easily
n Correspondence to: Department of Nuclear Engineering and Management, Faculty of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan. Tel.: +81 3 5841 7007; fax: +81 3 3818 3455. E-mail address: [email protected] (H. Takahashi).
0168-9002/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nima.2013.04.089
etched with hydrogen fluoride to form holes. Finally, a Cr/Cu layer is sputtered as an electrode for the G-GEM. The use of glass material produces a self-standing, outgas-free substrate. The outgassing problem is particularly important for sealed counters such as those in neutron detectors and gas photomultipliers. Another favorable feature of photosensitive glass is its compatibility with photolithography techniques. Thus, we can obtain a very fine electrode pattern on the glass material. In fact, we have made patterns that form guard rings to suppress the effective capacitance around the holes; however, in the present study, we focus on a normal structure. Table 1 summarizes the characteristics of the glass substrates. We selected PEG3 because it shows a rather low volume resistivity, which is favorable for removing the surface charge. Another potential advantage of this material is its opaqueness; therefore, we could realize a see-through structure if indium tin oxide (ITO) electrodes are used with this glass. 2. Design and fabrication of G-GEMs Because the fabrication process of the photosensitive G-GEM substrate relies entirely on the photolithography technique, we can define a very fine structure for the holes. However, we selected a hole diameter of 120–170 μm and a hole pitch of 360 μm as our standard dimensions considering the yield and requirements for neutron detection applications. To prevent metal spilling into holes, we filled the holes with resist before applying the metal layers. After coating the substrate with metal, we removed the resist.
2
H. Takahashi et al. / Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
Fig. 1. Fabrication process of G-GEMs. Table 1 Charactersitics of substrate materials.
Thermal conductivity (W/mK) Young's modulus (Gpa) Relative permittivity Volume resistivity (Ω cm)
PEG3
PEG3C
Polyimide
0.795 79.7 6.28 8.5 1012
2.72 90.3 5.26 4.5 1014
0.3 18.6 3.55 1018
The effective area of our G-GEM is set to 100 100 mm2. However, the maximum size of the glass is currently limited by the size of the ingots of the glass material, which we hope can be extended to 300 300 mm2. The yield of G-GEM fabrication has been continuously improved; it is currently about 80–90%. Fig. 2 shows a photograph of the G-GEM, and Table 2 lists the characteristics of the five G-GEMs that we fabricated and tested. The resistances are measured across the G-GEM substrates and greatly affected by the thicknesses of G-GEMs. For example resistances of G-GEM 1 and G-GEM 2 are almost identical because their thickness is same. However, the thickness of G-GEM3 is smaller than that of G-GEM 1, therefore, G-GEM3 provides lower resistance compared with G-GEM1. We tested the abovementioned G-GEMs with Ar (70%)+CH4 (30%) at atmospheric pressure in a gas flow mode. Fig. 3 shows the experimental setup. Our custom-designed CMOS ASIC preamps [equivalent noise charge: 880 electrons, full width at halfmaximum (FWHM)] were used. Pulse signals were obtained from a single planar anode as a function of the voltage across the GEM layer. The pulse height was measured using a general-purpose shaping amplifier and multi-channel analyzer (MPA3 from FastComtec). An 55 Fe source was used for the gas gain measurement. The drift field was set to 42 V/mm, and the induction field was set to 250 V/mm for G-GEMs 1–3 and 500 V/mm for G-GEMs 4 and 5. Fig. 4 shows the measured gas gains for these G-GEMs. Although G-GEMs 3 and 4 had the same dimensions, G-GEM 4 showed a higher gas gain. The highest gas gain of 6700 was achieved by G-GEM 5. This is attributed to the lower voltage across the G-GEM substrate needed to realize a higher electric field thanks to its thickness. Generally speaking, lower
Gas Gain
10000
Fig. 2. Photograph of a G-GEM.
1000
Table 2 Five different G-GEMs. Number Metal
#1 #2 #3 #4 #5
Copper Copper Copper Chromium Chromium
Thickness (μm)
Hole size Resistance between the (μm) front and the rear surfaces (MΩ)
700 700 580 580 420
170 140 120 120 120
280 315 180 60 50
100
1300
1500
1700 1900 2100 Applied Voltage (V)
No.2
No.1 No.4
No.3 No.5
Fig. 4. Gas gain and applied voltage.
5mm
2mm
Fig. 3. Experimental setup.
2300
H. Takahashi et al. / Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
1900V
1950V 2000V
Fig. 5. Energy spectra for an 55Fe source obtained with G-GEM 1 (uniform irradiation of entire effective area, induction field: 250 V/mm).
2050V
2100V
2150V 2190V
Fig. 6. Energy spectra for an 55Fe source obtained with G-GEM 2 (uniform irradiation of entire effective area, induction field: 250 V/mm).
3
voltage operation is favorable. If the GEM shows a sufficient gas gain at a lower applied voltage, then, the electric field around the edge of the GEM holes can be lower, which is in turn favorable for highergain operation. Assuming the same maximum tolerable voltage across the G-GEM substrate, a smaller voltage across the substrate is generally preferable. If gas amplification starts at a smaller voltage, we have considerable room to increase the voltage and achieve a higher gas gain. The gas gain of G-GEM 2 was the lowest among the tested prototypes. Because its hole size is small, and the thickness of the substrate and the aspect ratio are large; some charge collection loss might have degraded the gain. Figs. 5–7 show the energy spectra of G-GEMs 1, 2, and 4, correspondingly. These spectra were obtained with uniform irradiation from an 55Fe source. Assuming the same process tolerence, the relative tolerance of a structure depends on the size of the structure itself. Therefore, a larger hole size yields better tolerance, which in turn contributes to smaller relative variation in the hole sizes. G-GEM 1, which has the largest hole diameter, exhibited the best FWHM energy resolution, 18.8%. This is considered to be attributable to the good uniformity of the hole size over the entire effective area. For demonstration purposes, we took X-ray transmission images of a 0.5-mm-thick copper plate with holes using a charge division method. We placed a thin double-layer ITO-coated glass plate behind the G-GEM and obtained an X-ray image (see Fig. 8) by irradiating the detector with the 55Fe source. The measurement time was about 30 min owing to the slow data transfer rate of our measurement system, which uses a general-purpose 4-ch waveform digitizer (Agilent L4534A) to obtain the waveform data. The gas gain was 3000 in this measurement. We successfully obtained an X-ray image with our G-GEM.
3. Conclusions 1800V
1800V 1900V
Fig. 7. Energy spectra for an 55Fe source obtained with G-GEM 4 (uniform irradiation of entire effective area, induction field: 250 V/mm).
We designed and fabricated G-GEMs and tested their basic performance. The solid substrate of a G-GEM was easy to handle even without any support frame for tested plates. We successfully operated prototype detectors with effective areas of 100 100 mm2. The highest gas gain of 6700 was achieved by G-GEM 5. The best energy resolution of G-GEM1 for an 55Fe source was 18.8% under uniform irradiation. Note that these two results were obtained from two G-GEMs with very different materials and geometries. G-GEMs based on a photo-etchable glass technique have promise as micropattern gaseous detectors that yield high gain and good energy resolution with a single plate. A demonstration X-ray image was obtained successfully.
Fig. 8. Optical and X-ray transmission images “UT”.
4
H. Takahashi et al. / Nuclear Instruments and Methods in Physics Research A 724 (2013) 1–4
References [1] F. Sauli, Nuclear Instruments and Methods A386 (1997) 531. [2] B. Ketzer, et al., Nuclear Instruments and Methods A535 (2004) 314.
[3] A. Sharma, Nuclear Instruments and Methods A666 (2012) 98. [4] H. Ohshita, S. Uno, T. Otomo, et al., Nuclear Instruments and Methods A623 (1) (2010) 126. [5] R. Chechik, A. Breskin, C. Shalem, D. Mörmann, Nuclear Instruments and Methods A535 (2004) 303.