- Email: [email protected]
Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions
TSF-33757; No of Pages 5 Thin Solid Films xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions Kazufumi Azuma a, Satoko Ueno a, Yoshiyuki Konishi b, Kazuhiro Takahashi c a b c
Technology Research Laboratory, Shimadzu Corp., Kanagawa 259-1304, Japan Semiconductor Equipment Division, Shimadzu Corp., Kanagawa 259-1304, Japan Analytical & Measuring Instruments Division, Shimadzu Corp., Kanagawa 259-1304, Japan
a r t i c l e
i n f o
Article history: Received 5 June 2014 Received in revised form 24 September 2014 Accepted 26 September 2014 Available online xxxx Keyword: Surface wave plasma Chemical vapor deposition Silicon nitride Low temperature XPS RBS FTIR
a b s t r a c t Highly transparent silicon nitride films with a low absorption coefficient of only 200 cm-1 or lower were prepared under high NH3/SiH4 source gas ratio conditions at 80 °C or lower temperature using surface wave plasma chemical vapor deposition (SWP-CVD). Rutherford backscattering measurements indicated that a silicon nitride structure Si3Nx (x N 5) with excess nitrogen could be prepared with the SWP-CVD method under high NH3/SiH4 source gas ratio conditions. X-ray photon spectroscopy (XPS) analysis provided verification that the excess nitrogen combined with the oxygen contained in the SiNx film during low-temperature film formation and that the atomic ratio of Si and N was almost stoichiometric, i.e., Si3N4. XPS study also revealed that the Si3Nb4 structure contained suboxides, which presence may reduce the transparency of the films. In contrast, suboxides were not observed in the Si3N4 structure obtained under high NH3/SiH4 source gas ratio conditions. Fourier-transform infrared spectroscopy study confirmed that the SiNx film becomes more stable when the SiNx structure approaches the stoichiometric ratio. Achieving a near-transparent Si3N4 structure requires a sufficient amount of N atoms in the periphery of the Si atom, i.e., an adequate amount of NH3 is necessary in the presence of the SiH4. © 2014 Elsevier B.V. All rights reserved.
1. Introduction High-performance transparent barrier films have recently attracted much attention for flexible applications such as organic light emitting diode (OLED) displays [1]. Organic materials are susceptible to water, oxygen, and other environmental elements present in ambient conditions [2–4]. Many studies report on encapsulation methods or materials for increasing OLED display lifetimes. Transparency is particularly critical for encapsulating films that are applied to top-emitting OLED displays. The water vapor transmission rate requirement for OLED encapsulation is less than 10-6 g/m2/day [5]. Silicon nitride (SiNx) films are the strongest candidates for a high-performance barrier film. There are many reports that discuss the transparency of SiNx films [6–9]. These reports examine the influence of the Si/N ratio on optical characteristics, but few ones discuss the bonding states of transparent SiNx films. For this study, we prepared SiNx films under low temperature (less than 80 °C) conditions using a surface wave plasma chemical vapor deposition (SWP-CVD) system and elucidated the relationship between the Si/N elemental ratio and the Si\N bonding states using
E-mail address: [email protected] (K. Azuma).
Rutherford backscattering spectroscopy (RBS) and X-ray photoelectron spectroscopy (XPS). 2. Experimental methods Fig. 1 shows a schematic diagram of the SWP-CVD system [10]. This system has a plasma source with a slot antenna. Microwaves with a frequency of 2.45 GHz were introduced through the slot antenna and alumina dielectric window into the reactor. Ar, SiH4 and NH3 gases were fed from the upper and lower gas inlets. The distance between the dielectric window and the substrate stage was fixed at 200 mm. The films were prepared by exposing 6-inch- diameter FZ Si (100) wafers and quartz substrates to Ar, SiH4 and NH3 gases that were introduced from the gas inlets at flow rates of 350, 70, and 90–500 sccm, respectively. The total gas pressure was maintained at 10 Pa. The microwave power density was fixed at 1.57 W/cm2. The deposition time was fixed at 100 s for each sample. The deposition rate was almost the same when the SiH4 gas flow rate was constant (2.0 nm/s for the case of SiH4 70 sccm). The substrate and the gas lines were not positively heated. The surface temperature of the substrate increased from ambient temperature and saturated at 80 °C during 100 s of deposition time. The optical absorption coefficients of the SiNx films were determined by transmittance and reflectance (T–R) measurement in the 300– 800 nm range with a Shimadzu UV2600 spectrophotometer.
http://dx.doi.org/10.1016/j.tsf.2014.09.068 0040-6090/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
2
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx
Fig. 1. Experimental apparatus of surface wave plasma chemical vapor deposition system.
Film structures were evaluated with a Shimadzu FTIR-8400 Fourier transform infrared (FTIR) spectrometer and by XPS. XPS analysis was performed using a Shimadzu/Kratos AXIS-Nova spectrometer with a Vision 2 data system. The excitation source was a monochromated Al-Kα X-ray with 225 W-power, the size of the analysis area was 700 × 300 μm, and the emission angle of the photoelectron with respect to the sample normal was zero degrees. No ion sputtering was performed prior to the XPS analysis. For XPS data processing of Si 2p peaks, a linear background was subtracted before peak synthesis, which was performed with component peaks having a Gaussian (70%)–Lorentzian (30%) product peak shape. Without any constraints on peak position, peak height and peak width between component peaks were set during peak synthesis process. The position of the peaks was calibrated so that the highest Si 2p component appeared at 101.80 eV. The films were measured by XPS within 30 to 60 min after removal of the samples to ambient air from the vacuum tool after film formation. The elemental ratio of the SiNx films was also analyzed by RBS (NEC 3S-R10 and CEA RBS-400). RBS analysis can provide elemental bulk information of a film for the first several hundreds of nanometers of the surface. Measurements were performed with an incident energy of 2.275 MeV 4He++, a beam diameter of 1–2 mm, and detection angles of 160° (normal angle) and 119° (grazing angle).
Fig. 3. Optical band gap (Eg opt) as a function of source gas flow ratio. The conditions for each reference are as follows: [Ref. 6] RF PE-CVD, SiH4 + NH3 at 300 °C. [Ref. 7] RF PE-CVD, SiH4 + N2 at 300 °C. [Ref. 8] RF PE-CVD, SiH4 + NH3 at 200 °C. [Ref. 9] ECR PE-CVD, SiH4 + N2 at 60 °C.
3. Results and discussion 3.1. Optical properties of SiNx films First, the optical absorption coefficient of SiNx films was calculated from the Beer–Lambert law for the films prepared with various SiH4/ NH3 gas ratios. Fig. 2 shows the relationship between the absorption coefficient of the films at the wavelength of 400 nm and a supplied gas ratio of NH3 and SiH4. There is a noteworthy drop in the absorption coefficient down to a few hundred cm-1 or less at a NH3/(SiH4 + NH3) ratio of 0.7 and higher. An optical band gap energy E04 was defined to be the spectra wavelength (energy) for which the absorption coefficient α = 104 cm-1 in wavelength-dispersive optical absorption [11,12]. The optical band gap was plotted as a function of the NH3/(SiH4 + NH3) ratio in Fig. 3. Films with an optical absorption of lower than 104 cm-1 in the measured range were not plotted. Data from previous reports [6–9] were also plotted in the figure. Refs. [6,8] used RF plasma enhanced chemical vapor deposition (PE-CVD) with SiH4 and NH3 as source gases. The deposition temperature was 200–300 °C. Ref. [7] used RF PE-CVD with SiH4 and N2. The deposition temperature was 200 °C. Ref. [9] used electron cyclotron resonance (ECR) PE-CVD with SiH4 and N2 and with a deposition temperature of 60 °C. Refs. [6,8] determined the optical band gap using E04 and Refs. [7,9] determined the same with Tauc's relation [13]. According to Fig. 3, the optical band gap increases with an increase in the NH3 or N2 ratio, indicating that adequate NH3 or N2 is needed for transparent SiNx films. Since the PE-CVD process is a predominantly surface reaction, the usual SiNx film preparation involved heating at 200 °C. Since heating promotes Si\N bond formation, the SiNx structure approaches the stoichiometric ratio, which leads to a lower absorption
Table 1 Film preparation conditions (flow rates of source gases).
Fig. 2. Relationship between absorption coefficient of the SiN x films and NH 3 / (SiH 4 + NH 3) source gas ratio.
Sample
SiH4 (sccm)
NH3 (sccm)
1 2 3 4 5 6
70 70 70 70 70 70
650 500 350 175 153 90
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx Table 2 Result of RBS and XPS analysis. Sample
1 2 3 4 5 6 a b
3
Table 3 Peak fitting of the Si 2p photoelectron signals.
RBS
XPS
Si (%)
N (%)
34.0 34.0 35.6 39.8
66.0 61.0 59.7 60.2
43.7
45.9
Si (%)a
Sample
Bonding state
Concentration of each component (%)
Peak position (eV)
2
Si3N4 O\Si\N SiO2 Si2O Si3N4 O\Si\N SiO2 Si2O Si3N4 O\Si\N SiO2 Si2O
64.7 23.7 11.6 – 57.9 28.7 13.4 – 54.7 30.4 11.3 3.6
101.80 102.65 103.79 – 101.80 102.65 103.81 – 101.80 102.67 103.91 100.42
N (%)b
32.8
43.2
33.6 36.2
40.7 40.8
Narrow scan intensity of the N-related Si 2p peak. Narrow scan intensity of the Si-related N1s peak.
property. A study of SiNx film formation by the pyrolysis of a mixture of SiH4 and NH3 indicated that the optical band gap (E04) is approximately 5.3 eV [14]. The optical band gap of our study and Ref. [9] rose steeply, exceeding 5 eV at a lower NH3 or N2 flow ratio even though the process temperatures were 60–80 °C. Both methods used microwave-excited plasma and generated higher electron density than other RF PE-CVDs. The supposition is that the high-density plasma promotes the decomposition of source gases and forms Si\N bonding at low temperatures. This results in N-rich SiNx structures close to Si3N4 which have band gaps of over 5 eV.
Fig. 4. Elemental ratio (N/Si) as a function of NH3/(SiH4 + NH3) source gas ratio analyzed by XPS and RBS.
5
6
3.2. Structure of transparent SiNx films: comparison of RBS and XPS results The structures of the films were subsequently investigated by RBS and XPS. Evaluated samples are listed in Table 1. Samples 1–6 were prepared using a fixed SiH4 gas flow rate and varied NH3 flow rate. The film thickness was around 200 nm for all samples. RBS analysis was performed for Samples 1–4 and 6. XPS analysis was performed for Samples 2, 5, and 6. XPS analysis provides elemental and chemical information on the surface of the material. The depth for which information can be obtained is typically around 10 nm for the instrumental configuration used here. RBS analysis can obtain elemental information of a thickness of several hundreds of nanometers of the film from the surface. Therefore, if we assume that film growth is stable at a film thickness of 200 nm, then the bulk structure can be analogized by the surface structure with the exception of the surface oxidation. In other words, only Si and N were taken into account for the XPS analysis and compared with the RBS results. Table 2 shows the Si and N atom content ratio analyzed by RBS and XPS. RBS analysis provides information on the ratio of Si and N elements. For XPS analysis, Si content was calculated as a sum of N-related Si 2p peaks and N content was calculated as a sum of Si-related N1s peaks. The Si and N atomic ratio of the XPS analysis provides Si\N bonding information. These results were plotted as a function of the NH3/(SiH4 + NH3) flow ratio in Fig. 4. Fig. 4 shows that Si3NN4 was observed in the RBS analysis. XPS analysis data shows that SiNx approaches the stoichiometric structure in the region of higher NH3 ratios. For both RBS and XPS analyses, the N/Si ratio increases with the increase of NH3/(SiH4 + NH3) flow ratio. This means that the Si3NN4 the samples from the RBS analysis have an excess of N atoms that do not bond to Si in the SiNx films. A
Fig. 5. Peak fit analysis of Si2p X-ray photoelectron spectra of SiNx films prepared with 70 sccm of SiH4 and 650, 153, and 90 sccm of NH3 in the fed gas, corresponding to Samples (2), (5), and (6) in Table 1.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
4
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx
Fig. 8. Assumed mechanism of SiNx under exposure to rich NH3 atmosphere, which reduces Si-Si bond or Si dangling bond and increases Si\N bond.
3.3. Stability of SiNx films
Fig. 6. Concentration of each component (Si3N4, SiO2 + OSiN) calculated from deconvoluted XPS Si2p peaks as a function of N/Si elemental ratio.
previous RBS study indicated that the N/Si ratio increases with an increase of the NH3/SiH4 ratio, but that the N/Si ratio reached saturation at around N/Si = 1.7 [15]. We next looked at where the excess N bonds in the Si3Nx N 4 film. To clarify this, peak fit analysis was performed on the Si 2p peak. The result is shown in Fig. 5 and in Table 3. Samples 2 and 5 consist of three peaks each, i.e., Si3N4, O\Si\N, and SiO2. In addition to the three peaks from Samples 2 and 5, Sample 6 exhibited an additional Si2O (suboxide) peak for a total of four peaks. The assignment of the suboxide was referred to the peak position of the intermediate oxide observed in Si oxidation process [16]. Although a Si\Si bond could not be observed, the appearance of Si2O suggests the existence of Si\Si bonding, which lowers the visible transparency. It becomes evident from Fig. 5 that Si\O bonding (SiO2 and O\Si\N) is not negligible although oxygen was not intentionally added. This is because the low deposition temperature of 80 °C caused a large sticking coefficient condition, causing oxygen contamination from the reactor wall in the form of adsorbed O2. Excess nitrogen may combine in the form of O\Si\N in the SiNx film. Fig. 6 shows the XPS intensity of Si3N4, SiO2 + O\Si\N versus the N/Si atomic ratio. Si3N4 peak increased with an increase in the atomic ratio of N/Si. At the same time, the sum of SiO2 and O\Si\N peaks decreased with an increase in the N/Si ratio. When the N/Si ratio increased and came close to Si3N4, the oxygen in the film decreased, indicating that the number of Si\O bonds decreased with the increase in Si\N bonding.
The film's susceptibility to oxidation depends mostly on the N/Si atomic ratio. Samples 2 and 6 from Table 1 were analyzed with FT-IR to evaluate stability against oxidation. Fig. 7 shows the results. A Si\N stretch peak appears at 830–890 cm-1, a Si\O stretch peak appears at 1050 cm-1, and a N\H bend peak appears at 1175 cm-1 [17]. Sample 2 did not exhibit a Si\O stretching mode peak increase after 3 days under 60 °C 90% RH atmosphere. In contrast, the Si\O peak increased after only 3 days at 25 °C 60% RH in Sample 6. Sample 6 was Si-rich SiNx and there probably were many sites that could react with oxygen. XPS analysis indicates the presence of suboxide in Sample 6, so a Si\Si bond or a Si dangling bond may be included in the structure. An assumed mechanism is illustrated in Fig. 8 for SiNx in an NH3-rich atmosphere, which reduces Si\Si bonds or Si dangling bonds and increases Si\N bonds. 4. Conclusions We have described the structure of SiNx films using RBS, XPS, and FTIR. A transparent SiNx film with an optical band gap of over 5 eV could be obtained with high NH3/(SiH4 + NH3) source gas ratio. The structure of this transparent SiNx film was determined as Si3N5.82 by RBS, and Si3N3.95 by XPS, which indicates that excess nitrogen exists without bonding to Si atom at around the stoichiometric structure. This excess nitrogen may bind to contaminated oxygen inside the SiNx film. According to Si 2p peak deconvolution analysis, suboxide (Si2O) existed in the Si-rich film. Abrupt oxidation after film deposition was also observed by FTIR for the Si-rich film. These observations indicate that Si-rich SiNx film retains much Si\Si bonding which leads to optical absorption and instability of the film. Acknowledgments This study was partially supported by New Energy and Industrial Technology Development Organization (NEDO) (P08011).
Fig. 7. FTIR absorption spectra of Sample (2) Si3N3.95 and Sample (6) Si3N3.39.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx
References [1] S.R. Forrest, The path to ubiquitous and low-cost organic electronic appliances on plastic, Nature 428 (2004) 911. [2] K.K. Lien, S.J. Chua, S.F. Lim, Influence of electrical stress voltage on cathode degradation of organic light-emitting devices, J. Appl. Phys. 90 (2001) 976. [3] P.E. Burrows, V. Bulovic, S.R. Forrest, L.S. Sapochak, D.M. McCarty, M.E. Thompson, Reliability and degradation of organic light emitting devices, Appl. Phys. Lett. 65 (1994) 2922. [4] K. Yamashita, T. Mori, T. Mitzutani, Encapsulation of organic light-emitting diode using thermal chemical-vapour-deposition polymer film, J. Phys. D. Appl. Phys. 34 (2001) 740. [5] P.E. Burrows, G.L. Graff, M.E. Gross, P.M. Martin, M.K. Shi, M. Hall, E. Mast, C. Bonham, W. Bennet, M.B. Sullivan, Ultra barrier flexible substrates for flat panel displays, Displays 22 (2001) 65. [6] I. Ay, H. Tolunay, Optical transmission measurements on glow-discharge amorphous silicon nitride films, Turk. J. Phys. 25 (2001) 215. [7] M. Lipinski, Silicon nitride for photovoltaic application, Arch. Mater. Sci. Eng. 46 (2010) 69. [8] F. Demichelis, G. Crovini, F. Giorgis, C.F. Pirri, E. Tresso, Hydrogenated amorphous silicon-nitrogen alloys, a-SiNx:H-y: a wide band gap material for optoelectronic devices, J. Appl. Phys. 79 (1996) 1730.
5
[9] Y. Manabe, T. Mitsuyu, Silicon nitride thin films prepared by the electron cyclotron resonance plasma chemical vapor deposition method, J. Appl. Phys. 66 (1989) 2475. [10] K. Azuma, S. Ueno, M. Suzuki, Y. Konishi, S. Ishida, Novel surface-wave-excited plasma-enhanced CVD system with reciprocating substrate motion, ECS Trans. 28 (15) (2010) 27. [11] A. Madan, M.P. Shaw, The Physics and Applications of Amorphous Semiconductors, Academic Press, 1988, p. 152. [12] F. Giorgis, P. Mandracci, L.D. Negro, C. Mazzoleni, L. Pavesi, Optical absorption and luminescence properties of wide-band gap amorphous silicon based alloys, J. Non-Cryst. Solids 266–269 (2000) 588. [13] J. Tauc, Amorphous and Liquid Semiconductor, Plenum, New York, 1974. p. 159. [14] H.R. Philipp, Optical properties of silicon nitride [Formulation of bonding model], J. Electrochem. Soc. 120 (2) (1973) 295. [15] S.A. Almeida, S.R.P. Silva, Stoichiometric limitations of RF plasma deposited amorphous silicon–nitrogen alloys, Thin Solid Films 311 (1997) 133. [16] M.P. Seah, Ultrathin SiO2 on Si. VI. Evaluation of uncertainties in thickness measurement using XPS, Surf. Interface Anal. 37 (2005) 300. [17] D.V. Tsu, G. Lucovsky, M.J. Mantini, Local atomic structure in thin films of silicon nitride and silicon diimide produced by remote plasma-enhanced chemical-vapor deposition, Phys. Rev. B 33 (1986) 70697076.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
Contents lists available at ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions Kazufumi Azuma a, Satoko Ueno a, Yoshiyuki Konishi b, Kazuhiro Takahashi c a b c
Technology Research Laboratory, Shimadzu Corp., Kanagawa 259-1304, Japan Semiconductor Equipment Division, Shimadzu Corp., Kanagawa 259-1304, Japan Analytical & Measuring Instruments Division, Shimadzu Corp., Kanagawa 259-1304, Japan
a r t i c l e
i n f o
Article history: Received 5 June 2014 Received in revised form 24 September 2014 Accepted 26 September 2014 Available online xxxx Keyword: Surface wave plasma Chemical vapor deposition Silicon nitride Low temperature XPS RBS FTIR
a b s t r a c t Highly transparent silicon nitride films with a low absorption coefficient of only 200 cm-1 or lower were prepared under high NH3/SiH4 source gas ratio conditions at 80 °C or lower temperature using surface wave plasma chemical vapor deposition (SWP-CVD). Rutherford backscattering measurements indicated that a silicon nitride structure Si3Nx (x N 5) with excess nitrogen could be prepared with the SWP-CVD method under high NH3/SiH4 source gas ratio conditions. X-ray photon spectroscopy (XPS) analysis provided verification that the excess nitrogen combined with the oxygen contained in the SiNx film during low-temperature film formation and that the atomic ratio of Si and N was almost stoichiometric, i.e., Si3N4. XPS study also revealed that the Si3Nb4 structure contained suboxides, which presence may reduce the transparency of the films. In contrast, suboxides were not observed in the Si3N4 structure obtained under high NH3/SiH4 source gas ratio conditions. Fourier-transform infrared spectroscopy study confirmed that the SiNx film becomes more stable when the SiNx structure approaches the stoichiometric ratio. Achieving a near-transparent Si3N4 structure requires a sufficient amount of N atoms in the periphery of the Si atom, i.e., an adequate amount of NH3 is necessary in the presence of the SiH4. © 2014 Elsevier B.V. All rights reserved.
1. Introduction High-performance transparent barrier films have recently attracted much attention for flexible applications such as organic light emitting diode (OLED) displays [1]. Organic materials are susceptible to water, oxygen, and other environmental elements present in ambient conditions [2–4]. Many studies report on encapsulation methods or materials for increasing OLED display lifetimes. Transparency is particularly critical for encapsulating films that are applied to top-emitting OLED displays. The water vapor transmission rate requirement for OLED encapsulation is less than 10-6 g/m2/day [5]. Silicon nitride (SiNx) films are the strongest candidates for a high-performance barrier film. There are many reports that discuss the transparency of SiNx films [6–9]. These reports examine the influence of the Si/N ratio on optical characteristics, but few ones discuss the bonding states of transparent SiNx films. For this study, we prepared SiNx films under low temperature (less than 80 °C) conditions using a surface wave plasma chemical vapor deposition (SWP-CVD) system and elucidated the relationship between the Si/N elemental ratio and the Si\N bonding states using
E-mail address: [email protected] (K. Azuma).
Rutherford backscattering spectroscopy (RBS) and X-ray photoelectron spectroscopy (XPS). 2. Experimental methods Fig. 1 shows a schematic diagram of the SWP-CVD system [10]. This system has a plasma source with a slot antenna. Microwaves with a frequency of 2.45 GHz were introduced through the slot antenna and alumina dielectric window into the reactor. Ar, SiH4 and NH3 gases were fed from the upper and lower gas inlets. The distance between the dielectric window and the substrate stage was fixed at 200 mm. The films were prepared by exposing 6-inch- diameter FZ Si (100) wafers and quartz substrates to Ar, SiH4 and NH3 gases that were introduced from the gas inlets at flow rates of 350, 70, and 90–500 sccm, respectively. The total gas pressure was maintained at 10 Pa. The microwave power density was fixed at 1.57 W/cm2. The deposition time was fixed at 100 s for each sample. The deposition rate was almost the same when the SiH4 gas flow rate was constant (2.0 nm/s for the case of SiH4 70 sccm). The substrate and the gas lines were not positively heated. The surface temperature of the substrate increased from ambient temperature and saturated at 80 °C during 100 s of deposition time. The optical absorption coefficients of the SiNx films were determined by transmittance and reflectance (T–R) measurement in the 300– 800 nm range with a Shimadzu UV2600 spectrophotometer.
http://dx.doi.org/10.1016/j.tsf.2014.09.068 0040-6090/© 2014 Elsevier B.V. All rights reserved.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
2
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx
Fig. 1. Experimental apparatus of surface wave plasma chemical vapor deposition system.
Film structures were evaluated with a Shimadzu FTIR-8400 Fourier transform infrared (FTIR) spectrometer and by XPS. XPS analysis was performed using a Shimadzu/Kratos AXIS-Nova spectrometer with a Vision 2 data system. The excitation source was a monochromated Al-Kα X-ray with 225 W-power, the size of the analysis area was 700 × 300 μm, and the emission angle of the photoelectron with respect to the sample normal was zero degrees. No ion sputtering was performed prior to the XPS analysis. For XPS data processing of Si 2p peaks, a linear background was subtracted before peak synthesis, which was performed with component peaks having a Gaussian (70%)–Lorentzian (30%) product peak shape. Without any constraints on peak position, peak height and peak width between component peaks were set during peak synthesis process. The position of the peaks was calibrated so that the highest Si 2p component appeared at 101.80 eV. The films were measured by XPS within 30 to 60 min after removal of the samples to ambient air from the vacuum tool after film formation. The elemental ratio of the SiNx films was also analyzed by RBS (NEC 3S-R10 and CEA RBS-400). RBS analysis can provide elemental bulk information of a film for the first several hundreds of nanometers of the surface. Measurements were performed with an incident energy of 2.275 MeV 4He++, a beam diameter of 1–2 mm, and detection angles of 160° (normal angle) and 119° (grazing angle).
Fig. 3. Optical band gap (Eg opt) as a function of source gas flow ratio. The conditions for each reference are as follows: [Ref. 6] RF PE-CVD, SiH4 + NH3 at 300 °C. [Ref. 7] RF PE-CVD, SiH4 + N2 at 300 °C. [Ref. 8] RF PE-CVD, SiH4 + NH3 at 200 °C. [Ref. 9] ECR PE-CVD, SiH4 + N2 at 60 °C.
3. Results and discussion 3.1. Optical properties of SiNx films First, the optical absorption coefficient of SiNx films was calculated from the Beer–Lambert law for the films prepared with various SiH4/ NH3 gas ratios. Fig. 2 shows the relationship between the absorption coefficient of the films at the wavelength of 400 nm and a supplied gas ratio of NH3 and SiH4. There is a noteworthy drop in the absorption coefficient down to a few hundred cm-1 or less at a NH3/(SiH4 + NH3) ratio of 0.7 and higher. An optical band gap energy E04 was defined to be the spectra wavelength (energy) for which the absorption coefficient α = 104 cm-1 in wavelength-dispersive optical absorption [11,12]. The optical band gap was plotted as a function of the NH3/(SiH4 + NH3) ratio in Fig. 3. Films with an optical absorption of lower than 104 cm-1 in the measured range were not plotted. Data from previous reports [6–9] were also plotted in the figure. Refs. [6,8] used RF plasma enhanced chemical vapor deposition (PE-CVD) with SiH4 and NH3 as source gases. The deposition temperature was 200–300 °C. Ref. [7] used RF PE-CVD with SiH4 and N2. The deposition temperature was 200 °C. Ref. [9] used electron cyclotron resonance (ECR) PE-CVD with SiH4 and N2 and with a deposition temperature of 60 °C. Refs. [6,8] determined the optical band gap using E04 and Refs. [7,9] determined the same with Tauc's relation [13]. According to Fig. 3, the optical band gap increases with an increase in the NH3 or N2 ratio, indicating that adequate NH3 or N2 is needed for transparent SiNx films. Since the PE-CVD process is a predominantly surface reaction, the usual SiNx film preparation involved heating at 200 °C. Since heating promotes Si\N bond formation, the SiNx structure approaches the stoichiometric ratio, which leads to a lower absorption
Table 1 Film preparation conditions (flow rates of source gases).
Fig. 2. Relationship between absorption coefficient of the SiN x films and NH 3 / (SiH 4 + NH 3) source gas ratio.
Sample
SiH4 (sccm)
NH3 (sccm)
1 2 3 4 5 6
70 70 70 70 70 70
650 500 350 175 153 90
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx Table 2 Result of RBS and XPS analysis. Sample
1 2 3 4 5 6 a b
3
Table 3 Peak fitting of the Si 2p photoelectron signals.
RBS
XPS
Si (%)
N (%)
34.0 34.0 35.6 39.8
66.0 61.0 59.7 60.2
43.7
45.9
Si (%)a
Sample
Bonding state
Concentration of each component (%)
Peak position (eV)
2
Si3N4 O\Si\N SiO2 Si2O Si3N4 O\Si\N SiO2 Si2O Si3N4 O\Si\N SiO2 Si2O
64.7 23.7 11.6 – 57.9 28.7 13.4 – 54.7 30.4 11.3 3.6
101.80 102.65 103.79 – 101.80 102.65 103.81 – 101.80 102.67 103.91 100.42
N (%)b
32.8
43.2
33.6 36.2
40.7 40.8
Narrow scan intensity of the N-related Si 2p peak. Narrow scan intensity of the Si-related N1s peak.
property. A study of SiNx film formation by the pyrolysis of a mixture of SiH4 and NH3 indicated that the optical band gap (E04) is approximately 5.3 eV [14]. The optical band gap of our study and Ref. [9] rose steeply, exceeding 5 eV at a lower NH3 or N2 flow ratio even though the process temperatures were 60–80 °C. Both methods used microwave-excited plasma and generated higher electron density than other RF PE-CVDs. The supposition is that the high-density plasma promotes the decomposition of source gases and forms Si\N bonding at low temperatures. This results in N-rich SiNx structures close to Si3N4 which have band gaps of over 5 eV.
Fig. 4. Elemental ratio (N/Si) as a function of NH3/(SiH4 + NH3) source gas ratio analyzed by XPS and RBS.
5
6
3.2. Structure of transparent SiNx films: comparison of RBS and XPS results The structures of the films were subsequently investigated by RBS and XPS. Evaluated samples are listed in Table 1. Samples 1–6 were prepared using a fixed SiH4 gas flow rate and varied NH3 flow rate. The film thickness was around 200 nm for all samples. RBS analysis was performed for Samples 1–4 and 6. XPS analysis was performed for Samples 2, 5, and 6. XPS analysis provides elemental and chemical information on the surface of the material. The depth for which information can be obtained is typically around 10 nm for the instrumental configuration used here. RBS analysis can obtain elemental information of a thickness of several hundreds of nanometers of the film from the surface. Therefore, if we assume that film growth is stable at a film thickness of 200 nm, then the bulk structure can be analogized by the surface structure with the exception of the surface oxidation. In other words, only Si and N were taken into account for the XPS analysis and compared with the RBS results. Table 2 shows the Si and N atom content ratio analyzed by RBS and XPS. RBS analysis provides information on the ratio of Si and N elements. For XPS analysis, Si content was calculated as a sum of N-related Si 2p peaks and N content was calculated as a sum of Si-related N1s peaks. The Si and N atomic ratio of the XPS analysis provides Si\N bonding information. These results were plotted as a function of the NH3/(SiH4 + NH3) flow ratio in Fig. 4. Fig. 4 shows that Si3NN4 was observed in the RBS analysis. XPS analysis data shows that SiNx approaches the stoichiometric structure in the region of higher NH3 ratios. For both RBS and XPS analyses, the N/Si ratio increases with the increase of NH3/(SiH4 + NH3) flow ratio. This means that the Si3NN4 the samples from the RBS analysis have an excess of N atoms that do not bond to Si in the SiNx films. A
Fig. 5. Peak fit analysis of Si2p X-ray photoelectron spectra of SiNx films prepared with 70 sccm of SiH4 and 650, 153, and 90 sccm of NH3 in the fed gas, corresponding to Samples (2), (5), and (6) in Table 1.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
4
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx
Fig. 8. Assumed mechanism of SiNx under exposure to rich NH3 atmosphere, which reduces Si-Si bond or Si dangling bond and increases Si\N bond.
3.3. Stability of SiNx films
Fig. 6. Concentration of each component (Si3N4, SiO2 + OSiN) calculated from deconvoluted XPS Si2p peaks as a function of N/Si elemental ratio.
previous RBS study indicated that the N/Si ratio increases with an increase of the NH3/SiH4 ratio, but that the N/Si ratio reached saturation at around N/Si = 1.7 [15]. We next looked at where the excess N bonds in the Si3Nx N 4 film. To clarify this, peak fit analysis was performed on the Si 2p peak. The result is shown in Fig. 5 and in Table 3. Samples 2 and 5 consist of three peaks each, i.e., Si3N4, O\Si\N, and SiO2. In addition to the three peaks from Samples 2 and 5, Sample 6 exhibited an additional Si2O (suboxide) peak for a total of four peaks. The assignment of the suboxide was referred to the peak position of the intermediate oxide observed in Si oxidation process [16]. Although a Si\Si bond could not be observed, the appearance of Si2O suggests the existence of Si\Si bonding, which lowers the visible transparency. It becomes evident from Fig. 5 that Si\O bonding (SiO2 and O\Si\N) is not negligible although oxygen was not intentionally added. This is because the low deposition temperature of 80 °C caused a large sticking coefficient condition, causing oxygen contamination from the reactor wall in the form of adsorbed O2. Excess nitrogen may combine in the form of O\Si\N in the SiNx film. Fig. 6 shows the XPS intensity of Si3N4, SiO2 + O\Si\N versus the N/Si atomic ratio. Si3N4 peak increased with an increase in the atomic ratio of N/Si. At the same time, the sum of SiO2 and O\Si\N peaks decreased with an increase in the N/Si ratio. When the N/Si ratio increased and came close to Si3N4, the oxygen in the film decreased, indicating that the number of Si\O bonds decreased with the increase in Si\N bonding.
The film's susceptibility to oxidation depends mostly on the N/Si atomic ratio. Samples 2 and 6 from Table 1 were analyzed with FT-IR to evaluate stability against oxidation. Fig. 7 shows the results. A Si\N stretch peak appears at 830–890 cm-1, a Si\O stretch peak appears at 1050 cm-1, and a N\H bend peak appears at 1175 cm-1 [17]. Sample 2 did not exhibit a Si\O stretching mode peak increase after 3 days under 60 °C 90% RH atmosphere. In contrast, the Si\O peak increased after only 3 days at 25 °C 60% RH in Sample 6. Sample 6 was Si-rich SiNx and there probably were many sites that could react with oxygen. XPS analysis indicates the presence of suboxide in Sample 6, so a Si\Si bond or a Si dangling bond may be included in the structure. An assumed mechanism is illustrated in Fig. 8 for SiNx in an NH3-rich atmosphere, which reduces Si\Si bonds or Si dangling bonds and increases Si\N bonds. 4. Conclusions We have described the structure of SiNx films using RBS, XPS, and FTIR. A transparent SiNx film with an optical band gap of over 5 eV could be obtained with high NH3/(SiH4 + NH3) source gas ratio. The structure of this transparent SiNx film was determined as Si3N5.82 by RBS, and Si3N3.95 by XPS, which indicates that excess nitrogen exists without bonding to Si atom at around the stoichiometric structure. This excess nitrogen may bind to contaminated oxygen inside the SiNx film. According to Si 2p peak deconvolution analysis, suboxide (Si2O) existed in the Si-rich film. Abrupt oxidation after film deposition was also observed by FTIR for the Si-rich film. These observations indicate that Si-rich SiNx film retains much Si\Si bonding which leads to optical absorption and instability of the film. Acknowledgments This study was partially supported by New Energy and Industrial Technology Development Organization (NEDO) (P08011).
Fig. 7. FTIR absorption spectra of Sample (2) Si3N3.95 and Sample (6) Si3N3.39.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068
K. Azuma et al. / Thin Solid Films xxx (2014) xxx–xxx
References [1] S.R. Forrest, The path to ubiquitous and low-cost organic electronic appliances on plastic, Nature 428 (2004) 911. [2] K.K. Lien, S.J. Chua, S.F. Lim, Influence of electrical stress voltage on cathode degradation of organic light-emitting devices, J. Appl. Phys. 90 (2001) 976. [3] P.E. Burrows, V. Bulovic, S.R. Forrest, L.S. Sapochak, D.M. McCarty, M.E. Thompson, Reliability and degradation of organic light emitting devices, Appl. Phys. Lett. 65 (1994) 2922. [4] K. Yamashita, T. Mori, T. Mitzutani, Encapsulation of organic light-emitting diode using thermal chemical-vapour-deposition polymer film, J. Phys. D. Appl. Phys. 34 (2001) 740. [5] P.E. Burrows, G.L. Graff, M.E. Gross, P.M. Martin, M.K. Shi, M. Hall, E. Mast, C. Bonham, W. Bennet, M.B. Sullivan, Ultra barrier flexible substrates for flat panel displays, Displays 22 (2001) 65. [6] I. Ay, H. Tolunay, Optical transmission measurements on glow-discharge amorphous silicon nitride films, Turk. J. Phys. 25 (2001) 215. [7] M. Lipinski, Silicon nitride for photovoltaic application, Arch. Mater. Sci. Eng. 46 (2010) 69. [8] F. Demichelis, G. Crovini, F. Giorgis, C.F. Pirri, E. Tresso, Hydrogenated amorphous silicon-nitrogen alloys, a-SiNx:H-y: a wide band gap material for optoelectronic devices, J. Appl. Phys. 79 (1996) 1730.
5
[9] Y. Manabe, T. Mitsuyu, Silicon nitride thin films prepared by the electron cyclotron resonance plasma chemical vapor deposition method, J. Appl. Phys. 66 (1989) 2475. [10] K. Azuma, S. Ueno, M. Suzuki, Y. Konishi, S. Ishida, Novel surface-wave-excited plasma-enhanced CVD system with reciprocating substrate motion, ECS Trans. 28 (15) (2010) 27. [11] A. Madan, M.P. Shaw, The Physics and Applications of Amorphous Semiconductors, Academic Press, 1988, p. 152. [12] F. Giorgis, P. Mandracci, L.D. Negro, C. Mazzoleni, L. Pavesi, Optical absorption and luminescence properties of wide-band gap amorphous silicon based alloys, J. Non-Cryst. Solids 266–269 (2000) 588. [13] J. Tauc, Amorphous and Liquid Semiconductor, Plenum, New York, 1974. p. 159. [14] H.R. Philipp, Optical properties of silicon nitride [Formulation of bonding model], J. Electrochem. Soc. 120 (2) (1973) 295. [15] S.A. Almeida, S.R.P. Silva, Stoichiometric limitations of RF plasma deposited amorphous silicon–nitrogen alloys, Thin Solid Films 311 (1997) 133. [16] M.P. Seah, Ultrathin SiO2 on Si. VI. Evaluation of uncertainties in thickness measurement using XPS, Surf. Interface Anal. 37 (2005) 300. [17] D.V. Tsu, G. Lucovsky, M.J. Mantini, Local atomic structure in thin films of silicon nitride and silicon diimide produced by remote plasma-enhanced chemical-vapor deposition, Phys. Rev. B 33 (1986) 70697076.
Please cite this article as: K. Azuma, et al., Transparent silicon nitride films prepared by surface wave plasma chemical vapor deposition under low temperature conditions, Thin Solid Films (2014), http://dx.doi.org/10.1016/j.tsf.2014.09.068