- Email: [email protected]
Texture development in Ti–Si–N nanocomposite thin films
Materials Science and Engineering A 423 (2006) 111–115
Texture development in Ti–Si–N nanocomposite thin films R. Chandra a , Davinder Kaur b,∗ , Amit Kumar Chawla a , N. Phinichka c , Z.H. Barber d a
Institute Instrumentation Centre, Indian Institute of Technology Roorkee, Roorkee 247 667, India b Department of Physics, Indian Institute of Technology Roorkee, Roorkee 247 667, India c Faculty of Science, Srinakharinwirot University, Bangkok, Thailand d Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge CB2 3QZ, UK Received 8 July 2005; received in revised form 15 September 2005; accepted 30 September 2005
Abstract Nanocomposite thin films of titanium silicon nitride were deposited by sputtering on R-plane sapphire substrates. The effects of silicon addition and negative substrate bias on the texture development of the films were studied systematically by varying the bias voltage in the range −20 to −200 V. The accompanying changes in the microstructure and growth morphology of the phases in these films were investigated in detail using X-ray diffraction and a atomic force microscopy. In addition, the effect of texture on the mechanical properties of the films was also investigated using nanoindentation technique. Pure TiN films deposited without Si exhibit a strong (1 1 1) preferred orientation, while with addition of Si, the orientation of the films changes from (1 1 1) to (2 0 0). Meanwhile the surface morphology of these films changed from a pronounced columnar microstructure to a dense, fine-grained structure. The effect of negative substrate bias voltage applied during deposition also resulted in a similar change of film orientation and microstructure and leads to the increase in hardness of the films from 21 to 40 GPa, respectively. © 2006 Elsevier B.V. All rights reserved. Keywords: Texture; Nanocomposite; Thin films; Ti–Si–N films
1. Introduction Hard coatings with tailored properties are increasingly important for applications in many different areas of engineering and industry such as coatings for cutting tools under dry and high speed machining conditions protective coatings for turbine blades and engine parts to improve their durability [1–5]. A nanocomposite coating comprises of at least two phases: a nanocrystalline phase and an amorphous phase, or two nanocrystalline phases. The addition of the second phase not only prevents grain growth but also suppresses grain boundary sliding (for grain size <10 nm), and hence improve the mechanical properties [6–10]. The hardness of these coatings can be tailored depending on the design and application. Thin films with preferential crystallographic orientation are desirable for particular applications because of their anisotropic nature. For example, the mechanical behaviour of TiN films is governed by the preferred growth orientations, since TiN is an ∗
Corresponding author. E-mail address: [email protected] (D. Kaur).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2005.09.132
anisotropic material with E1 0 0 > E1 1 1 (where E1 0 0 and E1 1 1 are the elastic moduli along the [1 0 0] and [1 1 1] crystallographic directions, respectively) [11]. The preferred orientation of TiN films may be dependent upon the competition between the surface free energy and the strain energy [12]. By assuming that the strain energy in the film increases linearly with thickness, it has been postulated that, when the film thickness is sufficiently small, the film orientation is the result of minimization of surface energy (observed to be (1 0 0) for TiN films) [12]. However, in other work [13], an increase in internal stress from 0.3 to 2 GPa as a result of increased substrate bias voltage during film growth, resulted in a change in preferred TiN orientation from (1 1 1) to (2 0 0), also in Ref. [14] the orientation changes to (2 0 0) from (1 1 1) by varying the incident ion/metal flux ratio from 1 to ≥5 keeping N2 + ion energy constant at ∼20 eV which suggests that the strain energy might not be the main cause for the (1 1 1) preferred orientation. In this paper Ti1−x Six N thin films were deposited by ionised magnetron sputter deposition (IMSD). The variations of film structure and properties as a function of Si addition and substrate bias voltage were investigated.
112
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
2. Experimental Titanium silicon nitride (Ti1−x –Six N) films were deposited by IMSD, using separate Ti (99.96%) and Si (99.999%) dc magnetron targets (55 mm × 35 mm × 2 mm). Film stoichiometry, Ti1−x :Six (where x is the atomic fraction of Si: x = Si/[Ti + Si]) could be precisely controlled with the power to each target. The target power was typically set at 5.2 W/cm2 for Ti and 1.8 W/cm2 for the Si. A three-turn rf coil between targets and substrates generated an additional plasma, through which the depositing flux passed. The base pressure of the deposition chamber was 10−8 Pa, and the depositions were carried out in 1.4 Pa Ar/N2 (50:50): the gases both of 99.9999% purity were pre-mixed in a reservoir and bled into the system via a leak valve. By setting the leak and gate valve to the main pump, gas flow rate and pressure were controlled. Substrates rested on a platinum strip heater and the dc substrate bias was varied over the range of 0 to −200 V. Further details of the deposition system have been reported elsewhere [15]. X-ray diffraction (XRD) patterns were recorded using a conventional X’Pert Philips diffractometer (generator current and voltage 40 mA and 40 kV, respectively). The average grain size was calculated using Scherrer equation [16]. The titanium to silicon ratio in the deposited films was measured using a energy dispersive X-ray analysis (EDX) and was found to be Ti0.84 Si0.16 N. The surface morphology of the films was observed by atomic force microscopy (AFM), in tapping mode. Film thickness (measured by profilometry of a step on a masked substrate) was typically 3–3.5 m. Film hardness, H, was measured with nanoindentation (Nanotest 600, Micro Materials Ltd., Wrexham, UK, fitted with Berkovich indenter: a three-sided pyramid with the same area-to-depth ratio as a Vickers indenter). The intrinsic mechanical properties of these films were measured at indentation depths between 200 and 330 nm, which is less than 10% of the film thickness, and averaging over a total of five indents. The influence of the mechanical properties of the substrate on the measurement is therefore avoided. A test load of 30 mN was typically used and the loading and unloading speed was kept
Fig. 1. XRD patterns of Ti1−x Six N films deposited at various substrate bias voltages along with the pure TiN film.
at 0.8 nm s−1 . The Oliver–Pharr method was used to analyze the loading and unloading curves [17]. To minimize creep and thermal drift effects the hold period was kept at 10 s at maximum depth and 60 s during unloading at 20% of the maximum load. 3. Results and discussion Fig. 1 shows XRD spectra of the Ti1−x Six N films deposited at 400 ◦ C with varying negative substrate bias voltages, Vb , from −20, to −200 V (along with a reference TiN film prepared with a bias voltage of −80 V). Only diffraction peaks assigned to crystalline TiN are observed, with no indication of the presence of crystalline Si3 N4 or titanium silicide phases. These observations are in agreement with earlier findings from nc-TiN/a-Si3 N4 films deposited by PVD [18–21], suggesting that the silicon is present in an amorphous phase. The undoped TiN film shows strong preferred (1 1 1) orientation and a narrow line width corresponding to a grain size of approximately 30 nm. For the Ti1−x Six N films, an increase in negative substrate bias gradually changes the preferred film orientation of TiN (in Ti1−x Six N) from (1 1 1) to mixed, and finally to (2 0 0), with broadening of the peaks. Peak broadening is generally attributed to reduction in the coherent
Fig. 2. AFM images of (a) TiN and (b) Ti0.84 Si0.16 N film at bias voltage of −80 V.
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
113
Fig. 3. AFM images of Ti1−x Six N films deposited at different substrate bias voltage: (a) Ti1−x Six N, Vb = −20 V; (b) Ti1−x Six N, Vb = −80 V; (c) Ti1−x Six N, Vb = −150 V; (d) Ti1−x Six N, Vb = −200 V; (e) Ti1−x Six N, Vb = −200 V.
114
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
diffracting domain size (related to grain size) and/or microstrains due to defects. The negative bias on the substrate influences ion bombardment parameters (momentum, kinetic energy etc.) and hence the microstructure of the growing film. As shown in Fig. 1, the preferred orientation of TiN (1 1 1) changes to (2 0 0) orientation with increase in negative bias from −20 to −200 V. Similar changes in film orientation have also been observed by varying other processing parameters e.g. nature of the sputtering gas in our earlier report [22]. Generally, the (1 1 1) plane in the TiN crystal has the lowest strain energy [12], while the (2 0 0) plane has the lowest surface energy. It is assumed that the growing coating develops a crystallographic texture in order to minimize the total energy of the system. AFM images of TiN and Ti0.84 Si0.16 N films deposited at negative bias voltage of −80 V are shown in Fig. 2. The addition of silicon changes the surface morphology from a pronounced columnar microstructure with faceted grains to a dense, finegrained structure. Fig. 3 shows the AFM images of the films deposited at varying negative substrate bias voltages Vb (−20, −80, −150 and −200 V), The effect of increasing the bias voltage, Vb from −20 to −200 V results in decrease of grain size as observed in AFM images. In general with increase in higher negative substrate bias the energy of the incident ions increases that leads to increase in the grain size because of the increase of adatom mobility. However in our case due to the presence of amorphous matrix of Si3 N4 , the grain growth of TiN phase is suppressed resulting in more surface defects and consequently a higher density of nucleation sites on the substrate. The surface roughness of Ti1−x Six N films is also affected by the negative substrate bias during deposition. A decrease in surface roughness upon increasing negative substrate bias has been observed as shown in Fig. 3(b and c) [at Vb = −80 V (RMS, 2.55 nm), and at Vb = −150 V (RMS 1.35 nm respectively)]. The increase in Vb above −150 V results in the increase of surface roughness due to resputtering of the growing film. Although the surface roughness of the films deposited at Vb −200 V at a scale of m is very small (RMS, 0.548 nm) with a very small crystallite size (≈5 nm), however overall surface of the films at mm scale becomes very rough leading to deterioration of mechanical properties. This result could be explained by the XRD results of Fig. 1. From XRD patterns, the peak broadening, indicating the size reduction of TiN (30 nm for pure TiN; 18–20 nm (Vb = −80 V), 10 nm (Vb = −150 V) and 5 nm (Vb = −200 V) of Ti1−x Six N films), and the microstructure changes from strong (1 1 1) preferred orientation to mixed one and finally to highly (2 0 0) orientation with increase in substrate bias voltage. It is understood that surface roughness is related with the size and orientation of TiN crystallite and is well known that the increase of Vb leads to the enhancement of the adatom surface mobility as well as the atomic diffusion, thereby enhancing grain growth, which contributes to the increase of grain size and surface roughness as shown in Fig. 2(a) for undoped TiN film. But the presence of the amorphous matrix of Si3 N4 , suppresses the grain growth of TiN nanocrystals, leading to a finely grained dense microstructure.
Fig. 4. Hardness values of Ti1−x Six N films as a function of negative substrate bias voltage, along with crystallite size variation.
Fig. 4 shows hardness, H, of Ti1−x Six N films as a function of substrate bias (Vb ). The hardness of the films increases with increasing negative bias Vb, and reached a maximum value of 40 GPa at −150 V and dropped again with further increase in Vb . Above dc bias of −150 V, the resputtering of the growing film by bombardment of energetic particles results in large surface roughness and hence a decrease in hardness value. The high hardness value is attributed to the nanocomposite structure described by Veprek [6] that leads to change in grain size, orientation, and increase in density and elimination of voids. All these effects are related to the adatom mobility and the ion–solid interactions near the growing film surface under negative bias. The increase in Vb increases the mobility of adatoms promoting closed packed structure in near thermodynamic equilibrium conditions. Thus, for high adatom mobility Ti1−x Six N films are expected to grow along the (2 0 0) orientation corresponding to the lowest surface free energy [23]. On the other hand, for low adatom mobility the preferred orientation is the (1 1 1) in which the highest number of atoms per unit area can be incorporated at low energy sites. The amounts of grains grown along the (2 0 0) and (1 1 1) orientations play a significant role on the mechanical properties of Ti1−x Six N films. The crystallite size showed a monotonous decrease with increasing Vb , but the maximum hardness was obtained for a Vb = −150 V. We have observed that in addition to the decrease in grain size, the negative bias of the substrate during deposition strongly influences the preferred orientation of the reactively sputtered Ti1−x Six N films. Development of a preferred orientation in non-epitaxial thin films (where the substrate plays a relatively minor role in the texture determining process) is primarily governed by the surface and strain energy considerations. Thermodynamics derives the system towards the minimum possible sum of surface and strain energies under the restrictions imposed by kinetics. In TiN (2 0 0) planes or the square surfaces of the cubic unit cell have the lowest surface energy [23]. The increase in Vb increases the energy of the adatoms to diffuse on the substrate surface. Thus, the adatoms are no longer kinetically constrained to metastable high-energy configurations, but relax to the lowest surface energy configurations of (2 0 0) planes. However due to much higher kinetic energy imparted due to large negative substrate bias (−200 V) leads to resputtering of the deposited film leading to deterioration of surface and hence the mechanical properties.
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
4. Conclusions Nanocomposite thin films of Ti1−x Six N were deposited on R-plane sapphire substrates using ionised magnetron sputtering. The effect of processing parameters on the microstructure of these films was investigated using XRD and AFM. During deposition of the Ti1−x Six N films, the increase in negative bias voltage Vb , results in a structure change from (1 1 1) orientation with large TiN crystallite to (2 0 0) with a smaller size. The surface morphology of the films is also affected with increase in negative bias Vb . The roughness of the films is found to decrease with increase in Vb upto −150 V and with further increase in bias voltage to −200 V, the surface roughness increases due to resputtering which leads to deterioration of the mechanical properties of Ti1−x Six N films. The best mechanical properties (H = 40 GPa) of the Ti1−x Six N thin films for the present study were obtained at a negative dc substrate bias Vb of −150 V. Acknowledgements One of the authors R. Chandra, would like to acknowledge the University Grants Commission, New Delhi (India) and Association of Commonwealth Universities, UK, for providing Commonwealth Fellowship to visit and work at University of Cambridge, UK. Dr. Nadia Stelmashenko for technical help in getting AFM images is also acknowledged. References [1] J. Musil, Surf. Coat. Technol. 125 (2000) 322. [2] R.A. Andrievski, Mater. Trans. 42 (2001) 1471.
115
[3] P.J. Martin, A. Bendavid, Surf. Coat. Technol. 163–164 (2003) 245. [4] S. Veprek, J. Vac. Sci. Technol. A 17 (1999) 2401. [5] M. Diserens, J. Patscheider, F. Levy, Surf. Coat. Technol. 108–109 (1998) 241–246. [6] S. Veprek, Surf. Coat. Technol. 97 (1997) 15. [7] M. Diserens, J. Patscheider, F. Levy, Surf. Coat. Technol. 121 (1999) 158. [8] J. Musil, J. Vlcek, Surf. Coat. Technol. 142–144 (2001) 557–566. [9] W.J. Meng, X.D. Zhang, B. Shi, J.C. Jiang, L.E. Rehn, P.M. Baldo, R.C. Tillsworth, Surf. Coat. Technol. 163–164 (2003) 251–259. [10] S. Veprek, A. Niederhofer, K. Moto, P. Nesladek, H.-D. Mannling, M. Jilek, Surf. Coat. Technol. 133–134 (2000) 152. [11] J.O. Kim, J.D. Allenbach, P.B. Mirkarimi, S.A. Barnett, Phys. Rev. B 48 (1993) 1726. [12] J. Pelleg, L.Z. Zevin, S. Lungo, Thin Solid Films 197 (1991) 117. [13] P. Patsalas, C. Charitidis, S. Logothetidis, Surf. Coat. Technol. 125 (2000) 335. [14] J.E. Greene, J.E. Sundgren, L. Hultman, I. Petrov, D.B. Bergstrom, Appl. Phys. Lett. 67 (1995) 2928. [15] N. Phinichka, R. Chandra, Z.H. Barber, J. Vac. Sci. Technol. A 22 (2004) 477. [16] B.D. Cullity, Elements of X-ray Diffraction, Addison–Wesley Publishing Co, Reading, MA, 1978. [17] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564. [18] F. Vaz, L. Rebouta, P. Goudeau, J. Pacaud, H. Garem, J.P. Rivere, A. Cavaleiro, E. Alves, Surf. Coat. Technol. 133–134 (2000) 307–313. [19] F. Vaz, L. Rebouta, Ph. Goudeau, T. Girardeau, J.P. Rivere, A. Traverse, Surf. Coat. Technol. 146–147 (2001) 274–279. [20] S. Zhang, D. Sun, Y. Fu, H. Du, Surf. Coat. Technol. 167 (2003) 113–119. [21] X. Hu, Z. Han, G. Li, M. Gu, J. Vac. Sci. Technol. A 20 (2002) 1921. [22] R. Banerjee, R. Chandra, P. Ayyub, Thin Solid Films 405 (2002) 64. [23] L. Hultman, J.E. Sundgren, J.E. Greene, J. Appl. Phys. 66 (1989) 536.
Texture development in Ti–Si–N nanocomposite thin films R. Chandra a , Davinder Kaur b,∗ , Amit Kumar Chawla a , N. Phinichka c , Z.H. Barber d a
Institute Instrumentation Centre, Indian Institute of Technology Roorkee, Roorkee 247 667, India b Department of Physics, Indian Institute of Technology Roorkee, Roorkee 247 667, India c Faculty of Science, Srinakharinwirot University, Bangkok, Thailand d Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge CB2 3QZ, UK Received 8 July 2005; received in revised form 15 September 2005; accepted 30 September 2005
Abstract Nanocomposite thin films of titanium silicon nitride were deposited by sputtering on R-plane sapphire substrates. The effects of silicon addition and negative substrate bias on the texture development of the films were studied systematically by varying the bias voltage in the range −20 to −200 V. The accompanying changes in the microstructure and growth morphology of the phases in these films were investigated in detail using X-ray diffraction and a atomic force microscopy. In addition, the effect of texture on the mechanical properties of the films was also investigated using nanoindentation technique. Pure TiN films deposited without Si exhibit a strong (1 1 1) preferred orientation, while with addition of Si, the orientation of the films changes from (1 1 1) to (2 0 0). Meanwhile the surface morphology of these films changed from a pronounced columnar microstructure to a dense, fine-grained structure. The effect of negative substrate bias voltage applied during deposition also resulted in a similar change of film orientation and microstructure and leads to the increase in hardness of the films from 21 to 40 GPa, respectively. © 2006 Elsevier B.V. All rights reserved. Keywords: Texture; Nanocomposite; Thin films; Ti–Si–N films
1. Introduction Hard coatings with tailored properties are increasingly important for applications in many different areas of engineering and industry such as coatings for cutting tools under dry and high speed machining conditions protective coatings for turbine blades and engine parts to improve their durability [1–5]. A nanocomposite coating comprises of at least two phases: a nanocrystalline phase and an amorphous phase, or two nanocrystalline phases. The addition of the second phase not only prevents grain growth but also suppresses grain boundary sliding (for grain size <10 nm), and hence improve the mechanical properties [6–10]. The hardness of these coatings can be tailored depending on the design and application. Thin films with preferential crystallographic orientation are desirable for particular applications because of their anisotropic nature. For example, the mechanical behaviour of TiN films is governed by the preferred growth orientations, since TiN is an ∗
Corresponding author. E-mail address: [email protected] (D. Kaur).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2005.09.132
anisotropic material with E1 0 0 > E1 1 1 (where E1 0 0 and E1 1 1 are the elastic moduli along the [1 0 0] and [1 1 1] crystallographic directions, respectively) [11]. The preferred orientation of TiN films may be dependent upon the competition between the surface free energy and the strain energy [12]. By assuming that the strain energy in the film increases linearly with thickness, it has been postulated that, when the film thickness is sufficiently small, the film orientation is the result of minimization of surface energy (observed to be (1 0 0) for TiN films) [12]. However, in other work [13], an increase in internal stress from 0.3 to 2 GPa as a result of increased substrate bias voltage during film growth, resulted in a change in preferred TiN orientation from (1 1 1) to (2 0 0), also in Ref. [14] the orientation changes to (2 0 0) from (1 1 1) by varying the incident ion/metal flux ratio from 1 to ≥5 keeping N2 + ion energy constant at ∼20 eV which suggests that the strain energy might not be the main cause for the (1 1 1) preferred orientation. In this paper Ti1−x Six N thin films were deposited by ionised magnetron sputter deposition (IMSD). The variations of film structure and properties as a function of Si addition and substrate bias voltage were investigated.
112
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
2. Experimental Titanium silicon nitride (Ti1−x –Six N) films were deposited by IMSD, using separate Ti (99.96%) and Si (99.999%) dc magnetron targets (55 mm × 35 mm × 2 mm). Film stoichiometry, Ti1−x :Six (where x is the atomic fraction of Si: x = Si/[Ti + Si]) could be precisely controlled with the power to each target. The target power was typically set at 5.2 W/cm2 for Ti and 1.8 W/cm2 for the Si. A three-turn rf coil between targets and substrates generated an additional plasma, through which the depositing flux passed. The base pressure of the deposition chamber was 10−8 Pa, and the depositions were carried out in 1.4 Pa Ar/N2 (50:50): the gases both of 99.9999% purity were pre-mixed in a reservoir and bled into the system via a leak valve. By setting the leak and gate valve to the main pump, gas flow rate and pressure were controlled. Substrates rested on a platinum strip heater and the dc substrate bias was varied over the range of 0 to −200 V. Further details of the deposition system have been reported elsewhere [15]. X-ray diffraction (XRD) patterns were recorded using a conventional X’Pert Philips diffractometer (generator current and voltage 40 mA and 40 kV, respectively). The average grain size was calculated using Scherrer equation [16]. The titanium to silicon ratio in the deposited films was measured using a energy dispersive X-ray analysis (EDX) and was found to be Ti0.84 Si0.16 N. The surface morphology of the films was observed by atomic force microscopy (AFM), in tapping mode. Film thickness (measured by profilometry of a step on a masked substrate) was typically 3–3.5 m. Film hardness, H, was measured with nanoindentation (Nanotest 600, Micro Materials Ltd., Wrexham, UK, fitted with Berkovich indenter: a three-sided pyramid with the same area-to-depth ratio as a Vickers indenter). The intrinsic mechanical properties of these films were measured at indentation depths between 200 and 330 nm, which is less than 10% of the film thickness, and averaging over a total of five indents. The influence of the mechanical properties of the substrate on the measurement is therefore avoided. A test load of 30 mN was typically used and the loading and unloading speed was kept
Fig. 1. XRD patterns of Ti1−x Six N films deposited at various substrate bias voltages along with the pure TiN film.
at 0.8 nm s−1 . The Oliver–Pharr method was used to analyze the loading and unloading curves [17]. To minimize creep and thermal drift effects the hold period was kept at 10 s at maximum depth and 60 s during unloading at 20% of the maximum load. 3. Results and discussion Fig. 1 shows XRD spectra of the Ti1−x Six N films deposited at 400 ◦ C with varying negative substrate bias voltages, Vb , from −20, to −200 V (along with a reference TiN film prepared with a bias voltage of −80 V). Only diffraction peaks assigned to crystalline TiN are observed, with no indication of the presence of crystalline Si3 N4 or titanium silicide phases. These observations are in agreement with earlier findings from nc-TiN/a-Si3 N4 films deposited by PVD [18–21], suggesting that the silicon is present in an amorphous phase. The undoped TiN film shows strong preferred (1 1 1) orientation and a narrow line width corresponding to a grain size of approximately 30 nm. For the Ti1−x Six N films, an increase in negative substrate bias gradually changes the preferred film orientation of TiN (in Ti1−x Six N) from (1 1 1) to mixed, and finally to (2 0 0), with broadening of the peaks. Peak broadening is generally attributed to reduction in the coherent
Fig. 2. AFM images of (a) TiN and (b) Ti0.84 Si0.16 N film at bias voltage of −80 V.
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
113
Fig. 3. AFM images of Ti1−x Six N films deposited at different substrate bias voltage: (a) Ti1−x Six N, Vb = −20 V; (b) Ti1−x Six N, Vb = −80 V; (c) Ti1−x Six N, Vb = −150 V; (d) Ti1−x Six N, Vb = −200 V; (e) Ti1−x Six N, Vb = −200 V.
114
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
diffracting domain size (related to grain size) and/or microstrains due to defects. The negative bias on the substrate influences ion bombardment parameters (momentum, kinetic energy etc.) and hence the microstructure of the growing film. As shown in Fig. 1, the preferred orientation of TiN (1 1 1) changes to (2 0 0) orientation with increase in negative bias from −20 to −200 V. Similar changes in film orientation have also been observed by varying other processing parameters e.g. nature of the sputtering gas in our earlier report [22]. Generally, the (1 1 1) plane in the TiN crystal has the lowest strain energy [12], while the (2 0 0) plane has the lowest surface energy. It is assumed that the growing coating develops a crystallographic texture in order to minimize the total energy of the system. AFM images of TiN and Ti0.84 Si0.16 N films deposited at negative bias voltage of −80 V are shown in Fig. 2. The addition of silicon changes the surface morphology from a pronounced columnar microstructure with faceted grains to a dense, finegrained structure. Fig. 3 shows the AFM images of the films deposited at varying negative substrate bias voltages Vb (−20, −80, −150 and −200 V), The effect of increasing the bias voltage, Vb from −20 to −200 V results in decrease of grain size as observed in AFM images. In general with increase in higher negative substrate bias the energy of the incident ions increases that leads to increase in the grain size because of the increase of adatom mobility. However in our case due to the presence of amorphous matrix of Si3 N4 , the grain growth of TiN phase is suppressed resulting in more surface defects and consequently a higher density of nucleation sites on the substrate. The surface roughness of Ti1−x Six N films is also affected by the negative substrate bias during deposition. A decrease in surface roughness upon increasing negative substrate bias has been observed as shown in Fig. 3(b and c) [at Vb = −80 V (RMS, 2.55 nm), and at Vb = −150 V (RMS 1.35 nm respectively)]. The increase in Vb above −150 V results in the increase of surface roughness due to resputtering of the growing film. Although the surface roughness of the films deposited at Vb −200 V at a scale of m is very small (RMS, 0.548 nm) with a very small crystallite size (≈5 nm), however overall surface of the films at mm scale becomes very rough leading to deterioration of mechanical properties. This result could be explained by the XRD results of Fig. 1. From XRD patterns, the peak broadening, indicating the size reduction of TiN (30 nm for pure TiN; 18–20 nm (Vb = −80 V), 10 nm (Vb = −150 V) and 5 nm (Vb = −200 V) of Ti1−x Six N films), and the microstructure changes from strong (1 1 1) preferred orientation to mixed one and finally to highly (2 0 0) orientation with increase in substrate bias voltage. It is understood that surface roughness is related with the size and orientation of TiN crystallite and is well known that the increase of Vb leads to the enhancement of the adatom surface mobility as well as the atomic diffusion, thereby enhancing grain growth, which contributes to the increase of grain size and surface roughness as shown in Fig. 2(a) for undoped TiN film. But the presence of the amorphous matrix of Si3 N4 , suppresses the grain growth of TiN nanocrystals, leading to a finely grained dense microstructure.
Fig. 4. Hardness values of Ti1−x Six N films as a function of negative substrate bias voltage, along with crystallite size variation.
Fig. 4 shows hardness, H, of Ti1−x Six N films as a function of substrate bias (Vb ). The hardness of the films increases with increasing negative bias Vb, and reached a maximum value of 40 GPa at −150 V and dropped again with further increase in Vb . Above dc bias of −150 V, the resputtering of the growing film by bombardment of energetic particles results in large surface roughness and hence a decrease in hardness value. The high hardness value is attributed to the nanocomposite structure described by Veprek [6] that leads to change in grain size, orientation, and increase in density and elimination of voids. All these effects are related to the adatom mobility and the ion–solid interactions near the growing film surface under negative bias. The increase in Vb increases the mobility of adatoms promoting closed packed structure in near thermodynamic equilibrium conditions. Thus, for high adatom mobility Ti1−x Six N films are expected to grow along the (2 0 0) orientation corresponding to the lowest surface free energy [23]. On the other hand, for low adatom mobility the preferred orientation is the (1 1 1) in which the highest number of atoms per unit area can be incorporated at low energy sites. The amounts of grains grown along the (2 0 0) and (1 1 1) orientations play a significant role on the mechanical properties of Ti1−x Six N films. The crystallite size showed a monotonous decrease with increasing Vb , but the maximum hardness was obtained for a Vb = −150 V. We have observed that in addition to the decrease in grain size, the negative bias of the substrate during deposition strongly influences the preferred orientation of the reactively sputtered Ti1−x Six N films. Development of a preferred orientation in non-epitaxial thin films (where the substrate plays a relatively minor role in the texture determining process) is primarily governed by the surface and strain energy considerations. Thermodynamics derives the system towards the minimum possible sum of surface and strain energies under the restrictions imposed by kinetics. In TiN (2 0 0) planes or the square surfaces of the cubic unit cell have the lowest surface energy [23]. The increase in Vb increases the energy of the adatoms to diffuse on the substrate surface. Thus, the adatoms are no longer kinetically constrained to metastable high-energy configurations, but relax to the lowest surface energy configurations of (2 0 0) planes. However due to much higher kinetic energy imparted due to large negative substrate bias (−200 V) leads to resputtering of the deposited film leading to deterioration of surface and hence the mechanical properties.
R. Chandra et al. / Materials Science and Engineering A 423 (2006) 111–115
4. Conclusions Nanocomposite thin films of Ti1−x Six N were deposited on R-plane sapphire substrates using ionised magnetron sputtering. The effect of processing parameters on the microstructure of these films was investigated using XRD and AFM. During deposition of the Ti1−x Six N films, the increase in negative bias voltage Vb , results in a structure change from (1 1 1) orientation with large TiN crystallite to (2 0 0) with a smaller size. The surface morphology of the films is also affected with increase in negative bias Vb . The roughness of the films is found to decrease with increase in Vb upto −150 V and with further increase in bias voltage to −200 V, the surface roughness increases due to resputtering which leads to deterioration of the mechanical properties of Ti1−x Six N films. The best mechanical properties (H = 40 GPa) of the Ti1−x Six N thin films for the present study were obtained at a negative dc substrate bias Vb of −150 V. Acknowledgements One of the authors R. Chandra, would like to acknowledge the University Grants Commission, New Delhi (India) and Association of Commonwealth Universities, UK, for providing Commonwealth Fellowship to visit and work at University of Cambridge, UK. Dr. Nadia Stelmashenko for technical help in getting AFM images is also acknowledged. References [1] J. Musil, Surf. Coat. Technol. 125 (2000) 322. [2] R.A. Andrievski, Mater. Trans. 42 (2001) 1471.
115
[3] P.J. Martin, A. Bendavid, Surf. Coat. Technol. 163–164 (2003) 245. [4] S. Veprek, J. Vac. Sci. Technol. A 17 (1999) 2401. [5] M. Diserens, J. Patscheider, F. Levy, Surf. Coat. Technol. 108–109 (1998) 241–246. [6] S. Veprek, Surf. Coat. Technol. 97 (1997) 15. [7] M. Diserens, J. Patscheider, F. Levy, Surf. Coat. Technol. 121 (1999) 158. [8] J. Musil, J. Vlcek, Surf. Coat. Technol. 142–144 (2001) 557–566. [9] W.J. Meng, X.D. Zhang, B. Shi, J.C. Jiang, L.E. Rehn, P.M. Baldo, R.C. Tillsworth, Surf. Coat. Technol. 163–164 (2003) 251–259. [10] S. Veprek, A. Niederhofer, K. Moto, P. Nesladek, H.-D. Mannling, M. Jilek, Surf. Coat. Technol. 133–134 (2000) 152. [11] J.O. Kim, J.D. Allenbach, P.B. Mirkarimi, S.A. Barnett, Phys. Rev. B 48 (1993) 1726. [12] J. Pelleg, L.Z. Zevin, S. Lungo, Thin Solid Films 197 (1991) 117. [13] P. Patsalas, C. Charitidis, S. Logothetidis, Surf. Coat. Technol. 125 (2000) 335. [14] J.E. Greene, J.E. Sundgren, L. Hultman, I. Petrov, D.B. Bergstrom, Appl. Phys. Lett. 67 (1995) 2928. [15] N. Phinichka, R. Chandra, Z.H. Barber, J. Vac. Sci. Technol. A 22 (2004) 477. [16] B.D. Cullity, Elements of X-ray Diffraction, Addison–Wesley Publishing Co, Reading, MA, 1978. [17] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564. [18] F. Vaz, L. Rebouta, P. Goudeau, J. Pacaud, H. Garem, J.P. Rivere, A. Cavaleiro, E. Alves, Surf. Coat. Technol. 133–134 (2000) 307–313. [19] F. Vaz, L. Rebouta, Ph. Goudeau, T. Girardeau, J.P. Rivere, A. Traverse, Surf. Coat. Technol. 146–147 (2001) 274–279. [20] S. Zhang, D. Sun, Y. Fu, H. Du, Surf. Coat. Technol. 167 (2003) 113–119. [21] X. Hu, Z. Han, G. Li, M. Gu, J. Vac. Sci. Technol. A 20 (2002) 1921. [22] R. Banerjee, R. Chandra, P. Ayyub, Thin Solid Films 405 (2002) 64. [23] L. Hultman, J.E. Sundgren, J.E. Greene, J. Appl. Phys. 66 (1989) 536.