- Email: [email protected]
Silicon surface passivation by hot-wire CVD Si thin films studied by in situ surface spectroscopy
Thin Solid Films 517 (2009) 3456–3460
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t s f
Silicon surface passivation by hot-wire CVD Si thin films studied by in situ surface spectroscopy J.J.H. Gielis, B. Hoex, P.J. van den Oever, M.C.M. van de Sanden, W.M.M. Kessels ⁎ Department of Applied Physics, Eindhoven University of Technology, The Netherlands
a r t i c l e
i n f o
Available online 7 February 2009 Keywords: Hot-wire deposition Amorphous silicon Epitaxial silicon Silicon wafer Surface passivation In situ optical studies
a b s t r a c t Silicon thin films can provide an excellent surface passivation of crystalline silicon (c-Si) which is of importance for high efficiency heterojunction solar cells or diffused emitter solar cells with well-passivated rear surfaces. Hot-wire chemical vapor deposition (hot-wire CVD) is an attractive method to synthesize Si thin films for these applications as the method is ion-bombardment free yielding good quality films over a wide range of deposition rates. The properties of the interface between hot-wire CVD Si thin films and H-terminated c-Si substrates have been studied during film growth by three complementary in situ techniques. Spectroscopic ellipsometry has been used to determine the optical properties and thickness of the films, whereas information on the H-bonding modes and H-depth profile has been obtained by attenuated total reflection infrared spectroscopy. Second-harmonic generation (SHG), a nonlinear optical technique sensitive to surface and interface states, has been used to probe two-photon resonances related to modified Si–Si bonds at the interface. By correlating the observations with ex situ lifetime spectroscopy experiments the growth and surface passivation mechanism of the Si films are discussed. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Surface passivation, i.e., the minimization of surface recombination losses, is essential for the performance of high efficiency c-Si solar cells. The importance of surface passivation is even increasing due to a cost-driven reduction of overall solar cell thickness leading to an increase in the surface-to-volume ratio. The reduction of charge carrier recombination losses at surfaces and interfaces becomes therefore a decisive factor for the final solar cell efficiency. In general, surface passivation can be achieved by a reduction of surface recombination centers, referred to as chemical passivation, and by shielding of charge carriers by internal electric fields, also known as field-effect passivation. Surface passivation of crystalline Si (c-Si) can be provided by several functional thin films such as SiO2 [1], hydrogenated amorphous silicon nitride (a-SiNx:H) [2], Al2O3 [3], and hydrogenated amorphous silicon (a-Si:H). Thin films of a-Si:H are used, for example in Si heterojunction (SHJ) solar cells [4,5] and in diffused emitter solar cells for rear-surface passivation [6]. The performance of these devices is highly determined by interface quality and an abrupt and atomically flat interface between c-Si and the a-Si:H film is required to reduce surface recombination losses. Hot-wire chemical vapor deposition (hot-wire CVD) is an attractive method to synthesize Si thin films for these applications as this technique yields good quality films over a wide range of deposition
parameters while providing a high atomic hydrogen flux and being ion-bombardment-free. A good to excellent level of surface passivation by a-Si:H films deposited by hot-wire CVD has been reported in some brief minority carrier lifetime studies [7,8], while most of the work in this respect has concentrated on the synthesis of complete SHJ solar cells [8–10]. Efficiencies as high as 18.2% have recently been reported for hot-wire CVD a-Si:H front and back contacts in p-type SHJ solar cells [11]. Recently we reported on the simultaneous application of three complementary optical diagnostics during hot-wire CVD of Si thin films [12,13]: spectroscopic ellipsometry (SE) was implemented to provide information on the optical properties and thickness of the Si thin films, attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) to reflect the H content in the films, and the nonlinear optical technique of second-harmonic generation (SHG) to reveal properties of the surface and the buried film/substrate interface [14]. In this paper results obtained from these diagnostics are combined with results on the surface passivation performance of the hot-wire CVD a-Si:H films as investigated by ex situ lifetime spectroscopy. The growth and surface passivation mechanism of the Si films are discussed focusing on the role of the substrate temperature, deposition rate, epitaxial film growth, defect states and built-in electric fields. 2. Experimental
⁎ Corresponding author. E-mail address: [email protected] (W.M.M. Kessels). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.01.076
Si thin films were deposited using hot-wire CVD in a setup with a base pressure of b10− 9 mbar. A 0.45-mm-diameter coiled tungsten
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
3457
Physics Tsunami) tunable in the 1.35–1.75 eV photon energy range. An intracavity doubled Nd:YVO4 laser (Spectra Physics Millennia Vsj) was used to pump the oscillator. To separate the fundamental laser radiation from the SHG radiation generated at the sample, color filters and a Pellin Broca dispersing prism were applied. A photomultiplier tube connected to photon counting electronics was used to detect the SHG signal. The time resolution of SHG setup was 0.1 s. Both the incident fundamental and the detected SHG radiation were p-polarized. 3. Results and discussion 3.1. Passivation properties
Fig. 1. Surface recombination velocity Seff,max as a function of deposition temperature for Si thin films deposited with hot-wire CVD, plasma-enhanced CVD, and the expanding thermal plasma (ETP) technique.
filament was resistively heated by a 10 A dc to about 2000 °C to dissociate SiH4 gas (purity of over 99.995%) at a pressure of 8×10− 3 mbar. The filament was placed at 11–13 cm from the substrate, which resulted in deposition rates of 0.3–0.6 Å/s. The substrate temperature was actively controlled by a radiative heater at the rear side of the sample. The Si thin films were deposited on Si(100) substrates; either standard n-type wafers, undoped trapezoidally-shaped substrates with 45° angled bevels for experiments involving ATR-FTIR, or 2 in. n-type float zone wafers (P-doped, resistivity of 1.9 Ω cm, thickness of 275 μm) for lifetime spectroscopy. Prior to deposition the substrates were cleaned by an ultrasound ethanol bath and subsequent immersion in a 2% HF solution for 2 min or by standard RCA I and RCA II procedures using a buffered 1% NH4F/HF solution (BHF) with pH 4. Both cleaning methods removed the native oxide and terminated the Si surface with H. The passivation properties of the deposited Si thin films were determined by measuring the surface recombination velocity of charge carriers using carrier lifetime spectroscopy. These photoconductance experiments were carried out ex situ using a lifetime tester (Sinton Consulting WCT-100) in both quasisteady-state and transient mode [15]. The surface recombination velocity was calculated assuming an infinite bulk lifetime for the float zone c-Si wafer, which results in an upper limit for the surface recombination velocity Seff,max. The optical characterization of the Si thin films was performed in real time during deposition. Only the key aspects are briefly discussed here as the optical techniques are described in detail elsewhere [13,16,17]. The real-time SE data were acquired with a rotating compensator spectroscopic ellipsometer (J.A. Woollam M-2000U, 1.24–5.0 eV photon energy range) at a time resolution of 3.5 s. The SE data are expressed in terms of the pseudodielectric function bɛN = bɛ1N + ibɛ2N, which can be calculated by treating the film and the substrate as a single semi-infinite material. The ATR-FTIR measurements were performed using a Fourier transform infrared spectrometer (Bruker Vector 22 with a liquid nitrogen cooled HgCdTe detector) focusing on the frequency range of 1800–2250 cm− 1, that is, the region where radiation is absorbed due to Si–H stretching vibrations. The infrared radiation underwent ∼ 25 total internal reflections at each side of the sample, enhancing the sensitivity with a factor of ∼ 50 compared to single transmission. The spectral resolution of the ATR-FTIR experiments was 4 cm− 1 and the time resolution was 6.8 s. The ATR-FTIR data are expressed in terms of absorbance A, which is calculated from the transmission T and the background transmission T0 obtained prior to deposition by using A = −log(T/T0). The laser radiation for the SHG experiments had a pulse duration of ∼90 fs and was provided by a Ti:sapphire oscillator (Spectra
Fig. 1 shows the surface recombination velocity Seff,max as a function of deposition temperature for a-Si:H films deposited with hot-wire CVD at a deposition rate of 0.3 Å/s. The surface passivation properties significantly improve with increasing deposition temperature up to ∼ 140 °C and a Seff,max value as low as 17 cm/s is reached. For higher deposition temperatures the surface passivation properties strongly deteriorate again. The abrupt deterioration of the surface passivation properties for higher temperatures, can be attributed to the formation of epitaxial Si [8,12,13,18–20]. This effect causes a defect-rich interface between the c-Si substrate and the film. The formation of epitaxial material can be monitored with SE [21]. In Fig. 2 the imaginary part of the pseudodielectric function bɛ2N is shown as a function of photon energy for films deposited at a substrate temperature of ∼ 150 °C. The spectra in Fig. 2(a) clearly indicate the deposition of a-Si:H; with increasing film thickness the c-Si E′0/E1 and E2 critical point (CP) transitions broaden and ultimately cannot be separately distinguished anymore. The spectra in Fig. 2(b), however, hardly change with increasing film thickness. This reveals that these films developed epitaxially. In Fig. 2(c), an intermediate situation is displayed, that is, the formation of mixed phase material. In this case, the spectra also broaden with increasing thickness, however, not as rapidly as for purely amorphous material. Also, the characteristic c-Si E′0/E1 and E2
Fig. 2. Imaginary part of the pseudodielectric function bɛ2N as a function of photon energy obtained with in situ SE during deposition of (a) a-Si:H, (b) epitaxial Si, and (c) mixed phase material on H terminated Si(100). The substrate temperature was ∼150 °C for all films. Dashed lines indicate bɛ2N of H terminated Si(100) and ɛ2 of a-Si:H as deduced from the data.
3458
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
CP transitions remain present for thicker films. These results demonstrate that a-Si:H, epi-Si, and mixed phase formation can occur at similar processing conditions when the substrate temperature is at or close to the transition temperature from a-Si:H to epitaxial film growth. From the data in Fig. 1 it can be concluded that a high deposition temperature is beneficial for the surface passivation properties of a-Si:H. However, it is essential to prevent the formation of epitaxial Si. The latter can be accomplished by increasing the deposition rate [22–24] as is also illustrated in Fig. 1 for films deposited by radiofrequency plasma-enhanced CVD (PECVD) [25,26] and by the expanding thermal plasma (ETP) [27]. The PECVD data obtained for a deposition rate of 5 Å/s show the same trend as the hot-wire CVD data; an improvement of the surface passivation properties up to, in this case, ∼225 °C and a deterioration for higher temperatures. With the ETP deposited films low surface recombination velocities can be reached even at higher substrate temperatures when employing deposition rates of 10 Å/s (250 °C) and 30 Å/s (400 °C). These data clearly show that excellent passivation properties (Seff,max = 6 cm/s) can be obtained at deposition temperatures as high as 400 °C when using high deposition rates. Note that, similar to hot-wire CVD, ion bombardment of the substrate is negligible for the ETP technique [27].
3.2. Passivation mechanism As discussed above, surface passivation can be obtained by two mechanisms: chemical passivation and field-effect passivation. To investigate the surface passivation mechanism of a-Si:H we used two optical diagnostics: ATR-FTIR and SHG. In Fig. 3 the absorbance of a-Si:H and mixed phase Si thin films is shown as a function of frequency as measured with ATR-FTIR. For the a-Si:H film the absorbance increases strongly with increasing film thickness, in addition, the maximum in the absorbance shifts to lower wavenumbers with increasing film thickness. For mixed phase material the absorbance remains very low. Quantitative information about the hydrogen bonding in the film as a function of film thickness can be obtained by deconvolution of the spectra into several Gaussian peaks, corresponding to SiH and SiH2 bulk stretching modes and SiHx surface stretching modes [16,28,29]. The center positions of these Gaussian peaks are also indicated in Fig. 3. In Fig. 4(a) the integrated absorbance of the SiH and SiH2 bulk modes and the sum of the SiHx surface modes resulting from the deconvolution of the ATR-FTIR spectra are shown. For the a-Si:H film a strong
Fig. 3. ATR-FTIR spectra for a-Si:H (top) and mixed phase Si (bottom) deposited on H terminated Si(100). Data are obtained in real time during deposition and selected spectra for film thicknesses of 38, 100, 236, and 329 Å are shown. The dashed vertical lines indicate the position of the different SiH bulk and surface stretching modes.
Fig. 4. (a) Integrated absorbance by SiH and SiH2 bulk stretching modes and the sum of the surface SiHx modes for a-Si:H (closed symbols) and mixed phase Si films (open symbols) deposited on H terminated Si(100) as a function film thickness. (b) SiH and SiH2 bulk modes for a-Si:H and mixed phase Si displayed for a larger thickness range.
increase in absorbance by SiHx surface modes is observed at the onset of film growth. Subsequently, the absorption by the SiH2 bulk mode is dominant. After ∼40 Å of film growth, which corresponds to ∼100 s of film deposition, the absorbance is governed by the SiH bulk mode. The initial dominance of the SiHx surface modes and the SiH2 bulk modes indicates that the film growth starts off H rich. This H-rich interface region can be very important for the passivation properties of the a-Si:H films, as H can cause a reduction of interface defects and thereby prevent the recombination of charge carriers. Fig. 4(a) also shows that thin films of mixed phase Si display a very low integrated absorbance, which indicates that the H content, and thus the a-Si:H fraction, in these films is very low. As discussed above, the passivation properties of these films are marginal. Due to the low absorbance for these films, SiHx surface stretching modes could even not be distinguished from the bulk modes. In Fig. 4(b) the integrated absorbance for a-Si:H and mixed phase Si is shown over a larger thickness range for the SiH and SiH2 bulk modes. For a-Si:H the integrated absorbance for the SiH mode increases linearly as a function of film thickness indicating a constant H incorporation during bulk film growth. This constant H content can be calculated to be ∼12.5 at.%. For the mixed phase material the absorbance for the SiH mode increases more than linearly with thickness, indicating that the H content of the film and, hence, the a-Si:H content of the material increase with film thickness. The ATR-FTIR data provide evidence that chemical passivation is playing a role in the c-Si passivation mechanism of a-Si:H thin films. To investigate the possible additional influence of field-effect passivation, the nonlinear optical technique of second-harmonic generation was applied. Fig. 5(a) shows SHG spectra for a-Si:H films of 38, 100, and 727 Å. For comparison, also the SHG spectrum for a pristine H terminated Si(100) substrate is displayed. All spectra were obtained at room temperature. The spectra for the ultrathin films of 38 and 100 Å display a strongly enhanced SHG intensity compared to the spectrum for the H terminated substrate and have a sharp maximum at 3.3 eV. The spectrum for the thicker film of 727 Å shows a much lower SHG intensity, especially in the higher photon energy range. The combination of the spectra for the three films gives clear indications of both the macroscopic origin (i.e., the geometrical origin) and the microscopic origin (i.e., the origin on the atomic scale) of the SHG response. The decrease in SHG intensity for the thicker film can be explained by SHG radiation that is generated at the buried interface between the a-Si:H film and the c-Si substrate and that is absorbed while propagating through the film. The a-Si:H
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
3459
For the 38 and 100 Å thick a-Si:H films on H terminated Si(100) a shift in peak position from 3.4 eV to 3.3 eV due to reasons such as absorption or interference of multiple SHG contributions is not likely. a-Si:H is strongly absorbing SHG radiation, especially in the higher photon energy range, however, both the 38 and the 100 Å thick films display a maximum in SHG intensity at 3.3 eV irrespective of their thickness. Furthermore, the a-Si:H film on native oxide covered Si(100) also indicates that ∼100 Å of a-Si:H is too thin to effectively redshift a resonance from 3.4 eV to 3.3 eV. Interference of multiple SHG contributions would result in asymmetric features, which is not observed for the 38 and 100 Å thick films in Fig. 5(a). In summary, the sharp resonance at 3.3 eV is most likely governed by Si–Si bonds in the c-Si modified due to the presence of the interface with the a-Si:H. The absence of a EFISH contribution for this case indicates that no strong electric field is present in the a-Si:H/c-Si system. This leads to the conclusion that the influence of field-effect passivation on the c-Si surface passivation mechanism of a-Si:H is marginal. In this respect, the surface passivation properties of a-Si:H differ from dielectrics such as a-SiNx:H and Al2O3 that also provide an excellent level of surface passivation. For these materials field-effect passivation is predominant due to their very high density (1012– 1013 cm− 2) of fixed charges [2,31]. 4. Conclusion
Fig. 5. (a) SHG intensity as a function of SHG photon energy for a-Si:H films with a thickness of 38, 100, and 727 Å deposited on H terminated Si(100). In addition, the SHG spectrum for H terminated Si(100) is also shown. (b) SHG intensity as a function of SHG photon energy for 100 Å a-Si:H on deposited H terminated Si(100) and 97 Å a-Si:H deposited on native oxide (NO) covered Si(100). All data were obtained at room temperature. The dashed lines indicate the peak positions for the 100 and 97 Å thick a-Si:H films.
is strongly absorbing radiation at the SHG photon energy, especially in the higher photon energy range [13,17]. Equivalently, the decrease in SHG intensity for the thicker film shows that the influence of the surface and the bulk of the a-Si:H films is very limited. The strong increase of the SHG signal for the ultrathin films of 38 and 100 Å with respect to the spectrum for the substrate indicates that the possible influence of contributions from the substrate bulk is also limited. The sharp resonance at 3.3 eV as observed for the 38 and 100 Å thick films is an indication that the SHG response is related to the E′0/E1 critical point (CP) transition in the c-Si. This observation on the microscopic origin is in line with the conclusion that the SHG response is originating macroscopically from the buried a-Si:H/c-Si interface. In general, SHG contributions near the E'0/E1 CP energy might be related to Si–Si bonds in c-Si modified due to the presence of an interface or to electric-field-induced second-harmonic generation (EFISH, sensitive to electric fields N105 V/cm) [30,31]. However, from the location of the resonance it can be inferred that a dominant EFISH contribution is not likely. EFISH reflects the spectroscopic properties of the Si space-charge region. Consequently, at room temperature this effect would result in a resonance centered at 3.4 eV. This is clearly not the case for the 38 and 100 Å thick a-Si:H films deposited on H terminated Si(100), which have a strong resonance at 3.3 eV. For comparison, Fig. 5(b) shows the room temperature SHG spectrum obtained from a 97 Å thick a-Si:H film deposited on native oxide covered Si(100). For native oxide covered Si, the SHG response near the E′0/E1 CP energy is known to be originating from EFISH to a large extent, as related to a positively charged Si spacecharge region [32]. As Fig. 5(b) shows, for native oxide covered Si(100) with a 97 Å a-Si:H film a strong resonance at 3.4 eV is observed, hence, suggesting a strong influence of EFISH for this sample.
Thin films of a-Si:H result in an excellent level of c-Si surface passivation. In general, a high deposition temperature is beneficial for the surface passivation properties of a-Si:H thin films. However, it is essential to prevent the formation of epitaxial Si, which, at high temperatures, requires high deposition rates. Considering the fact that with hot-wire CVD a-Si:H films with good film properties can also be obtained at high deposition rates [33], hot-wire CVD is a very promising method to provide a high level of surface passivation of c-Si for solar cell applications. The ATR-FTIR and SHG experiments indicate that the c-Si surface passivation mechanism of a-Si:H is governed by chemical passivation with the influence of field-effect passivation being marginal. However, when doped a-Si:H layers are applied, for example in SHJ solar cells, space charge regions can exist and field-effect passivation can be of influence. This effect will improve the excellent surface passivation properties of a-Si:H even further. Acknowledgments The authors thank M. J. F. van de Sande, J. F. C. Jansen, J. J. A. Zeebregts, and H. M. M. de Jong for their skillful technical assistance. This work was supported by the Netherlands Foundation for Fundamental Research on Matter (FOM). References [1] M.J. Kerr, A. Cuevas, Semicond. Sci. Tech. 17 (2002) 35. [2] R. Hezel, R. Schorner, J. Appl. Phys. 52 (1981) 3076. [3] B. Hoex, S.B.S. Heil, E. Langereis, M.C.M. van de Sanden, W.M.M. Kessels, Appl. Phys. Lett. 89 (2006) 042112. [4] M. Taguchi, K. Kawamoto, S. Tsuge, T. Baba, H. Sakata, M. Morizane, K. Uchihashi, N. Nakamura, S. Kiyama, O. Oota, Prog. Photovolt. 8 (2000) 503. [5] S. Taira, Y. Yoshimine, T. Baba, M. Taguchi, H. Kanno, T. Kinoshita, H. Sakata, E. Maruyama, M. Tanaka, Proceedings of the 22nd European Photovoltaic Solar Energy Conference, Milan, Italy, 2007, p. 932. [6] M. Schaper, J. Schmidt, H. Plagwitz, R. Brendel, Prog. Photovolt. 13 (2005) 381. [7] C. Voz, I. Martin, A. Orpella, J. Puigdollers, M. Vetter, R. Alcubilla, D. Soler, M. Fonrodona, J. Bertomeu, J. Andreu, Thin Solid Films 430 (2003) 270. [8] T.H. Wang, E. Iwaniczko, M.R. Page, D.H. Levi, Y. Yan, H.M. Branz, Q. Wang, Thin Solid Films 501 (2006) 284. [9] D. Muñoz, C. Voz, A. Orpella, J. Puigdollers, R. Alcubilla, F. Villar, J. Bertomeu, J. Andreu, J. Doman-Lacoste, P. Roca i Cabarrocas, Thin Solid Films 516 (2008) 761. [10] S.-Y. Lien, B.-R. Wu, J.-C. Liu, D.-S. Wuu, Thin Solid Films 516 (2008) 747. [11] H.M. Branz, C.W. Teplin, D.L. Young, M.R. Page, E. Iwaniczko, L. Roybal, R. Bauer, A.H. Mahan, Y. Xu, P. Stradins, T. Wang, Q. Wang, Thin Solid Films 516 (2008) 743.
3460
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
[12] J.J.H. Gielis, P.J. van den Oever, M.C.M. van de Sanden, W.M.M. Kessels, Appl. Phys. Lett. 90 (2007) 202108. [13] J.J.H. Gielis, P.J. van den Oever, B. Hoex, M.C.M. van de Sanden, W.M.M. Kessels, Phys. Rev. B 77 (2008) 205329. [14] T.F. Heinz, Second-order Nonlinear Optical Effects at Surfaces and Interfaces, in: H.E. Ponath, G.I. Stegeman (Eds.), Nonlinear Surface Electromagnetic Phenomena, Elsevier, Amsterdam, 1991. [15] R.A. Sinton, A. Cuevas, Appl. Phys. Lett. 69 (1996) 2510. [16] P.J. van den Oever, J.J.H. Gielis, M.C.M. van de Sanden, W.M.M. Kessels, Thin Solid Films 516 (2008) 511. [17] J.J.H. Gielis, P.M. Gevers, A.A.E. Stevens, H.C.W. Beijerinck, M.C.M. van de Sanden, W.M.M. Kessels, Phys. Rev. B 74 (2006) 165311. [18] Y. Yan, M. Page, T.H. Wang, M.M. Al-Jassim, H.M. Branz, Q. Wang, Appl. Phys. Lett. 88 (2006) 121925. [19] S. De Wolf, M. Kondo, Appl. Phys. Lett. 90 (2007) 042111. [20] H. Fujiwara, M. Kondo, Appl. Phys. Lett. 90 (2007) 013503. [21] At the onset of film growth, however, SE cannot resolve the film morphology due to a lack in contrast between c-Si substrates and Si films, see Ref. 13. [22] D.J. Eaglesham, H.-J. Gossmann, M. Cerullo, Phys. Rev. Lett. 65 (1990) 1227. [23] H. Jorke, H. Kibbel, F. Schäffler, H.-J. Herzog, Thin Solid Films 183 (1989) 307.
[24] J. Thiesen, H.M. Branz, R.S. Crandall, Appl. Phys. Lett. 77 (2000) 3589. [25] S. Dauwe, J. Schmidt, R. Hezel, Proceedings of the 29th IEEE Photovolt. Specialists Conference, New Orleans, 2002, p. 1246. [26] The displayed surface recombination velocity for these films results from a new calculation assuming an infinite bulk lifetime. [27] W.M.M. Kessels, R.J. Severens, A.H.M. Smets, B.A. Korevaar, G.J. Adriaenssens, D.C. Schram, M.C.M. van de Sanden, J. Appl. Phys. 89 (2001) 2404. [28] H. Fujiwara, M. Kondo, Appl. Phys. Lett. 86 (2005) 032112. [29] D.C. Marra, W.M.M. Kessels, M.C.M. van de Sanden, K. Kashefizadeh, E.S. Aydil, Surf. Sci. 530 (2003) 1. [30] O.A. Aktsipetrov, A.A. Fedyanin, E.D. Mishina, A.N. Rubtsov, C.W. van Hasselt, M.A.C. Devillers, Th. Rasing, Phys. Rev. B 54 (1996) 1825. [31] J.J.H. Gielis, B. Hoex, M.C.M. van de Sanden, W.M.M. Kessels, J. Appl. Phys. 104 (2008) 073701. [32] A. Rumpel, B. Manschwetus, G. Lilienkamp, H. Schmidt, W. Daum, Phys. Rev. B 74 (2006) 081303. [33] A.H. Mahan, Y. Xu, B.P. Nelson, R.S. Crandall, J.D. Cohen, K.C. Palinginis, A.C. Gallagher, Appl. Phys. Lett. 78 (2001) 3788.
Contents lists available at ScienceDirect
Thin Solid Films j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / t s f
Silicon surface passivation by hot-wire CVD Si thin films studied by in situ surface spectroscopy J.J.H. Gielis, B. Hoex, P.J. van den Oever, M.C.M. van de Sanden, W.M.M. Kessels ⁎ Department of Applied Physics, Eindhoven University of Technology, The Netherlands
a r t i c l e
i n f o
Available online 7 February 2009 Keywords: Hot-wire deposition Amorphous silicon Epitaxial silicon Silicon wafer Surface passivation In situ optical studies
a b s t r a c t Silicon thin films can provide an excellent surface passivation of crystalline silicon (c-Si) which is of importance for high efficiency heterojunction solar cells or diffused emitter solar cells with well-passivated rear surfaces. Hot-wire chemical vapor deposition (hot-wire CVD) is an attractive method to synthesize Si thin films for these applications as the method is ion-bombardment free yielding good quality films over a wide range of deposition rates. The properties of the interface between hot-wire CVD Si thin films and H-terminated c-Si substrates have been studied during film growth by three complementary in situ techniques. Spectroscopic ellipsometry has been used to determine the optical properties and thickness of the films, whereas information on the H-bonding modes and H-depth profile has been obtained by attenuated total reflection infrared spectroscopy. Second-harmonic generation (SHG), a nonlinear optical technique sensitive to surface and interface states, has been used to probe two-photon resonances related to modified Si–Si bonds at the interface. By correlating the observations with ex situ lifetime spectroscopy experiments the growth and surface passivation mechanism of the Si films are discussed. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Surface passivation, i.e., the minimization of surface recombination losses, is essential for the performance of high efficiency c-Si solar cells. The importance of surface passivation is even increasing due to a cost-driven reduction of overall solar cell thickness leading to an increase in the surface-to-volume ratio. The reduction of charge carrier recombination losses at surfaces and interfaces becomes therefore a decisive factor for the final solar cell efficiency. In general, surface passivation can be achieved by a reduction of surface recombination centers, referred to as chemical passivation, and by shielding of charge carriers by internal electric fields, also known as field-effect passivation. Surface passivation of crystalline Si (c-Si) can be provided by several functional thin films such as SiO2 [1], hydrogenated amorphous silicon nitride (a-SiNx:H) [2], Al2O3 [3], and hydrogenated amorphous silicon (a-Si:H). Thin films of a-Si:H are used, for example in Si heterojunction (SHJ) solar cells [4,5] and in diffused emitter solar cells for rear-surface passivation [6]. The performance of these devices is highly determined by interface quality and an abrupt and atomically flat interface between c-Si and the a-Si:H film is required to reduce surface recombination losses. Hot-wire chemical vapor deposition (hot-wire CVD) is an attractive method to synthesize Si thin films for these applications as this technique yields good quality films over a wide range of deposition
parameters while providing a high atomic hydrogen flux and being ion-bombardment-free. A good to excellent level of surface passivation by a-Si:H films deposited by hot-wire CVD has been reported in some brief minority carrier lifetime studies [7,8], while most of the work in this respect has concentrated on the synthesis of complete SHJ solar cells [8–10]. Efficiencies as high as 18.2% have recently been reported for hot-wire CVD a-Si:H front and back contacts in p-type SHJ solar cells [11]. Recently we reported on the simultaneous application of three complementary optical diagnostics during hot-wire CVD of Si thin films [12,13]: spectroscopic ellipsometry (SE) was implemented to provide information on the optical properties and thickness of the Si thin films, attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) to reflect the H content in the films, and the nonlinear optical technique of second-harmonic generation (SHG) to reveal properties of the surface and the buried film/substrate interface [14]. In this paper results obtained from these diagnostics are combined with results on the surface passivation performance of the hot-wire CVD a-Si:H films as investigated by ex situ lifetime spectroscopy. The growth and surface passivation mechanism of the Si films are discussed focusing on the role of the substrate temperature, deposition rate, epitaxial film growth, defect states and built-in electric fields. 2. Experimental
⁎ Corresponding author. E-mail address: [email protected] (W.M.M. Kessels). 0040-6090/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2009.01.076
Si thin films were deposited using hot-wire CVD in a setup with a base pressure of b10− 9 mbar. A 0.45-mm-diameter coiled tungsten
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
3457
Physics Tsunami) tunable in the 1.35–1.75 eV photon energy range. An intracavity doubled Nd:YVO4 laser (Spectra Physics Millennia Vsj) was used to pump the oscillator. To separate the fundamental laser radiation from the SHG radiation generated at the sample, color filters and a Pellin Broca dispersing prism were applied. A photomultiplier tube connected to photon counting electronics was used to detect the SHG signal. The time resolution of SHG setup was 0.1 s. Both the incident fundamental and the detected SHG radiation were p-polarized. 3. Results and discussion 3.1. Passivation properties
Fig. 1. Surface recombination velocity Seff,max as a function of deposition temperature for Si thin films deposited with hot-wire CVD, plasma-enhanced CVD, and the expanding thermal plasma (ETP) technique.
filament was resistively heated by a 10 A dc to about 2000 °C to dissociate SiH4 gas (purity of over 99.995%) at a pressure of 8×10− 3 mbar. The filament was placed at 11–13 cm from the substrate, which resulted in deposition rates of 0.3–0.6 Å/s. The substrate temperature was actively controlled by a radiative heater at the rear side of the sample. The Si thin films were deposited on Si(100) substrates; either standard n-type wafers, undoped trapezoidally-shaped substrates with 45° angled bevels for experiments involving ATR-FTIR, or 2 in. n-type float zone wafers (P-doped, resistivity of 1.9 Ω cm, thickness of 275 μm) for lifetime spectroscopy. Prior to deposition the substrates were cleaned by an ultrasound ethanol bath and subsequent immersion in a 2% HF solution for 2 min or by standard RCA I and RCA II procedures using a buffered 1% NH4F/HF solution (BHF) with pH 4. Both cleaning methods removed the native oxide and terminated the Si surface with H. The passivation properties of the deposited Si thin films were determined by measuring the surface recombination velocity of charge carriers using carrier lifetime spectroscopy. These photoconductance experiments were carried out ex situ using a lifetime tester (Sinton Consulting WCT-100) in both quasisteady-state and transient mode [15]. The surface recombination velocity was calculated assuming an infinite bulk lifetime for the float zone c-Si wafer, which results in an upper limit for the surface recombination velocity Seff,max. The optical characterization of the Si thin films was performed in real time during deposition. Only the key aspects are briefly discussed here as the optical techniques are described in detail elsewhere [13,16,17]. The real-time SE data were acquired with a rotating compensator spectroscopic ellipsometer (J.A. Woollam M-2000U, 1.24–5.0 eV photon energy range) at a time resolution of 3.5 s. The SE data are expressed in terms of the pseudodielectric function bɛN = bɛ1N + ibɛ2N, which can be calculated by treating the film and the substrate as a single semi-infinite material. The ATR-FTIR measurements were performed using a Fourier transform infrared spectrometer (Bruker Vector 22 with a liquid nitrogen cooled HgCdTe detector) focusing on the frequency range of 1800–2250 cm− 1, that is, the region where radiation is absorbed due to Si–H stretching vibrations. The infrared radiation underwent ∼ 25 total internal reflections at each side of the sample, enhancing the sensitivity with a factor of ∼ 50 compared to single transmission. The spectral resolution of the ATR-FTIR experiments was 4 cm− 1 and the time resolution was 6.8 s. The ATR-FTIR data are expressed in terms of absorbance A, which is calculated from the transmission T and the background transmission T0 obtained prior to deposition by using A = −log(T/T0). The laser radiation for the SHG experiments had a pulse duration of ∼90 fs and was provided by a Ti:sapphire oscillator (Spectra
Fig. 1 shows the surface recombination velocity Seff,max as a function of deposition temperature for a-Si:H films deposited with hot-wire CVD at a deposition rate of 0.3 Å/s. The surface passivation properties significantly improve with increasing deposition temperature up to ∼ 140 °C and a Seff,max value as low as 17 cm/s is reached. For higher deposition temperatures the surface passivation properties strongly deteriorate again. The abrupt deterioration of the surface passivation properties for higher temperatures, can be attributed to the formation of epitaxial Si [8,12,13,18–20]. This effect causes a defect-rich interface between the c-Si substrate and the film. The formation of epitaxial material can be monitored with SE [21]. In Fig. 2 the imaginary part of the pseudodielectric function bɛ2N is shown as a function of photon energy for films deposited at a substrate temperature of ∼ 150 °C. The spectra in Fig. 2(a) clearly indicate the deposition of a-Si:H; with increasing film thickness the c-Si E′0/E1 and E2 critical point (CP) transitions broaden and ultimately cannot be separately distinguished anymore. The spectra in Fig. 2(b), however, hardly change with increasing film thickness. This reveals that these films developed epitaxially. In Fig. 2(c), an intermediate situation is displayed, that is, the formation of mixed phase material. In this case, the spectra also broaden with increasing thickness, however, not as rapidly as for purely amorphous material. Also, the characteristic c-Si E′0/E1 and E2
Fig. 2. Imaginary part of the pseudodielectric function bɛ2N as a function of photon energy obtained with in situ SE during deposition of (a) a-Si:H, (b) epitaxial Si, and (c) mixed phase material on H terminated Si(100). The substrate temperature was ∼150 °C for all films. Dashed lines indicate bɛ2N of H terminated Si(100) and ɛ2 of a-Si:H as deduced from the data.
3458
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
CP transitions remain present for thicker films. These results demonstrate that a-Si:H, epi-Si, and mixed phase formation can occur at similar processing conditions when the substrate temperature is at or close to the transition temperature from a-Si:H to epitaxial film growth. From the data in Fig. 1 it can be concluded that a high deposition temperature is beneficial for the surface passivation properties of a-Si:H. However, it is essential to prevent the formation of epitaxial Si. The latter can be accomplished by increasing the deposition rate [22–24] as is also illustrated in Fig. 1 for films deposited by radiofrequency plasma-enhanced CVD (PECVD) [25,26] and by the expanding thermal plasma (ETP) [27]. The PECVD data obtained for a deposition rate of 5 Å/s show the same trend as the hot-wire CVD data; an improvement of the surface passivation properties up to, in this case, ∼225 °C and a deterioration for higher temperatures. With the ETP deposited films low surface recombination velocities can be reached even at higher substrate temperatures when employing deposition rates of 10 Å/s (250 °C) and 30 Å/s (400 °C). These data clearly show that excellent passivation properties (Seff,max = 6 cm/s) can be obtained at deposition temperatures as high as 400 °C when using high deposition rates. Note that, similar to hot-wire CVD, ion bombardment of the substrate is negligible for the ETP technique [27].
3.2. Passivation mechanism As discussed above, surface passivation can be obtained by two mechanisms: chemical passivation and field-effect passivation. To investigate the surface passivation mechanism of a-Si:H we used two optical diagnostics: ATR-FTIR and SHG. In Fig. 3 the absorbance of a-Si:H and mixed phase Si thin films is shown as a function of frequency as measured with ATR-FTIR. For the a-Si:H film the absorbance increases strongly with increasing film thickness, in addition, the maximum in the absorbance shifts to lower wavenumbers with increasing film thickness. For mixed phase material the absorbance remains very low. Quantitative information about the hydrogen bonding in the film as a function of film thickness can be obtained by deconvolution of the spectra into several Gaussian peaks, corresponding to SiH and SiH2 bulk stretching modes and SiHx surface stretching modes [16,28,29]. The center positions of these Gaussian peaks are also indicated in Fig. 3. In Fig. 4(a) the integrated absorbance of the SiH and SiH2 bulk modes and the sum of the SiHx surface modes resulting from the deconvolution of the ATR-FTIR spectra are shown. For the a-Si:H film a strong
Fig. 3. ATR-FTIR spectra for a-Si:H (top) and mixed phase Si (bottom) deposited on H terminated Si(100). Data are obtained in real time during deposition and selected spectra for film thicknesses of 38, 100, 236, and 329 Å are shown. The dashed vertical lines indicate the position of the different SiH bulk and surface stretching modes.
Fig. 4. (a) Integrated absorbance by SiH and SiH2 bulk stretching modes and the sum of the surface SiHx modes for a-Si:H (closed symbols) and mixed phase Si films (open symbols) deposited on H terminated Si(100) as a function film thickness. (b) SiH and SiH2 bulk modes for a-Si:H and mixed phase Si displayed for a larger thickness range.
increase in absorbance by SiHx surface modes is observed at the onset of film growth. Subsequently, the absorption by the SiH2 bulk mode is dominant. After ∼40 Å of film growth, which corresponds to ∼100 s of film deposition, the absorbance is governed by the SiH bulk mode. The initial dominance of the SiHx surface modes and the SiH2 bulk modes indicates that the film growth starts off H rich. This H-rich interface region can be very important for the passivation properties of the a-Si:H films, as H can cause a reduction of interface defects and thereby prevent the recombination of charge carriers. Fig. 4(a) also shows that thin films of mixed phase Si display a very low integrated absorbance, which indicates that the H content, and thus the a-Si:H fraction, in these films is very low. As discussed above, the passivation properties of these films are marginal. Due to the low absorbance for these films, SiHx surface stretching modes could even not be distinguished from the bulk modes. In Fig. 4(b) the integrated absorbance for a-Si:H and mixed phase Si is shown over a larger thickness range for the SiH and SiH2 bulk modes. For a-Si:H the integrated absorbance for the SiH mode increases linearly as a function of film thickness indicating a constant H incorporation during bulk film growth. This constant H content can be calculated to be ∼12.5 at.%. For the mixed phase material the absorbance for the SiH mode increases more than linearly with thickness, indicating that the H content of the film and, hence, the a-Si:H content of the material increase with film thickness. The ATR-FTIR data provide evidence that chemical passivation is playing a role in the c-Si passivation mechanism of a-Si:H thin films. To investigate the possible additional influence of field-effect passivation, the nonlinear optical technique of second-harmonic generation was applied. Fig. 5(a) shows SHG spectra for a-Si:H films of 38, 100, and 727 Å. For comparison, also the SHG spectrum for a pristine H terminated Si(100) substrate is displayed. All spectra were obtained at room temperature. The spectra for the ultrathin films of 38 and 100 Å display a strongly enhanced SHG intensity compared to the spectrum for the H terminated substrate and have a sharp maximum at 3.3 eV. The spectrum for the thicker film of 727 Å shows a much lower SHG intensity, especially in the higher photon energy range. The combination of the spectra for the three films gives clear indications of both the macroscopic origin (i.e., the geometrical origin) and the microscopic origin (i.e., the origin on the atomic scale) of the SHG response. The decrease in SHG intensity for the thicker film can be explained by SHG radiation that is generated at the buried interface between the a-Si:H film and the c-Si substrate and that is absorbed while propagating through the film. The a-Si:H
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
3459
For the 38 and 100 Å thick a-Si:H films on H terminated Si(100) a shift in peak position from 3.4 eV to 3.3 eV due to reasons such as absorption or interference of multiple SHG contributions is not likely. a-Si:H is strongly absorbing SHG radiation, especially in the higher photon energy range, however, both the 38 and the 100 Å thick films display a maximum in SHG intensity at 3.3 eV irrespective of their thickness. Furthermore, the a-Si:H film on native oxide covered Si(100) also indicates that ∼100 Å of a-Si:H is too thin to effectively redshift a resonance from 3.4 eV to 3.3 eV. Interference of multiple SHG contributions would result in asymmetric features, which is not observed for the 38 and 100 Å thick films in Fig. 5(a). In summary, the sharp resonance at 3.3 eV is most likely governed by Si–Si bonds in the c-Si modified due to the presence of the interface with the a-Si:H. The absence of a EFISH contribution for this case indicates that no strong electric field is present in the a-Si:H/c-Si system. This leads to the conclusion that the influence of field-effect passivation on the c-Si surface passivation mechanism of a-Si:H is marginal. In this respect, the surface passivation properties of a-Si:H differ from dielectrics such as a-SiNx:H and Al2O3 that also provide an excellent level of surface passivation. For these materials field-effect passivation is predominant due to their very high density (1012– 1013 cm− 2) of fixed charges [2,31]. 4. Conclusion
Fig. 5. (a) SHG intensity as a function of SHG photon energy for a-Si:H films with a thickness of 38, 100, and 727 Å deposited on H terminated Si(100). In addition, the SHG spectrum for H terminated Si(100) is also shown. (b) SHG intensity as a function of SHG photon energy for 100 Å a-Si:H on deposited H terminated Si(100) and 97 Å a-Si:H deposited on native oxide (NO) covered Si(100). All data were obtained at room temperature. The dashed lines indicate the peak positions for the 100 and 97 Å thick a-Si:H films.
is strongly absorbing radiation at the SHG photon energy, especially in the higher photon energy range [13,17]. Equivalently, the decrease in SHG intensity for the thicker film shows that the influence of the surface and the bulk of the a-Si:H films is very limited. The strong increase of the SHG signal for the ultrathin films of 38 and 100 Å with respect to the spectrum for the substrate indicates that the possible influence of contributions from the substrate bulk is also limited. The sharp resonance at 3.3 eV as observed for the 38 and 100 Å thick films is an indication that the SHG response is related to the E′0/E1 critical point (CP) transition in the c-Si. This observation on the microscopic origin is in line with the conclusion that the SHG response is originating macroscopically from the buried a-Si:H/c-Si interface. In general, SHG contributions near the E'0/E1 CP energy might be related to Si–Si bonds in c-Si modified due to the presence of an interface or to electric-field-induced second-harmonic generation (EFISH, sensitive to electric fields N105 V/cm) [30,31]. However, from the location of the resonance it can be inferred that a dominant EFISH contribution is not likely. EFISH reflects the spectroscopic properties of the Si space-charge region. Consequently, at room temperature this effect would result in a resonance centered at 3.4 eV. This is clearly not the case for the 38 and 100 Å thick a-Si:H films deposited on H terminated Si(100), which have a strong resonance at 3.3 eV. For comparison, Fig. 5(b) shows the room temperature SHG spectrum obtained from a 97 Å thick a-Si:H film deposited on native oxide covered Si(100). For native oxide covered Si, the SHG response near the E′0/E1 CP energy is known to be originating from EFISH to a large extent, as related to a positively charged Si spacecharge region [32]. As Fig. 5(b) shows, for native oxide covered Si(100) with a 97 Å a-Si:H film a strong resonance at 3.4 eV is observed, hence, suggesting a strong influence of EFISH for this sample.
Thin films of a-Si:H result in an excellent level of c-Si surface passivation. In general, a high deposition temperature is beneficial for the surface passivation properties of a-Si:H thin films. However, it is essential to prevent the formation of epitaxial Si, which, at high temperatures, requires high deposition rates. Considering the fact that with hot-wire CVD a-Si:H films with good film properties can also be obtained at high deposition rates [33], hot-wire CVD is a very promising method to provide a high level of surface passivation of c-Si for solar cell applications. The ATR-FTIR and SHG experiments indicate that the c-Si surface passivation mechanism of a-Si:H is governed by chemical passivation with the influence of field-effect passivation being marginal. However, when doped a-Si:H layers are applied, for example in SHJ solar cells, space charge regions can exist and field-effect passivation can be of influence. This effect will improve the excellent surface passivation properties of a-Si:H even further. Acknowledgments The authors thank M. J. F. van de Sande, J. F. C. Jansen, J. J. A. Zeebregts, and H. M. M. de Jong for their skillful technical assistance. This work was supported by the Netherlands Foundation for Fundamental Research on Matter (FOM). References [1] M.J. Kerr, A. Cuevas, Semicond. Sci. Tech. 17 (2002) 35. [2] R. Hezel, R. Schorner, J. Appl. Phys. 52 (1981) 3076. [3] B. Hoex, S.B.S. Heil, E. Langereis, M.C.M. van de Sanden, W.M.M. Kessels, Appl. Phys. Lett. 89 (2006) 042112. [4] M. Taguchi, K. Kawamoto, S. Tsuge, T. Baba, H. Sakata, M. Morizane, K. Uchihashi, N. Nakamura, S. Kiyama, O. Oota, Prog. Photovolt. 8 (2000) 503. [5] S. Taira, Y. Yoshimine, T. Baba, M. Taguchi, H. Kanno, T. Kinoshita, H. Sakata, E. Maruyama, M. Tanaka, Proceedings of the 22nd European Photovoltaic Solar Energy Conference, Milan, Italy, 2007, p. 932. [6] M. Schaper, J. Schmidt, H. Plagwitz, R. Brendel, Prog. Photovolt. 13 (2005) 381. [7] C. Voz, I. Martin, A. Orpella, J. Puigdollers, M. Vetter, R. Alcubilla, D. Soler, M. Fonrodona, J. Bertomeu, J. Andreu, Thin Solid Films 430 (2003) 270. [8] T.H. Wang, E. Iwaniczko, M.R. Page, D.H. Levi, Y. Yan, H.M. Branz, Q. Wang, Thin Solid Films 501 (2006) 284. [9] D. Muñoz, C. Voz, A. Orpella, J. Puigdollers, R. Alcubilla, F. Villar, J. Bertomeu, J. Andreu, J. Doman-Lacoste, P. Roca i Cabarrocas, Thin Solid Films 516 (2008) 761. [10] S.-Y. Lien, B.-R. Wu, J.-C. Liu, D.-S. Wuu, Thin Solid Films 516 (2008) 747. [11] H.M. Branz, C.W. Teplin, D.L. Young, M.R. Page, E. Iwaniczko, L. Roybal, R. Bauer, A.H. Mahan, Y. Xu, P. Stradins, T. Wang, Q. Wang, Thin Solid Films 516 (2008) 743.
3460
J.J.H. Gielis et al. / Thin Solid Films 517 (2009) 3456–3460
[12] J.J.H. Gielis, P.J. van den Oever, M.C.M. van de Sanden, W.M.M. Kessels, Appl. Phys. Lett. 90 (2007) 202108. [13] J.J.H. Gielis, P.J. van den Oever, B. Hoex, M.C.M. van de Sanden, W.M.M. Kessels, Phys. Rev. B 77 (2008) 205329. [14] T.F. Heinz, Second-order Nonlinear Optical Effects at Surfaces and Interfaces, in: H.E. Ponath, G.I. Stegeman (Eds.), Nonlinear Surface Electromagnetic Phenomena, Elsevier, Amsterdam, 1991. [15] R.A. Sinton, A. Cuevas, Appl. Phys. Lett. 69 (1996) 2510. [16] P.J. van den Oever, J.J.H. Gielis, M.C.M. van de Sanden, W.M.M. Kessels, Thin Solid Films 516 (2008) 511. [17] J.J.H. Gielis, P.M. Gevers, A.A.E. Stevens, H.C.W. Beijerinck, M.C.M. van de Sanden, W.M.M. Kessels, Phys. Rev. B 74 (2006) 165311. [18] Y. Yan, M. Page, T.H. Wang, M.M. Al-Jassim, H.M. Branz, Q. Wang, Appl. Phys. Lett. 88 (2006) 121925. [19] S. De Wolf, M. Kondo, Appl. Phys. Lett. 90 (2007) 042111. [20] H. Fujiwara, M. Kondo, Appl. Phys. Lett. 90 (2007) 013503. [21] At the onset of film growth, however, SE cannot resolve the film morphology due to a lack in contrast between c-Si substrates and Si films, see Ref. 13. [22] D.J. Eaglesham, H.-J. Gossmann, M. Cerullo, Phys. Rev. Lett. 65 (1990) 1227. [23] H. Jorke, H. Kibbel, F. Schäffler, H.-J. Herzog, Thin Solid Films 183 (1989) 307.
[24] J. Thiesen, H.M. Branz, R.S. Crandall, Appl. Phys. Lett. 77 (2000) 3589. [25] S. Dauwe, J. Schmidt, R. Hezel, Proceedings of the 29th IEEE Photovolt. Specialists Conference, New Orleans, 2002, p. 1246. [26] The displayed surface recombination velocity for these films results from a new calculation assuming an infinite bulk lifetime. [27] W.M.M. Kessels, R.J. Severens, A.H.M. Smets, B.A. Korevaar, G.J. Adriaenssens, D.C. Schram, M.C.M. van de Sanden, J. Appl. Phys. 89 (2001) 2404. [28] H. Fujiwara, M. Kondo, Appl. Phys. Lett. 86 (2005) 032112. [29] D.C. Marra, W.M.M. Kessels, M.C.M. van de Sanden, K. Kashefizadeh, E.S. Aydil, Surf. Sci. 530 (2003) 1. [30] O.A. Aktsipetrov, A.A. Fedyanin, E.D. Mishina, A.N. Rubtsov, C.W. van Hasselt, M.A.C. Devillers, Th. Rasing, Phys. Rev. B 54 (1996) 1825. [31] J.J.H. Gielis, B. Hoex, M.C.M. van de Sanden, W.M.M. Kessels, J. Appl. Phys. 104 (2008) 073701. [32] A. Rumpel, B. Manschwetus, G. Lilienkamp, H. Schmidt, W. Daum, Phys. Rev. B 74 (2006) 081303. [33] A.H. Mahan, Y. Xu, B.P. Nelson, R.S. Crandall, J.D. Cohen, K.C. Palinginis, A.C. Gallagher, Appl. Phys. Lett. 78 (2001) 3788.