- Email: [email protected]
Ion-induced changes in semiconductor properties of hydrogenated amorphous silicon
Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
Ion-induced changes in semiconductor properties of hydrogenated amorphous silicon Shin-ichiro Sato ⇑, Takeshi Ohshima Japan Atomic Energy Agency (JAEA), 1233 Watanuki, Takasaki, Gunma 370-1292, Japan
a r t i c l e
i n f o
Article history: Received 30 November 2012 Received in revised form 19 April 2013 Accepted 3 May 2013 Available online 18 June 2013 Keywords: Hydrogenated amorphous silicon Electric conductivity Seebeck effect Radiation effect Self-ion irradiation
a b s t r a c t Undoped, phosphorus doped (n-type), and boron doped (p-type) hydrogenated amorphous silicon (a-Si:H) thin films are irradiated with 3.0 MeV protons, 100 keV protons, and 2.8 MeV silicon ions, and the electric conductivity variations as a function of ion fluence are investigated. The Seebeck coefficient variations as a function of ion fluence are also investigated and are compared to the electric conductivity variations. As a result, a systematic interpretation of radiation effects on a-Si:H semiconductors is obtained. In the fluence regime of below 106 dpa, the increase in electric conductivity and the emergence of negative Seebeck effect are observed in the undoped a-Si:H because of donor-center generation. In the fluence regime from 106 dpa to 104 dpa, the decrease in electric conductivity and the decrease in absolute value of Seebeck coefficient are observed in the doped a-Si:H, since the carrier removal effect is caused by radiation defects, which are thought to be mainly dangling bonds. In the fluence regime of above 104 dpa, the increase in electric conductivity caused by the enhancement of hopping transport via localized states is observed. The absolute value of Seebeck coefficient of doped a-Si:H decreases in this fluence regime, whereas no Seebeck effect is observed in the undoped a-Si:H. Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction Because of the development of high energy physics and space technologies, high radiation-tolerance is increasingly required for electronic devices utilized in spacecrafts and nuclear and accelerator facilities [1–5]. Variations of semiconductor properties depending on irradiation fluence are fundamental and important knowledge to understand behavior of electronic devices operated under radiation environments. The knowledge is also essential to design the radiation-hardened devices, and thus variation of electric conductivity of crystalline silicon (c-Si) due to radiation exposure has been extensively studied [6,7]. On the other hand, investigation for radiation effects on amorphous type semiconductors including hydrogenated amorphous silicon (a-Si:H) is insufficient at the present stage, even though a-Si:H semiconductors are utilized as a material of thin-film transistors (TFTs), solar cells, and photo-detectors. Since a-Si:H devices are generally known to have higher radiation-tolerance than c-Si devices [8–10], a-Si:H-based devices are one of the strong candidates of radiationtolerant photoelectric devices. Owing to these facts, radiation effects on a-Si:H thin films have been investigated by several research groups [11–14]. However, the details are still unclear. In particular, little is known about vari⇑ Corresponding author. Tel.: +81 27 346 9421; fax: +81 27 346 9687. E-mail address: [email protected] (S.-i. Sato). 0168-583X/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nimb.2013.05.042
ations of semiconductor properties of a-Si:H depending on irradiation fluence even though comprehensive study is required to obtain a systematic interpretation. Accordingly, we have investigated electric conductivity and photoconductivity of a-Si:H irradiated with charged particles and have obtained systematic understandings of the radiation effects on a-Si:H [15,16]. In this paper, electric conductivity and Seebeck coefficient variations of a-Si:H depending on irradiation fluence are comprehensively investigated. Ion fluence and ion species as well as impurity-doping conditions of samples are used as experimental parameters. In particular, continuous variations of Seebeck coefficient of a-Si:H as a function of ion fluence are clarified for the first time. Since the Seebeck coefficient depends on the majority carrier concentration and the electronic transport mechanism, a systematic understanding of radiation effects on semiconductor properties is obtained from the relationship between variations of electric conductivity and Seebeck coefficient.
2. Experimental The samples used in this study were device-grade undoped, n-type (phosphorous doped) and p-type (boron doped) a-Si:H thin films fabricated on glass substrates by plasma enhanced chemical vapor deposition (PECVD). The excitation frequency was 13.56 MHz. The substrate temperature during deposition and the
154
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
gas flow rates were 180 °C and SiH4/H2 = 20/100 sccm for undoped samples, 195 °C and SiH4/H2/PH3 = 20/80/23 sccm for n-type samples, and 200 °C and SiH4/H2/B2H6 = 10/100/30 sccm for p-type samples, respectively (PH3 and B2H6 are 5000 ppm mixtures with hydrogen balance gas). Interdigitated type and coplanar type aluminum Ohmic electrodes were formed on the samples for conductivity measurement and for thermoelectric power measurement, respectively. The thicknesses were 0.30 lm for undoped, 0.27 lm for n-type, and 0.21 lm for p-type a-Si:Hs. The majority carrier concentrations of 6.2 1017 cm3 for n-type and 6.6 1013 cm3 for p-type a-Si:Hs were determined by Hall measurement. The samples were irradiated with 3.0 MeV protons, 100 keV protons, and 2.8 MeV silicon (Si) ions at room temperature (RT). The electric conductivity variation and the Seebeck coefficient variation as a function of irradiation fluence were investigated in situ in an irradiation vacuum chamber. Ion irradiation was performed at the Takasaki Ion Accelerators of advanced Radiation Application (TIARA), Japan Atomic Energy Agency (JAEA). A raster beam scanning system was used for uniform irradiation of the whole area of a sample. The fluctuation of beam uniformity was estimated to be within ±5%. The projected ranges of all the ions used in this study are beyond the thickness of the a-Si:H film and deposit their energy almost uniformly through the film, according to the Monte Carlo simulation code, TRIM [17]. Thus, no passivation by the implanted hydrogen atoms of dangling bonds in the a-Si:H films is expected. Also, a unit of displacement per atom (dpa) was applied in order to compare results in different irradiation conditions. The conversion ratios are 1 dpa = 8.1 1019 cm2, 4.5 1018 cm2, and 9.8 1015 cm2 for 3.0 MeV protons, 100 keV protons, and 2.8 MeV Si ions, respectively. In the TRIM calculation, the mass density of 2.3 g/cm3 and the hydrogen concentration of 11.6% were used for undoped a-Si:H. These values were experimentally determined by using Rutherford backscattering spectroscopy (RBS) and elastic recoil detection analysis (ERDA). The default values of displacement energy installed in TRIM were used: 15 eV for Si and 10 eV for H. The same values were also applied for the analysis of n-type and p-type a-Si:Hs, since the difference was sufficiently small. The current–voltage (I-V) characteristics and the thermoelectric power of the samples were measured in situ under dark conditions. The conductivity was derived from the slope of I-V characteristics and the Seebeck coefficient was derived from the ratio of the thermoelectric power divided by the temperature difference. The uncertainty of obtained data is estimated to be ±3% for electric conductivity and ±10% for Seebeck coefficient, which were mainly due to the uncertainty of temperature. The ion irradiation was paused during measurement and was resumed after completion of the measurement. Details of the experimental procedure are described elsewhere [18,19].
No Seebeck effect was observed in the undoped a-Si:H without irradiation.
Fig. 1. Seebeck coefficient variations of a-Si:H due to 3.0 MeV proton irradiation. Open squares, shaded circles, and shaded triangles denote the results of undoped, n-type, and p-type a-Si:Hs, respectively. The abscissa axis in the upper part is converted from fluence to a unit of displacement per atom (dpa). Solid lines are to guide the eye.
Comparison between the electric conductivity and the Seebeck coefficient of the undoped a-Si:H irradiated with 3.0 MeV protons are shown in Fig. 2. The electric conductivity drastically increased with increasing fluence and reached 1.5 105 S/cm at the fluence of 5.0 1012 cm2 (6.2 108 dpa), and after that decreased with further irradiation. However, the electric conductivity even at the fluence of 1.0 1014 cm2 (1.2 106 dpa) was around two hundred times higher than that before irradiation. As shown in Fig. 2, the Seebeck effect was observed in the fluence regime where the electric conductivity was enhanced. The electric conductivity decreased and the absolute value of Seebeck coefficient increased at above 1013 cm2 (1.2 107 dpa). The drastic increase in electric conductivity at the fluence of around 5 1012 cm2 was not caused by a charging effect of protons in the sample. The conductivity variation of a dummy sample (glass substrate with
Undoped a-Si:H
3. Results Fig. 1 shows the Seebeck coefficient variations of a-Si:H as a function of 3.0 MeV proton fluence. Not observed in the undoped a-Si:H before irradiation, the negative Seebeck effect appeared after the irradiation of 1.0 1011 cm2 (1.2 109 dpa). The absolute value of Seebeck coefficient increased at the fluence of above 1013 cm2 and could be observed at the fluence up to 2.0 1014 cm2 (2.5 106 dpa). The absolute value of Seebeck coefficient of n-type a-Si:H similarly increased at the fluence of above 1013 cm2 (1.2 107 dpa). On the other hand, the absolute value of Seebeck coefficient of p-type a-Si:H gradually increased with increasing fluence and could not be observed at the fluence of above 1014 cm2 (1.2 106 dpa).
Fig. 2. Variations of electric conductivity (closed squares) and Seebeck coefficient (open squares) of undoped a-Si:H due to 3.0 MeV proton irradiation. Error bars on the data of electric conductivity are not shown since these are sufficiently smaller than the displaying symbol size. The abscissa axis in the upper part is converted from fluence to dpa. Solid lines are to guide the eye. The same is true in Figs. 3–6.
155
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
n-type a-Si:H
Undoped a-Si:H
Fig. 3. Variations of electric conductivity (closed squares) and Seebeck coefficient (open squares) of undoped a-Si:H due to 100 keV proton irradiation. The fluence regime in which the Seebeck effect was observable is shown as shaded zone.
Fig. 5. Variations of electric conductivity (closed circles) and Seebeck coefficient (open circles) of n-type a-Si:H due to 2.8 MeV Si irradiation.
interdigitated Al electrodes) during proton irradiation was investigated and the result showed that no significant change was observed. Therefore, the increase in electric conductivity is attributed to the radiation effects of a-Si:H. Fig. 3 shows the electric conductivity and the Seebeck coefficient variations of undoped a-Si:H due to 100 keV proton irradiation. The negative Seebeck effect was observed only in the fluence regime where the conductivity had a small peak (shaded zone between 1011 cm2 and 1013 cm2 in Fig. 3). The electric conductivity increased at above 1.0 1014 cm2 (2.2 105 dpa), although no Seebeck effect was observed in this fluence regime. The results of n-type a-Si:H irradiated with 100 keV proton are shown in Fig. 4. The electric conductivity decreased and the absolute value of Seebeck coefficient increased at the fluence of above 1012 cm2 (2.2 107 dpa). However, the absolute value of Seebeck coefficient reached the maximum at the fluence of 2.0 1015 cm2 (4.4 104 dpa) and then decreased with increasing the fluence. Here, radiation effects on the glass substrate should be considered, since generally the insulation property of glass materials degrades due to irradiation. In other words, the electric conductivity of glass substrate increases with increasing
radiation defects. This phenomenon is called radiation induced electrical degradation (RIED) [20] and RIED of the glass substrate might affect the electric conductivity measurement of a-Si:H thin film. In order to investigate RIED of the glass substrate, the dummy sample was irradiated with 100 keV protons at the fluence of 1.0 1016 cm2 and the electric conductivity was measured. As a result, the electric conductivity of glass substrate was less than the detection limit of this study, suggesting that the glass substrates kept its electrical insulation in the high fluence regime and the increase in electric conductivity definitely resulted from the radiation effects of a-Si:H. The results of n-type and p-type a-Si:Hs irradiated with 2.8 MeV Si ions are shown in Figs. 5 and 6, respectively. In both cases, the electric conductivity decreased and the absolute value of Seebeck coefficient increased in the fluence regime up to 1.0 1012 cm2 (1.0 104 dpa). However, these changes with increasing the fluence were reversed at the fluence of above 1.0 1012 cm2 (1.0 104 dpa) and eventually the Seebeck effect was unobservable at the fluence above 2 1015 cm2 (0.2 dpa). Also, no Seebeck effect appeared in the undoped a-Si:H irradiated with 2.8 MeV Si ions.
n-type a-Si:H
Fig. 4. Variations of electric conductivity (closed circles) and Seebeck coefficient (open circles) of n-type a-Si:H due to 100 keV proton irradiation.
p-type a-Si:H
Fig. 6. Variations of electric conductivity (closed triangles) and Seebeck coefficient (open triangles) of p-type a-Si:H due to 2.8 MeV Si irradiation.
156
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
4. Discussion In this section, fundamental understandings of radiation effects on electric conductivity and Seebeck effect in semiconductor materials are briefly explained. After that, these variations of a-Si:H as a function of ion fluence are discussed. Electric conductivity of semiconductors is represented as products of the elementary charge, the mobility of majority carriers, and the majority carrier concentration. When exposed to radiations, structural disorders such as atomic displacement and defects are produced in semiconductor materials because of the displacement damage effect. Radiation defects usually act as deep levels near to the intrinsic Fermi level and may also provide a decrease in majority carrier concentration of impurity-doped semiconductors by compensating the donor or acceptor levels. This is called carrier removal effect [21]. Since the carrier scattering mechanism does not change in the low fluence regime (e.g. less than 5 106 dpa in the case of crystalline silicon [7]), the change in electric conductivity due to radiations is mainly attributed to the change in majority carrier concentration. Next, the Seebeck coefficient of semiconductors, S is represented as follows:
S¼
DV k B N ¼ In þ A n DT e
ð1Þ
where DV, DT, kB, N, n, and A are the potential difference between electrodes (V), the temperature difference between electrodes (K), the Boltzmann’s constant, the effective density of state (cm3), majority carrier concentration (cm3), and the parameter related to carrier transport. The Seebeck coefficient of an n-type semiconductor is negative and a p-type is positive. The absolute value of Seebeck coefficient is higher as the majority carrier concentration is higher. Accordingly, the absolute value of Seebeck coefficient increases with increasing radiation defects. However, the absolute value of Seebeck coefficient decreases in the case that the density of localized states within bandgap (mobility-edges) extremely increases, since the parameter related to carrier transport decreases when the hopping transport via the localized states is dominant in the electric conduction. In Fig. 1, the absolute value of Seebeck coefficient of n-type and p-type a-Si:Hs increased with increasing 3.0 MeV proton fluence. Main radiation defects in a-Si:H are dangling bonds [22,23] and they provide the carrier removal effect which induces the decrease in majority carrier concentration. In the case of undoped a-Si:H, the negative Seebeck effect was detected at the fluence of above 1011 cm2, although no Seebeck effect was observed before irradiation. This result strongly suggests that donor-centers as well as dangling bonds were created by proton irradiation. This concept is supported by the results of Figs. 2 and 3. The fluence regime in which the negative Seebeck effect emerged corresponded to that in which the electric conductivity increased. Both can be attributed to the increase in electron carrier concentration. The effects of donor-center generation on the n-type a-Si:H was less pronounced than that on the undoped a-Si:H, since the initial majority carrier concentration was much higher (6.2 1017 cm3 for Hall measurement). However, the slight increase in electric conductivity was observed in the fluence regime of around 1012 cm2 and was also due to the donor-center generation [24]. The drastic decrease in electric conductivity and the increase in absolute value of Seebeck coefficient which were observed at above 1.0 1012 cm2 (2.2 107 dpa) were due to the carrier removal effect. On the other hand, no evidence of donor-center generation was observed in the case of 2.8 MeV Si ion irradiation (see Figs. 5 and 6). Only the carrier removal effect was observed at the fluence up to 1012 cm2 (104 dpa). However, both the variations of
electric conductivity and Seebeck coefficient as a function of Si ion fluence were reversed at above 1012 cm2 (104 dpa). These change inversions mean that the band transport, which is the main electronic transport mechanism in semiconductors, is undermined by the increase in localized states within mobility-edges while the hopping transport becomes dominant. The transition from the band transport to the hopping transport provides the decrease in parameter related to carrier transport, A, and results in the decrease in absolute value of Seebeck coefficient. In addition, the electronic transport transition was also observed in the case of 100 keV proton irradiation. As shown in Fig. 3, no Seebeck effect was detected in the high fluence regime, since the drastic increase in electric conductivity at above 1.0 1014 cm2 (2.2 105 dpa) was not due to the increase in majority carrier concentration, but due to the enhancement in hopping transport. The decrease in absolute value of Seebeck coefficient at above 2.0 1015 cm2 (4.4 104 dpa) which was observed in Fig. 4 could be interpreted as the decrease in carrier transport parameter, A. It can be concluded from the above that the variations of electric conductivity and Seebeck coefficient due to ion irradiation are systematically classified as follows. Firstly, in the case of undoped and n-type a-Si:Hs irradiated with protons, the increase in electric conductivity and the emergence of negative Seebeck effect are observed at the low fluence regime (around 108 to 106 dpa) because of the donor-center generation, which is thought to be caused by the electronic excitation effect of incident protons [24,25]. Secondly, both the decrease in electric conductivity and the increase in absolute value of Seebeck coefficient are observed in the middle fluence regime between 106 and 104 dpa. These are attributed to the carrier removal effect of accumulated dangling bonds. Thirdly, the density of localized states created by dangling bonds becomes extremely high at the fluence regime of above 104 dpa and it results in the enhancement of hopping transport. In this fluence regime, the dangling bond density of irradiated a-Si:H is estimated to increase to around 1020 cm3, since the electric conduction in hydrogen-free amorphous silicon, which contains around 1020 cm3 of dangling bonds, is dominated by the hopping transport [26]. The enhancement of hopping transport provides the increase in electric conductivity and the decrease in absolute value of Seebeck coefficient. Finally, the Seebeck effect disappears almost completely at above 0.2 dpa even in the case of impurity doped a-Si:H.
5. Conclusion The electric conductivity and the Seebeck coefficient variations of a-Si:H due to proton and Si ion irradiation were investigated using the in situ measurement techniques. These results were compared and interpreted systematically. In the case of undoped a-Si:H, the Seebeck effect is observed only in the fluence regime where the electric conductivity is enhanced. This is due to donor-centers generated by proton irradiation. Further irradiation induces the decrease in electric conductivity based on the carrier removal effect. However, the increase in electric conductivity associated with the enhancement of hopping transport appears in the high fluence regime of above 104 dpa. In the high fluence regime, no Seebeck effect is observed. In the case of the doped a-Si:H, the Seebeck coefficient shows an opposite change from the conductivity. That is, the decrease in electric conductivity and the increase in absolute value of Seebeck coefficient are observed in the fluence regime of below 104 dpa, and the increase in electric conductivity and the decrease in absolute value of Seebeck coefficient are observed at above 104 dpa. The former one is due to the carrier removal effect and the latter one is caused by the enhancement of hopping transport. Both
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
phenomena are simply attributed to the accumulation of danglingbonds (radiation defects). Acknowledgements The authors would like to thank Dr. Hitoshi Sai and Dr. Michio Kondo of National Institute of Advanced Industrial Science and Technology (AIST) for fabricating the a-Si:H samples. The authors also would like to thank Dr. Kazunori Shimazaki and Dr. Mitsuru Imaizumi of Japan Aerospace Exploration Agency (JAXA) for their technical support and fruitful discussion. References
[10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
[1] P.J. Sellin, J. Vaitkus, Nucl. Instr. Meth. A 557 (2006) 479–489. [2] W. Fernando, Nucl. Instr. Meth. A 596 (2008) 58–62. [3] R. Velazco, P. Fouillat, R. Reis, Radiation Effects on Embedded Systems, Springer, Dordrecht, 2007. p. 11. [4] A. Holmes-Siedle, L. Adams, Handbook of Radiation Effects, Second ed., Oxford University Press, Oxford, 2007. p. 17. [5] M. Imaizumi, S. Matsuda, S. Kawakita, T. Sumita, T. Takamoto, T. Ohshima, M. Yamaguchi , Prog. Photovolt: Res. Appl. 13 (2005) 529–543. [6] H. Amekura, N. Kishimoto, K. Kono, J. Appl. Phys. 84 (1998) 4834–4841. [7] H. Amekura, N. Kishimoto, K. Kono, IEEE Trans. Nucl. Sci. 45 (1998) 1508–1513. [8] S. Sato, H. Sai, T. Ohshima, M. Imaizumi, K. Shimazaki, M. Kondo, Prog. Photovolt. Res. Appl. (2012). [9] A. Banerjee, X. Xu, K. Beernink, F. Liu, K. Lord, G. DeMaggio, B. Yan, T. Su, G. Pietka, C. Worrel, S. Ehlert, D. Beglau, J. Yang, S. Guha, Proceedings of 35th
[20] [21] [22] [23] [24] [25] [26]
157
Photovoltaic Specialists Conference (PVSC-35), Philadelphia, USA, 2010, p. 002651. N. Wrysch, C. Miazza, S. Dunand, C. Ballif, A. Shah, M. Despeisse, D. Moraes, F. Powolny, P. Jarron, J. Non-Cryst. Solids 352 (2006) 1797–1800. E. Morgado, R. Schwarz, T. Braz, C. Casteleiro, A. Maçarico, M. Vieira, E. Alves, J. Non-Cryst. Solids 352 (2006) 1071–1074. P. Danesh, B. Pantchev, E. Vlaikova, Nucl. Instr. Meth. B 239 (2005) 370–374. N. Kishimoto, H. Amekura, K. Kono, C.G. Lee, J. Non-Cryst. Solids 227–230 (1998) 238–242. S. Coffa, F. Priolo, A. Battaglia, Phys. Rev. Lett. 70 (1993) 3756–3759. S. Sato, H. Sai, T. Ohshima, M. Imaizumi, K. Shimazaki, M. Kondo, Nucl. Instr. Meth. B 286 (2012) 29–34. S. Sato, H. Sai, T. Ohshima, M. Imaizumi, K. Shimazaki, M. Kondo, J. Non-Cryst. Solids 356 (2012) 2039–2043.
Contents lists available at SciVerse ScienceDirect
Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb
Ion-induced changes in semiconductor properties of hydrogenated amorphous silicon Shin-ichiro Sato ⇑, Takeshi Ohshima Japan Atomic Energy Agency (JAEA), 1233 Watanuki, Takasaki, Gunma 370-1292, Japan
a r t i c l e
i n f o
Article history: Received 30 November 2012 Received in revised form 19 April 2013 Accepted 3 May 2013 Available online 18 June 2013 Keywords: Hydrogenated amorphous silicon Electric conductivity Seebeck effect Radiation effect Self-ion irradiation
a b s t r a c t Undoped, phosphorus doped (n-type), and boron doped (p-type) hydrogenated amorphous silicon (a-Si:H) thin films are irradiated with 3.0 MeV protons, 100 keV protons, and 2.8 MeV silicon ions, and the electric conductivity variations as a function of ion fluence are investigated. The Seebeck coefficient variations as a function of ion fluence are also investigated and are compared to the electric conductivity variations. As a result, a systematic interpretation of radiation effects on a-Si:H semiconductors is obtained. In the fluence regime of below 106 dpa, the increase in electric conductivity and the emergence of negative Seebeck effect are observed in the undoped a-Si:H because of donor-center generation. In the fluence regime from 106 dpa to 104 dpa, the decrease in electric conductivity and the decrease in absolute value of Seebeck coefficient are observed in the doped a-Si:H, since the carrier removal effect is caused by radiation defects, which are thought to be mainly dangling bonds. In the fluence regime of above 104 dpa, the increase in electric conductivity caused by the enhancement of hopping transport via localized states is observed. The absolute value of Seebeck coefficient of doped a-Si:H decreases in this fluence regime, whereas no Seebeck effect is observed in the undoped a-Si:H. Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction Because of the development of high energy physics and space technologies, high radiation-tolerance is increasingly required for electronic devices utilized in spacecrafts and nuclear and accelerator facilities [1–5]. Variations of semiconductor properties depending on irradiation fluence are fundamental and important knowledge to understand behavior of electronic devices operated under radiation environments. The knowledge is also essential to design the radiation-hardened devices, and thus variation of electric conductivity of crystalline silicon (c-Si) due to radiation exposure has been extensively studied [6,7]. On the other hand, investigation for radiation effects on amorphous type semiconductors including hydrogenated amorphous silicon (a-Si:H) is insufficient at the present stage, even though a-Si:H semiconductors are utilized as a material of thin-film transistors (TFTs), solar cells, and photo-detectors. Since a-Si:H devices are generally known to have higher radiation-tolerance than c-Si devices [8–10], a-Si:H-based devices are one of the strong candidates of radiationtolerant photoelectric devices. Owing to these facts, radiation effects on a-Si:H thin films have been investigated by several research groups [11–14]. However, the details are still unclear. In particular, little is known about vari⇑ Corresponding author. Tel.: +81 27 346 9421; fax: +81 27 346 9687. E-mail address: [email protected] (S.-i. Sato). 0168-583X/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nimb.2013.05.042
ations of semiconductor properties of a-Si:H depending on irradiation fluence even though comprehensive study is required to obtain a systematic interpretation. Accordingly, we have investigated electric conductivity and photoconductivity of a-Si:H irradiated with charged particles and have obtained systematic understandings of the radiation effects on a-Si:H [15,16]. In this paper, electric conductivity and Seebeck coefficient variations of a-Si:H depending on irradiation fluence are comprehensively investigated. Ion fluence and ion species as well as impurity-doping conditions of samples are used as experimental parameters. In particular, continuous variations of Seebeck coefficient of a-Si:H as a function of ion fluence are clarified for the first time. Since the Seebeck coefficient depends on the majority carrier concentration and the electronic transport mechanism, a systematic understanding of radiation effects on semiconductor properties is obtained from the relationship between variations of electric conductivity and Seebeck coefficient.
2. Experimental The samples used in this study were device-grade undoped, n-type (phosphorous doped) and p-type (boron doped) a-Si:H thin films fabricated on glass substrates by plasma enhanced chemical vapor deposition (PECVD). The excitation frequency was 13.56 MHz. The substrate temperature during deposition and the
154
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
gas flow rates were 180 °C and SiH4/H2 = 20/100 sccm for undoped samples, 195 °C and SiH4/H2/PH3 = 20/80/23 sccm for n-type samples, and 200 °C and SiH4/H2/B2H6 = 10/100/30 sccm for p-type samples, respectively (PH3 and B2H6 are 5000 ppm mixtures with hydrogen balance gas). Interdigitated type and coplanar type aluminum Ohmic electrodes were formed on the samples for conductivity measurement and for thermoelectric power measurement, respectively. The thicknesses were 0.30 lm for undoped, 0.27 lm for n-type, and 0.21 lm for p-type a-Si:Hs. The majority carrier concentrations of 6.2 1017 cm3 for n-type and 6.6 1013 cm3 for p-type a-Si:Hs were determined by Hall measurement. The samples were irradiated with 3.0 MeV protons, 100 keV protons, and 2.8 MeV silicon (Si) ions at room temperature (RT). The electric conductivity variation and the Seebeck coefficient variation as a function of irradiation fluence were investigated in situ in an irradiation vacuum chamber. Ion irradiation was performed at the Takasaki Ion Accelerators of advanced Radiation Application (TIARA), Japan Atomic Energy Agency (JAEA). A raster beam scanning system was used for uniform irradiation of the whole area of a sample. The fluctuation of beam uniformity was estimated to be within ±5%. The projected ranges of all the ions used in this study are beyond the thickness of the a-Si:H film and deposit their energy almost uniformly through the film, according to the Monte Carlo simulation code, TRIM [17]. Thus, no passivation by the implanted hydrogen atoms of dangling bonds in the a-Si:H films is expected. Also, a unit of displacement per atom (dpa) was applied in order to compare results in different irradiation conditions. The conversion ratios are 1 dpa = 8.1 1019 cm2, 4.5 1018 cm2, and 9.8 1015 cm2 for 3.0 MeV protons, 100 keV protons, and 2.8 MeV Si ions, respectively. In the TRIM calculation, the mass density of 2.3 g/cm3 and the hydrogen concentration of 11.6% were used for undoped a-Si:H. These values were experimentally determined by using Rutherford backscattering spectroscopy (RBS) and elastic recoil detection analysis (ERDA). The default values of displacement energy installed in TRIM were used: 15 eV for Si and 10 eV for H. The same values were also applied for the analysis of n-type and p-type a-Si:Hs, since the difference was sufficiently small. The current–voltage (I-V) characteristics and the thermoelectric power of the samples were measured in situ under dark conditions. The conductivity was derived from the slope of I-V characteristics and the Seebeck coefficient was derived from the ratio of the thermoelectric power divided by the temperature difference. The uncertainty of obtained data is estimated to be ±3% for electric conductivity and ±10% for Seebeck coefficient, which were mainly due to the uncertainty of temperature. The ion irradiation was paused during measurement and was resumed after completion of the measurement. Details of the experimental procedure are described elsewhere [18,19].
No Seebeck effect was observed in the undoped a-Si:H without irradiation.
Fig. 1. Seebeck coefficient variations of a-Si:H due to 3.0 MeV proton irradiation. Open squares, shaded circles, and shaded triangles denote the results of undoped, n-type, and p-type a-Si:Hs, respectively. The abscissa axis in the upper part is converted from fluence to a unit of displacement per atom (dpa). Solid lines are to guide the eye.
Comparison between the electric conductivity and the Seebeck coefficient of the undoped a-Si:H irradiated with 3.0 MeV protons are shown in Fig. 2. The electric conductivity drastically increased with increasing fluence and reached 1.5 105 S/cm at the fluence of 5.0 1012 cm2 (6.2 108 dpa), and after that decreased with further irradiation. However, the electric conductivity even at the fluence of 1.0 1014 cm2 (1.2 106 dpa) was around two hundred times higher than that before irradiation. As shown in Fig. 2, the Seebeck effect was observed in the fluence regime where the electric conductivity was enhanced. The electric conductivity decreased and the absolute value of Seebeck coefficient increased at above 1013 cm2 (1.2 107 dpa). The drastic increase in electric conductivity at the fluence of around 5 1012 cm2 was not caused by a charging effect of protons in the sample. The conductivity variation of a dummy sample (glass substrate with
Undoped a-Si:H
3. Results Fig. 1 shows the Seebeck coefficient variations of a-Si:H as a function of 3.0 MeV proton fluence. Not observed in the undoped a-Si:H before irradiation, the negative Seebeck effect appeared after the irradiation of 1.0 1011 cm2 (1.2 109 dpa). The absolute value of Seebeck coefficient increased at the fluence of above 1013 cm2 and could be observed at the fluence up to 2.0 1014 cm2 (2.5 106 dpa). The absolute value of Seebeck coefficient of n-type a-Si:H similarly increased at the fluence of above 1013 cm2 (1.2 107 dpa). On the other hand, the absolute value of Seebeck coefficient of p-type a-Si:H gradually increased with increasing fluence and could not be observed at the fluence of above 1014 cm2 (1.2 106 dpa).
Fig. 2. Variations of electric conductivity (closed squares) and Seebeck coefficient (open squares) of undoped a-Si:H due to 3.0 MeV proton irradiation. Error bars on the data of electric conductivity are not shown since these are sufficiently smaller than the displaying symbol size. The abscissa axis in the upper part is converted from fluence to dpa. Solid lines are to guide the eye. The same is true in Figs. 3–6.
155
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
n-type a-Si:H
Undoped a-Si:H
Fig. 3. Variations of electric conductivity (closed squares) and Seebeck coefficient (open squares) of undoped a-Si:H due to 100 keV proton irradiation. The fluence regime in which the Seebeck effect was observable is shown as shaded zone.
Fig. 5. Variations of electric conductivity (closed circles) and Seebeck coefficient (open circles) of n-type a-Si:H due to 2.8 MeV Si irradiation.
interdigitated Al electrodes) during proton irradiation was investigated and the result showed that no significant change was observed. Therefore, the increase in electric conductivity is attributed to the radiation effects of a-Si:H. Fig. 3 shows the electric conductivity and the Seebeck coefficient variations of undoped a-Si:H due to 100 keV proton irradiation. The negative Seebeck effect was observed only in the fluence regime where the conductivity had a small peak (shaded zone between 1011 cm2 and 1013 cm2 in Fig. 3). The electric conductivity increased at above 1.0 1014 cm2 (2.2 105 dpa), although no Seebeck effect was observed in this fluence regime. The results of n-type a-Si:H irradiated with 100 keV proton are shown in Fig. 4. The electric conductivity decreased and the absolute value of Seebeck coefficient increased at the fluence of above 1012 cm2 (2.2 107 dpa). However, the absolute value of Seebeck coefficient reached the maximum at the fluence of 2.0 1015 cm2 (4.4 104 dpa) and then decreased with increasing the fluence. Here, radiation effects on the glass substrate should be considered, since generally the insulation property of glass materials degrades due to irradiation. In other words, the electric conductivity of glass substrate increases with increasing
radiation defects. This phenomenon is called radiation induced electrical degradation (RIED) [20] and RIED of the glass substrate might affect the electric conductivity measurement of a-Si:H thin film. In order to investigate RIED of the glass substrate, the dummy sample was irradiated with 100 keV protons at the fluence of 1.0 1016 cm2 and the electric conductivity was measured. As a result, the electric conductivity of glass substrate was less than the detection limit of this study, suggesting that the glass substrates kept its electrical insulation in the high fluence regime and the increase in electric conductivity definitely resulted from the radiation effects of a-Si:H. The results of n-type and p-type a-Si:Hs irradiated with 2.8 MeV Si ions are shown in Figs. 5 and 6, respectively. In both cases, the electric conductivity decreased and the absolute value of Seebeck coefficient increased in the fluence regime up to 1.0 1012 cm2 (1.0 104 dpa). However, these changes with increasing the fluence were reversed at the fluence of above 1.0 1012 cm2 (1.0 104 dpa) and eventually the Seebeck effect was unobservable at the fluence above 2 1015 cm2 (0.2 dpa). Also, no Seebeck effect appeared in the undoped a-Si:H irradiated with 2.8 MeV Si ions.
n-type a-Si:H
Fig. 4. Variations of electric conductivity (closed circles) and Seebeck coefficient (open circles) of n-type a-Si:H due to 100 keV proton irradiation.
p-type a-Si:H
Fig. 6. Variations of electric conductivity (closed triangles) and Seebeck coefficient (open triangles) of p-type a-Si:H due to 2.8 MeV Si irradiation.
156
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
4. Discussion In this section, fundamental understandings of radiation effects on electric conductivity and Seebeck effect in semiconductor materials are briefly explained. After that, these variations of a-Si:H as a function of ion fluence are discussed. Electric conductivity of semiconductors is represented as products of the elementary charge, the mobility of majority carriers, and the majority carrier concentration. When exposed to radiations, structural disorders such as atomic displacement and defects are produced in semiconductor materials because of the displacement damage effect. Radiation defects usually act as deep levels near to the intrinsic Fermi level and may also provide a decrease in majority carrier concentration of impurity-doped semiconductors by compensating the donor or acceptor levels. This is called carrier removal effect [21]. Since the carrier scattering mechanism does not change in the low fluence regime (e.g. less than 5 106 dpa in the case of crystalline silicon [7]), the change in electric conductivity due to radiations is mainly attributed to the change in majority carrier concentration. Next, the Seebeck coefficient of semiconductors, S is represented as follows:
S¼
DV k B N ¼ In þ A n DT e
ð1Þ
where DV, DT, kB, N, n, and A are the potential difference between electrodes (V), the temperature difference between electrodes (K), the Boltzmann’s constant, the effective density of state (cm3), majority carrier concentration (cm3), and the parameter related to carrier transport. The Seebeck coefficient of an n-type semiconductor is negative and a p-type is positive. The absolute value of Seebeck coefficient is higher as the majority carrier concentration is higher. Accordingly, the absolute value of Seebeck coefficient increases with increasing radiation defects. However, the absolute value of Seebeck coefficient decreases in the case that the density of localized states within bandgap (mobility-edges) extremely increases, since the parameter related to carrier transport decreases when the hopping transport via the localized states is dominant in the electric conduction. In Fig. 1, the absolute value of Seebeck coefficient of n-type and p-type a-Si:Hs increased with increasing 3.0 MeV proton fluence. Main radiation defects in a-Si:H are dangling bonds [22,23] and they provide the carrier removal effect which induces the decrease in majority carrier concentration. In the case of undoped a-Si:H, the negative Seebeck effect was detected at the fluence of above 1011 cm2, although no Seebeck effect was observed before irradiation. This result strongly suggests that donor-centers as well as dangling bonds were created by proton irradiation. This concept is supported by the results of Figs. 2 and 3. The fluence regime in which the negative Seebeck effect emerged corresponded to that in which the electric conductivity increased. Both can be attributed to the increase in electron carrier concentration. The effects of donor-center generation on the n-type a-Si:H was less pronounced than that on the undoped a-Si:H, since the initial majority carrier concentration was much higher (6.2 1017 cm3 for Hall measurement). However, the slight increase in electric conductivity was observed in the fluence regime of around 1012 cm2 and was also due to the donor-center generation [24]. The drastic decrease in electric conductivity and the increase in absolute value of Seebeck coefficient which were observed at above 1.0 1012 cm2 (2.2 107 dpa) were due to the carrier removal effect. On the other hand, no evidence of donor-center generation was observed in the case of 2.8 MeV Si ion irradiation (see Figs. 5 and 6). Only the carrier removal effect was observed at the fluence up to 1012 cm2 (104 dpa). However, both the variations of
electric conductivity and Seebeck coefficient as a function of Si ion fluence were reversed at above 1012 cm2 (104 dpa). These change inversions mean that the band transport, which is the main electronic transport mechanism in semiconductors, is undermined by the increase in localized states within mobility-edges while the hopping transport becomes dominant. The transition from the band transport to the hopping transport provides the decrease in parameter related to carrier transport, A, and results in the decrease in absolute value of Seebeck coefficient. In addition, the electronic transport transition was also observed in the case of 100 keV proton irradiation. As shown in Fig. 3, no Seebeck effect was detected in the high fluence regime, since the drastic increase in electric conductivity at above 1.0 1014 cm2 (2.2 105 dpa) was not due to the increase in majority carrier concentration, but due to the enhancement in hopping transport. The decrease in absolute value of Seebeck coefficient at above 2.0 1015 cm2 (4.4 104 dpa) which was observed in Fig. 4 could be interpreted as the decrease in carrier transport parameter, A. It can be concluded from the above that the variations of electric conductivity and Seebeck coefficient due to ion irradiation are systematically classified as follows. Firstly, in the case of undoped and n-type a-Si:Hs irradiated with protons, the increase in electric conductivity and the emergence of negative Seebeck effect are observed at the low fluence regime (around 108 to 106 dpa) because of the donor-center generation, which is thought to be caused by the electronic excitation effect of incident protons [24,25]. Secondly, both the decrease in electric conductivity and the increase in absolute value of Seebeck coefficient are observed in the middle fluence regime between 106 and 104 dpa. These are attributed to the carrier removal effect of accumulated dangling bonds. Thirdly, the density of localized states created by dangling bonds becomes extremely high at the fluence regime of above 104 dpa and it results in the enhancement of hopping transport. In this fluence regime, the dangling bond density of irradiated a-Si:H is estimated to increase to around 1020 cm3, since the electric conduction in hydrogen-free amorphous silicon, which contains around 1020 cm3 of dangling bonds, is dominated by the hopping transport [26]. The enhancement of hopping transport provides the increase in electric conductivity and the decrease in absolute value of Seebeck coefficient. Finally, the Seebeck effect disappears almost completely at above 0.2 dpa even in the case of impurity doped a-Si:H.
5. Conclusion The electric conductivity and the Seebeck coefficient variations of a-Si:H due to proton and Si ion irradiation were investigated using the in situ measurement techniques. These results were compared and interpreted systematically. In the case of undoped a-Si:H, the Seebeck effect is observed only in the fluence regime where the electric conductivity is enhanced. This is due to donor-centers generated by proton irradiation. Further irradiation induces the decrease in electric conductivity based on the carrier removal effect. However, the increase in electric conductivity associated with the enhancement of hopping transport appears in the high fluence regime of above 104 dpa. In the high fluence regime, no Seebeck effect is observed. In the case of the doped a-Si:H, the Seebeck coefficient shows an opposite change from the conductivity. That is, the decrease in electric conductivity and the increase in absolute value of Seebeck coefficient are observed in the fluence regime of below 104 dpa, and the increase in electric conductivity and the decrease in absolute value of Seebeck coefficient are observed at above 104 dpa. The former one is due to the carrier removal effect and the latter one is caused by the enhancement of hopping transport. Both
S.-i. Sato, T. Ohshima / Nuclear Instruments and Methods in Physics Research B 314 (2013) 153–157
phenomena are simply attributed to the accumulation of danglingbonds (radiation defects). Acknowledgements The authors would like to thank Dr. Hitoshi Sai and Dr. Michio Kondo of National Institute of Advanced Industrial Science and Technology (AIST) for fabricating the a-Si:H samples. The authors also would like to thank Dr. Kazunori Shimazaki and Dr. Mitsuru Imaizumi of Japan Aerospace Exploration Agency (JAXA) for their technical support and fruitful discussion. References
[10] [11] [12] [13] [14] [15] [16] [17] [18] [19]
[1] P.J. Sellin, J. Vaitkus, Nucl. Instr. Meth. A 557 (2006) 479–489. [2] W. Fernando, Nucl. Instr. Meth. A 596 (2008) 58–62. [3] R. Velazco, P. Fouillat, R. Reis, Radiation Effects on Embedded Systems, Springer, Dordrecht, 2007. p. 11. [4] A. Holmes-Siedle, L. Adams, Handbook of Radiation Effects, Second ed., Oxford University Press, Oxford, 2007. p. 17. [5] M. Imaizumi, S. Matsuda, S. Kawakita, T. Sumita, T. Takamoto, T. Ohshima, M. Yamaguchi , Prog. Photovolt: Res. Appl. 13 (2005) 529–543. [6] H. Amekura, N. Kishimoto, K. Kono, J. Appl. Phys. 84 (1998) 4834–4841. [7] H. Amekura, N. Kishimoto, K. Kono, IEEE Trans. Nucl. Sci. 45 (1998) 1508–1513. [8] S. Sato, H. Sai, T. Ohshima, M. Imaizumi, K. Shimazaki, M. Kondo, Prog. Photovolt. Res. Appl. (2012). [9] A. Banerjee, X. Xu, K. Beernink, F. Liu, K. Lord, G. DeMaggio, B. Yan, T. Su, G. Pietka, C. Worrel, S. Ehlert, D. Beglau, J. Yang, S. Guha, Proceedings of 35th
[20] [21] [22] [23] [24] [25] [26]
157
Photovoltaic Specialists Conference (PVSC-35), Philadelphia, USA, 2010, p. 002651. N. Wrysch, C. Miazza, S. Dunand, C. Ballif, A. Shah, M. Despeisse, D. Moraes, F. Powolny, P. Jarron, J. Non-Cryst. Solids 352 (2006) 1797–1800. E. Morgado, R. Schwarz, T. Braz, C. Casteleiro, A. Maçarico, M. Vieira, E. Alves, J. Non-Cryst. Solids 352 (2006) 1071–1074. P. Danesh, B. Pantchev, E. Vlaikova, Nucl. Instr. Meth. B 239 (2005) 370–374. N. Kishimoto, H. Amekura, K. Kono, C.G. Lee, J. Non-Cryst. Solids 227–230 (1998) 238–242. S. Coffa, F. Priolo, A. Battaglia, Phys. Rev. Lett. 70 (1993) 3756–3759. S. Sato, H. Sai, T. Ohshima, M. Imaizumi, K. Shimazaki, M. Kondo, Nucl. Instr. Meth. B 286 (2012) 29–34. S. Sato, H. Sai, T. Ohshima, M. Imaizumi, K. Shimazaki, M. Kondo, J. Non-Cryst. Solids 356 (2012) 2039–2043.