- Email: [email protected]
Beryllium implantation induced deep levels in 6H-silicon carbide
Physica B 308–310 (2001) 718–721
Beryllium implantation induced deep levels in 6H-silicon carbide X.D. Chena,*, S. Funga, C.D. Belinga, M. Gongb, T. Henkelc, H. Tanouec, N. Kobayashic a
Department of Physics, University of Hong Kong, Hong Kong, People’s Republic of China b Department of Physics, Sichuan University, Chengdu, People’s Republic of China c Electrotechnical Laboratory, Tsukuba, Ibaraki 305-8568, Japan
Abstract Beryllium has been implanted into both n- and p-type 6H-silicon carbide (SiC) with high and low doses. Upon subsequent annealing at 16001C, Beryllium implantation induced deep levels have been investigated by deep level transient spectroscopy. Five deep level centers labeled as BE1–BE5 were detected from high dose beryllium implantation produced pn junctions. A comparative study of low dose beryllium implanted n-type 6H-SiC sample proved that the BE1–BE3 centers were electron traps located at 0.34, 0.44, and 0.53 eV, respectively below the conduction band edge. At the same time, the BE4 and BE5 centers were found to be hole traps situated at 0.64 and 0.73 eV, respectively, above the valence band edge. In the case of beryllium implanted p-type 6H-SiC, four hole traps labeled as BEP1, BEP2, BEP3, and BEP4 have been observed. The observed levels of the hole traps BEP1 and BEP2 at 0.41 and 0.60 eV, respectively, above the valence band agree well with those from the Hall effect data from material with beryllium acting as doubly charged acceptor. The other hole traps BEP3 and BEP4 at 0.76 and 0.88 eV, above the valence band, respectively, are thought to be due to beryllium implantation induced defects or complexes. r 2001 Elsevier Science B.V. All rights reserved. Keywords: Deep level defect; Beryllium; 6H-silicon carbide; Implantation
1. Introduction
2. Experiment and results
Beryllium (Be) is known as a doubly charged acceptor in silicon carbide (SiC). The acceptor levels, 0.42 and 0.60 eV were determined at the first attempt by Hall measurements on 6H-SiC bulk material [1]. Recently, Be was successfully applied in the fabrication of SiC diodes. A better forward characteristic was obtained in comparison with pn junctions produced by boron (B) and aluminum (Al) implantation [2]. On the other hand, Be may also act as a donor in SiC when residing on the interstitial sites, although, the donor levels are not known [3]. However, the knowledge about the physics of the Be dopant in SiC is still very limited.
The 6H-SiC used in this work was purchased from CREE Research Inc. in the form of 10 mm thick nitrogen doped (0 0 0 1) oriented epilayers grown on n+ type SiC substrates. The nitrogen donor concentrations were 1 1016 and 8 1017 cm3 in the epilayer and the substrate, respectively. 50–590 keV Be implantation (termed as high-dose Be implantation hereafter) was carried out to obtain a box-shaped profile with a mean Be concentration of about 1 1019 cm3. To repair the implantation induced damage and to electrically activate the dopant, samples were annealed in flowing argon gas at 16001C for 1 min using a rapid thermal annealing (RTA) system. For a comparative study, 6H-SiC samples with a lower Be concentration of about 1 1018 cm3 (termed
*Corresponding author. Fax: +852-28598972. E-mail address: [email protected] (S. Fung).
0921-4526/01/$ - see front matter r 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 2 6 ( 0 1 ) 0 0 8 8 0 - 8
X.D. Chen et al. / Physica B 308–310 (2001) 718–721
Fig. 1. DLTS spectrum recorded on the high-dose Be implanted 6H-SiC pn junction samples.
low-dose Be implantation hereafter) were implanted into n- and p-type 6H-SiC samples, respectively. The same multiple energy implantation schedule as for the highdose implantation was applied. RTA was performed at 16001C for 30 s under similar conditions as stated above. A typical majority carrier spectrum is shown in Fig. 1. Five peaks labeled as BE1, BE2, BE3, BE4, and BE5 were observed in the deep level transient spectroscopy (DLTS) measurement temperature range of 100–450 K. To check whether these traps were induced by Be implantation, a control measurement was performed on a schottky contact on unimplanted n-type epilayer. None of the above traps were detected indicating that all the traps described above were induced by the Be implantation process. It should be pointed out that for the high-dose Be implanted samples, the Hall measurements indicated a weak p-type conduction in the Be implanted layer, the free hole concentration in the p-region and the free electron concentration in the n-region of the pn junction were of the same order of magnitude. Under reverse bias conditions it is thus reasonable to expect that the width of the depletion layer in the p-region is comparable with the one in the n-region. Consequently, the observed DLTS peaks may arise either from hole traps in the p-region or electron traps in the n-region of the pn junction. A definitive assignment is not possible without further information about these traps. Therefore, the n-type low-dose Be implanted samples was additionally prepared. A typical DLTS spectrum of these samples is shown in Fig. 2, all the three levels are attributed to electron traps. Since the energy level and the temperature position of the DLTS peak BE10 (see Fig. 2) are in excellent agreement with those of BE1 (see Fig. 1), BE10
719
Fig. 2. DLTS spectrum recorded on the low-dose Be implanted n-type 6H-SiC.
Fig. 3. DLTS spectrum recorded on the low-dose Be implanted p-type 6H-SiC.
and BE1 are attributed to the same defect center. Likewise BE20 and BE30 are assumed to be related to the overlapping peaks BE2 and BE3 because of their very similar energy levels. Hence, it is unambiguously clear that three (BE1, BE2, and BE3) of the five deep level traps observed in the high-dose Be implanted sample are electron traps arising from the n-region while the other traps (BE4, BE5) are assumed to be hole traps from the p-region of the pn junction. Fig. 3 shows the DLTS spectrum of Be implanted ptype 6H-SiC. Four deep levels labeled as BEP1, BEP2, BEP3, and BEP4 as hole traps are observed in a
720
X.D. Chen et al. / Physica B 308–310 (2001) 718–721
Table 1 Energy levels ET ; capture cross sections sT ; and concentrations NT of the deep level traps determined using DLTS data of the high- and low-dose Be implanted 6H-SiC samples
BE1 BE2 BE3 BE4 BE5 BEP1 BEP2 BEP3 BEP4
ET (eV)
sT (cm2)
NT (cm3)
Ec 0:34 Ec 0:44 Ec 0:53 EV þ 0:64 EV þ 0:73 EV þ 0:41 EV þ 0:60 EV þ 0:76 EV þ 0:88
B1013 B1014 B1014 B1016 B1016 B1013 B1015 B1016 B1017
4–10 1013 1–6 1013 3–8 1013 4–6 1014 2–7 1014 4–10 1014 1–6 1013 1–3 1013 2–6 1013
temperature range of 180–450 K. At temperatures below T ¼ 180 K, the DLTS peak shape is strongly affected by the increased series resistance due to the free carrier freeze-out of the shallow Al dopant. The energy levels, capture cross sections, and concentrations of the defect centers determined are summarized in Table 1.
3. Discussion Recent DLTS studies on electron-irradiated [4–6] and deuterium-implanted [5] n-type 6H-SiC have demonstrated the existence of two overlapping levels at EC 0:33 and 0:34 eV that were designated as E1 and E2 ; respectively. Our previous DLTS investigations on electron-irradiated 6H-SiC revealed several deep level traps. Two of them (labeled as ED3 and ED4) were attributed to the E1 =E2 center [6]. The fact that the BE1 peak is so close in energy to those of the E1 =E2 doublet is suggestive that the two are one and the same defect with the perturbations present on the center close to the pn junction perhaps causing the doublet to smear into a single peak. Such a view, however, meets with some difficulty when it is realized that the peak maximum temperatures of BE1 and ED3 differ by 50 K and the calculated cross-section by two orders of magnitude. This makes it unlikely that BE1 and E1 =E2 are the same defect. An alternative explanation for BE1 is that it is associated with interstitial Be. It is known that Be may act as a donor when residing on interstitial positions [3]. Further, there is strong evidence indicating that Be diffuses via interstitial sites [3,7,8]. Since a deep diffusion tail into the epilayer is observed after RTA, there is definite evidence for fast interstitial diffusion of Be and thus by inference the related donor site [7,8]. It may tentatively be argued that Be is residing on an interstitial site that leads to the BE1 center. Clearly, further theoretical investigations will be necessary to test this possibility.
The trap BE3, the energy level of which has been determined as EC 0:53 eV, is similar to the defect level was previously found in electron irradiated 6H-SiC by several groups [4,5]. On the other hand, Dalibor et al. observed a trap labeled as ID7 at EC 0:50 eV after vanadium as well as titanium implantation followed by an annealing process at 17001C [9]. This defect level was ascribed to an implantation induced intrinsic defect center. Due to the different annealing behavior, the defect associated with the ID7 center must be different from the one observed in electron irradiated material. We believe that the BE3 and ID7 levels originate from the same defect, i.e., an intrinsic defect center induced by the implantation process. The peak BE2 in the DLTS spectrum is very weak and overlapped by the peak BE3, which limits the precision of the calculated parameters. Therefore, the question whether the trap BE2 is to be associated with an implantation induced intrinsic defect or a Be induced defect cannot be answered at present. In the case of the low-dose Be implanted Al doped ptype 6H-SiC samples, the Fermi-level should mainly be controlled by the acceptor level of Al dopant. Therefore, the doubly charged states of Be may be observed in Be implanted p-type samples by using DLTS technique. The energy levels of BEP1 and BEP2, respectively at EV þ 0:41 and 0:60 eV determined from the current experiment are consistent with the Hall effect result determined in Be diffused 6H-SiC samples [1]. Thus, we can conclude that the BEP1 and BEP2 traps are attributed to the doubly charged states of substitutional Be acting as acceptor. On the other hand, the BE4 center located at EV þ 0:64 eV from the p-side of Be implanted pn junction may also arise from the second charge state of the Be acceptor. The different measurement techniques and purities of the samples are assumed to be responsible for the small discrepancy between the DLTS and the Hall result (EV þ 0:60 eV) reported in the past [1]. Considering the hole trap BEP3, energy level was determined to be 0.76 eV above the valence band. Though comprehensive investigations on intrinsic defects in p-type 6H-SiC are still lacking, a defect level labeled as H2 at a similar energy position was found in electron irradiated material by our group [10]. It was demonstrated that the defect associated with H2 anneals out at about 3501C. As stated earlier, the current Be implanted 6H-SiC samples were annealed at temperature T ¼ 16001C, the different thermal annealing behaviors seem to argue against the assignment of BEP3 as H2 in the p-type 6H-SiC. However, it should be pointed out that the DLTS signal of BEP3 is very weak (as shown in Fig. 3), and the low-dose Be implanted ptype samples were annealed at 16001C for 30 s. It does not preclude the possibility that BEP3 may arise from an intrinsic defect due to the very short annealing duration used for the low-dose Be implanted SiC samples.
X.D. Chen et al. / Physica B 308–310 (2001) 718–721
The BEP4 at EV þ 0:88 eV determined from p-type sample and the BE5 at EV þ 0:73 eV detected from the p-side of pn junction are determined as hole traps induced by Be implantation. Since there was no similar energy levels observed from electron irradiated or boron and gallium implanted p-type 6H-SiC experiments, we assign the two levels to be beryllium-related defects. Recently, several deep Be related centers were found by electron paramagnetic resonance (EPR) investigation on 6H-SiC crystal doped by diffusion technique [11]. The BEP4 and BE5 centers may be associated with these defects which were found to resemble the deep B center in terms of electronic and magnetic properties [12]. Acknowledgements The work described in this paper was partially supported by the grants from the Research Grant Council of the Hong Kong Special Administrative Region, China (under project No. HKU7137/99P). References [1] Y.P. Maslakovets, E.N. Mokhov, Y.A. Vodakov, G.A. Lomakina, Sov. Phys. Solid State 10 (1968) 634.
721
[2] N. Ramungul, Y.P. Zheng, R. Rupal, T.P. Chow, IEEE Trans. Electron Devices ED-46 (1999) 465. [3] Y.A. Vodakov, G.A. Lomakina, E.N. Mokhov, V.G. Oding, E.I. Radovanova, Sov. Phys. Solid State 20 (1978) 258. [4] C. Hemmingsson, N.T. Son, O. Kordina, E. Janzen, J. Appl. Phys. 84 (1998) 704. [5] M.O. Aboelfotoh, J.P. Doyle, Phys. Rev. B 59 (10) (1999) 823. [6] M. Gong, S. Fung, C.D. Beling, Z. You, J. Appl. Phys. 85 (1999) 7604. [7] X.D. Chen, S. Fung, C.D. Beling, M. Gong, T. Henkel, Y. Tanaka, N. Kobayashi, H. Tanoue, J. Appl. Phys. 88 (2000) 4558. [8] T. Henkel, Y. Tanaka, N. Kobayashi, H. Tanoue, S. Hishita, Appl. Phys. Lett. 78 (2001) 231. [9] T. Dalibor, G. Pensl, H. Matsunami, T. Kimoto, W.J. Choyke, A. Schoener, N. Nordell, Phys. Stat. Sol. A 162 (1997) 199. [10] M. Gong, S. Fung, C.D. Beling, Z. You, J. Appl. Phys. 85 (1999) 7120. [11] A. van Duijn-Arnold, J. Schmidt, O.G. Poluektov, P.G. Baranov, E.N. Mokhov, Phys. Rev. B 60 (1999) 15799. [12] A. van Duijn-Arnold, T. Ikoma, O.G. Poluektov, P.G. Baranov, J. Schmidt, Phys. Rev. B 57 (1998) 1607.
Beryllium implantation induced deep levels in 6H-silicon carbide X.D. Chena,*, S. Funga, C.D. Belinga, M. Gongb, T. Henkelc, H. Tanouec, N. Kobayashic a
Department of Physics, University of Hong Kong, Hong Kong, People’s Republic of China b Department of Physics, Sichuan University, Chengdu, People’s Republic of China c Electrotechnical Laboratory, Tsukuba, Ibaraki 305-8568, Japan
Abstract Beryllium has been implanted into both n- and p-type 6H-silicon carbide (SiC) with high and low doses. Upon subsequent annealing at 16001C, Beryllium implantation induced deep levels have been investigated by deep level transient spectroscopy. Five deep level centers labeled as BE1–BE5 were detected from high dose beryllium implantation produced pn junctions. A comparative study of low dose beryllium implanted n-type 6H-SiC sample proved that the BE1–BE3 centers were electron traps located at 0.34, 0.44, and 0.53 eV, respectively below the conduction band edge. At the same time, the BE4 and BE5 centers were found to be hole traps situated at 0.64 and 0.73 eV, respectively, above the valence band edge. In the case of beryllium implanted p-type 6H-SiC, four hole traps labeled as BEP1, BEP2, BEP3, and BEP4 have been observed. The observed levels of the hole traps BEP1 and BEP2 at 0.41 and 0.60 eV, respectively, above the valence band agree well with those from the Hall effect data from material with beryllium acting as doubly charged acceptor. The other hole traps BEP3 and BEP4 at 0.76 and 0.88 eV, above the valence band, respectively, are thought to be due to beryllium implantation induced defects or complexes. r 2001 Elsevier Science B.V. All rights reserved. Keywords: Deep level defect; Beryllium; 6H-silicon carbide; Implantation
1. Introduction
2. Experiment and results
Beryllium (Be) is known as a doubly charged acceptor in silicon carbide (SiC). The acceptor levels, 0.42 and 0.60 eV were determined at the first attempt by Hall measurements on 6H-SiC bulk material [1]. Recently, Be was successfully applied in the fabrication of SiC diodes. A better forward characteristic was obtained in comparison with pn junctions produced by boron (B) and aluminum (Al) implantation [2]. On the other hand, Be may also act as a donor in SiC when residing on the interstitial sites, although, the donor levels are not known [3]. However, the knowledge about the physics of the Be dopant in SiC is still very limited.
The 6H-SiC used in this work was purchased from CREE Research Inc. in the form of 10 mm thick nitrogen doped (0 0 0 1) oriented epilayers grown on n+ type SiC substrates. The nitrogen donor concentrations were 1 1016 and 8 1017 cm3 in the epilayer and the substrate, respectively. 50–590 keV Be implantation (termed as high-dose Be implantation hereafter) was carried out to obtain a box-shaped profile with a mean Be concentration of about 1 1019 cm3. To repair the implantation induced damage and to electrically activate the dopant, samples were annealed in flowing argon gas at 16001C for 1 min using a rapid thermal annealing (RTA) system. For a comparative study, 6H-SiC samples with a lower Be concentration of about 1 1018 cm3 (termed
*Corresponding author. Fax: +852-28598972. E-mail address: [email protected] (S. Fung).
0921-4526/01/$ - see front matter r 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 4 5 2 6 ( 0 1 ) 0 0 8 8 0 - 8
X.D. Chen et al. / Physica B 308–310 (2001) 718–721
Fig. 1. DLTS spectrum recorded on the high-dose Be implanted 6H-SiC pn junction samples.
low-dose Be implantation hereafter) were implanted into n- and p-type 6H-SiC samples, respectively. The same multiple energy implantation schedule as for the highdose implantation was applied. RTA was performed at 16001C for 30 s under similar conditions as stated above. A typical majority carrier spectrum is shown in Fig. 1. Five peaks labeled as BE1, BE2, BE3, BE4, and BE5 were observed in the deep level transient spectroscopy (DLTS) measurement temperature range of 100–450 K. To check whether these traps were induced by Be implantation, a control measurement was performed on a schottky contact on unimplanted n-type epilayer. None of the above traps were detected indicating that all the traps described above were induced by the Be implantation process. It should be pointed out that for the high-dose Be implanted samples, the Hall measurements indicated a weak p-type conduction in the Be implanted layer, the free hole concentration in the p-region and the free electron concentration in the n-region of the pn junction were of the same order of magnitude. Under reverse bias conditions it is thus reasonable to expect that the width of the depletion layer in the p-region is comparable with the one in the n-region. Consequently, the observed DLTS peaks may arise either from hole traps in the p-region or electron traps in the n-region of the pn junction. A definitive assignment is not possible without further information about these traps. Therefore, the n-type low-dose Be implanted samples was additionally prepared. A typical DLTS spectrum of these samples is shown in Fig. 2, all the three levels are attributed to electron traps. Since the energy level and the temperature position of the DLTS peak BE10 (see Fig. 2) are in excellent agreement with those of BE1 (see Fig. 1), BE10
719
Fig. 2. DLTS spectrum recorded on the low-dose Be implanted n-type 6H-SiC.
Fig. 3. DLTS spectrum recorded on the low-dose Be implanted p-type 6H-SiC.
and BE1 are attributed to the same defect center. Likewise BE20 and BE30 are assumed to be related to the overlapping peaks BE2 and BE3 because of their very similar energy levels. Hence, it is unambiguously clear that three (BE1, BE2, and BE3) of the five deep level traps observed in the high-dose Be implanted sample are electron traps arising from the n-region while the other traps (BE4, BE5) are assumed to be hole traps from the p-region of the pn junction. Fig. 3 shows the DLTS spectrum of Be implanted ptype 6H-SiC. Four deep levels labeled as BEP1, BEP2, BEP3, and BEP4 as hole traps are observed in a
720
X.D. Chen et al. / Physica B 308–310 (2001) 718–721
Table 1 Energy levels ET ; capture cross sections sT ; and concentrations NT of the deep level traps determined using DLTS data of the high- and low-dose Be implanted 6H-SiC samples
BE1 BE2 BE3 BE4 BE5 BEP1 BEP2 BEP3 BEP4
ET (eV)
sT (cm2)
NT (cm3)
Ec 0:34 Ec 0:44 Ec 0:53 EV þ 0:64 EV þ 0:73 EV þ 0:41 EV þ 0:60 EV þ 0:76 EV þ 0:88
B1013 B1014 B1014 B1016 B1016 B1013 B1015 B1016 B1017
4–10 1013 1–6 1013 3–8 1013 4–6 1014 2–7 1014 4–10 1014 1–6 1013 1–3 1013 2–6 1013
temperature range of 180–450 K. At temperatures below T ¼ 180 K, the DLTS peak shape is strongly affected by the increased series resistance due to the free carrier freeze-out of the shallow Al dopant. The energy levels, capture cross sections, and concentrations of the defect centers determined are summarized in Table 1.
3. Discussion Recent DLTS studies on electron-irradiated [4–6] and deuterium-implanted [5] n-type 6H-SiC have demonstrated the existence of two overlapping levels at EC 0:33 and 0:34 eV that were designated as E1 and E2 ; respectively. Our previous DLTS investigations on electron-irradiated 6H-SiC revealed several deep level traps. Two of them (labeled as ED3 and ED4) were attributed to the E1 =E2 center [6]. The fact that the BE1 peak is so close in energy to those of the E1 =E2 doublet is suggestive that the two are one and the same defect with the perturbations present on the center close to the pn junction perhaps causing the doublet to smear into a single peak. Such a view, however, meets with some difficulty when it is realized that the peak maximum temperatures of BE1 and ED3 differ by 50 K and the calculated cross-section by two orders of magnitude. This makes it unlikely that BE1 and E1 =E2 are the same defect. An alternative explanation for BE1 is that it is associated with interstitial Be. It is known that Be may act as a donor when residing on interstitial positions [3]. Further, there is strong evidence indicating that Be diffuses via interstitial sites [3,7,8]. Since a deep diffusion tail into the epilayer is observed after RTA, there is definite evidence for fast interstitial diffusion of Be and thus by inference the related donor site [7,8]. It may tentatively be argued that Be is residing on an interstitial site that leads to the BE1 center. Clearly, further theoretical investigations will be necessary to test this possibility.
The trap BE3, the energy level of which has been determined as EC 0:53 eV, is similar to the defect level was previously found in electron irradiated 6H-SiC by several groups [4,5]. On the other hand, Dalibor et al. observed a trap labeled as ID7 at EC 0:50 eV after vanadium as well as titanium implantation followed by an annealing process at 17001C [9]. This defect level was ascribed to an implantation induced intrinsic defect center. Due to the different annealing behavior, the defect associated with the ID7 center must be different from the one observed in electron irradiated material. We believe that the BE3 and ID7 levels originate from the same defect, i.e., an intrinsic defect center induced by the implantation process. The peak BE2 in the DLTS spectrum is very weak and overlapped by the peak BE3, which limits the precision of the calculated parameters. Therefore, the question whether the trap BE2 is to be associated with an implantation induced intrinsic defect or a Be induced defect cannot be answered at present. In the case of the low-dose Be implanted Al doped ptype 6H-SiC samples, the Fermi-level should mainly be controlled by the acceptor level of Al dopant. Therefore, the doubly charged states of Be may be observed in Be implanted p-type samples by using DLTS technique. The energy levels of BEP1 and BEP2, respectively at EV þ 0:41 and 0:60 eV determined from the current experiment are consistent with the Hall effect result determined in Be diffused 6H-SiC samples [1]. Thus, we can conclude that the BEP1 and BEP2 traps are attributed to the doubly charged states of substitutional Be acting as acceptor. On the other hand, the BE4 center located at EV þ 0:64 eV from the p-side of Be implanted pn junction may also arise from the second charge state of the Be acceptor. The different measurement techniques and purities of the samples are assumed to be responsible for the small discrepancy between the DLTS and the Hall result (EV þ 0:60 eV) reported in the past [1]. Considering the hole trap BEP3, energy level was determined to be 0.76 eV above the valence band. Though comprehensive investigations on intrinsic defects in p-type 6H-SiC are still lacking, a defect level labeled as H2 at a similar energy position was found in electron irradiated material by our group [10]. It was demonstrated that the defect associated with H2 anneals out at about 3501C. As stated earlier, the current Be implanted 6H-SiC samples were annealed at temperature T ¼ 16001C, the different thermal annealing behaviors seem to argue against the assignment of BEP3 as H2 in the p-type 6H-SiC. However, it should be pointed out that the DLTS signal of BEP3 is very weak (as shown in Fig. 3), and the low-dose Be implanted ptype samples were annealed at 16001C for 30 s. It does not preclude the possibility that BEP3 may arise from an intrinsic defect due to the very short annealing duration used for the low-dose Be implanted SiC samples.
X.D. Chen et al. / Physica B 308–310 (2001) 718–721
The BEP4 at EV þ 0:88 eV determined from p-type sample and the BE5 at EV þ 0:73 eV detected from the p-side of pn junction are determined as hole traps induced by Be implantation. Since there was no similar energy levels observed from electron irradiated or boron and gallium implanted p-type 6H-SiC experiments, we assign the two levels to be beryllium-related defects. Recently, several deep Be related centers were found by electron paramagnetic resonance (EPR) investigation on 6H-SiC crystal doped by diffusion technique [11]. The BEP4 and BE5 centers may be associated with these defects which were found to resemble the deep B center in terms of electronic and magnetic properties [12]. Acknowledgements The work described in this paper was partially supported by the grants from the Research Grant Council of the Hong Kong Special Administrative Region, China (under project No. HKU7137/99P). References [1] Y.P. Maslakovets, E.N. Mokhov, Y.A. Vodakov, G.A. Lomakina, Sov. Phys. Solid State 10 (1968) 634.
721
[2] N. Ramungul, Y.P. Zheng, R. Rupal, T.P. Chow, IEEE Trans. Electron Devices ED-46 (1999) 465. [3] Y.A. Vodakov, G.A. Lomakina, E.N. Mokhov, V.G. Oding, E.I. Radovanova, Sov. Phys. Solid State 20 (1978) 258. [4] C. Hemmingsson, N.T. Son, O. Kordina, E. Janzen, J. Appl. Phys. 84 (1998) 704. [5] M.O. Aboelfotoh, J.P. Doyle, Phys. Rev. B 59 (10) (1999) 823. [6] M. Gong, S. Fung, C.D. Beling, Z. You, J. Appl. Phys. 85 (1999) 7604. [7] X.D. Chen, S. Fung, C.D. Beling, M. Gong, T. Henkel, Y. Tanaka, N. Kobayashi, H. Tanoue, J. Appl. Phys. 88 (2000) 4558. [8] T. Henkel, Y. Tanaka, N. Kobayashi, H. Tanoue, S. Hishita, Appl. Phys. Lett. 78 (2001) 231. [9] T. Dalibor, G. Pensl, H. Matsunami, T. Kimoto, W.J. Choyke, A. Schoener, N. Nordell, Phys. Stat. Sol. A 162 (1997) 199. [10] M. Gong, S. Fung, C.D. Beling, Z. You, J. Appl. Phys. 85 (1999) 7120. [11] A. van Duijn-Arnold, J. Schmidt, O.G. Poluektov, P.G. Baranov, E.N. Mokhov, Phys. Rev. B 60 (1999) 15799. [12] A. van Duijn-Arnold, T. Ikoma, O.G. Poluektov, P.G. Baranov, J. Schmidt, Phys. Rev. B 57 (1998) 1607.