A. Structure

Figure 1 illustrates the XRD patterns of Ti films, which are deposited under different conditions. It is found from the figure that the films exhibit crystalline structure with only α-Ti phase and hcp structure (JCPDS 44-1294). Actually, the diffraction patterns show that the films exhibit characteristic peaks of α-Ti crystal planes, i.e., (002), (100), (101), (102), (110), and (103). Also, strong (002) preferential orientation in the Ti films could be seen, which indicates crystalline structure with a preferential orientation along “c” axis. Furthermore, the 2θ position of the 002 peak in all the Ti films is higher than the value of 38.42° reported for the bulk Ti.^{Reference Sailor and McCarthy19} This shift in the peak position is due to the uniform distribution of the internal stress in the Ti films.^{Reference Cullity20} Such a relative shift in the 002 peak position ranges from 38.70° to 39.47° for different deposition conditions.

FIG. 1.

X-ray diffraction patterns of 2.4 µm Ti films deposited onto FTO glass substrate as a function of (a) substrate temperature and deposition rate (deposited at fixed Ar pressure of 0.8 Pa) and (b) substrate temperature and Ar pressure (deposited at fixed deposition rate of 66 nm/min) as tabulated in Table I.

Figure 1(a) shows the XRD patterns of Ti films deposited at a fixed Ar gas pressure of 0.8 Pa, but at different substrate temperatures (RT and 250 °C) and different deposition rates (66 and 90 nm/min). Table II shows the calculated values of lattice constants a and c for the films, which are in good agreement with those reported for a bulk α-Ti phase crystal.^{Reference Andersson, Collen, Kuylenstierna, Magneli, Magnéli, Pestmalis and Åsbrink21} By comparing the XRD pattern of the samples shown in Fig. 1(a), it can be found that from sample T1 to T3 (i.e., with increasing the substrate temperature, see Table I) the position of 002 peak is shifted to higher angles (from 38.80° to 39.12°). The same result keeps when we increase the deposition rate (from T1 to T2). We notice that the intensity ratio of (002) preferred orientation becomes stronger by increasing substrate temperature and/or increasing deposition rate [see Fig. 1(a)]. This could be due to the increase in the energy of incident atoms which can improve the crystal growth. Such a behavior is also reported by other research groups.^{Reference Sonoda, Watazu, Zhu, Shi, Kato and Asahina22,Reference Zhang, Zhong, Zhou and Sun23} Also, Fig. 1(b) shows the XRD patterns of Ti films deposited at the fixed deposition rate of 66 nm/min but at different substrate temperatures (RT and 250 °C) and also at different Ar gas pressures (0.8 and 1.6 Pa). It is clear from Fig. 1(b) that the position of 002 peak is shifted to lower angles (from 38.80° to 38.15°) by increasing the Ar pressure (from 0.8 to 1.6 Pa) indicating an increase in the lattice constant (see Table II). Actually, by increasing the Ar pressure (from sample T1 to T5 and from sample T2 to T6) the lattice constant increases (from 0.46530 to 0.46732 nm and from 0.46115 to 0.46363 nm, respectively). Furthermore, at lower substrate temperature, the intensity ratio of the refractive peaks for lower pressure is stronger than that of the higher pressure samples. Obviously, this could be due to the increase in the Ar pressure decreases the energy of incident atoms evaporated at low substrate temperatures, and therefore, leads to a reduction in the crystal growth of the film.^{Reference Klabunde, Löhmann, Bläsing and Drüsedau24,Reference Sasaki and Nafarizal25} In contrast, there is no clear change in the peak intensity ratio with increasing the Ar pressure at higher substrate temperature. One also can easily distinguish the peaks associated with SnO₂ (JCPDS 46-1088), component of the FTO glass substrate, where are marked by filled circles in the figure. Beyond these extra peaks, the spectra show no characteristic peaks of any impurities or other phases of Ti.

TABLE II.

Lattice constant, RMS roughness, FWHM, grain size, roughness exponent, correlation length, and mean fractional volume of Ti deposited films at different sputtering conditions.

Bulk values of lattice constants; a = 0.29505 nm, c = 0.46826 nm.

To assess the crystalline quality of the deposited films, the crystalline grain size (d) is evaluated according to the Williamson–Hall's formula^{Reference Williamson and Hall26}:

(1)$${{\beta {\rm{cos}}\,\theta } \over \lambda } = {1 \over d} + \eta {{{\rm{sin}}\,\theta } \over \lambda }\quad ,$$

where β is the full width at half maximum (FWHM) of XRD peaks (in radians) (Table II), λ is the x-ray wave length (Cu K_α = 0.154 nm), θ is the Bragg diffraction angle, η is the lattice stress in the sample, and d is the average grain size which is the unknown parameter and should be calculated from this formula. To determine the effect of sputtering conditions on the crystalline grain size, the Williamson–Hall plot is performed and shown in Fig. 2. The solid lines in the figure present the linear fits to the data; therefore, the grain size values can be obtained from the plot through the intercept of the fitted lines with the vertical axis. The estimated grain sizes are summarized in Table II. According to these results, the deposited Ti films onto FTO glass contain nanocrystalline grains of about 48–163 nm. The film deposited at 250 °C, 66 nm/min, and 1.6 Pa (sample T6) has the minimum FWHM (0.014°) and the maximum d (163.93 nm) and is expected to possess the best crystalline quality. As Fig. 2 shows, sample T6 has the strongest (002) diffraction peak, that is, the stronger the (002) diffraction, the better the crystal quality.

FIG. 2.

Williamson–Hall plot to determine the crystalline size of Ti films deposited at different (a) deposition rate and substrate temperature (deposited at fixed Ar pressure of 0.8 Pa) and (b) substrate temperature and Ar gas pressure (deposited at fixed deposition rate of 66 nm/min).

Moreover, it can be observed that the substrate temperature and the deposition rate have direct effect on the grain size, which is in agreement with the reports of Refs. Reference Thornton13, Reference Liu and Yang27, and Reference Savaloni and Kangarloo28. It can also be concluded from Table II that the crystalline grain size monotonically increases with temperature and/or deposition rate. Such a dependence of the grain size on the substrate temperature can be consistently described by the activation energy model proposed by Grovenor et al.^{Reference Grovenor, Hentzell and Smith29} However, for the films deposited at the constant temperature, an increment of the deposition rate from 66 to 90 nm/min leads to the strong coalescence of the neighboring grains which in turn results in larger grains.

From Table II, it can be also seen that for the samples deposited at 0.8 Pa (T1 to T4) the estimated grain size ranges from 79 to 96 nm. These values are approximately half of the grain size values (ranging from 104 to 163 nm) for T5 and T6 which are deposited at 1.6 Pa. Therefore, one may concluded that the grain size increases linearly with the Ar pressure.

B. Surface morphology

Figure 3(a) shows a cross-sectional view of a Ti film deposited at RT, 66 nm/min, and 0.8 Pa Ar atmosphere onto FTO glass substrate (sample T1). As the image shows the Ti film thickness is about 2.4 µm. Furthermore, void free microstructure with uniform and compact surface morphology without peeling and cracking with columnar structure can be observed in the film. The result of the chemical analysis by EDX spectroscopy [Fig. 3(b)] indicates that an almost pure Ti film is formed with a low oxygen contamination level (below 2%). Also, the EDX spectra reveal that the FTO glass substrate is composed of Sn and O, as it is clearly confirmed by the XRD patterns [Figs. 1(a) and 1(b)].

FIG. 3.

(a) Cross-sectional FESEM image of Ti thin films deposited at room temperature, 66 nm/min, and 0.8 Pa Ar atmosphere and the corresponding (b) EDX.

The surface morphology and the grain structure of the Ti thin film deposited onto FTO glass substrate at different sputtering conditions are characterized by AFM images (Fig. 4). The figure shows that the film is composed of grains with diameter of about 75–95 nm, which confirms the nanocrystalline nature of Ti films. The three-dimensional AFM images accompanying line profiles of the studied samples are shown in Fig. 5. The studied surfaces are equal to 5 µm × 5 µm in area. The 3D images of Ti films clearly indicate that the morphology is relatively uniform for all films. To obtain more information about the RMS surface roughness, in this study we performed a particle analysis. The values of RMS surface roughness of Ti films deposited at various substrate temperatures, deposition rates, and Ar gas pressures are computed by surface profile length (Fig. 5) and given in Table II. As the table shows, sample T3 deposited at 250 °C, 66 nm/min, and 0.8 Pa has smoother surface compared to the other samples. It is concluded that smoother films can be obtained at higher substrate temperatures and/or lower deposition rates and/or lower Ar pressures. Moreover, the quantitative information of the surface morphology can be extracted from the height–height correlation function H(r), defined as^{Reference Qi, Huang, Tang, Cheng, Shao and Fan30}:

(2)$$H\left( r \right) = {1 \over {N\left( {N - n} \right)}}\mathop\sum\limits_{j = 1}^N {\mathop\sum\limits_{k = 1}^{N - n} {{{\left[ {h\left( {k + n,j} \right) - h\left( {k,j} \right)} \right]}^2}} } \quad .$$

FIG. 4.

Grain structure of Ti thin films deposited at different sputtering conditions studied by AFM.

FIG. 5.

3D AFM images (the scanned area 5.0 µm × 5.0 µm) and accompanying line profiles of Ti thin films deposited at different sputtering conditions.

Here h(k,j) is the height of the measured surface by AFM at the point (k,j), N × N the total number of points at which the surface heights have been measured. For a self-affine surface, H(r) scales as^{Reference Als-Nielsen and McMorrow31}:

(3)$$H\left( r \right) = 2{\sigma ^2}\left\{ {1 - \exp \left[ { - {{\left( {{r \over \xi }} \right)}^{2\alpha }}} \right]} \right\}\quad ,$$

where σ is the RMS surface roughness, α is the roughness exponent, and ξ is the correlation length. When r is smaller than ξ, H(r) scales as H(r) ∼ r ^2α, while for r larger than ξ, H(r) approximately saturates to a constant value. Figure 6 shows the height–height correlation functions in the log–log scale for the samples. The roughness exponents α can be extracted from a linear fit to H(r) in the short-range part and are listed in Table II. With roughness exponent α and RMS surface roughness σ, the correlation length ξ can be obtained by fitting the curve of H(r) with Eq. (3) (Table II). These three surface parameters σ, α, and ξ are independent from each other and completely characterize roughness of a surface. In fact, when the Ti atoms approach the substrate surface, three major phenomena contribute in the depositing process: shadowing effect, surface diffusion effect, and grain growth. The competition between these three processes will determine the roughness exponent. Surface diffusion effect, which tends to smooth the surface, is the only process responsible for lowering the exponent, while the shadowing effect and oriented grain growth destabilize surface smoothness during the film deposition. Therefore, they will increase the roughness exponent.^{Reference Zhang, Ma and Zhang32}

FIG. 6.

Height–height correlation function H(r) plot to determine roughness exponent α and correlation length ξ of Ti films deposited at different (a) deposition rate and substrate temperature (deposited at fixed Ar pressure of 0.8 Pa) and (b) substrate temperature and Ar gas pressure (deposited at fixed deposition rate of 66 nm/min).

As Table II shows, from T1 to T3 (from T1 to T5) by increasing the substrate temperature (increasing Ar pressure), the roughness exponent decreases from 0.90 to 0.77 (from 0.90 to 0.84). For Ti film deposited at RT, the larger roughness exponent α of 0.90 can be explained based on the shadowing effect mechanism. Shadowing is a phenomenon arising from the geometric constraints imposed by the roughness of the growing film which leads to a rough surface with columnar growth structure.^{Reference Wang and Zhang33} With increasing the substrate temperature from RT to 250 °C, the Ti atoms will obtain higher energy. Therefore, the surface diffusion effect is dominant, yielding a relatively smoother surface with lower RMS surface roughness values and roughness exponents. On the other hand, increasing deposition rate from 66 to 90 nm/min at low Ar pressure (from T1 to T2 and/or T3 to T4) leads to a 1.1–3.1% increase in the roughness exponent. While this value is a 16.3% increase in the roughness exponent at high Ar pressure (1.6 Pa). In the films deposited at high deposition rate and high Ar gas pressure, the nucleation and fast grain growth overcome the surface diffusion of Ti atoms, and therefore, a higher roughness exponent can be achieved.

The correlation length can be interpreted as the average distance between two defects, or dislocations, or any other kind of imperfections in the crystal. Thus, ξ should reflect the crystalline quality.^{Reference Ganguli and Ingale34} In other words, a higher defect or dislocation density corresponds to a smaller value of the correlation length and vice versa. The relatively high values of ξ ∼ 4.2 × 10³–5.7 × 10³ nm for all the samples indicate the low density of defects or dislocations existing in the deposited Ti films. This is also confirmed by the XRD spectrum with a remarkably high intensity 002 peak [Figs. 1(a) and 1(b)] and the crystalline grain size (Table II). As Figs. 1(a) and 1(b) show sample T6 has the highest (002) peak intensity and mean while from Table II we notice that the highest value of crystalline grain size and correlation length is obtained for this sample.

Finally, considering both the XRD and AFM results, it is obvious that the mean diameter of the grains and RMS surface roughness is sensitive to substrate temperature, deposition rate, and Ar gas pressure. The influence of these deposition parameters on the mean grain size and the roughness characterization can be related to the adatom mobility during film deposition. It means that at higher substrate temperatures, surface adatoms have enough energy to migrate on the substrate, so the intergrain voids begin to be filled in and the structure transforms into a transition zone of tightly packed fibrous grains with smoother surface.^{Reference Chinmulgund, Inturi and Barnard12} Also, samples deposited with the higher deposition rate, show larger spherical grains with a relatively rough surface (see Table II). In fact, increasing the deposition rate leads to the strong coalescence of the neighboring grains, which in turn results in larger grains and rougher surface morphology.^{Reference Thornton13} Moreover, comparing samples deposited at different inert gas pressures indicates that increasing the Ar pressure increases the grain size and RMS surface roughness. Actually, increment in the Ar pressure leads to more numerous collisions between atoms on the film and ambient gas atoms, which accompanied with the adatom surface mobility, results in a greater roughness for the resulting surface.^{Reference Barnes, Kelly, MacKay, Mateeva, O'Brien and Lagally35} In principle, one controls the grain size and RMS surface roughness of the films by controlling the Ar pressure during deposition process.

C. Optical properties

In this subsection, we investigate the effect of sputtering conditions on the optical properties of the Ti films. The reflectivity of the films versus photon energy is shown in Fig. 7. As the figure shows that the reflectivity of all the films increases with decreasing photon energy. From Fig. 7(a), it can be seen that sample T4 (deposited at substrate temperature of 250 °C and deposition rate of 90 nm/min) has the highest reflectivity between the samples deposited at fixed Ar pressure of 0.8 Pa. In Fig. 7(b), we compare the samples deposited at fixed deposition rate of 66 nm/min. The figure indicates that sample T5 (deposited at substrate temperature of 250 °C, and Ar pressure of 1.6 Pa) has the highest value of the reflectivity compared to the samples deposited at RT and Ar pressure of 0.8 Pa. Therefore, it can be concluded that the decrement of the reflectivity for the samples deposited at low substrate temperature is related to the loss scattering which are the highest for these samples because of their high roughness (see Table II and related discussion).^{Reference Bennett, Bennett and Nagel36} To achieve more information about the effect of the deposition conditions and morphology on the optical properties of the samples, their refractive index, $\hat N\left( E \right) = n\left( E \right) - i\kappa \left( E \right)$, can be calculated from their reflectivity spectrum. In fact, the real and imaginary parts of the refractive index are evaluated by applying the Kramers–Kronig relations to the reflectance data.^{Reference Lucarini, Saarinen, Peiponen, Vartiainen and Rhodes37}

FIG. 7.

Reflectivity of Ti films deposited at different (a) deposition rate and substrate temperature (deposited at fixed Ar pressure of 0.8 Pa) and (b) substrate temperature and Ar gas pressure (deposited at fixed deposition rate of 66 nm/min).

Figure 8 shows the real and imaginary parts of the film refractive index versus photon energy for the different sputtering conditions. As the Figs. 8(a) and 8(b) represent the real part of the refractive index, n increases with increasing photon energy. Increasing the deposition rate (from sample T1 to T2, deposited at RT and 0.6 Pa) the real part of refractive index increases from 0.35 to 0.6 at 1.5 eV and from 0.78 to 0.95 at 5 eV [Fig. 8(a)]. Also, as the Fig. 8(b) shows, increasing the Ar pressure (from sample T1 to T5, deposited at RT and 60 nm/min) the real part of refractive index increases from 0.35 to 1.3 at 1.5 eV and from 0.8 to 0.9 at 5 eV. In the low photon energy region (below 2.0 eV), increasing the substrate temperature at constant Ar pressure and deposition rate increases the real part of refractive index, for instance, increasing the substrate temperature from sample T1 to T3 increases the real part of refractive index from 0.13 to 0.5. Whereas, in the high photon energy region (above 4 eV), increasing the substrate temperature at constant Ar pressure and deposition rate shows no clear effect [Figs. 8(a) and 8(b)].

FIG. 8.

Real and imaginary parts of refractive index of Ti films deposited at different (a) deposition rate and substrate temperature (deposited at fixed Ar pressure of 0.8 Pa) and (b) substrate temperature and Ar gas pressure (deposited at fixed deposition rate of 66 nm/min).

The relationship between deposition conditions and the extinction coefficient κ is also shown in Figs. 8(a) and 8(b). Ti films deposited at RT, 66 nm/min, and/or lower Ar pressure represent lower extinction coefficient than the films deposited at 250 °C, higher deposition rate, and/or higher Ar pressure [Figs. 8(a) and 8(b)]. In addition, by increasing the substrate temperature, an increase in the refractive index can be observed, as shown in Figs. 9(a) and 9(b). Any increase in the refractive index with increasing the substrate temperature attributes to the decreased surface roughness, which in turn decreases the surface optical scattering and optical loss.^{Reference Yoon, Kim, Kim, Lee and Yoon38–Reference Lai, Li, Wang, Hu, Wang, Hou, Song and Jiang40} It can be seen that the refractive index of the samples slightly increases with the increase of the deposition rate [see Fig. 9(a)]. We argue that the denser structure (see below for further discussion) is a reason for larger refractive index of the film deposited at higher deposition rate.

FIG. 9.

Refractive index of Ti films deposited at different (a) deposition rate and substrate temperature (deposited at fixed Ar pressure of 0.8 Pa) and (b) substrate temperature and Ar gas pressure (deposited at fixed deposition rate of 66 nm/min).

Moreover, it can be extracted from the Figs. 8(a) and 8(b) that the real and imaginary parts of the refractive index increase with any increment in the Ar pressure. This is also true for the total refractive index of the films [Fig. 9(b)]. Actually, at low sputtering gas pressures, more intensive collisions between the particles occur, leading to the formation of looser films. Hence, the Ti films deposited at lower Ar pressure are less dense than the films deposited at the higher pressure. This finding is in coincidence with the other reported results^{Reference Muraishi, Aizawa and Kuwahara16,Reference Chawla, Jayaganthan, Chawla and Chandra17,Reference Thornton41} where by increasing the substrate temperature, deposition rate, and/or Ar pressure an increase in the refractive index was observed. We use Maxwell–Garnett (MG) theory^{Reference Garnett42} to further clarify the effect of sputtering parameters and the morphology on the optical properties of the Ti film deposited onto FTO glass substrate. In the MG theory, the dielectric function $\hat \varepsilon = {\varepsilon _1} + i{\varepsilon _2}$ (where ε₁ = n ² + κ² and ε₂ = 2nκ) of the aggregated system is related to the dielectric function $\hat \xi$ of the metal, by the following relation^{Reference Garnett42}:

(4)$${{\hat \varepsilon - 1} \over {\hat \varepsilon + 2}} = q{{\hat \xi - 1} \over {\hat \xi + 2}}\quad ,$$

where q is the fractional volume of the metal in the medium and changes between 0 and 1. When q is equal to 1, it means that the system is a bulk metal and as the q decreases, the condensation of the metal in the medium decreases.^{Reference Lee, Chang, Ogi, Iskandar and Okuyama43} In the other words, the metal's porosity is defined as the pore space in the metal and can be calculated by 1 − q. Measuring the optical properties and calculating the values of q give useful information about roughness, condensation, and aggregation of the films. The values of q are calculated for the samples with the different sputtering conditions and plotted versus the photon energy in Fig. 10. Also, the mean value of q is tabulated in Table II. As Fig. 10 and Table II show, the value of q depends on the deposition conditions and morphology of the films. By increasing the substrate temperature from RT to 250 °C and/or increasing the deposition rate from 66 to 90 nm/min as well as increasing the Ar pressure from 0.8 to 1.6 Pa, the q value increases (Table II). This is in agreement with the argument that by increasing the substrate temperature, the deposition rate and/or sputtering pressure the separation of the metallic grains decreases and leads to reduction of the voids density or porosity in the deposited films.^{Reference Liu and Yang27,Reference Thornton41} It can be concluded that this increment in the q value (decrement in the porosity) is another reason for increasing the refractive index value by increasing substrate temperature, deposition rate, and/or sputtering gas pressure.

FIG. 10.

Fractional volume (q) of Ti films deposited at different sputtering conditions as listed in Table I.