Open Access
Add to BibSonomyAbstract
The method of elastic recoil detection (ERD) has been utilized to analyze the actual hydrogen concentration in hydrogenated silicon thin films and has been demonstrated to provide more stable and accurate analysis for quantification of the hydrogen concentration than does Fourier Transform Infrared Spectroscopy. In order to realize the quality of thin films, spectroscopic ellipsometry and the applied effective medium approximation (EMA) theory of the Maxwell-Garnett and Bruggeman models are applied to deducing the effective refractive index of the films. The relative void fractions corresponding to amorphous silicon films with lowest hydrogen flow could be obtained based on the EMA theory. Therefore, this study found an important linear relationship between the hydrogen concentration and structural relaxation as a consequence of the void fraction induced by the hydrogen flow.
©2013 Optical Society of America
1. Introduction
Hydrogenated silicon (Si:H) thin films have attracted a great deal of attention for thin film solar cell applications due to their superior properties and low manufacturing cost [1–3]. Controlling the hydrogen flow is one of the most important factors to achieving good quality silicon thin films, assisting in the formation of the film during the fabrication process [1]. The interaction between the hydrogen and silicon atoms can be affected by improving the crystalline silicon ratio [4]. The mechanism leading to the phase transition has been reported upon in a past study [5]. Controlling the hydrogen flow during the deposition of Si:H thin films affects the transition phase from amorphous to microcrystalline types. However, the relationship between the crystalline silicon volume fraction and corresponding hydrogen concentration has not yet been clearly clarified. Normally, the hydrogen concentration in a silicon thin film is confirmed using Fourier Transform Infrared Spectroscopy (FTIR) related to the absorption of the silicon-hydrogen bonds [6,7]. However, only the specific bonds or vibration mode can be detected, such as the rocking mode of the silicon-hydrogen bond [8] which leads to concomitant uncertainty. In order to provide the absolute value of hydrogen content, elastic recoil detection (ERD) can support this analysis [9–11] and even compensate the FTIR measurement [12,13]. There are also many materials that can be analyzed through ERD such as olivine crystals and diamond nanorod (DNR) thin films [14,15].
In this paper, we discuss application of the technique of ERD to calibrate the FTIR method. The Si:H films are then characterized by optical measurement through spectroscopic ellipsometry and the results compared to the Effective Medium Approximation (EMA) [16,17] results. We found a significant difference, mainly coming from the internal void fraction [4], which could fill with hydrogen [18,19] and even be connected with different hydrogen concentrations in this study during the formation of Si:H thin films.
2. Fabrication and measurement
Si:H thin films were fabricated by a pulse DC magnetron sputtering deposition system on B270 glass substrates. The substrates were cleaned using detergent and rinsed by de-ionized water to remove particles and organic matters on the surface. Therefore, the surface of the substrates could achieve small contact angles less than 8 degrees corresponding to good hydrophilic properties prior to the thin film deposition. Before sputtering, the chamber was pumped to below 8 × 10−6 Torr by a turbo molecular pump to reduce the residual gases in the chamber. The target for sputtering was a 3-inch silicon disk with a purity of 99.999%. Argon and hydrogen working gases were introduced into the chamber through mass flow controllers. The argon flow rate was fixed at 10 sccm, while the hydrogen partial pressure was varied from 2.8 mTorr to 15.8 mTorr by adjusting the hydrogen flow from 1 to 9 sccm; the tuning interval was 1 sccm. These deposited Si:H films are all controlled in physical thickness to around 250 nm. After the thin film deposition process, the hydrogen concentration was measured using FTIR and ERD.
The FTIR was a HORIBA FT-720 with a measurement wavenumber from 400 cm−1 to 4000 cm−1. The ERD setup is depicted in Fig. 1(a). An incident helium ion beam (energy of 2.5 MeV) was collimated and fell incident onto the silicon films. Hydrogen atoms knocked out of the films were detected by one surface barrier detector. The energy signals were amplified, and collected to show the energy spectrum with a fixed ion dose. The RUMP simulation software was applied to obtain the depth profile of the hydrogen composition, by simulating the energy-loss process from the sample surface; for example, see Fig. 1(b). In addition, the crystalline volume fraction of each silicon film was measured using Raman spectroscopy.
Fig. 1 (a) The setup for ERD; (b) hydrogen concentration in hydrogenated silicon thin film analyzed through ERD when hydrogen flow is 2 sccm.
3. Results and discussion
Silicon thin films fabricated utilizing different hydrogen flow rates contain different kinds of silicon-hydrogen bonds [20]. FTIR analysis is the method generally used to estimate the hydrogen concentration needed to achieve absorption of the silicon-hydrogen bonds at a wave number of 640 cm−1 [21]. However, we can only determine the relative hydrogen concentration from specific silicon-hydrogen bonds, not the overall hydrogen concentration actually existing in silicon thin films.
To overcome this problem, we applied the ERD technique, which is an ion beam analysis technique for quantitative analysis of light elements in solids. From the measured energy spectrum of the recoils a concentration depth profile can be calculated which can be used to obtain the major hydrogen atom concentration. In addition, the interaction between hydrogen and silicon atoms will transform the amorphous silicon into a microcrystalline state [5]. The crystalline volume fraction of silicon can be determined by Raman Spectroscopy [22,23]. The major crystalline orientation was the <111> plane located at diffraction angle (2θ) of 28.41° via X-ray diffraction (XRD) measurement [24]. For comparison, we also used FTIR to determine the hydrogen concentration in Si:H films [25,26]. Silicon-hydrogen bonds contribute to the absorption spectrum through three principle modes: stretching (~2000 cm−1), bending (~850 cm−1) and rocking (~640 cm−1). Most of the bonded hydrogen, however, contributes to the rocking mode, at about 640 cm−1, and an estimate of the bonded hydrogen content can probably be obtained from this peak integration [8,27].
The comprehensive results of the crystalline silicon volume fraction, ERD and FTIR hydrogen concentrations are compared in Table 1. An examination of Fig. 2 shows a similar trend for both ERD and FTIR, that the maximum hydrogen concentrations are located at similar hydrogen flow. However, with FTIR, the hydrogen concentration is greatly underestimated, because it detects only specific bonds and vibration modes.
Table 1. The crystalline volume fraction and hydrogen concentration in silicon thin films versus hydrogen flow.
Fig. 2 Comparison of the conventional FTIR method and proposed ERD method for determining the hydrogen concentration in silicon thin films versus increasing hydrogen flow.
Figure 3 shows the relationship between the crystalline volume fraction and the calculated refractive index with hydrogen flow rates from 1 sccm to 9 sccm. The crystalline volume fraction of each silicon film was measured using Raman spectroscopy. From the figure, we can see that the crystalline volume fraction increases with increasing hydrogen flow when the hydrogen flow is larger than the crystallization threshold, 4 sccm. Below this threshold, the silicon thin film is only amorphous. The refractive index is achieved using spectroscopic ellipsometry (GES-5 SOPRA) for silicon films with the hydrogen flow of 1 sccm. Therefore, we used 4.50 and 3.89 as the refractive indices for amorphous silicon and crystalline silicon, respectively. Based on the effective medium approximation (EMA) theory, we can calculate the refractive index for the Si:H thin films corresponding to the crystalline volume fraction, as determined by Raman Spectroscopy analysis. Since such silicon thin films contain amorphous and microcrystalline types, the EMA of the Maxwell-Garnett model [28,29] and Bruggeman model [29,30] can be separately applied for these two cases. If the film contains two compositions with large volume difference to each other, it could logically resemble a Si:H film of low crystalline volume fraction such that one composition, crystalline silicon, nc, has a small fraction, fc, compared to the surrounding host amorphous silicon, na, which has a large fraction, fa. Therefore, the effective refractive index, neff, can be applied in the Maxwell-Garnett model, which can be expressed as
Fig. 3 Crystalline volume fraction of silicon and relative calculated refractive index of silicon thin films versus increasing hydrogen flow.
Moreover, the Bruggeman model is suitable for modeling microcrystalline materials. The amorphous and crystalline optical constants are mixed together when the fraction of crystalline silicon in the films is great enough that the effective refractive index, neff, can be deduced as follows:
Since crystalline silicon has a low refractive index, the effective refractive index obtained for Si:H thin films should decrease when the crystalline volume fraction of silicon is increased, as in Fig. 3. After deduction of the effective refractive index in Fig. 3, all of the samples with different hydrogen flow rates were measured by spectroscopic ellipsometry. The hydrogenated silicon films were analyzed by ellipsometric measurements using the Forouhi Bloomer Model [31] to determine optical constants directly. The film thickness was around 250 nm. All of the determined results indicate that the standard deviations are all about 2x 10−3, thereby ensuring more precise and trusted values. Figure 4(a) shows the results for hydrogen flows of 1, 3, 5, 7 and 9 sccm. The trend shows a surprising decrease in the refractive index for hydrogen flows of 1, 3, 5 sccm, but an increase for 7 and 9 sccm. Figure 4(b) shows the refractive indices of all the samples at photon energy of 2 eV. The trend is totally different to that observed in Fig. 3. To explain this result, we suggest that the mechanism could be due to the different silicon thin film densities that occur due to hydrogen assistance. Under the crystallization threshold, the refractive index should remain almost the same, because of the existence of only one material, amorphous silicon. However, in our measurements, the refractive indices of the thin films decreased even when only the amorphous type was present. This could be caused by structural relaxation induced by the generation of void fractions in these films. Above the threshold, silicon crystallization starts to occur so the films become denser with increasing hydrogen flow.
Fig. 4 (a) Refractive index of silicon thin films with hydrogen flows from 1, 3, 5, 7 and 9 sccm analyzed by spectroscopic ellipsometry; (b) the refractive index at a photon energy of 2 eV for silicon thin films with hydrogen flows from 1 to 9 sccm.
Figure 5 shows a Si:H film that is assumed to be 100% dense with a void fraction of 0% when the hydrogen flow rate is 1 sccm. The relative void fractions of other samples calculated through the EMA theory are proportional to the hydrogen flow before the start of silicon crystallization. After crystallization, this decreases so the films become denser.
Fig. 5 Relative void fraction of silicon thin films for a hydrogen flow of 1 sccm versus hydrogen flows from 1 to 9 sccm.
Figure 6 shows an illustration to help visualize the above results concerning structural variation of Si:H thin films with increasing hydrogen flow. Before crystallization, the hydrogen passivated dangling bonds [1] bonded with Si atoms during film deposition, so the void fraction of the films could continuously increase. After the initiation of silicon crystallization (c-Si), the number of crystalline Si-Si bonds increases and the amount of hydrogen-silicon bonds is suppressed at the same time [32] which would result in a decreasing hydrogen concentration and void fraction in the films.
Fig. 6 Structural variation of silicon thin films from amorphous to microcrystalline types with increasing hydrogen flows. Meanwhile, the void fraction and interior hydrogen concentration increase before the start of crystallization. The void fraction begins to be saturated and even decreases as crystallization starts to be generated.
To analyze the relationship of the hydrogen concentration with the relative void fraction, Fig. 7 shows the double linear relationships found between the hydrogen concentration and relative void fraction that is directly reflected in structural relaxation. When crystallization is below 7%, the hydrogen concentration is linearly proportional to the relative void fraction, following a slope of 0.89. After crystallization exceeds 7%, the hydrogen concentration decreases linearly with reducing void fractions and also follows the same slope of 0.89. The only difference between both linear relations is a shift in the void fraction of 11.21% that may also include the extra grain boundaries generated between crystallizations. In addition, all of the Si:H films have hydrogen concentrations above 12.58% which is a good norm in our samples, since a hydrogen concentration below 8% generally leads to a high defect density and a reduced photo response [33].
Fig. 7 It can be seen from the relationship between the hydrogen concentration and structural relaxation induced by the relative void fraction that Si:H thin films (from amorphous to microcrystalline) follow the same slope variation with only a difference in the shift of the void fraction by 11.21%.
4. Conclusion
In summary, we have demonstrated how to use the ERD method to analyze the actual hydrogen concentration in silicon thin films. This method gives a more stable and accurate analysis than the commonly used FTIR method and can compensate for deficiencies in the FTIR results, especially when the hydrogen concentration is high. The accurate evaluation of the hydrogen concentration allows us to observe that hydrogen cannot only assist the formation of crystalline silicon but also affect the microstructure of thin films. By applying the EMA theories for effective refractive index calculation, the structural relaxation from induced void fractions in Si:H thin films has been obtained. Moreover, we found the linear relations between the hydrogen concentration and the relative void fraction that can be fitted very well and interpreted for Si:H thin films from amorphous to microcrystalline. Through this linear relationship, it can be applied to judge the quality of hydrogenated silicon films and even the final power conversion performance of solar cell devices.
Acknowledgments
The authors thank the National Science Council of Taiwan for the financial support of this research under Contract No. NSC 101-2221-E-008-052, 101-3113-P-008-009 and also express our appreciation to Chun-Yen Cheng (Nuclear Science and Technology Development Center, National Tsing-Hua University, Taiwan) for his great help with the elastic recoil detection (ERD) measurements.
References and links
1. B. Rolf, Thin-Film Crystalline Silicon Solar Cells: Physics and Technology (Weinheim: Wiley-VCH, 2003).
2. S. Klein, T. Repmann, and T. Brammer, “Microcrystalline silicon films and solar cells deposited by PECVD and HWCVD,” Sol. Energy 77(6), 893–908 (2004). [CrossRef]
3. J. Nelson, The Physics of Solar Cells (Imperial College Press, 2003).
4. O. Vetterl, F. Finger, R. Carius, P. Hapke, L. Houben, O. Kluth, A. Lambertz, A. Muck, B. Rech, and H. Wagner, “Intrinsic microcrystalline silicon: a new material for photovoltaics,” Sol. Energy Mater. Sol. Cells 62(1-2), 97–108 (2000). [CrossRef]
5. M. Katiyar and J. R. Abelson, “Investigation of hydrogen plasma induced phase transition from a-Si:H to mc-Si:H using real time infrared spectroscopy,” Mater. Sci. Eng. A 304–306, 349–352 (2001). [CrossRef]
6. T. Su, P. C. Taylor, G. Ganguly, and D. E. Carlson, “Direct role of hydrogen in the Staebler-Wronski effect in hydrogenated amorphous silicon,” Phys. Rev. Lett. 89(1), 015502–015506 (2002). [CrossRef] [PubMed]
7. Y. Kuo, Thin Film Transistors-Material and Processes, Volume 1 (Texas A&M University, U.S.A., 2004).
8. R. E. J. Schropp and M. Zeman, Amorphous and Microcrystalline Silicon Solar Cells: Modeling, Materials and Device Technology (Kluwer Academic Publishers, 1998).
9. N. Q. Khánh, M. Serényi, A. Csik, and C. Frigeri, “Determination of hydrogen concentration in a-Si and a-Ge layers by elastic recoil detection analysis,” Vacuum 86(6), 711–713 (2012). [CrossRef]
10. C. Frigeri, M. Serenyi, N. Q. Khanh, A. Csik, L. Nasi, Z. Erdelyi, D. L. Beke, and H. G. Boyen, “Hydrogen behaviour in amorphous Si/Ge nano-structures after annealing,” Appl. Surf. Sci. 267, 30–34 (2013). [CrossRef]
11. I. Crupi, F. S. Ruggeri, A. Grasso, F. Ruffino, G. Catania, A. M. Piro, S. Di Marco, S. Mirabella, F. Simone, and F. Priolo, “Influence of the electro-optical properties of an α-Si:H single layer on the performances of a pin solar cell,” Thin Solid Films 520(11), 4036–4040 (2012). [CrossRef]
12. Z. Said-Bacar, P. Prathap, C. Cayron, F. Mermet, Y. Leroy, F. Antoni, A. Slaoui, and E. Fogarassy, “CW laser induced crystallization of thin amorphous silicon films deposited by EBE and PECVD,” Appl. Surf. Sci. 258(23), 9359–9365 (2012). [CrossRef]
13. R. Garcia-Hernansanz, E. Garcia-Hemme, D. Pastor, A. del Prado, I. Martil, G. Gonzalez-Diaz, J. Olea, and D. Pastor, “Hydrogenated amorphous silicon deposited by high pressure sputtering for HIT solar cells,” in Proceedings of IEEE conference on Electron Devices (CDE), (Institute of Electrical and Electronics Engineers, New York, 2013), pp. 337–340. [CrossRef]
14. A. C. Withers, H. Bureau, C. Raepsaet, and M. M. Hirschmann, “Calibration of infrared spectroscopy by elastic recoil detection analysis of H in synthetic olivine,” Chem. Geol. 334, 92–98 (2012).
15. A. R. Sobia, S. Adnan, A. Mukhtiar, A. A. Khurram, A. A. Turab, A. Awais, A. Naveed, Q. J. Faisal, H. Javaid, and G. J. Yu, “Effect of nitrogen addition on hydrogen incorporation in diamond nanorod thin films,” Curr. Appl. Phys. 12(3), 712–717 (2012). [CrossRef]
16. D. Stroud, “The effective medium approximations: Some recent developments,” Superlattices Microstruct. 23(3–4), 567–573 (1998). [CrossRef]
17. S. Torquato and S. Hyun, “Effective-medium approximation for composite media: Realizable single-scale dispersions,” J. Appl. Phys. 89(3), 1725–1729 (2001). [CrossRef]
18. C. Garozzo, R. A. Puglisi, C. Bongiorno, C. Spinella, S. Mirabella, R. Reitano, S. D. Marco, M. Foti, and S. Lombardo, “Evolution of SiHx hydrides during the phase transition from amorphous to nanocrystalline silicon films,” J. Appl. Phys. 111(4), 043510 (2012). [CrossRef]
19. D. Gracina, Z. Siketic, K. Juraic, and M. Ceh, “Analysis of amorphous-nanocrystalline silicon thin films by time-of-flight elastic recoil detection analysis and high-resolution electron microscopy,” Appl. Surf. Sci. 275, 19–22 (2013). [CrossRef]
20. H. Vach, Q. Brulin, N. Chaabane, T. Novikova, P. Roca i Cabarrocas, B. Kalache, K. Hassouni, S. Botti, and L. Reining, “Growth dynamics of hydrogenated silicon nanoparticles under realistic conditions of a plasma reactor,” Comput. Mater. Sci. 35(3), 216–222 (2006). [CrossRef]
21. V. Verlaan, C. H. M. van der Werf, W. M. Arnoldbik, H. D. Goldbach, and R. E. I. Schropp, “Unambiguous determination of Fourier-transform infrared spectroscopy proportionality factors: The case of silicon nitride,” Phys. Rev. B 73(19), 195333 (2006). [CrossRef]
22. P. Wurfel, Physics of Solar Cells from Principles to New Concepts (Wiley-VCH Verlag GmbH, 2005).
23. S. C. Saha, A. K. Barua, and S. Ray, “The Role of Hydrogen Dilution and Radio Frequency Power in the Formation of Microcrystallinity of n-type Si:H Thin Films,” J. Appl. Phys. 74(9), 5561–5568 (1993). [CrossRef]
24. W. Du, X. Yang, H. Povolny, X. Liao, and X. Deng, “Impact of hydrogen dilution on microstructure and optoelectronic properties of silicon films deposited using trisilane,” J. Phys. D Appl. Phys. 38(6), 838–842 (2005). [CrossRef]
25. H. M. Brodsky, M. Cardona, and J. J. Couomoet, “Infrared and Raman spectra of the silicon-hydrogen bonds in amorphous silicon prepared by glow discharge and sputtering,” Phys. Rev. B 16(8), 3556–3571 (1977). [CrossRef]
26. P. J. Zanzucchi, C. R. Wronski, and D. E. Carlson, “Optical and photoconductive properties of discharge-produced amorphous silicon,” J. Appl. Phys. 48(12), 5227–5236 (1977). [CrossRef]
27. M. Cardona, “Vibrational spectra of hydrogen in silicon and germanium,” Phys. Status Solidi B 118(2), 463–481 (1983). [CrossRef]
28. R. Ruppin, “Evaluation of extended Maxwell-Garnett theories,” Opt. Commun. 182(4-6), 273–279 (2000). [CrossRef]
29. S. H. Chen, H. W. Wang, and T. W. Chang, “Absorption coefficient modeling of microcrystalline silicon thin film using Maxwell-Garnett effective medium theory,” Opt. Express 20(S2Suppl 2), A197–A204 (2012). [CrossRef] [PubMed]
30. A. Bittar, S. Berthier, and J. Lafait, “Non metal-metal transition in Bruggeman optical theory for inhomogeneous media,” J. Phys. 45(3), 623–631 (1984). [CrossRef]
31. A. R. Forouhi and I. Bloomer, “Optical dispersion relations for amorphous semiconductors and amorphous dielectrics,” Phys. Rev. B Condens. Matter 34(10), 7018–7026 (1986). [CrossRef] [PubMed]
32. J. M. Seo, M. C. Jeong, and J. M. Myoung, “Effects of hydrogen on poly-and nano-crystallization of a-Si: H prepared by RF magnetron sputtering,” J. Cryst. Growth 295(2), 119–123 (2006). [CrossRef]
33. A. H. Mahan, J. Carapella, B. P. Nelson, R. S. Crandall, and I. Balberg, “Deposition of device quality, low H content amorphous silicon,” J. Appl. Phys. 69(9), 6728–6730 (1991). [CrossRef]
