- Email: [email protected]
Correlation of the charge collection efficiency of GaAs particle detectors with material properties
Nuclear Instruments and Methods in Physics Research A 380 ( 1996) 46-49
NUCLEAR INSTRUMENTS &METHODS IN PHYS#CS RESEARCH Secrm A
ELSEYIER
Correlation of the charge collection efficiency of GaAs particle detectors with material properties K. Berwick”,
M.R. Brozela, CM. Buttarb, J.S. Poonia, P.J. Sellinb, S.M. Young” “Centre
for Electrvnic ofPhysics,
hDepartment
Materials, University
I/MIST,
P.O. Box 88, Manchester,
of Shefield,
Hounsjeld
M60
Road. Sheffield,
IQD,
UK
S3 7RH. UK
Abstract Measurements of the Charge Collection Efficiency (CCE) of radiation detectors fabricated from semi-insulating GaAs wafers which have different thermal histories have been undertaken. A correlation between charge collection efficiency and electrical resistivity has been observed. These are consistent with a dominant dependence of CCE with the concentration of the ionized deep donor, EL2.‘.
1. Introduction Radiation detectors for room temperature operation based on GaAs have been studied for many years. Devices based on LPE grown material have shown very good results, but it is difficult to produce thick layers of adequate purity to allow sufficient depletion for detection of minimum ionizing particles (mips) or gamma radiation [l]. An alternative that has been studied is the use of Semi-Insulating (SI) GaAs [2] which has the advantage of being readily available in thicknesses of several hundred microns and having a high resistivity (of the order of 10’ fl. cm). The SI nature of GaAs is a result of the compensation of residual acceptor impurities (primarily carbon) by intrinsic deep donor levels introduced as the crystal cools from its growth temperature or following post-growth anneals [3]. One of these electron levels, EL2, is dominant because it occurs in concentrations that far exceed all others. Since it has an ionization energy near the middle of the bandgap, it pins the Fermi level near midgap resulting in extremely high resistivity. Detectors based on these wafers suffer from losing some of the charge signal generated by the incident radiation [2,4]. It is believed that this degradation in performance is partly a result of carrier trapping by some or ail of the intrinsic deep level defects mentioned above, together with effects due to the incomplete penetration of the electric field through the detector [5,6]. In the present paper, we present data obtained from detectors fabricated on wafers that have undergone various heat treatments in an attempt to alter the concentrations of individual traps. We conclude that the degradation due to trapping is dominated by the EL2 defect in its positive charge state.
2. The GaAs samples We have investigated GaAs samples grown by the liquid encapsulated Czochralski (LEC) process and which contain different concentrations of carbon. Several of these samples have undergone different thermal treatments, Table 1 [7]. Such treatments not only result in different concentrations of native deep electronic levels, but also change the electrical resistivity by varying the position of the Fermi energy. The GaAs particle detectors used in this study consist of a Schottky barrier-semiconductor-Ohmic contact sandwich with the thickness of GaAs being approximately 500 km. Concentrations of defects were measured using thermally stimulated current spectroscopy (TSC) [IO]. This analysis requires optical excitation of deep levels and it is necessary to illuminate the semiconductor through the Schottky contact. Fortunately, the structures of samples for TSC defect measurements and particle detectors are identical. A circular semi-transparent Au Schottky contact 28 nm thick was evaporated onto each sample. The back ohmic contact consisted of either annealed Au/Ge or indium. Neither the CCE nor the TSC results depended on the composition or the thickness of the back contact.
3. Defect assessment In TSC the sample is cooled to 80 K and donors are filled with electrons by illumination with near bandedge light. At these low temperatures donors deeper than 0.17 eV remain filled after the illumination is terminated. The sample is then slowly warmed under reverse bias, As the temperature is increased the shallow traps empty first,
016%9002/96/$15.00 Copyright 01996 Elsevier Science B.V. All rights reserved PZI
SOi68-9002(96)00325-7
K. Eerwick
et al. I Nucl. Itrstr. and Me&.
in Ph.vs. Res. A 380 f 19%)
47
X-49
Table I Samples used in this study Sample no.
Supplier
Heat treatments
I
MCP MCP MCP MCP Nippon Mining MCP MCP
Ingot anneal 950°C for 5 h slow cool to room temp. Medium carbon concentration Ingot anneal 950°C for 5 h siow cool to room temp. Low carbon cont. (Seed end wafer) Ingot anneal 950°C for 5 h slow cool to room temp. Ingot anneal 950°C for 12 h slow cool to room temp. Multiple wafer anneal Ingot anneal 950°C for 5 h + multiple wafer anneal t t 130°C and 830°C) As far sample 2. Same ingot, sample wafer from middle
2 3 4 5 6 7
Table 2 Concentrations of deep electron traps for each sample in this study. A dash means not detected Sample no.
0.17 eV
0.23 eV
X10'" cm-'
X10'" cm-'
i
1.5
4.3
2 3 4 5 6 I
370 150 240 880 150 -
310 20 200 10
0.27 eV X 10" cm”’
0.32 eV X IO” cm->
0.38 eV X IO’” cm‘ ’
0.45 eV X IO” cm-’
0.51 ev XlO”cm
1.3
3.7
1.a
1.0 4.0
84 4.4 _ 40
30 9.7 26 19 I2 10
8.2 5 .o _
1.6 2.0 7.3
1.38 2.0
followed by progressively deeper levels. Therefore, peaks in sample conductivity occur in the current versus temperature spectrum, each corresponding to different trap energies. By analysing this spectrum it is possible to extract both the activation energy and the concentrations of the levels present. This technique is limited by the onset of thermal currents so the deepest level measured in this study was 0.51 eV [S]. The defect densities in these samples were very different. either as a result of heat treatments or different growth c(~nditions (Table 2). measurements of the concentrations of the dominant deep donor EL2, [EL21 were obtained from photoquenching the calibrated optical absorption band at a wavelength of I p,rn [9] (Table 2). In addition, the resistivities of these samples were measured using a non-c~)ntacting time dependent charge measurement technique f IO]. Because the resistivity of SI C&As is extremely sensitive to temperature and there appears to be no consensus amongst substrate suppliers about which temperature to use, we have standardized our measurements at 393iO.l K.
100 2x0
7.0 3.0
EL2 '
XlO"cm"
12 Il.5 11 7.5 8.5 12 12
charge to the charge deposited in the detector, assuming that the energy to create an electron-hole pair is 4.3 eV [I 11. The calibration is carried out using a pulser to inject charge into the input of the spectroscopy chain. We have attempted to find a correlation between CCE and the concentration of any observed electron trap. However, none is seen with any observed deep level including EL2. This failure to tind a correlation is consistent with a previous report [ 121. However, we find a good correlation between the CCE (at 200 V) and the resistivity of the sample as shown in Fig. I. The behaviour of CCE as a function of applied voltage for the sample of lowest resistivity (4 X IO’ fi. cm) and a standard LEC substrate is shown in Fig. 2.
4. Experimental Each detector was reverse biased at 200 V and the spectrum from alpha particles irradiating the Schottky contact was taken using a standard spectroscopy chain with a shaping time of 0.5 ps and recorded on a MCA. As a result of Ramo’s theorem, the current pulse in these circumstances is due almost entirely to the motion of electrons. The CCE is defined as the ratio of the measured
OI 0
100 200 300 400 500 600 700 Bins volt8fJnv
Pig. 1. The charge collection efficiency (WE%) plotted against wafer resistivity fat 293 K). for o particles i~diatjng the Schottky contact. The numbers correspond to the samples in Table 1 and Table 2.
I(a). VARIOUS MATERIALS
48
K. Berwick et al. / Nucl. Instr. and Meth. in Phys. Res. A 380 (1996) 46-49
lo.,.,. 0
1
,. 2
I.
I.
1.1,
3 4 5 RoslslWly x10’ ncm
6
8
7
of CCE measured for a particles irradiating the Schottky contact as a function of bias for two samples.
Fig. 2. Comparison
5. Discussion We have shown in previous studies that it is primarily the trapping of electrons that reduces the CCE in those circumstances where effects due to incomplete penetration of the electric field are small [13]. This previous data suggests an electron drift length of around 60 pm. However, the data of Fig. 2 showing a saturated CCE of - 80% is consistent with a drift length of no less than 350 Fm. Trapping of electrons must be due to ionised donors or neutral acceptors having an ionisation energy above the mid-gap Fermi-level, E,. The energy level of the only known neutral acceptor in SI GaAs, the so-called RC defect [14], is degenerate with the conduction band at room temperature and is, therefore, inactive. We, therefore, believe that only ionized donors are available for trapping. We have noted above that there is no apparent correlation between CCE and concentrations of individual deep donor levels in SI GaAs. The only parameter that is seen to correlate with CCE is sample conductivity (Fig. 1, noting that conductivity is the inverse of resistivity). In SI GaAs with constant carrier mobility, such an increase is due to a lifting of the Fermi energy. Raising E, has a further consequence. We note that total concentrations of EL2 defects are relatively constant from sample to sample, even after quite dramatic heat treatments (Table 2). Also, E, can be expressed in terms of the ratio of ionized to unionized EL2, by [EL2+]/[EL2]
= 1 - [(I + l/2 exp (E,,,
- E,)IkT)-‘1
where EEL2 is the energy to ionize neutral EL2 into its positive charge state and k is Boltzmann’s constant. It follows that raising E, reduces the value of [EL2+]/[EL2] and, since [EL21 is nearly constant, it reduces [EL2+]. Thus the increase in CCE associated with a reduction in sample resistivity can be directly related to a concomitant
diminution in [EL2+]. For these reasons we believe that the major trapping centre in these detectors is the EL2+ defect. Trapping by an ionized deep donor like EL2’ is hardly unexpected. However, we are aware that concentrations of EL2’ probably do not exceed lOI5 crnb3 in any of our material, a value that is comparable with concentrations of other ionized donors. Additionally, we note that published electron capture cross sections for EL2’ are of the order of 1O-m’3cm’, a value also comparable with those of other donors [ 151. Therefore, the great effectiveness for electron capture by EL2+ appears to be problematic. However, it has been demonstrated that electron capture by EL2’ is greatly enhanced in the presence of an electric field [16] and, indeed, such behaviour has recently been employed successfully in a model to simulate the penetration of the electric field into SI GaAs 1171. Thus, we propose that EL2’ defects are very efficient electron traps in the electric fields typical in these particle detectors [ 171 and that their concentration is critical in reducing CCE. Trapping of electrons by EL2+ defects is particularly important because de-trapping rates at room temperature are negligible compared to the typical integration time of particle detectors. Unfortunately, the performance of devices fabricated on “low resistivity” material is degraded by the relatively high voltage current. This has also made it impossible to perform electric field penetration measurements which would confirm whether the data of Fig. 2 is to be associated with electric field penetration or the drift length of electrons in a uniform field. We are currently performing measurements to clarify this situation.
6. Conclusions We conclude that EL2’ centres are crucial in the trapping of free electrons in SI GaAs under bias. A similar mechanism has been employed to model the electric field penetration in similar devices [ 171. Reducing the value of the electrical resistivity appears to be a potent way of improving CCE by reducing net concentrations of EL2’ defects.
Acknowledgments The authors are grateful to Miss J. Gilmore for the careful preparation of this paper. The authors acknowledge the support of PARC.
References [I] D. Alexiev and K.S.A. Butcher, Nucl. Instr. and Meth. A 317 (1992)
111.
K. Betwick
et al. / Nucl. Instr.
and Meth.
D.S. McGregor, G.F. Knoll, Y. Eisen and R. Brake. IEEE Trans. Nucl. Sci. NS-39 (1992) 1226. [3] G.M. Martin, Proc. Conf. Semi-Insulating B-V Mats.,
[2]
1980(Shiva, UK, 1980) p. 13. [4] R. Bertin, S. d’Auria. C. Del Papa, F. Fiori, B. Lis0wsksi.V. Oshea. P.G. Pelfer. K. Smith and A. Zichichi, Nucl. Instr. and Meth. A 294 (1990) 21 I. ]S] K. Berwick, M.R. Brozel, C. Buttar, M. Cowperthwaite. P. Sellin and Y. Hou, Inst. Phys. Conf. Ser. 135 (IOP, Bristol, 1994). 16) C. De1 Papa, Proc. Int. GaAs Workshop, Jostal, 1993. [7] J.S. Pooni and M.R. Brozel, Inst. Phys. Conf. Ser. 135 (IOP Bristol. 1994). [S] P. Blood and W. Orton, Electrical characterisation of semiconductors, majority carrier and electron states, Techniques in Physics (.I9923 p.14. [9] G.M. Martin, Appl. Phys. Lett. 39 (1981) 747. IO] R. Stibal. J. Windscheif and W. Jantz, Semicond. Sci. Technol. 6 (1991) 995.
in Php.
Res. A 380 (1996)
4’)
46-49
[I I] W. Bencivelli. R. Bertin, E. Bettolucci. U. Bottigli, S. d’ Auria, C. Del Papa, M.E. Fantacci, P. Randaccio. V. Ross0 and A. Stafanini, Nucl. Ins&. and Meth. A 338 (1994) 549. [12] F. Nava, C. Canah, A. Castaldini, A. Cavallini, S.D. Auria. C. Del Papa. C. Frigeri, L. Zanotti, A. Cetronio. C. Lanzieri and A. Zichichi, Nucl. Instr. and Meth. A 349 (1994) 156. [ 131 P.J. Sellin, C.M. Buttar, S. Manopoulos, K. Berwick, M.R. Brozel and M. Cowperthwaite. IEEE Trans. Nucl. Sci. NS42 f t995). [14] M.R. Brozel and S. Tuzemen. IOP Conf. Ser. No. 13.5 (1994) p. 187. [ 151 Properties of GaAs. 2nd Ed. (INSPEC. 1990) p. 309. [16] V.Ya. Prim and B.A. Bobylev, Sov. Phys. Semicond. 14 (1980) 1097. [ 171 D.S. McGregor. R.A. Rojeski. G.F. Knoll, F.L. Terry, J. East and Y. Eisen, J. Appl. Phys. 75 (1994) 7910.
I(a). VARIOlJS
MATERIALS
NUCLEAR INSTRUMENTS &METHODS IN PHYS#CS RESEARCH Secrm A
ELSEYIER
Correlation of the charge collection efficiency of GaAs particle detectors with material properties K. Berwick”,
M.R. Brozela, CM. Buttarb, J.S. Poonia, P.J. Sellinb, S.M. Young” “Centre
for Electrvnic ofPhysics,
hDepartment
Materials, University
I/MIST,
P.O. Box 88, Manchester,
of Shefield,
Hounsjeld
M60
Road. Sheffield,
IQD,
UK
S3 7RH. UK
Abstract Measurements of the Charge Collection Efficiency (CCE) of radiation detectors fabricated from semi-insulating GaAs wafers which have different thermal histories have been undertaken. A correlation between charge collection efficiency and electrical resistivity has been observed. These are consistent with a dominant dependence of CCE with the concentration of the ionized deep donor, EL2.‘.
1. Introduction Radiation detectors for room temperature operation based on GaAs have been studied for many years. Devices based on LPE grown material have shown very good results, but it is difficult to produce thick layers of adequate purity to allow sufficient depletion for detection of minimum ionizing particles (mips) or gamma radiation [l]. An alternative that has been studied is the use of Semi-Insulating (SI) GaAs [2] which has the advantage of being readily available in thicknesses of several hundred microns and having a high resistivity (of the order of 10’ fl. cm). The SI nature of GaAs is a result of the compensation of residual acceptor impurities (primarily carbon) by intrinsic deep donor levels introduced as the crystal cools from its growth temperature or following post-growth anneals [3]. One of these electron levels, EL2, is dominant because it occurs in concentrations that far exceed all others. Since it has an ionization energy near the middle of the bandgap, it pins the Fermi level near midgap resulting in extremely high resistivity. Detectors based on these wafers suffer from losing some of the charge signal generated by the incident radiation [2,4]. It is believed that this degradation in performance is partly a result of carrier trapping by some or ail of the intrinsic deep level defects mentioned above, together with effects due to the incomplete penetration of the electric field through the detector [5,6]. In the present paper, we present data obtained from detectors fabricated on wafers that have undergone various heat treatments in an attempt to alter the concentrations of individual traps. We conclude that the degradation due to trapping is dominated by the EL2 defect in its positive charge state.
2. The GaAs samples We have investigated GaAs samples grown by the liquid encapsulated Czochralski (LEC) process and which contain different concentrations of carbon. Several of these samples have undergone different thermal treatments, Table 1 [7]. Such treatments not only result in different concentrations of native deep electronic levels, but also change the electrical resistivity by varying the position of the Fermi energy. The GaAs particle detectors used in this study consist of a Schottky barrier-semiconductor-Ohmic contact sandwich with the thickness of GaAs being approximately 500 km. Concentrations of defects were measured using thermally stimulated current spectroscopy (TSC) [IO]. This analysis requires optical excitation of deep levels and it is necessary to illuminate the semiconductor through the Schottky contact. Fortunately, the structures of samples for TSC defect measurements and particle detectors are identical. A circular semi-transparent Au Schottky contact 28 nm thick was evaporated onto each sample. The back ohmic contact consisted of either annealed Au/Ge or indium. Neither the CCE nor the TSC results depended on the composition or the thickness of the back contact.
3. Defect assessment In TSC the sample is cooled to 80 K and donors are filled with electrons by illumination with near bandedge light. At these low temperatures donors deeper than 0.17 eV remain filled after the illumination is terminated. The sample is then slowly warmed under reverse bias, As the temperature is increased the shallow traps empty first,
016%9002/96/$15.00 Copyright 01996 Elsevier Science B.V. All rights reserved PZI
SOi68-9002(96)00325-7
K. Eerwick
et al. I Nucl. Itrstr. and Me&.
in Ph.vs. Res. A 380 f 19%)
47
X-49
Table I Samples used in this study Sample no.
Supplier
Heat treatments
I
MCP MCP MCP MCP Nippon Mining MCP MCP
Ingot anneal 950°C for 5 h slow cool to room temp. Medium carbon concentration Ingot anneal 950°C for 5 h siow cool to room temp. Low carbon cont. (Seed end wafer) Ingot anneal 950°C for 5 h slow cool to room temp. Ingot anneal 950°C for 12 h slow cool to room temp. Multiple wafer anneal Ingot anneal 950°C for 5 h + multiple wafer anneal t t 130°C and 830°C) As far sample 2. Same ingot, sample wafer from middle
2 3 4 5 6 7
Table 2 Concentrations of deep electron traps for each sample in this study. A dash means not detected Sample no.
0.17 eV
0.23 eV
X10'" cm-'
X10'" cm-'
i
1.5
4.3
2 3 4 5 6 I
370 150 240 880 150 -
310 20 200 10
0.27 eV X 10" cm”’
0.32 eV X IO” cm->
0.38 eV X IO’” cm‘ ’
0.45 eV X IO” cm-’
0.51 ev XlO”cm
1.3
3.7
1.a
1.0 4.0
84 4.4 _ 40
30 9.7 26 19 I2 10
8.2 5 .o _
1.6 2.0 7.3
1.38 2.0
followed by progressively deeper levels. Therefore, peaks in sample conductivity occur in the current versus temperature spectrum, each corresponding to different trap energies. By analysing this spectrum it is possible to extract both the activation energy and the concentrations of the levels present. This technique is limited by the onset of thermal currents so the deepest level measured in this study was 0.51 eV [S]. The defect densities in these samples were very different. either as a result of heat treatments or different growth c(~nditions (Table 2). measurements of the concentrations of the dominant deep donor EL2, [EL21 were obtained from photoquenching the calibrated optical absorption band at a wavelength of I p,rn [9] (Table 2). In addition, the resistivities of these samples were measured using a non-c~)ntacting time dependent charge measurement technique f IO]. Because the resistivity of SI C&As is extremely sensitive to temperature and there appears to be no consensus amongst substrate suppliers about which temperature to use, we have standardized our measurements at 393iO.l K.
100 2x0
7.0 3.0
EL2 '
XlO"cm"
12 Il.5 11 7.5 8.5 12 12
charge to the charge deposited in the detector, assuming that the energy to create an electron-hole pair is 4.3 eV [I 11. The calibration is carried out using a pulser to inject charge into the input of the spectroscopy chain. We have attempted to find a correlation between CCE and the concentration of any observed electron trap. However, none is seen with any observed deep level including EL2. This failure to tind a correlation is consistent with a previous report [ 121. However, we find a good correlation between the CCE (at 200 V) and the resistivity of the sample as shown in Fig. I. The behaviour of CCE as a function of applied voltage for the sample of lowest resistivity (4 X IO’ fi. cm) and a standard LEC substrate is shown in Fig. 2.
4. Experimental Each detector was reverse biased at 200 V and the spectrum from alpha particles irradiating the Schottky contact was taken using a standard spectroscopy chain with a shaping time of 0.5 ps and recorded on a MCA. As a result of Ramo’s theorem, the current pulse in these circumstances is due almost entirely to the motion of electrons. The CCE is defined as the ratio of the measured
OI 0
100 200 300 400 500 600 700 Bins volt8fJnv
Pig. 1. The charge collection efficiency (WE%) plotted against wafer resistivity fat 293 K). for o particles i~diatjng the Schottky contact. The numbers correspond to the samples in Table 1 and Table 2.
I(a). VARIOUS MATERIALS
48
K. Berwick et al. / Nucl. Instr. and Meth. in Phys. Res. A 380 (1996) 46-49
lo.,.,. 0
1
,. 2
I.
I.
1.1,
3 4 5 RoslslWly x10’ ncm
6
8
7
of CCE measured for a particles irradiating the Schottky contact as a function of bias for two samples.
Fig. 2. Comparison
5. Discussion We have shown in previous studies that it is primarily the trapping of electrons that reduces the CCE in those circumstances where effects due to incomplete penetration of the electric field are small [13]. This previous data suggests an electron drift length of around 60 pm. However, the data of Fig. 2 showing a saturated CCE of - 80% is consistent with a drift length of no less than 350 Fm. Trapping of electrons must be due to ionised donors or neutral acceptors having an ionisation energy above the mid-gap Fermi-level, E,. The energy level of the only known neutral acceptor in SI GaAs, the so-called RC defect [14], is degenerate with the conduction band at room temperature and is, therefore, inactive. We, therefore, believe that only ionized donors are available for trapping. We have noted above that there is no apparent correlation between CCE and concentrations of individual deep donor levels in SI GaAs. The only parameter that is seen to correlate with CCE is sample conductivity (Fig. 1, noting that conductivity is the inverse of resistivity). In SI GaAs with constant carrier mobility, such an increase is due to a lifting of the Fermi energy. Raising E, has a further consequence. We note that total concentrations of EL2 defects are relatively constant from sample to sample, even after quite dramatic heat treatments (Table 2). Also, E, can be expressed in terms of the ratio of ionized to unionized EL2, by [EL2+]/[EL2]
= 1 - [(I + l/2 exp (E,,,
- E,)IkT)-‘1
where EEL2 is the energy to ionize neutral EL2 into its positive charge state and k is Boltzmann’s constant. It follows that raising E, reduces the value of [EL2+]/[EL2] and, since [EL21 is nearly constant, it reduces [EL2+]. Thus the increase in CCE associated with a reduction in sample resistivity can be directly related to a concomitant
diminution in [EL2+]. For these reasons we believe that the major trapping centre in these detectors is the EL2+ defect. Trapping by an ionized deep donor like EL2’ is hardly unexpected. However, we are aware that concentrations of EL2’ probably do not exceed lOI5 crnb3 in any of our material, a value that is comparable with concentrations of other ionized donors. Additionally, we note that published electron capture cross sections for EL2’ are of the order of 1O-m’3cm’, a value also comparable with those of other donors [ 151. Therefore, the great effectiveness for electron capture by EL2+ appears to be problematic. However, it has been demonstrated that electron capture by EL2’ is greatly enhanced in the presence of an electric field [16] and, indeed, such behaviour has recently been employed successfully in a model to simulate the penetration of the electric field into SI GaAs 1171. Thus, we propose that EL2’ defects are very efficient electron traps in the electric fields typical in these particle detectors [ 171 and that their concentration is critical in reducing CCE. Trapping of electrons by EL2+ defects is particularly important because de-trapping rates at room temperature are negligible compared to the typical integration time of particle detectors. Unfortunately, the performance of devices fabricated on “low resistivity” material is degraded by the relatively high voltage current. This has also made it impossible to perform electric field penetration measurements which would confirm whether the data of Fig. 2 is to be associated with electric field penetration or the drift length of electrons in a uniform field. We are currently performing measurements to clarify this situation.
6. Conclusions We conclude that EL2’ centres are crucial in the trapping of free electrons in SI GaAs under bias. A similar mechanism has been employed to model the electric field penetration in similar devices [ 171. Reducing the value of the electrical resistivity appears to be a potent way of improving CCE by reducing net concentrations of EL2’ defects.
Acknowledgments The authors are grateful to Miss J. Gilmore for the careful preparation of this paper. The authors acknowledge the support of PARC.
References [I] D. Alexiev and K.S.A. Butcher, Nucl. Instr. and Meth. A 317 (1992)
111.
K. Betwick
et al. / Nucl. Instr.
and Meth.
D.S. McGregor, G.F. Knoll, Y. Eisen and R. Brake. IEEE Trans. Nucl. Sci. NS-39 (1992) 1226. [3] G.M. Martin, Proc. Conf. Semi-Insulating B-V Mats.,
[2]
1980(Shiva, UK, 1980) p. 13. [4] R. Bertin, S. d’Auria. C. Del Papa, F. Fiori, B. Lis0wsksi.V. Oshea. P.G. Pelfer. K. Smith and A. Zichichi, Nucl. Instr. and Meth. A 294 (1990) 21 I. ]S] K. Berwick, M.R. Brozel, C. Buttar, M. Cowperthwaite. P. Sellin and Y. Hou, Inst. Phys. Conf. Ser. 135 (IOP, Bristol, 1994). 16) C. De1 Papa, Proc. Int. GaAs Workshop, Jostal, 1993. [7] J.S. Pooni and M.R. Brozel, Inst. Phys. Conf. Ser. 135 (IOP Bristol. 1994). [S] P. Blood and W. Orton, Electrical characterisation of semiconductors, majority carrier and electron states, Techniques in Physics (.I9923 p.14. [9] G.M. Martin, Appl. Phys. Lett. 39 (1981) 747. IO] R. Stibal. J. Windscheif and W. Jantz, Semicond. Sci. Technol. 6 (1991) 995.
in Php.
Res. A 380 (1996)
4’)
46-49
[I I] W. Bencivelli. R. Bertin, E. Bettolucci. U. Bottigli, S. d’ Auria, C. Del Papa, M.E. Fantacci, P. Randaccio. V. Ross0 and A. Stafanini, Nucl. Ins&. and Meth. A 338 (1994) 549. [12] F. Nava, C. Canah, A. Castaldini, A. Cavallini, S.D. Auria. C. Del Papa. C. Frigeri, L. Zanotti, A. Cetronio. C. Lanzieri and A. Zichichi, Nucl. Instr. and Meth. A 349 (1994) 156. [ 131 P.J. Sellin, C.M. Buttar, S. Manopoulos, K. Berwick, M.R. Brozel and M. Cowperthwaite. IEEE Trans. Nucl. Sci. NS42 f t995). [14] M.R. Brozel and S. Tuzemen. IOP Conf. Ser. No. 13.5 (1994) p. 187. [ 151 Properties of GaAs. 2nd Ed. (INSPEC. 1990) p. 309. [16] V.Ya. Prim and B.A. Bobylev, Sov. Phys. Semicond. 14 (1980) 1097. [ 171 D.S. McGregor. R.A. Rojeski. G.F. Knoll, F.L. Terry, J. East and Y. Eisen, J. Appl. Phys. 75 (1994) 7910.
I(a). VARIOlJS
MATERIALS