- Email: [email protected]
Cellular structure formed by ion-implantation-induced point defect
ARTICLE IN PRESS
Physica B 376–377 (2006) 881–885 www.elsevier.com/locate/physb
Cellular structure formed by ion-implantation-induced point defect N. Nittaa,, M. Taniwakia, Y. Hayashib, T. Yoshiieb a
Department of Environmental Systems Engineering, Kochi University of Technology (KUT), Tosayamada, Kochi 782-8502, Japan b Research Reactor Institute, Kyoto University, Kumatori, Osaka 596-0821, Japan
Abstract The authors have found that a cellular defect structure is formed on the surface of Sn+ ion implanted GaSb at a low temperature and proposed its formation mechanism based on the movement of the induced point defects. This research was carried out in order to examine the validity of the mechanism by clarifying the effect of the mobility of the point defects on the defect formation. The defect structure on the GaSb surfaces implanted at cryogenic temperature and room temperature was investigated by scanning electron microscopy (SEM) and cross-sectional transmission electron microscopy (TEM) observation. In the sample implanted at room temperature, the sponge-like structure (a pileup of voids) was formed and the cellular structure, as observed at a low temperature, did not develop. This behavior was explained by the high mobility of the vacancies during implantation at room temperature, and the proposed idea that the defect formation process is dominated by the induced point defects was confirmed. r 2006 Elsevier B.V. All rights reserved. PACS: 61.72.Ji; 61.72.Qq; 61.72.Vv; 68.37.Hk; 68.37.Lp Keywords: GaSb; Ion-implantation; Cellular structure; Point defect; Radiation effect; Void; Cross-sectional transmission electron microscopy; Scanning electron microscopy
1. Introduction Defect structure formation in III–V compound semiconductors by ion-implantation is classified into two categories. In the case of GaAs and InP [1–3], a damaged layer is formed on the surface by ion-implantation; further implantation transforms the damaged layer to an amorphous structure. On the other hand, unusual behaviors, such as elevation and swelling, have been observed in GaSb and InSb when irradiated by energetic ions [4–7]. Taniwaki et al. observed an anomalous structure with a thickness 5 times the projected ion range on GaSb and InSb surfaces implanted with 60 keV Sn+ at a low temperature using cross-sectional TEM (transmission electron microscopy) [8]. Nitta et al. examined this defect structure in detail and clarified that it consists of many cells, like a honeycomb [9]. They showed that it was not created by sputtering nor by a vapor phase process, and proposed a defect formation Corresponding author. Tel.: +81 887 57 2504; fax: +81 887 57 2520.
E-mail address: [email protected] (N. Nitta). 0921-4526/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.physb.2005.12.220
mechanism based on movement of the point defects induced by ion implantation. The validity of this mechanism is to be examined further by controlling the quantity and behavior of the induced point defects. In the previous paper, we showed that the defect structure enlarged when increasing the ion dose (i.e. the amount of the induced point defects), which suggests strongly that the defect formation is dominated by point defects [10]. In the present work, the effect of the mobility of point defects on the defect structure is studied by implantation at different substrate temperatures. Implantation was carried out at cryogenic temperature (130–150 K) and room temperature. There is little knowledge on the mobility of the point defects in GaSb, so that their behavior is not yet clarified. However Eisen’s work on the recovery of electron irradiated GaSb at liquid nitrogen temperature [11,12] gives some suggestion. From this work, the long-range migration of the vacancies are considered not to occur at temperatures of 130–150 K where the cellular structure was formed, and the vacancies will be mobile in a long range during implantation at room temperature. Therefore it is
ARTICLE IN PRESS 882
N. Nitta et al. / Physica B 376–377 (2006) 881–885
expected, if the proposed idea is right, that the defect structure in the samples implanted at room temperature differs from that at a low temperature. 2. Experiments Te-doped n-type compound semiconductor GaSb polished wafers with (1 0 0) orientation were implanted with 60 keV Sn+ ions. Implantation was performed using low ion doses (about 4 1014) and high doses (about 8 1014 ions/cm2), at low temperatures (LT sample, 151 and 137 K) and at room temperature (RT sample). As-implanted surfaces were observed through a field emission-type scanning electron microscope (FE-SEM), JEOL JSM-6400F. Cross-sectional TEM samples, prepared from implanted wafers by mechanical polishing and Ar ion-milling, were observed using a field emission-type transmission electron microscope (FE-TEM), JEOL JEM-2010F. 3. Results 3.1. SEM observation Fig. 1 shows the FE-SEM images of the GaSb samples implanted at 151 K and at room temperature. Their ion doses were almost the same (4.0 1014 and 4.1 1014 ions/ cm2, respectively). The image of the LT sample is weak, because its dark contrast does not reflect the surface roughness, but reflect the cavities just under the surface, which will be clearly shown later by cross-sectional TEM observation. These images were considerably different. A high density of isolated roundish holes with about 30 nm diameters was created in the LT sample. The hollows observed in the RT sample were not roundish, and seem to
have been formed by coalescence of several round holes. The densities of the holes or hollows are 5 109 and 4.5 109 cm 2, and the thickness of the walls partitioning those were about 20 nm in the LT sample and about 10 nm in the RT sample. FE-SEM images of the highly dosed GaSb surfaces are shown in Fig. 2. The ion doses were 7.9 1014 ions/cm2 for the LT sample and 8.1 1014 ions/cm2 for the RT sample. The contrast of the several holes in the LT sample is strong compared with that in the lower-dosed sample. This shows that the top surfaces of the holes were eliminated. In the case of room temperature implantation, the image of the hollows did not change much, though they were coarsened to the density of 2.7 109 cm 2. 3.2. Cross-sectional TEM observation Fig. 3 shows the (1 1 0) cross-sectional TEM views (bright field) and the selected area electron diffraction patterns of the GaSb samples implanted with low doses at 151 K and at room temperature. The ion doses were 4.0 1014 and 4.1 1014 ions/cm2, respectively. Although characteristic defects were formed in the surfaces of both samples, their defects were different from each other. Under the surface of the LT sample, large ellipsoidal voids with axes of 50 and 70 nm were arranged in partitions with 10–20 nm thick walls. The top surfaces of the voids are not open, therefore the dark contrast in the SEM image (Fig. 1) shows the voids but not the surface roughness. The cavities are formed ranging from the surface to a 130 nm depth in the RT sample, which look to be a pileup of many voids with about a 20 nm diameter. The crystalline structure around the voids and the cavities is polycrystalline in both samples.
Fig. 1. FE-SEM images of the GaSb surfaces implanted with low doses at 151 K and at room temperature are shown. The ion doses were 4.0 1014 and 4.1 1014 ions/cm2, respectively. Most holes in the sample implanted at room temperature are not round, which seem to be formed by coalescence of several round holes.
ARTICLE IN PRESS N. Nitta et al. / Physica B 376–377 (2006) 881–885
883
Fig. 2. FE-SEM images of the GaSb surfaces implanted with high doses at 137 K and at room temperature. The ion doses were 7.9 1014 and 8.1 1014 ions/cm2, respectively. As compared with Fig. 1, the cavities were developed in the sample implanted at 150 K. However, the image of holes did not change much at room temperature.
Fig. 3. The cross-sectional TEM view (bright field) and the selected area electron diffraction pattern of GaSb surface implanted at 151 K and at room temperature. The ion doses were 4.0 1014 and 4.1 1014 ions/cm2, respectively. Under the surface of the sample implanted at 151 K, large ellipsoidal voids with axes of 50 and 100 nm are arranged. Cavities consisting of many voids with 20–30 nm diameter are formed and are observed to a 100 nm depth in the sample implanted at room temperature.
The (1 1 0) cross-sectional views (bright field) of the GaSb samples implanted with high doses are shown in Fig. 4. The ion doses were 8.9 1014 and 8.1 1014 ions/cm2, respectively. The cellular structure, with a large aspect ratio, had formed on the surface of the LT sample. The diameter and the height of the cells were about 50 and 250 nm respectively, and the thickness of the partitioning cell walls was about 10 nm. A heavy strain region with a 50 nm thickness (strong contrast) can be observed under the cells and its interface with matrix is serrate. Though the defect structure is also created on the surface in the RT sample, it differs significantly from that observed in the LT sample. It is a sponge-like layer with a 180 nm thickness which were formed by a pileup of many voids with about a 20–30 nm
diameter as well as the defect observed in the lower-dosed sample at room temperature. Heavy strain contrast was not observed under the defect structure in this sample, and alternatively the voids with about a 5–10 nm diameter were formed. The structure of the walls was polycrystalline when dosed at room temperature and were amorphous when dosed at a low temperature. 4. Discussion As we expected, different defect structures were formed by implantation at a low temperature and at room temperature; the cellular structure was developed from the voids in LT sample and the sponge-like structure was
ARTICLE IN PRESS 884
N. Nitta et al. / Physica B 376–377 (2006) 881–885
Fig. 4. The cross-sectional TEM view (bright field) and the selected area electron diffraction pattern of GaSb surface implanted at 153 K and at room temperature. Those ion doses were 8.9 1014 and 8.1 1014 ions/cm2, respectively. An sponge-like defect structure is observed on the surface in the sample implanted at room temperature, which is a pileup of many voids with 20–100 nm diameter.
formed by sequential void formation in the RT sample. We intend to explain this difference in the defect structure by the mobility of the implantation-induced point defects. Eisen’s recovery data shows that there are five stages in the recovery of the irradiated III–IV compound semiconductors (I; 120 K, II; 160 K, III, IV; 200 K V; 360 K for GaSb) [11,12]. He mentioned that stages I and II correspond to the recovery of close Frenkel pairs, very small stages III and IV do to the recombination of somewhat far separated pairs and stage V must represent the movement of defects over greater distances than in the other stages. According to Eisen’s interpretation, at the temperature of 130–150 K where the cellular structure is formed, long-range migration of point defects does not occur. And at room temperature, point defects are considered to move in a long range to some extent, where we consider that these point defects are vacancies. According to this idea, voids are formed all over the implanted region by the aggregation of migrating vacancies during implantation at room temperature. Subsequent implantation creates new vacancies under the voids, which also gather and form new voids, and finally lead to the sponge-like structure in the RT samples. In the LT samples, the formation of voids is depressed because of the poor mobility of the vacancies. However, when the induced vacancy concentration has exceeded a critical value determined by the elastic property of GaSb, the voids are abruptly formed at a depth nearly equal to the ion projected range (the highest vacancy concentration). The vacancies induced by further implantation are absorbed by the preformed voids, which grow in the direction of the surface. The cellular defect structure of the heavily implanted sample at a low temperature is formed in this manner. In the above discussion, the difference of the defect structure in the samples implanted at different substrate temperatures was reasonably explained by the mobility of vacancies. Though the behavior of the interstitials was not
discussed here, we noted that the cellular structure was highly developed in the LT sample in Fig. 4. This suggests that the implantation-induced interstitials also significantly contribute to the formation of the cellular structure, which will be examined elsewhere. 5. Conclusion In order to examine the validity of the proposed formation mechanism of the cellular structure on GaSb surface implanted at a low temperature, the effect of substrate temperature on the defect structure was studied using SEM and TEM. At a low temperature, the voids formed in the early stage of implantation grow perpendicular to the surface absorbing newly created vacancies, so that the cell structure develops by increasing ion dose. At room temperature, the voids formed sequentially due to the high mobility of the vacancies leading to the formation of the sponge-like structure and the cellular structure does not develop. The results support our proposed mechanism based on the movement of the point defects for the formation of the cellular structure in ion-implanted GaSb surfaces at a low temperature. Acknowledgements We would like to thank Mr. N. Baba and Ms. S. Komatsu of Kochi Casio Co., Ltd. for their kind help with scanning electron microscope observation. References [1] M. Taniwaki, H. Koide, N. Yoshimoto, T. Yoshiie, S. Ohnuki, M. Maeda, K. Sassa, J. Appl. Phys. 67 (1990) 4036. [2] M. Taniwaki, H. Koide, T. Yoshiie, Y. Hayashi, H. Yoshida, J. NonCryst. Solids 117/118 (1990) 745. [3] M. Taniwaki, T. Yoshiie, H. Koide, M. Ichihasi, N. Yoshimoto, H. Yoshida, Y. Hayashi, J. Appl. Phys. (1989) 66.
ARTICLE IN PRESS N. Nitta et al. / Physica B 376–377 (2006) 881–885 [4] D. Kleitman, H.J. Yearian, Phys. Rev. 108 (1957) 901. [5] Y. Homma, J. Vac. Sci. Technol. A 5 (1987) 321. [6] R. Callec, P.N. Favennce, M. Salvi, H. L’Haridon, M. Gauneau, Appl. Phys. Lett. 59 (1991) 1872. [7] R. Callec, A. Poudoulec, J. Appl. Phys. 73 (1993) 4831. [8] M. Taniwaki, Y. Hayashi, T. Yoshiie, in: Proceedings of the International Conference on Solid–Solid Phase Transformations ’99, 1999, p. 425.
885
[9] N. Nitta, M. Taniwaki, T. Suzuki, Y. Hayashi, Y. Satoh, T. Yoshiie, Mater. Trans. 43 (2002) 674. [10] N. Nitta, M. Taniwaki, Y. Hayashi, T. Yoshiie, J. Appl. Phys. 92 (2002) 1799. [11] F.H. Eisen, Phys. Rev. 123 (1961) 736. [12] F.H. Eisen, in: J.W. Corbett, G.D. Watkins (Eds.), Radiation Effects in Semiconductors, Gordon and Breach, New York, 1971, p. 273.
Physica B 376–377 (2006) 881–885 www.elsevier.com/locate/physb
Cellular structure formed by ion-implantation-induced point defect N. Nittaa,, M. Taniwakia, Y. Hayashib, T. Yoshiieb a
Department of Environmental Systems Engineering, Kochi University of Technology (KUT), Tosayamada, Kochi 782-8502, Japan b Research Reactor Institute, Kyoto University, Kumatori, Osaka 596-0821, Japan
Abstract The authors have found that a cellular defect structure is formed on the surface of Sn+ ion implanted GaSb at a low temperature and proposed its formation mechanism based on the movement of the induced point defects. This research was carried out in order to examine the validity of the mechanism by clarifying the effect of the mobility of the point defects on the defect formation. The defect structure on the GaSb surfaces implanted at cryogenic temperature and room temperature was investigated by scanning electron microscopy (SEM) and cross-sectional transmission electron microscopy (TEM) observation. In the sample implanted at room temperature, the sponge-like structure (a pileup of voids) was formed and the cellular structure, as observed at a low temperature, did not develop. This behavior was explained by the high mobility of the vacancies during implantation at room temperature, and the proposed idea that the defect formation process is dominated by the induced point defects was confirmed. r 2006 Elsevier B.V. All rights reserved. PACS: 61.72.Ji; 61.72.Qq; 61.72.Vv; 68.37.Hk; 68.37.Lp Keywords: GaSb; Ion-implantation; Cellular structure; Point defect; Radiation effect; Void; Cross-sectional transmission electron microscopy; Scanning electron microscopy
1. Introduction Defect structure formation in III–V compound semiconductors by ion-implantation is classified into two categories. In the case of GaAs and InP [1–3], a damaged layer is formed on the surface by ion-implantation; further implantation transforms the damaged layer to an amorphous structure. On the other hand, unusual behaviors, such as elevation and swelling, have been observed in GaSb and InSb when irradiated by energetic ions [4–7]. Taniwaki et al. observed an anomalous structure with a thickness 5 times the projected ion range on GaSb and InSb surfaces implanted with 60 keV Sn+ at a low temperature using cross-sectional TEM (transmission electron microscopy) [8]. Nitta et al. examined this defect structure in detail and clarified that it consists of many cells, like a honeycomb [9]. They showed that it was not created by sputtering nor by a vapor phase process, and proposed a defect formation Corresponding author. Tel.: +81 887 57 2504; fax: +81 887 57 2520.
E-mail address: [email protected] (N. Nitta). 0921-4526/$ - see front matter r 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.physb.2005.12.220
mechanism based on movement of the point defects induced by ion implantation. The validity of this mechanism is to be examined further by controlling the quantity and behavior of the induced point defects. In the previous paper, we showed that the defect structure enlarged when increasing the ion dose (i.e. the amount of the induced point defects), which suggests strongly that the defect formation is dominated by point defects [10]. In the present work, the effect of the mobility of point defects on the defect structure is studied by implantation at different substrate temperatures. Implantation was carried out at cryogenic temperature (130–150 K) and room temperature. There is little knowledge on the mobility of the point defects in GaSb, so that their behavior is not yet clarified. However Eisen’s work on the recovery of electron irradiated GaSb at liquid nitrogen temperature [11,12] gives some suggestion. From this work, the long-range migration of the vacancies are considered not to occur at temperatures of 130–150 K where the cellular structure was formed, and the vacancies will be mobile in a long range during implantation at room temperature. Therefore it is
ARTICLE IN PRESS 882
N. Nitta et al. / Physica B 376–377 (2006) 881–885
expected, if the proposed idea is right, that the defect structure in the samples implanted at room temperature differs from that at a low temperature. 2. Experiments Te-doped n-type compound semiconductor GaSb polished wafers with (1 0 0) orientation were implanted with 60 keV Sn+ ions. Implantation was performed using low ion doses (about 4 1014) and high doses (about 8 1014 ions/cm2), at low temperatures (LT sample, 151 and 137 K) and at room temperature (RT sample). As-implanted surfaces were observed through a field emission-type scanning electron microscope (FE-SEM), JEOL JSM-6400F. Cross-sectional TEM samples, prepared from implanted wafers by mechanical polishing and Ar ion-milling, were observed using a field emission-type transmission electron microscope (FE-TEM), JEOL JEM-2010F. 3. Results 3.1. SEM observation Fig. 1 shows the FE-SEM images of the GaSb samples implanted at 151 K and at room temperature. Their ion doses were almost the same (4.0 1014 and 4.1 1014 ions/ cm2, respectively). The image of the LT sample is weak, because its dark contrast does not reflect the surface roughness, but reflect the cavities just under the surface, which will be clearly shown later by cross-sectional TEM observation. These images were considerably different. A high density of isolated roundish holes with about 30 nm diameters was created in the LT sample. The hollows observed in the RT sample were not roundish, and seem to
have been formed by coalescence of several round holes. The densities of the holes or hollows are 5 109 and 4.5 109 cm 2, and the thickness of the walls partitioning those were about 20 nm in the LT sample and about 10 nm in the RT sample. FE-SEM images of the highly dosed GaSb surfaces are shown in Fig. 2. The ion doses were 7.9 1014 ions/cm2 for the LT sample and 8.1 1014 ions/cm2 for the RT sample. The contrast of the several holes in the LT sample is strong compared with that in the lower-dosed sample. This shows that the top surfaces of the holes were eliminated. In the case of room temperature implantation, the image of the hollows did not change much, though they were coarsened to the density of 2.7 109 cm 2. 3.2. Cross-sectional TEM observation Fig. 3 shows the (1 1 0) cross-sectional TEM views (bright field) and the selected area electron diffraction patterns of the GaSb samples implanted with low doses at 151 K and at room temperature. The ion doses were 4.0 1014 and 4.1 1014 ions/cm2, respectively. Although characteristic defects were formed in the surfaces of both samples, their defects were different from each other. Under the surface of the LT sample, large ellipsoidal voids with axes of 50 and 70 nm were arranged in partitions with 10–20 nm thick walls. The top surfaces of the voids are not open, therefore the dark contrast in the SEM image (Fig. 1) shows the voids but not the surface roughness. The cavities are formed ranging from the surface to a 130 nm depth in the RT sample, which look to be a pileup of many voids with about a 20 nm diameter. The crystalline structure around the voids and the cavities is polycrystalline in both samples.
Fig. 1. FE-SEM images of the GaSb surfaces implanted with low doses at 151 K and at room temperature are shown. The ion doses were 4.0 1014 and 4.1 1014 ions/cm2, respectively. Most holes in the sample implanted at room temperature are not round, which seem to be formed by coalescence of several round holes.
ARTICLE IN PRESS N. Nitta et al. / Physica B 376–377 (2006) 881–885
883
Fig. 2. FE-SEM images of the GaSb surfaces implanted with high doses at 137 K and at room temperature. The ion doses were 7.9 1014 and 8.1 1014 ions/cm2, respectively. As compared with Fig. 1, the cavities were developed in the sample implanted at 150 K. However, the image of holes did not change much at room temperature.
Fig. 3. The cross-sectional TEM view (bright field) and the selected area electron diffraction pattern of GaSb surface implanted at 151 K and at room temperature. The ion doses were 4.0 1014 and 4.1 1014 ions/cm2, respectively. Under the surface of the sample implanted at 151 K, large ellipsoidal voids with axes of 50 and 100 nm are arranged. Cavities consisting of many voids with 20–30 nm diameter are formed and are observed to a 100 nm depth in the sample implanted at room temperature.
The (1 1 0) cross-sectional views (bright field) of the GaSb samples implanted with high doses are shown in Fig. 4. The ion doses were 8.9 1014 and 8.1 1014 ions/cm2, respectively. The cellular structure, with a large aspect ratio, had formed on the surface of the LT sample. The diameter and the height of the cells were about 50 and 250 nm respectively, and the thickness of the partitioning cell walls was about 10 nm. A heavy strain region with a 50 nm thickness (strong contrast) can be observed under the cells and its interface with matrix is serrate. Though the defect structure is also created on the surface in the RT sample, it differs significantly from that observed in the LT sample. It is a sponge-like layer with a 180 nm thickness which were formed by a pileup of many voids with about a 20–30 nm
diameter as well as the defect observed in the lower-dosed sample at room temperature. Heavy strain contrast was not observed under the defect structure in this sample, and alternatively the voids with about a 5–10 nm diameter were formed. The structure of the walls was polycrystalline when dosed at room temperature and were amorphous when dosed at a low temperature. 4. Discussion As we expected, different defect structures were formed by implantation at a low temperature and at room temperature; the cellular structure was developed from the voids in LT sample and the sponge-like structure was
ARTICLE IN PRESS 884
N. Nitta et al. / Physica B 376–377 (2006) 881–885
Fig. 4. The cross-sectional TEM view (bright field) and the selected area electron diffraction pattern of GaSb surface implanted at 153 K and at room temperature. Those ion doses were 8.9 1014 and 8.1 1014 ions/cm2, respectively. An sponge-like defect structure is observed on the surface in the sample implanted at room temperature, which is a pileup of many voids with 20–100 nm diameter.
formed by sequential void formation in the RT sample. We intend to explain this difference in the defect structure by the mobility of the implantation-induced point defects. Eisen’s recovery data shows that there are five stages in the recovery of the irradiated III–IV compound semiconductors (I; 120 K, II; 160 K, III, IV; 200 K V; 360 K for GaSb) [11,12]. He mentioned that stages I and II correspond to the recovery of close Frenkel pairs, very small stages III and IV do to the recombination of somewhat far separated pairs and stage V must represent the movement of defects over greater distances than in the other stages. According to Eisen’s interpretation, at the temperature of 130–150 K where the cellular structure is formed, long-range migration of point defects does not occur. And at room temperature, point defects are considered to move in a long range to some extent, where we consider that these point defects are vacancies. According to this idea, voids are formed all over the implanted region by the aggregation of migrating vacancies during implantation at room temperature. Subsequent implantation creates new vacancies under the voids, which also gather and form new voids, and finally lead to the sponge-like structure in the RT samples. In the LT samples, the formation of voids is depressed because of the poor mobility of the vacancies. However, when the induced vacancy concentration has exceeded a critical value determined by the elastic property of GaSb, the voids are abruptly formed at a depth nearly equal to the ion projected range (the highest vacancy concentration). The vacancies induced by further implantation are absorbed by the preformed voids, which grow in the direction of the surface. The cellular defect structure of the heavily implanted sample at a low temperature is formed in this manner. In the above discussion, the difference of the defect structure in the samples implanted at different substrate temperatures was reasonably explained by the mobility of vacancies. Though the behavior of the interstitials was not
discussed here, we noted that the cellular structure was highly developed in the LT sample in Fig. 4. This suggests that the implantation-induced interstitials also significantly contribute to the formation of the cellular structure, which will be examined elsewhere. 5. Conclusion In order to examine the validity of the proposed formation mechanism of the cellular structure on GaSb surface implanted at a low temperature, the effect of substrate temperature on the defect structure was studied using SEM and TEM. At a low temperature, the voids formed in the early stage of implantation grow perpendicular to the surface absorbing newly created vacancies, so that the cell structure develops by increasing ion dose. At room temperature, the voids formed sequentially due to the high mobility of the vacancies leading to the formation of the sponge-like structure and the cellular structure does not develop. The results support our proposed mechanism based on the movement of the point defects for the formation of the cellular structure in ion-implanted GaSb surfaces at a low temperature. Acknowledgements We would like to thank Mr. N. Baba and Ms. S. Komatsu of Kochi Casio Co., Ltd. for their kind help with scanning electron microscope observation. References [1] M. Taniwaki, H. Koide, N. Yoshimoto, T. Yoshiie, S. Ohnuki, M. Maeda, K. Sassa, J. Appl. Phys. 67 (1990) 4036. [2] M. Taniwaki, H. Koide, T. Yoshiie, Y. Hayashi, H. Yoshida, J. NonCryst. Solids 117/118 (1990) 745. [3] M. Taniwaki, T. Yoshiie, H. Koide, M. Ichihasi, N. Yoshimoto, H. Yoshida, Y. Hayashi, J. Appl. Phys. (1989) 66.
ARTICLE IN PRESS N. Nitta et al. / Physica B 376–377 (2006) 881–885 [4] D. Kleitman, H.J. Yearian, Phys. Rev. 108 (1957) 901. [5] Y. Homma, J. Vac. Sci. Technol. A 5 (1987) 321. [6] R. Callec, P.N. Favennce, M. Salvi, H. L’Haridon, M. Gauneau, Appl. Phys. Lett. 59 (1991) 1872. [7] R. Callec, A. Poudoulec, J. Appl. Phys. 73 (1993) 4831. [8] M. Taniwaki, Y. Hayashi, T. Yoshiie, in: Proceedings of the International Conference on Solid–Solid Phase Transformations ’99, 1999, p. 425.
885
[9] N. Nitta, M. Taniwaki, T. Suzuki, Y. Hayashi, Y. Satoh, T. Yoshiie, Mater. Trans. 43 (2002) 674. [10] N. Nitta, M. Taniwaki, Y. Hayashi, T. Yoshiie, J. Appl. Phys. 92 (2002) 1799. [11] F.H. Eisen, Phys. Rev. 123 (1961) 736. [12] F.H. Eisen, in: J.W. Corbett, G.D. Watkins (Eds.), Radiation Effects in Semiconductors, Gordon and Breach, New York, 1971, p. 273.