- Email: [email protected]
Preparation of single-crystal TiC (111) by radio frequency magnetron sputtering at low temperature
Thin Solid Films 520 (2012) 6882–6887
Contents lists available at SciVerse ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Preparation of single-crystal TiC (111) by radio frequency magnetron sputtering at low temperature Q. Qi, W.Z. Zhang, L.Q. Shi ⁎, W.Y. Zhang, W. Zhang, B. Zhang Applied Ion Beam Physics Laboratory, Institute of Modern Physics, Fudan University, Shanghai 200433, People's Republic of China
a r t i c l e
i n f o
Article history: Received 26 February 2012 Received in revised form 10 July 2012 Accepted 13 July 2012 Available online 20 July 2012 Keywords: Titanium carbide Radio-frequency magnetron sputtering Rutherford backscattering spectroscopy X-ray diffraction
a b s t r a c t Single-crystal films of TiC (111) have been synthesized at room temperature on Al2O3 (0001) substrates by radio frequency magnetron sputtering using a compound Ti–C target. The substrate temperature and bias were varied to explore the influence of deposition parameters on the crystal structure. Both Al2O3 (0001) and Si (100) substrates were used for epitaxial growth of TiC films. A series of characterizations of TiC films were carried out, including Rutherford backscattering spectroscopy, X-ray diffraction, Raman and X-ray photoelectron spectroscopy. Single-crystal films of TiC (111) on the Al2O3 (0001) were demonstrated. © 2012 Elsevier B.V. All rights reserved.
2. Experimental details
Al2O3 wafers. The wafers were cleaned several times in an ultrasonic bath containing a solution of ethanol and then were dried and loaded into the sputtering chamber immediately to avoid contamination. The compound target was fabricated by a Ti disk (purity 99.99% and 75 mm in diameter) and sector carbon sheets with the Ti/C area ratio of 2.429. The target was mounted on the RF electrode and located 50 mm from the substrate. The substrate was mounted in a stainless steel holder which was biased negatively to ground. The base pressure before deposition was less than 3 × 10 −5 Pa, using a turbo-molecular pump. In order to decrease doping of H from H2O adsorbing on the wall, the chamber was baked at about 200 °C for 3 h. The working gas used for discharge was argon (99.999% purity). In order to investigate the effects of the substrate and deposition temperature on the TiC films, both Al2O3 (0001) and Si (100) substrates were used and a series of substrate temperatures from room temperature to 700 °C were performed. The total pressure of the sputtered gas was controlled at 0.1 Pa in order to obtain stable plasma. In this experiment, the flux of Ar was 1.8 sccm (standard cubic centimeter per minute). The applied bias during the deposition was fixed at −60 V except for when the effects of bias voltage were studied. The RF power for all depositions was 100 W. The film thickness and deposition rate were about 200 nm and 3 nm/min respectively.
2.1. Specimen preparation
2.2. Ion beam analysis
The TiC films were prepared by a radio frequency (RF) magnetron sputtering method. The films were deposited on 10 cm by 15 cm
The composition and thickness of the TiC films were measured by using non-Rutherford scattering (non-RBS) of 3.65 MeV He+ ion beam. At this energy the scattering cross section of C is significantly enhanced over the standard Rutherford cross section. The incident angle of the beam was 0°, while He+ ions backscattered by target were
1. Introduction TiC possesses many special properties, such as high melting point, high thermal stability, exceptional hardness, outstanding wear resistance and good corrosion resistance, which make it of interest in many fields including wear-resistant coatings on cutting and grinding tools, thermal barriers and corrosion resistance on metallic structure. TiC thin films have been synthesized by various methods such as chemical vapor deposition [1,2], pulsed laser deposition [3], arc ion plating [4], and sputtering [5–8]. The disadvantage of these techniques, however, is that the microstructure of the film is highly dependent on the deposition temperature. Synthesis of single-crystal TiC films requires high temperature. Lower deposition temperature usually leads to amorphous or polycrystalline films. Recently it has been demonstrated that epitaxial carbide films could be deposited by co-evaporation of C60 and Ti at very low temperature [9,10]. Up until now no report on synthesis of single-crystal TiC films at room temperature has been made. This paper will introduce a method of growing single-crystal TiC films at room temperature and explore the effects of various parameters on crystal microstructure.
⁎ Corresponding author. Tel.: +86 21 65642292. E-mail address: [email protected] (L.Q. Shi). 0040-6090/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2012.07.040
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
6883
detected at a scattering angle of 165°. All the non-RBS spectra were simulated by the SIMNRA 6.0 codes. Non-RBS measurements were performed at the NEC 9SDH-2 tandem accelerator at the Institute of Modern Physics of Fudan University, China.
can be determined. In this study the Raman spectra were measured over the range from 200 to 1800 cm−1 using a Jobin Yvon HR800 confocal Raman with the laser wavelength of 632.8 nm. The spectra were record with acquisition time of 100 s.
2.3. X-ray diffraction (XRD) analysis
2.5. X-ray photoelectron spectroscopy (XPS) analysis
The crystal structure was investigated by XRD. The measurement was performed at a D/MAX 2550 V X-ray diffractometer made by Rigaknu, Japan. This diffractometer with Cu Kα source was used in a 2θ mode, 2θ varying from 20° to 68° with a 0.02° step.
XPS was recorded using a Kratos AXIS ULTRA DLD instrument with the source of an Al Kα monochromatic X-ray. The gun was operated at 15 kV and 10 mA. Survey spectra were acquired from 0 to 1200 eV in steps of 1 eV for the pass energy of 160 eV. Region scans from selected ranges were obtained using the pass energy of 20 eV. The peaks were acquired from the selected elements Ti 2p, C 1s and O 1s. The base pressure during measurement is lower than 10−7 Pa. In order to avoid the effects of oxides, the samples were etched by Ar+ for 3 min using the source of 2 kV and 30 mA over an area of 3 × 3 mm2. The spectra were record for both unetched and etched samples.
2.4. Raman spectroscopy analysis Raman spectroscopy is based on Raman effect, in which the incident photons are scattered inelastically by molecules or crystal lattices and collected. From the Raman spectra the details of the crystal structure
Fig. 1. RBS spectra of TiC films deposited on (a) Al2O3 substrate and (b) Si substrate with the deposition power of 100 W and bias voltage of −60 V.
6884
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
3. Results and discussion To allow the deposited atoms to diffuse across the substrate the deposition rate of the films was limited to a maximum rate of approximately 3 nm/min through the RF power. Typical non-RBS spectra of the TiC films deposited on the Al2O3 and Si wafers by RF magnetron sputtering are shown in Fig. 1(a) and (b). The composition of the films can be acquired from the best fit simulation to the spectra. The Ti to C ratio for all deposited films on the two types of substrates is between 0.8 and 1.2 due to different deposition temperatures. The substrate temperature is a key parameter in the synthesis of single-crystal TiC films, as it is the controlling feature of atomic mobility. At all substrate temperatures used in this study there is only one peak in the XRD patterns at 35.9° (apart from the substrate Al2O3 peak at 41.71°) which corresponds to TiC (111) (Fig. 2). This indicates that single crystal TiC (111) grows epitaxially on the Al2O3 (0001) substrates by RF magnetron sputtering from room temperature to 700 °C. The spectra also suggest that at high substrate temperature the peak intensity is enhanced and the full width at half maximum (FWHM) of the TiC peak becomes narrower with increasing temperature except for room temperature (Fig. 3). The high crystallinity of the film deposited at room temperature may be attributed to the film composition approaching to stoichiometric TiC so that the unreacted carbon or titanium in the film is less. However further investigation of this effect is under way. In general the crystal structure of TiC films tends be more ordered at higher temperature. The XRD patterns of TiC films grown on Al2O3 (0001) substrate with different bias voltages at the same deposition temperature of 700 °C are shown in Fig. 4. At the bias voltage of −25 V the TiC (111) peak almost disappears which implies that at this low bias voltage the deposited atoms and the impinging Ar+ ions do not have sufficient energy to promote atomic diffusion of surface atoms hence the degree of crystallinity is poor and the crystal microstructure tends to be amorphous. At a moderate bias voltage of −60 V there is only one sharp peak of TiC which demonstrates that a suitable bias in synthesizing single-crystal TiC films on Al2O3 (0001) substrate was around −60 V. At an applied higher bias voltage of −100 V, however, the intensity of the peak is weaker and the FWHM is broader than that at −60 V. The reason may be that the more energetic ions penetrate the film to create defects and produce compressive stress which causes the distortion of the lattice and a damaged film.
Fig. 3. Full width at half maximum (FWHM) and intensity from XRD peaks of TiC (111) deposited on Al2O3 at different temperatures.
In order to further explore the role of the substrate on epitaxial growth by RF sputtering, the structure formed at different deposition temperatures for TiC films on Si substrates is identified by XRD (Fig. 5). At a high temperature of 700 °C a relative strong TiC (111) peak and a very weak TiC (220) peak as well can be observed in diffraction pattern, suggesting relatively high preferred (111) orientation and better order. As the temperature was reduced to 500 °C, only a relative weak TiC (200) peak appears in the pattern and the crystal exhibits a fully (200) orientation and a poorer degree of crystallinity. At a deposition temperature of 200 °C, a very weak diffraction from polycrystalline TiC is observed, indicating very poor order. The major factors for epitaxial growth are lattice mismatch, temperature and deposition rate. The lattice mismatch could be characterized by the parameter: Δ¼
jas −ae j ae
where Δ is the lattice mismatch value, and as and ae are the lattice parameters of the substrate and epitaxial film respectively. For TiC film deposited on Al2O3 or Si substrates, ae = 4.327 Å, as = 4.758 Å for Al2O3
Fig. 2. XRD patterns of TiC films deposited on Al2O3 with different temperatures.
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
Fig. 4. XRD patterns of TiC films deposited on Al2O3 in different bias. The deposition power is 100 W.
6885
Fig. 6. Raman spectra of TiC films deposited on the Al2O3 with different temperatures. The deposition power is 100 W and the bias voltage is −60 V.
and as =5.429 Å for Si which results in the mismatch parameters of ΔAl2O3 = 0.0996, ΔSi = 0.25. The large difference in the mismatch parameters is the primary reason that the crystal quality of epitaxial films deposited on Al2O3 is better than that on Si given all other parameters equal. In order to determine the crystal structure quality of TiC, Raman spectroscopy was performed in the range between 200 and 1800 cm−1 [11–13]. In this range high purity titanium sheet does not have Raman active vibrational modes so no peaks can be observed for the pure metal. For a high purity carbon wafer two strong peaks at 1320 cm−1 and 1598 cm−1 can be detected, which are associated with the A1g and E2g vibrational modes of graphite. Stoichiometric TiC does not have Raman active vibrational modes, however, peaks arising from carbon vacancies can be observed for TiCx (where x b 1) as the vacancies can destroy the inversion symmetry of nearby atoms and induce first-order Raman scattering [14,15]. The peaks at 264, 349, 600, and 658 cm−1 (Fig. 6) correspond to nonstoichiometric TiC. The width and intensity of these peaks decrease with increasing temperature which indicates that the quality of the crystal has improved. At a substrate temperature of 700 °C, the spectrum has a broaden peak from 1300 to 1600 cm−1 which is caused by superposed spectra of the A1g and E2g vibrational modes of graphite however these two peaks do not appear in the spectra at other temperatures. The reason is that at
higher temperature excess carbon atoms aggregate to form graphite which has Raman active vibrational modes that produce the peaks. This case was further demonstrated in Fig. 7, which presents the spectra of the same samples before and after annealing. The higher intensity and smaller FWHM of the A1g and E2g vibrational modes at 1320 and 1598 cm−1 peaks are indicative of the formation of graphite in the samples. For the sample TiC deposited at 700 °C, X-ray photoelectron spectra (XPS) for C 1 s and Ti 2p before sputter etching with Ar + were shown in Fig. 8(a) and (b) respectively and after etching were presented in Fig. 9(a) and (b). For the region scans for C 1s, the peak at binding energy of 281.6 eV shown in Fig. 8(a) was attributed to the Ti\C bond and the peak at 284.6 eV was to the C\C bond [16,17]. There are also two other peaks at higher binding energies of 286 eV and 288.1 eV which are responding to C\O\R bond and C_O bond respectively. However, after etching as shown in Fig. 9(a), there are only two peaks: one is a sharp peak at 281.8 eV corresponding to Ti\C bond, the other is a small peak associated to C\C bond. The peaks corresponding to oxide and contamination bonds disappear. Figs. 8(b) and 9(b) show the region scans of Ti 2p. The Ti 2p XPS spectra present one spin orbit doublet Ti 2p3/2 and 2p1/2. The peaks of binding energy at 454.76 eV and 458.3 eV represent Ti 2p3/2 carbide and 2p3/2 oxide TiO2 and the peaks at 460.8 eV and 463.9 eV are associated to Ti 2p1/2 carbide and Ti1/2
Fig. 5. XRD patterns of TiC films grown on Si substrate with different temperatures. The deposition power is 100 W and the bias voltage is −60 V.
Fig. 7. Raman spectra of annealed and unannealed TiC films.
6886
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
Fig. 8. (a) C 1s and (b) Ti 2p XPS spectra of the TiC films deposited on Al2O3 before etching with the deposition power of 100 W and bias voltage of −60 V.
oxide TiO2. After sputter etching (Fig. 9(b)) Ti 2p oxides presented in Fig. 8(b) have almost disappeared and only Ti 2p3/2 carbide and Ti 2p1/2 carbide exist. From the spectra before and after sputter etching it is suggested that there is a contamination on the sample surface assigned to C_O bond in the as‐prepared sample but after cleaning nonstoichiometric TiC and a small amount of free carbon remained near the surface. This is consistent with the Raman analysis. 4. Conclusion Single-crystal TiC films have been synthesized by RF sputtering at room temperature on Al2O3 substrate from the Ti/C compound target and the film composition was determined by RBS. There are three key features in the successful growth of a single crystal TiC film: - Temperature is a key factor in the synthesis of single-crystal films as the crystallinity improves at high substrate temperature. Both Raman spectra and XRD demonstrate that films grown at higher temperature result in better crystal quality. - Al2O3 is a better substrate for the growth of TiC films than Si due to the smaller lattice mismatch between the TiC and Al2O3. - There is an optimum bias voltage at which the best TiC single crystal films are grown. XPS reveals that the as‐prepared surface is contaminated and that below this surface layer of contamination relatively pure TiC (with some carbon inclusions) is present.
Fig. 9. (a) C 1s and (b) Ti 2p XPS spectra of the same samples as Fig. 8 after sputter etching.
Acknowledgment The authors are grateful to the staff of the tandem accelerator in The Key Lab of Applied Ion Beam Physics at Fudan University, for their cooperation during ion beam analysis experiments. Our work was supported by the National Nature Science Foundation of China under Grant No. 10975035, and by the National Magnetic Confinement Fusion Science Program under contract No. 2010 GB104002. References [1] I. Zergioti, A. Hatziapostolou, E. Hontzopoulos, A. Zervaki, G.N. Haidemenopoulos, Thin Solid Films 271 (1995) 96. [2] I.Yu. Konyashin, Thin Solid Films 278 (1996) 37. [3] F. Santerre, M.A. El Khakani, M. Chaker, J.P. Dodelet, Appl. Surf. Sci. 148 (1999) 24. [4] H. Randhawa, Thin Solid Films 153 (1987) 209. [5] E. Kusano, A. Satoh, M. Kitagawa, H. Nanto, A. Kinbara, Thin Solid Films 343 (1999) 254. [6] M. Stüber, H. Leiste, S. Ulrich, H. Holleck, D. Schild, Surf. Coat. Technol. 150 (2002) 218. [7] S. Inoue, Y. Wada, K. Koterazawa, Vacuum 59 (2000) 735. [8] E. Kusano, A. Sato, N. Kikuchi, H. Nanto, A. Kinbara, Surf. Coat. Technol. 120 (1999) 378. [9] L. Norin, S. McGinnis, U. Jansson, J.-O. Carlsson, J. Vac. Sci. Technol. A 15 (1997) 3082. [10] H. Högberg, L. Norin, J. Lu, J.O. Malm, U. Jansson, J. Mater. Res. 14 (1999) 1589. [11] M. Rester, J. Neidhardt, P. Eklund, J. Emmerlich, H. Ljungcrantz, L. Hultman, C. Mitterer, Mater. Sci. Eng., A 429 (2006) 90. [12] B.H. Lohse, A. Calka, D. Wexler, J. Appl. Phys. 97 (2005) 114912.
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887 [13] S. Intarasiri, L.D. Yu, S. Singkarat, A. Hallén, J. Lu, M. Ottosson, J. Jensen, G. Possnert, J. Appl. Phys. 101 (2007) 084311. [14] M. Amer, M.W. Barsoum, T. EI-Raghy, I. Weiss, S. Leclair, D. Liptak, J. Appl. Phys. 84 (1998) 5817. [15] M.V. Klein, John A. Holy, W.S. Williams, Phys. Rev., B 17 (1978) 1546.
6887
[16] E.H. Kisi, J.A.A. Crossley, S. Myhra, M.W. Barsoum, J. Phys. Chem Solids. 59 (1998) 1437. [17] D.P. Riley, D.J. O'Connor, P. Dastoor, N. Brack, P.J. Pigram, J. Phys. D: Appl. Phys. 35 (2002) 1603.
Contents lists available at SciVerse ScienceDirect
Thin Solid Films journal homepage: www.elsevier.com/locate/tsf
Preparation of single-crystal TiC (111) by radio frequency magnetron sputtering at low temperature Q. Qi, W.Z. Zhang, L.Q. Shi ⁎, W.Y. Zhang, W. Zhang, B. Zhang Applied Ion Beam Physics Laboratory, Institute of Modern Physics, Fudan University, Shanghai 200433, People's Republic of China
a r t i c l e
i n f o
Article history: Received 26 February 2012 Received in revised form 10 July 2012 Accepted 13 July 2012 Available online 20 July 2012 Keywords: Titanium carbide Radio-frequency magnetron sputtering Rutherford backscattering spectroscopy X-ray diffraction
a b s t r a c t Single-crystal films of TiC (111) have been synthesized at room temperature on Al2O3 (0001) substrates by radio frequency magnetron sputtering using a compound Ti–C target. The substrate temperature and bias were varied to explore the influence of deposition parameters on the crystal structure. Both Al2O3 (0001) and Si (100) substrates were used for epitaxial growth of TiC films. A series of characterizations of TiC films were carried out, including Rutherford backscattering spectroscopy, X-ray diffraction, Raman and X-ray photoelectron spectroscopy. Single-crystal films of TiC (111) on the Al2O3 (0001) were demonstrated. © 2012 Elsevier B.V. All rights reserved.
2. Experimental details
Al2O3 wafers. The wafers were cleaned several times in an ultrasonic bath containing a solution of ethanol and then were dried and loaded into the sputtering chamber immediately to avoid contamination. The compound target was fabricated by a Ti disk (purity 99.99% and 75 mm in diameter) and sector carbon sheets with the Ti/C area ratio of 2.429. The target was mounted on the RF electrode and located 50 mm from the substrate. The substrate was mounted in a stainless steel holder which was biased negatively to ground. The base pressure before deposition was less than 3 × 10 −5 Pa, using a turbo-molecular pump. In order to decrease doping of H from H2O adsorbing on the wall, the chamber was baked at about 200 °C for 3 h. The working gas used for discharge was argon (99.999% purity). In order to investigate the effects of the substrate and deposition temperature on the TiC films, both Al2O3 (0001) and Si (100) substrates were used and a series of substrate temperatures from room temperature to 700 °C were performed. The total pressure of the sputtered gas was controlled at 0.1 Pa in order to obtain stable plasma. In this experiment, the flux of Ar was 1.8 sccm (standard cubic centimeter per minute). The applied bias during the deposition was fixed at −60 V except for when the effects of bias voltage were studied. The RF power for all depositions was 100 W. The film thickness and deposition rate were about 200 nm and 3 nm/min respectively.
2.1. Specimen preparation
2.2. Ion beam analysis
The TiC films were prepared by a radio frequency (RF) magnetron sputtering method. The films were deposited on 10 cm by 15 cm
The composition and thickness of the TiC films were measured by using non-Rutherford scattering (non-RBS) of 3.65 MeV He+ ion beam. At this energy the scattering cross section of C is significantly enhanced over the standard Rutherford cross section. The incident angle of the beam was 0°, while He+ ions backscattered by target were
1. Introduction TiC possesses many special properties, such as high melting point, high thermal stability, exceptional hardness, outstanding wear resistance and good corrosion resistance, which make it of interest in many fields including wear-resistant coatings on cutting and grinding tools, thermal barriers and corrosion resistance on metallic structure. TiC thin films have been synthesized by various methods such as chemical vapor deposition [1,2], pulsed laser deposition [3], arc ion plating [4], and sputtering [5–8]. The disadvantage of these techniques, however, is that the microstructure of the film is highly dependent on the deposition temperature. Synthesis of single-crystal TiC films requires high temperature. Lower deposition temperature usually leads to amorphous or polycrystalline films. Recently it has been demonstrated that epitaxial carbide films could be deposited by co-evaporation of C60 and Ti at very low temperature [9,10]. Up until now no report on synthesis of single-crystal TiC films at room temperature has been made. This paper will introduce a method of growing single-crystal TiC films at room temperature and explore the effects of various parameters on crystal microstructure.
⁎ Corresponding author. Tel.: +86 21 65642292. E-mail address: [email protected] (L.Q. Shi). 0040-6090/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.tsf.2012.07.040
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
6883
detected at a scattering angle of 165°. All the non-RBS spectra were simulated by the SIMNRA 6.0 codes. Non-RBS measurements were performed at the NEC 9SDH-2 tandem accelerator at the Institute of Modern Physics of Fudan University, China.
can be determined. In this study the Raman spectra were measured over the range from 200 to 1800 cm−1 using a Jobin Yvon HR800 confocal Raman with the laser wavelength of 632.8 nm. The spectra were record with acquisition time of 100 s.
2.3. X-ray diffraction (XRD) analysis
2.5. X-ray photoelectron spectroscopy (XPS) analysis
The crystal structure was investigated by XRD. The measurement was performed at a D/MAX 2550 V X-ray diffractometer made by Rigaknu, Japan. This diffractometer with Cu Kα source was used in a 2θ mode, 2θ varying from 20° to 68° with a 0.02° step.
XPS was recorded using a Kratos AXIS ULTRA DLD instrument with the source of an Al Kα monochromatic X-ray. The gun was operated at 15 kV and 10 mA. Survey spectra were acquired from 0 to 1200 eV in steps of 1 eV for the pass energy of 160 eV. Region scans from selected ranges were obtained using the pass energy of 20 eV. The peaks were acquired from the selected elements Ti 2p, C 1s and O 1s. The base pressure during measurement is lower than 10−7 Pa. In order to avoid the effects of oxides, the samples were etched by Ar+ for 3 min using the source of 2 kV and 30 mA over an area of 3 × 3 mm2. The spectra were record for both unetched and etched samples.
2.4. Raman spectroscopy analysis Raman spectroscopy is based on Raman effect, in which the incident photons are scattered inelastically by molecules or crystal lattices and collected. From the Raman spectra the details of the crystal structure
Fig. 1. RBS spectra of TiC films deposited on (a) Al2O3 substrate and (b) Si substrate with the deposition power of 100 W and bias voltage of −60 V.
6884
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
3. Results and discussion To allow the deposited atoms to diffuse across the substrate the deposition rate of the films was limited to a maximum rate of approximately 3 nm/min through the RF power. Typical non-RBS spectra of the TiC films deposited on the Al2O3 and Si wafers by RF magnetron sputtering are shown in Fig. 1(a) and (b). The composition of the films can be acquired from the best fit simulation to the spectra. The Ti to C ratio for all deposited films on the two types of substrates is between 0.8 and 1.2 due to different deposition temperatures. The substrate temperature is a key parameter in the synthesis of single-crystal TiC films, as it is the controlling feature of atomic mobility. At all substrate temperatures used in this study there is only one peak in the XRD patterns at 35.9° (apart from the substrate Al2O3 peak at 41.71°) which corresponds to TiC (111) (Fig. 2). This indicates that single crystal TiC (111) grows epitaxially on the Al2O3 (0001) substrates by RF magnetron sputtering from room temperature to 700 °C. The spectra also suggest that at high substrate temperature the peak intensity is enhanced and the full width at half maximum (FWHM) of the TiC peak becomes narrower with increasing temperature except for room temperature (Fig. 3). The high crystallinity of the film deposited at room temperature may be attributed to the film composition approaching to stoichiometric TiC so that the unreacted carbon or titanium in the film is less. However further investigation of this effect is under way. In general the crystal structure of TiC films tends be more ordered at higher temperature. The XRD patterns of TiC films grown on Al2O3 (0001) substrate with different bias voltages at the same deposition temperature of 700 °C are shown in Fig. 4. At the bias voltage of −25 V the TiC (111) peak almost disappears which implies that at this low bias voltage the deposited atoms and the impinging Ar+ ions do not have sufficient energy to promote atomic diffusion of surface atoms hence the degree of crystallinity is poor and the crystal microstructure tends to be amorphous. At a moderate bias voltage of −60 V there is only one sharp peak of TiC which demonstrates that a suitable bias in synthesizing single-crystal TiC films on Al2O3 (0001) substrate was around −60 V. At an applied higher bias voltage of −100 V, however, the intensity of the peak is weaker and the FWHM is broader than that at −60 V. The reason may be that the more energetic ions penetrate the film to create defects and produce compressive stress which causes the distortion of the lattice and a damaged film.
Fig. 3. Full width at half maximum (FWHM) and intensity from XRD peaks of TiC (111) deposited on Al2O3 at different temperatures.
In order to further explore the role of the substrate on epitaxial growth by RF sputtering, the structure formed at different deposition temperatures for TiC films on Si substrates is identified by XRD (Fig. 5). At a high temperature of 700 °C a relative strong TiC (111) peak and a very weak TiC (220) peak as well can be observed in diffraction pattern, suggesting relatively high preferred (111) orientation and better order. As the temperature was reduced to 500 °C, only a relative weak TiC (200) peak appears in the pattern and the crystal exhibits a fully (200) orientation and a poorer degree of crystallinity. At a deposition temperature of 200 °C, a very weak diffraction from polycrystalline TiC is observed, indicating very poor order. The major factors for epitaxial growth are lattice mismatch, temperature and deposition rate. The lattice mismatch could be characterized by the parameter: Δ¼
jas −ae j ae
where Δ is the lattice mismatch value, and as and ae are the lattice parameters of the substrate and epitaxial film respectively. For TiC film deposited on Al2O3 or Si substrates, ae = 4.327 Å, as = 4.758 Å for Al2O3
Fig. 2. XRD patterns of TiC films deposited on Al2O3 with different temperatures.
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
Fig. 4. XRD patterns of TiC films deposited on Al2O3 in different bias. The deposition power is 100 W.
6885
Fig. 6. Raman spectra of TiC films deposited on the Al2O3 with different temperatures. The deposition power is 100 W and the bias voltage is −60 V.
and as =5.429 Å for Si which results in the mismatch parameters of ΔAl2O3 = 0.0996, ΔSi = 0.25. The large difference in the mismatch parameters is the primary reason that the crystal quality of epitaxial films deposited on Al2O3 is better than that on Si given all other parameters equal. In order to determine the crystal structure quality of TiC, Raman spectroscopy was performed in the range between 200 and 1800 cm−1 [11–13]. In this range high purity titanium sheet does not have Raman active vibrational modes so no peaks can be observed for the pure metal. For a high purity carbon wafer two strong peaks at 1320 cm−1 and 1598 cm−1 can be detected, which are associated with the A1g and E2g vibrational modes of graphite. Stoichiometric TiC does not have Raman active vibrational modes, however, peaks arising from carbon vacancies can be observed for TiCx (where x b 1) as the vacancies can destroy the inversion symmetry of nearby atoms and induce first-order Raman scattering [14,15]. The peaks at 264, 349, 600, and 658 cm−1 (Fig. 6) correspond to nonstoichiometric TiC. The width and intensity of these peaks decrease with increasing temperature which indicates that the quality of the crystal has improved. At a substrate temperature of 700 °C, the spectrum has a broaden peak from 1300 to 1600 cm−1 which is caused by superposed spectra of the A1g and E2g vibrational modes of graphite however these two peaks do not appear in the spectra at other temperatures. The reason is that at
higher temperature excess carbon atoms aggregate to form graphite which has Raman active vibrational modes that produce the peaks. This case was further demonstrated in Fig. 7, which presents the spectra of the same samples before and after annealing. The higher intensity and smaller FWHM of the A1g and E2g vibrational modes at 1320 and 1598 cm−1 peaks are indicative of the formation of graphite in the samples. For the sample TiC deposited at 700 °C, X-ray photoelectron spectra (XPS) for C 1 s and Ti 2p before sputter etching with Ar + were shown in Fig. 8(a) and (b) respectively and after etching were presented in Fig. 9(a) and (b). For the region scans for C 1s, the peak at binding energy of 281.6 eV shown in Fig. 8(a) was attributed to the Ti\C bond and the peak at 284.6 eV was to the C\C bond [16,17]. There are also two other peaks at higher binding energies of 286 eV and 288.1 eV which are responding to C\O\R bond and C_O bond respectively. However, after etching as shown in Fig. 9(a), there are only two peaks: one is a sharp peak at 281.8 eV corresponding to Ti\C bond, the other is a small peak associated to C\C bond. The peaks corresponding to oxide and contamination bonds disappear. Figs. 8(b) and 9(b) show the region scans of Ti 2p. The Ti 2p XPS spectra present one spin orbit doublet Ti 2p3/2 and 2p1/2. The peaks of binding energy at 454.76 eV and 458.3 eV represent Ti 2p3/2 carbide and 2p3/2 oxide TiO2 and the peaks at 460.8 eV and 463.9 eV are associated to Ti 2p1/2 carbide and Ti1/2
Fig. 5. XRD patterns of TiC films grown on Si substrate with different temperatures. The deposition power is 100 W and the bias voltage is −60 V.
Fig. 7. Raman spectra of annealed and unannealed TiC films.
6886
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887
Fig. 8. (a) C 1s and (b) Ti 2p XPS spectra of the TiC films deposited on Al2O3 before etching with the deposition power of 100 W and bias voltage of −60 V.
oxide TiO2. After sputter etching (Fig. 9(b)) Ti 2p oxides presented in Fig. 8(b) have almost disappeared and only Ti 2p3/2 carbide and Ti 2p1/2 carbide exist. From the spectra before and after sputter etching it is suggested that there is a contamination on the sample surface assigned to C_O bond in the as‐prepared sample but after cleaning nonstoichiometric TiC and a small amount of free carbon remained near the surface. This is consistent with the Raman analysis. 4. Conclusion Single-crystal TiC films have been synthesized by RF sputtering at room temperature on Al2O3 substrate from the Ti/C compound target and the film composition was determined by RBS. There are three key features in the successful growth of a single crystal TiC film: - Temperature is a key factor in the synthesis of single-crystal films as the crystallinity improves at high substrate temperature. Both Raman spectra and XRD demonstrate that films grown at higher temperature result in better crystal quality. - Al2O3 is a better substrate for the growth of TiC films than Si due to the smaller lattice mismatch between the TiC and Al2O3. - There is an optimum bias voltage at which the best TiC single crystal films are grown. XPS reveals that the as‐prepared surface is contaminated and that below this surface layer of contamination relatively pure TiC (with some carbon inclusions) is present.
Fig. 9. (a) C 1s and (b) Ti 2p XPS spectra of the same samples as Fig. 8 after sputter etching.
Acknowledgment The authors are grateful to the staff of the tandem accelerator in The Key Lab of Applied Ion Beam Physics at Fudan University, for their cooperation during ion beam analysis experiments. Our work was supported by the National Nature Science Foundation of China under Grant No. 10975035, and by the National Magnetic Confinement Fusion Science Program under contract No. 2010 GB104002. References [1] I. Zergioti, A. Hatziapostolou, E. Hontzopoulos, A. Zervaki, G.N. Haidemenopoulos, Thin Solid Films 271 (1995) 96. [2] I.Yu. Konyashin, Thin Solid Films 278 (1996) 37. [3] F. Santerre, M.A. El Khakani, M. Chaker, J.P. Dodelet, Appl. Surf. Sci. 148 (1999) 24. [4] H. Randhawa, Thin Solid Films 153 (1987) 209. [5] E. Kusano, A. Satoh, M. Kitagawa, H. Nanto, A. Kinbara, Thin Solid Films 343 (1999) 254. [6] M. Stüber, H. Leiste, S. Ulrich, H. Holleck, D. Schild, Surf. Coat. Technol. 150 (2002) 218. [7] S. Inoue, Y. Wada, K. Koterazawa, Vacuum 59 (2000) 735. [8] E. Kusano, A. Sato, N. Kikuchi, H. Nanto, A. Kinbara, Surf. Coat. Technol. 120 (1999) 378. [9] L. Norin, S. McGinnis, U. Jansson, J.-O. Carlsson, J. Vac. Sci. Technol. A 15 (1997) 3082. [10] H. Högberg, L. Norin, J. Lu, J.O. Malm, U. Jansson, J. Mater. Res. 14 (1999) 1589. [11] M. Rester, J. Neidhardt, P. Eklund, J. Emmerlich, H. Ljungcrantz, L. Hultman, C. Mitterer, Mater. Sci. Eng., A 429 (2006) 90. [12] B.H. Lohse, A. Calka, D. Wexler, J. Appl. Phys. 97 (2005) 114912.
Q. Qi et al. / Thin Solid Films 520 (2012) 6882–6887 [13] S. Intarasiri, L.D. Yu, S. Singkarat, A. Hallén, J. Lu, M. Ottosson, J. Jensen, G. Possnert, J. Appl. Phys. 101 (2007) 084311. [14] M. Amer, M.W. Barsoum, T. EI-Raghy, I. Weiss, S. Leclair, D. Liptak, J. Appl. Phys. 84 (1998) 5817. [15] M.V. Klein, John A. Holy, W.S. Williams, Phys. Rev., B 17 (1978) 1546.
6887
[16] E.H. Kisi, J.A.A. Crossley, S. Myhra, M.W. Barsoum, J. Phys. Chem Solids. 59 (1998) 1437. [17] D.P. Riley, D.J. O'Connor, P. Dastoor, N. Brack, P.J. Pigram, J. Phys. D: Appl. Phys. 35 (2002) 1603.