- Email: [email protected]
porous silicon complexes
Applied Surface Science 253 (2007) 4566–4569 www.elsevier.com/locate/apsusc
Study of photoconductivity and photoluminescence of organic/porous silicon complexes Yue Zhao a,b,*, Dongsheng Li b, Wenbin Sang a, Deren Yang b, Minhua Jiang c a
School of Materials Science and Engineering, Shanghai University, Shanghai 200072, PR China b State Key Laboratory of Silicon Materials, Zhejiang University, Hangzhou 310027, PR China c State Key Laboratory of Crystal Materials, Shandong University, Jinan 250100, PR China Received 17 May 2006; received in revised form 25 September 2006; accepted 8 October 2006 Available online 13 November 2006
Abstract In this paper, time-varying photoconductivity (PC) and the photoluminescence (PL) of different complexes were studied. Due to thick polymer layer hindering light penetrating into porous silicon (PS) layer, intrinsic PS luminescence in polymer/PS system disappeared. The physical origin of PL may be related to the recombination mechanisms involving surface defect states such as silicon oxide, siloxene. Due to carrier transfer controlled by different energy barrier, different devices prepared from different doped Si wafer showed opposite current–voltage characteristic. # 2006 Elsevier B.V. All rights reserved. Keywords: Porous silicon; Complexes; Photoconductivity; Photoluminescence
1. Introduction The discovery of room temperature visible photoluminescence (PL) from PS prepared in hydrofluoric acid has stimulated a great of interest in recent years because of its potential application in solar cells, light emitting diodes, optical sensors, moisture detector, interference filters, waveguide, SOI structures, and biomedical applications, etc. [1]. But up to now, the nature of porous silicon (PS) luminescence has not yet been confirmed. The photoconductivity (PC) is an important tool for the study of photo-generated carriers transfer in PS and also gives an insight into recombination mechanism. The PS luminescence is controlled by the radiative combination process, but the PC mainly is controlled by irradiative combination process, which can give complimentary information for understanding of carriers transfer and origin of PS luminescence. Many papers were published to study the relationship between PC and PL spectroscopy [2–9]. Some author considered that the transition states in PS are consistent of a series of energy levels and surface states formed in hydrogen-rich layer and SiOX, which produced PS lumines-
* Corresponding author. E-mail address: [email protected] (Y. Zhao). 0169-4332/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2006.10.007
cence. Mehra researched the influence of etching parameters on PC spectra of PS. They found that the optical activity of PS is based on quantum effect in nano-crystalline with disorder electronic boundaries with the surface states, and combination process in PS is found to have contribution from both monomolecular and bimolecular processes. Other author found that time evolution of PC indicated the presence of two competing mechanisms, one is related to photo-induced creation of charge carriers in silicon substrate followed by diffusion into PS layer, and the other is associated to removing of hydrogen from PS. Furthermore, in previous articles, several research groups have investigated the optical properties of complexes made by PS layer capped with organic materials [10–18] to increase efficiency and carrier injection of PS based devices. Rendu researched the PS/poly ( p-phenylene vinylene) (PPV) system and found that there is no interaction between both components and an energy transfer from PS to PPV is generated. A light emitting diode made by contacting an n-type PS film with chemically polymerized polyaniline showed a rectifying I–V characteristic and emitted red light under a forward bias voltage exceeding 3 V. The addition of the polymer layer into PS devices, which have been fabricated from n-type Si substrates using indium tin oxide, hole transporting poly (9-vinyl carbazole) and p-type nickel oxide film as hole injecting
Y. Zhao et al. / Applied Surface Science 253 (2007) 4566–4569
4567
contacts, led to an increase in device quantum efficiency of two orders of magnitude. In our paper, the complexes made by PS capped with PPV or Alq3 were investigated. The influence of doping level on PC and PL of PS-based devices were obtained. Furthermore, SEM observations were carried out in this article. 2. Experiments PS films prepared by electrochemically etched use two type silicon substrates: one is (1 0 0) n-type silicon wafer with a resistivity of 1–10 V cm, the other is (1 0 0) n-type epitaxial silicon (ESi) wafer with resistivities of 0.01–0.03 V cm for epitaxial layer and 1–10 V cm for bulk layer. The Si wafers were first cleaned using aqueous HF acid for 10 s to remove natural oxide film on Si surface. The back sides of Si wafers were coated with an aluminum layer and then were annealed at 400 8C for 4 min to provide a low resistance ohmic electric contact. The resulting Si wafers were electrochemically etched under a constant density of 30 mA/cm2 in a HF-ethanol (1:2 in volume) mixed solution for 1 min. During electrochemical etching, the wafers were illuminated by a 50 W tungsten lamp to produce holes. After etching, PS were rinsed with de-ionized water and dried in air. Before the polymer was spun onto PS, PS samples were dipped in aqueous HF to remove any native oxide and survival aluminum film on PS surface. The polymer solution was spun at 1000 rpm for 6 s and then at 3000 rmp for 30 s and then PS samples were subsequently baked at 100 8C for 30 min. A typical polymer formulation used is 10 mg poly (N-vinylcarbazole) (PVK) or 8-hydroxyquinoline aluminum (Alq3) in 10 ml tetrahydrofuran. The samples were characterized by PL spectroscopy, PC spectroscopy and scanning electron microscopy (SEM). The PL spectra excited by a 360 nm wavelength laser were measured using a HITACHI F-4500 fluorescence spectrophotometer. The PC measurements were carried out using a computer-interfaced KEITHLEY 4200 semiconductor characteristics system. The morphology observations were carried out using a FEI SIRION FESEM. All the measurements were carried out at room temperature. 3. Results and discussions Fig. 1 showed the PL spectra of polymer/PS system, in which PS was prepared from general Si (GPS) and from ESi (EPS). Furthermore, intrinsic PL spectra of polymer were also shown in Fig. 1c. From Fig. 1c, it can found the Alq3 peak was centered at 498 nm and the PL spectrum of PVK has a peak centered at 427 nm. It can be seen that PL spectrum of the asprepared PS prepared from both Si wafers exhibits a broad multiple peak from 420 to 470 nm and PL efficiency of GPS is higher than that of EPS. It was well known that PS luminescence is strongly depended on morphology and large specific surface of PS. Fig. 2 showed SEM images of surface morphology of GPS (Fig. 2a) and EPS (Fig. 2b). GPS porosity is higher than that of EPS, which led to stronger blue emission of GPS, as shown in
Fig. 1. PL spectra of polymers (c) and polymer/PS complexes prepared from general Si (a) or ESE (b).
Fig. 1. In addition, from the frequencies of Raman peak, the quantum size effects cannot be observed in both PS films. It was indicated that PL luminescence in blue band might come from the defects states on the surface of silicon rods and in the Si complexes including siloxene, Si oxide and Si hydrides [19]. In Ref. [11], authors researched PL of PPV/PS system and found PL spectrum of PS/PPV system showed all characteristics of the components with, in addition, a band located at 473 nm, which assumed to be related to the energy transfer from PS to PPV. But in Ref. [16], it was found that there is no interaction between PPV and PS, which also was proved by FTIR spectra and Raman measurements. Furthermore, PL spectrum of PS exhibited a peak at 620 nm while that of PPV has a broad multiple-peak structure between 510 and 550 nm. In our experiment, after the polymer was spun on PS surface, PL spectra of PS/polymer complexes only showed the characteristic of the polymer, as showed in Fig. 1a and b. In Ref. [19], the authors have reported that the PL spectra of PS/polymer complexes were strongly affected by the thickness of polymer.
4568
Y. Zhao et al. / Applied Surface Science 253 (2007) 4566–4569
Fig. 3. PC spectra of polymer/PS complexes prepared from general Si (a) or ESi (b).
Fig. 2. SEM images of PS prepared from general Si (a) or ESi (b).
The features in our experiment can be interpreted as thick polymer layer hindered light penetrating onto PS surface so intrinsic PS emission disappeared. Fig. 3 showed typical PC current versus time curves obtained with corresponding samples in Fig. 1. The time, at which the illumination on PS layers was initiated and at which the illumination on PS layers was terminated, has been marked with a vertical arrow, respectively. The device based on different PS showed a reverse current–voltage characteristics. When the measurements of PC carried out, the positive electrode of EPS device is at the side of bulk Si, but that of GPS device is at the side of polymer. This can be explained by energy band structure of our devices, as shown in Fig. 4. Fig. 4a showed the energy band structure of EPS-based device and Fig. 4b gave the energy band structure of GPS-based device. The band gap of GPS and EPS is 1.8 ev [20] and the band gap of PVK is 3.0 ev [21]. In addition, the band gap of crystal silicon is 1.12 ev. The relatively positions of objects in devices are rough shown in map. In EPS-based device, the energy barrier between polymer and EPS controlled carrier transfer process, but the energy barrier between PS and c-Si controlled carrier transfer process in GPS-based device, which led to the different electrical characteristic. In addition, the PC signals obtained from Alq3/PS system is higher than that obtained from PVK/PS
system under 100 W white tungsten lamp, as shown in Fig. 3. It may be said that the Alq/PS system is more efficient to adsorb light from the tungsten lamp. But from the Fig. 1, PL efficiency of PVK/PS system is higher than that of Alq3/PS system. It seemed to be contradiction between two results, which can be explained by the following model. The PL spectrum of PVK is close to that of PS, as shown in Fig. 1, and the excited wavelength at 360 nm is fit to PL measurements of PS samples, which may be adapted to excite PVK/PS system, but not be adapted to excite Alq3/PS system, which led to the lower efficiency of Alq3/PS system.
Fig. 4. Energy band structure of polymer/PS complexes prepared from general Si (a) or ESi (b).
Y. Zhao et al. / Applied Surface Science 253 (2007) 4566–4569
4. Summary In this paper, time-varying PC and PL of organic/PS complexes were studied. The carrier transfer was controlled by different energy barrier in different devices, which led to opposite current–voltage characteristic. The thick polymer layers hindered light penetrating into PS layer leading to the lack of intrinsic PL of PS in polymer/PS system. The origin of PL may be related to the defect states on PS surface. Reference [1] L.T. Canham, Appl. Phys. Lett. 57 (1990) 1046. [2] Baohui Wang, Dejun Wang, Lihua Zhang, Tiejin Li, J. Phys. Chem. Solids 58 (1) (1997) 25. [3] H. Shi, Y. Zheng, Y. Wang, R. Yuan, Appl. Phys. Lett. 63 (6) (1993) 770. [4] R.M. Mehra, V. Agarwal, V.K. Jain, P.C. Mathur, Thin Solid Films 315 (1998) 281. [5] D.W. Boeringer, R. Tsu, Appl. Phys. Lett. 65 (18) (1994) 2332. [6] T. Frello, E. Veje, O. Leistiko, J. Appl. Phys. 79 (2) (1996) 1027. [7] V. Duzhko, F. Koch, T. Dittrich, J. Appl. Phys. 91 (11) (2002) 9432.
4569
[8] B. Unal, S.C. Bayliss, D.T. Clarke, J. Appl. Phys. 87 (7) (2000) 3547. [9] J. Hiavka, J. Appl. Phys. 81 (3) (1997) 1404. [10] P.L. Rendu, T.P. Nguyen, K. Cheah, P. Joubert, Mater. Sci. Eng. C23 (2003) 847. [11] T.P. Nguyen, P.L. Rendu, K.W. Cheah, Phys. E 17 (2003) 664. [12] N. Koshida, H. Koyama, Y. Yamamoto, G.J. Collins, Appl. Phys. Lett. 63 (19) (1993) 2655. [13] J. Fan, M. Wan, D. Zhu, Synth. Met. 95 (1998) 119. [14] J.A. Lara, P. Kathirgamanathan, Synth. Met. 110 (2000) 233. [15] G. Wakefield, P.J. Dobson, Y.Y. Foo, A. Loni, A. Simons, J.L. Hutchison, Semicond. Sci. Technol. 12 (1997) 1304. [16] T.P. Nguyen, P.L. Rendu, M. Lakehal, P. Joubert, P. Destruel, Mater. Sci. Eng. B 69–70 (2000) 177. [17] A. Bsiesy, Y.F. Nicolau, A. Ermolieff, F. Muller, F. Gaspard, Thin Solid Films 255 (1995) 43. [18] D.P. Halliday, J.M. Eggleston, P.N. Adams, I.A. Pentland, A.P. Monkman, Synth. Met. 85 (1997) 1245. [19] Q.W. Chen, D.L. Zhu, C. Zhu, J. Wang, Y.G. Zhang, Appl. Phys. Lett. 82 (7) (2003) 1018. [20] C. Peng, K.D. Hirschman, P.M. Fauchet, J. Appl. Phys. 80 (1) (1996) 295. [21] Yong Quip, Lain Duan, Xeroxing Hun, Dewing Zhang, Min Zheng, Fenland Bay, Synth. Met. 123 (2001) 39.
Study of photoconductivity and photoluminescence of organic/porous silicon complexes Yue Zhao a,b,*, Dongsheng Li b, Wenbin Sang a, Deren Yang b, Minhua Jiang c a
School of Materials Science and Engineering, Shanghai University, Shanghai 200072, PR China b State Key Laboratory of Silicon Materials, Zhejiang University, Hangzhou 310027, PR China c State Key Laboratory of Crystal Materials, Shandong University, Jinan 250100, PR China Received 17 May 2006; received in revised form 25 September 2006; accepted 8 October 2006 Available online 13 November 2006
Abstract In this paper, time-varying photoconductivity (PC) and the photoluminescence (PL) of different complexes were studied. Due to thick polymer layer hindering light penetrating into porous silicon (PS) layer, intrinsic PS luminescence in polymer/PS system disappeared. The physical origin of PL may be related to the recombination mechanisms involving surface defect states such as silicon oxide, siloxene. Due to carrier transfer controlled by different energy barrier, different devices prepared from different doped Si wafer showed opposite current–voltage characteristic. # 2006 Elsevier B.V. All rights reserved. Keywords: Porous silicon; Complexes; Photoconductivity; Photoluminescence
1. Introduction The discovery of room temperature visible photoluminescence (PL) from PS prepared in hydrofluoric acid has stimulated a great of interest in recent years because of its potential application in solar cells, light emitting diodes, optical sensors, moisture detector, interference filters, waveguide, SOI structures, and biomedical applications, etc. [1]. But up to now, the nature of porous silicon (PS) luminescence has not yet been confirmed. The photoconductivity (PC) is an important tool for the study of photo-generated carriers transfer in PS and also gives an insight into recombination mechanism. The PS luminescence is controlled by the radiative combination process, but the PC mainly is controlled by irradiative combination process, which can give complimentary information for understanding of carriers transfer and origin of PS luminescence. Many papers were published to study the relationship between PC and PL spectroscopy [2–9]. Some author considered that the transition states in PS are consistent of a series of energy levels and surface states formed in hydrogen-rich layer and SiOX, which produced PS lumines-
* Corresponding author. E-mail address: [email protected] (Y. Zhao). 0169-4332/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2006.10.007
cence. Mehra researched the influence of etching parameters on PC spectra of PS. They found that the optical activity of PS is based on quantum effect in nano-crystalline with disorder electronic boundaries with the surface states, and combination process in PS is found to have contribution from both monomolecular and bimolecular processes. Other author found that time evolution of PC indicated the presence of two competing mechanisms, one is related to photo-induced creation of charge carriers in silicon substrate followed by diffusion into PS layer, and the other is associated to removing of hydrogen from PS. Furthermore, in previous articles, several research groups have investigated the optical properties of complexes made by PS layer capped with organic materials [10–18] to increase efficiency and carrier injection of PS based devices. Rendu researched the PS/poly ( p-phenylene vinylene) (PPV) system and found that there is no interaction between both components and an energy transfer from PS to PPV is generated. A light emitting diode made by contacting an n-type PS film with chemically polymerized polyaniline showed a rectifying I–V characteristic and emitted red light under a forward bias voltage exceeding 3 V. The addition of the polymer layer into PS devices, which have been fabricated from n-type Si substrates using indium tin oxide, hole transporting poly (9-vinyl carbazole) and p-type nickel oxide film as hole injecting
Y. Zhao et al. / Applied Surface Science 253 (2007) 4566–4569
4567
contacts, led to an increase in device quantum efficiency of two orders of magnitude. In our paper, the complexes made by PS capped with PPV or Alq3 were investigated. The influence of doping level on PC and PL of PS-based devices were obtained. Furthermore, SEM observations were carried out in this article. 2. Experiments PS films prepared by electrochemically etched use two type silicon substrates: one is (1 0 0) n-type silicon wafer with a resistivity of 1–10 V cm, the other is (1 0 0) n-type epitaxial silicon (ESi) wafer with resistivities of 0.01–0.03 V cm for epitaxial layer and 1–10 V cm for bulk layer. The Si wafers were first cleaned using aqueous HF acid for 10 s to remove natural oxide film on Si surface. The back sides of Si wafers were coated with an aluminum layer and then were annealed at 400 8C for 4 min to provide a low resistance ohmic electric contact. The resulting Si wafers were electrochemically etched under a constant density of 30 mA/cm2 in a HF-ethanol (1:2 in volume) mixed solution for 1 min. During electrochemical etching, the wafers were illuminated by a 50 W tungsten lamp to produce holes. After etching, PS were rinsed with de-ionized water and dried in air. Before the polymer was spun onto PS, PS samples were dipped in aqueous HF to remove any native oxide and survival aluminum film on PS surface. The polymer solution was spun at 1000 rpm for 6 s and then at 3000 rmp for 30 s and then PS samples were subsequently baked at 100 8C for 30 min. A typical polymer formulation used is 10 mg poly (N-vinylcarbazole) (PVK) or 8-hydroxyquinoline aluminum (Alq3) in 10 ml tetrahydrofuran. The samples were characterized by PL spectroscopy, PC spectroscopy and scanning electron microscopy (SEM). The PL spectra excited by a 360 nm wavelength laser were measured using a HITACHI F-4500 fluorescence spectrophotometer. The PC measurements were carried out using a computer-interfaced KEITHLEY 4200 semiconductor characteristics system. The morphology observations were carried out using a FEI SIRION FESEM. All the measurements were carried out at room temperature. 3. Results and discussions Fig. 1 showed the PL spectra of polymer/PS system, in which PS was prepared from general Si (GPS) and from ESi (EPS). Furthermore, intrinsic PL spectra of polymer were also shown in Fig. 1c. From Fig. 1c, it can found the Alq3 peak was centered at 498 nm and the PL spectrum of PVK has a peak centered at 427 nm. It can be seen that PL spectrum of the asprepared PS prepared from both Si wafers exhibits a broad multiple peak from 420 to 470 nm and PL efficiency of GPS is higher than that of EPS. It was well known that PS luminescence is strongly depended on morphology and large specific surface of PS. Fig. 2 showed SEM images of surface morphology of GPS (Fig. 2a) and EPS (Fig. 2b). GPS porosity is higher than that of EPS, which led to stronger blue emission of GPS, as shown in
Fig. 1. PL spectra of polymers (c) and polymer/PS complexes prepared from general Si (a) or ESE (b).
Fig. 1. In addition, from the frequencies of Raman peak, the quantum size effects cannot be observed in both PS films. It was indicated that PL luminescence in blue band might come from the defects states on the surface of silicon rods and in the Si complexes including siloxene, Si oxide and Si hydrides [19]. In Ref. [11], authors researched PL of PPV/PS system and found PL spectrum of PS/PPV system showed all characteristics of the components with, in addition, a band located at 473 nm, which assumed to be related to the energy transfer from PS to PPV. But in Ref. [16], it was found that there is no interaction between PPV and PS, which also was proved by FTIR spectra and Raman measurements. Furthermore, PL spectrum of PS exhibited a peak at 620 nm while that of PPV has a broad multiple-peak structure between 510 and 550 nm. In our experiment, after the polymer was spun on PS surface, PL spectra of PS/polymer complexes only showed the characteristic of the polymer, as showed in Fig. 1a and b. In Ref. [19], the authors have reported that the PL spectra of PS/polymer complexes were strongly affected by the thickness of polymer.
4568
Y. Zhao et al. / Applied Surface Science 253 (2007) 4566–4569
Fig. 3. PC spectra of polymer/PS complexes prepared from general Si (a) or ESi (b).
Fig. 2. SEM images of PS prepared from general Si (a) or ESi (b).
The features in our experiment can be interpreted as thick polymer layer hindered light penetrating onto PS surface so intrinsic PS emission disappeared. Fig. 3 showed typical PC current versus time curves obtained with corresponding samples in Fig. 1. The time, at which the illumination on PS layers was initiated and at which the illumination on PS layers was terminated, has been marked with a vertical arrow, respectively. The device based on different PS showed a reverse current–voltage characteristics. When the measurements of PC carried out, the positive electrode of EPS device is at the side of bulk Si, but that of GPS device is at the side of polymer. This can be explained by energy band structure of our devices, as shown in Fig. 4. Fig. 4a showed the energy band structure of EPS-based device and Fig. 4b gave the energy band structure of GPS-based device. The band gap of GPS and EPS is 1.8 ev [20] and the band gap of PVK is 3.0 ev [21]. In addition, the band gap of crystal silicon is 1.12 ev. The relatively positions of objects in devices are rough shown in map. In EPS-based device, the energy barrier between polymer and EPS controlled carrier transfer process, but the energy barrier between PS and c-Si controlled carrier transfer process in GPS-based device, which led to the different electrical characteristic. In addition, the PC signals obtained from Alq3/PS system is higher than that obtained from PVK/PS
system under 100 W white tungsten lamp, as shown in Fig. 3. It may be said that the Alq/PS system is more efficient to adsorb light from the tungsten lamp. But from the Fig. 1, PL efficiency of PVK/PS system is higher than that of Alq3/PS system. It seemed to be contradiction between two results, which can be explained by the following model. The PL spectrum of PVK is close to that of PS, as shown in Fig. 1, and the excited wavelength at 360 nm is fit to PL measurements of PS samples, which may be adapted to excite PVK/PS system, but not be adapted to excite Alq3/PS system, which led to the lower efficiency of Alq3/PS system.
Fig. 4. Energy band structure of polymer/PS complexes prepared from general Si (a) or ESi (b).
Y. Zhao et al. / Applied Surface Science 253 (2007) 4566–4569
4. Summary In this paper, time-varying PC and PL of organic/PS complexes were studied. The carrier transfer was controlled by different energy barrier in different devices, which led to opposite current–voltage characteristic. The thick polymer layers hindered light penetrating into PS layer leading to the lack of intrinsic PL of PS in polymer/PS system. The origin of PL may be related to the defect states on PS surface. Reference [1] L.T. Canham, Appl. Phys. Lett. 57 (1990) 1046. [2] Baohui Wang, Dejun Wang, Lihua Zhang, Tiejin Li, J. Phys. Chem. Solids 58 (1) (1997) 25. [3] H. Shi, Y. Zheng, Y. Wang, R. Yuan, Appl. Phys. Lett. 63 (6) (1993) 770. [4] R.M. Mehra, V. Agarwal, V.K. Jain, P.C. Mathur, Thin Solid Films 315 (1998) 281. [5] D.W. Boeringer, R. Tsu, Appl. Phys. Lett. 65 (18) (1994) 2332. [6] T. Frello, E. Veje, O. Leistiko, J. Appl. Phys. 79 (2) (1996) 1027. [7] V. Duzhko, F. Koch, T. Dittrich, J. Appl. Phys. 91 (11) (2002) 9432.
4569
[8] B. Unal, S.C. Bayliss, D.T. Clarke, J. Appl. Phys. 87 (7) (2000) 3547. [9] J. Hiavka, J. Appl. Phys. 81 (3) (1997) 1404. [10] P.L. Rendu, T.P. Nguyen, K. Cheah, P. Joubert, Mater. Sci. Eng. C23 (2003) 847. [11] T.P. Nguyen, P.L. Rendu, K.W. Cheah, Phys. E 17 (2003) 664. [12] N. Koshida, H. Koyama, Y. Yamamoto, G.J. Collins, Appl. Phys. Lett. 63 (19) (1993) 2655. [13] J. Fan, M. Wan, D. Zhu, Synth. Met. 95 (1998) 119. [14] J.A. Lara, P. Kathirgamanathan, Synth. Met. 110 (2000) 233. [15] G. Wakefield, P.J. Dobson, Y.Y. Foo, A. Loni, A. Simons, J.L. Hutchison, Semicond. Sci. Technol. 12 (1997) 1304. [16] T.P. Nguyen, P.L. Rendu, M. Lakehal, P. Joubert, P. Destruel, Mater. Sci. Eng. B 69–70 (2000) 177. [17] A. Bsiesy, Y.F. Nicolau, A. Ermolieff, F. Muller, F. Gaspard, Thin Solid Films 255 (1995) 43. [18] D.P. Halliday, J.M. Eggleston, P.N. Adams, I.A. Pentland, A.P. Monkman, Synth. Met. 85 (1997) 1245. [19] Q.W. Chen, D.L. Zhu, C. Zhu, J. Wang, Y.G. Zhang, Appl. Phys. Lett. 82 (7) (2003) 1018. [20] C. Peng, K.D. Hirschman, P.M. Fauchet, J. Appl. Phys. 80 (1) (1996) 295. [21] Yong Quip, Lain Duan, Xeroxing Hun, Dewing Zhang, Min Zheng, Fenland Bay, Synth. Met. 123 (2001) 39.