Effect of evaporation rate and substrate temperature on optical, structural, and electrical properties of ZnTe:Sb films deposited by thermal evaporation of Zn, Te, and Sb sources

XRD was used to determine the structures of the films. XRD patterns depicted in Figure 1 show that all samples have a cubic zinc blend structure, with interplanar spacing (d) agreeing with the data given by JCPDC No. 15-0746 for the cubic phase of ZnTe. Six different peaks indexed to (111), (220), (311), (400), (331), and (422) were observed for the ZnTe films with no Sb content. However, some of the peaks nearly disappeared when the films were deposited in the presence of Sb vapor, and the films became more oriented in the (111) direction. The cubic structure of ZnTe films with a preferred (111) orientation has been reported by many authors for ZnTe films deposited using various methods including films in which deposition was by sputtering [23], thermal evaporation [3], pulsed laser deposition [18], brush plated [11], and closed space sublimation [24]. It should be mentioned here that there were no peaks that could be attributed to antimony or antimony compounds, which might be due to the limited amount of Sb. The lattice constant of the films was ∼0.613 nm, which almost matches the value reported for the ZnTe powder. Moreover, there is no noticeable lattice constant difference between ZnTe:Sb and ZnTe because Sb and Te have similar ionic radii of 140 pm and 145 pm, respectively. The average crystallite size (calculated using Scherrer’s formula) of the films deposited at 350°C is 22 ± 2 nm. However, the grain size increases with increasing Sb evaporation rate and reaches 31 nm for Za4. This increase in the grain size is most likely attributed to an increase in the film thickness.

Figure 1

XRD patterns of the prepared films.

The EDX spectra presented in Figure 2 clearly show the presence of Sb in the films. In particular, sample Za4 contained large quantities of Sb. This sample was prepared with a high Sb source temperature at a substrate temperature of 250°C. Small oxygen and Si peaks, mainly from the substrate, were excluded from the calculations. The Zn, Te, and Sb (element %) ratios are listed in Table 2. For films deposited at 250°C, the increase in the Sb ratio was directly proportional to the Sb evaporation rate. However, films deposited at 350°C showed much lower Sb content, which might be due to the re-evaporation of Sb atoms from the surface of the substrate. This result agrees with Barati et al. [22] who reported that no Sb was detected in the films deposited at a substrate temperature of 420°C, even at high Sb evaporation rates. Romeo et al. explained that the high resistivity of films deposited at 300°C is due to re-evaporation of Sb from the substrate surface. In conclusion, a high Sb ratio could be easily achieved at a moderate substrate temperature of 250°C.

Figure 2

EDX spectra of films Za1 to Za4.

The SEM images in Figure 3 show a featureless surface with a pinhole-free, homogeneous, smooth, and uniform distribution. In addition, the films covered the entire substrate surface. Sample Za4, deposited at 250°C with a high Sb evaporation rate, shows some sort of surface randomness with clusters of varied sizes appearing at different points. This could be attributed to the nanostructure with densely packed connected fine grains [13] and the improvement in the crystalline size with doping. The SEM image of film number Za4 shows black flower-like spots, where the ratio of antimony inside these spots was 13.23% which is close to the ratio outside them (12.96 at%).

Figure 3

SEM micrographs of films Za4 (with two resolutions), Zb1, and Zb3.

The samples used for electrical measurements were prepared by cutting a square of 1 cm² of the sample and evaporating gold contacts of equal size. A silver paste was used to connect the wires to the sample. The dark DC conductivity (σ) was measured under vacuum in the temperature range 30–170°C. In this temperature range, DC conductivity can be expressed as [25] σ = σ(0) Exp[−(E _f – E _v )/(kT)]. Here, E _v is the critical energy at which the delocalization of states in the valence band occurs, k is the Boltzmann constant, σ(0) is the conductivity at 1/T = 0, and E _f is the Fermi energy level. The dark conductivity activation energy ( E a ≡ E f − E v ) ({E}_{a}\equiv {E}_{f}-{E}_{v}) was determined by the linear fitting of ln(σ) against 1/kT as exemplified in Figure 4 for sample Za3. The plots of ln(σ) versus 1/kT for all samples are shown in Figure 5. The values of the activation energy along with the room temperature resistivity of all samples are listed in Table 2. The dark conductivity activation energy of the deposited films, with no Sb content, was high (0.68 eV for Za1 and 0.76 eV for Zb1). Such high values are expected for highly stoichiometric ZnTe films [17]. In addition, the films exhibited high resistance in order of 10⁷ Ω.

Figure 4

A plot of ln(σ) against 1/kT along with the linear fit of the data.

Figure 5

A plot of ln(σ) against 1/kT of the films.

Increasing the Sb content led to a noticeable decrease in the activation energy and room-temperature resistivity. The activation energy in our case (0.68 eV) is close to that mentioned in the literature (0.65 eV) [16]. The minimum activation energy and room-temperature resistivity were found in sample Za4, which was deposited at 250°C with a high Sb evaporation rate. Relatively higher activation energies and resistivities were observed for the films deposited at 350°C. The decrease in the activation energy is attributed to the increase in Sb concentration, which led to strong interactions among the impurities, causing the Fermi level to shift closer to the valence band.

Films free of antimony are transparent and have a bright orange color. However, as the Sb ratio increased, the transparency of the films gradually decreased. This was clearly observed for the film deposited at 250°C compared to other films prepared at 350°C, as shown in Figure 6. The film thickness (d), mean square surface roughness (σ), and refractive index (n) were evaluated by fitting the experimental normal transmission (T) data to the following equations [26]: (1) T = A x B − C x cos ( φ ) + D x 2 . T=\frac{Ax}{B-Cx\hspace{.25em}\cos (\varphi )+D{x}^{2}}.

Figure 6

A plot of transmittance against wavelength of all deposited films.

Equation (1) represents the normal transmittance of light through a film deposited on a transparent substrate assuming k ² ≪ n ² , where (2) A = 16 n 2 n s e − 0.5 2 π σ ( 1 − n ) λ 2 , A=16{n}^{2}{n}_{s}{e}^{-0.5{\left(\frac{2\pi \sigma (1\left-n)}{\lambda }\right)}^{2}}, (3) B = ( n + 1 ) 3 ( n + n s 2 ) , B\left={(n+1)}^{3}(n\left+{n}_{s}^{2}), (4) C = 2 ( n 2 − 1 ) ( n 2 − n s 2 ) e − 2 2 π σ n λ 2 , C=2({n}^{2}-1)({n}^{2}-{n}_{s}^{2}){{\rm{e}}}^{-2{\left(\frac{2\pi \sigma n}{\lambda }\right)}^{2}}, (5) D = ( n − 1 ) 3 ( n − n s 2 ) e − 2 2 π σ n λ 2 , D={(n-1)}^{3}(n-{n}_{s}^{2}){{\rm{e}}}^{-2{\left(\frac{2\pi \sigma n}{\lambda }\right)}^{2}}, (6) ϕ = 4 π nd λ , \phi =\frac{4\pi {nd}}{\lambda }, (7) x = e − α d , x={{\rm{e}}}^{-\alpha d}, (8) k = α λ 4 π . k=\frac{\alpha \lambda }{4\pi }.

The symbols in equations (2)–(8) have the following meaning: n _s is the refractive index of the glass substrate, α is the absorption coefficient of the films, λ is the photon wavelength, and k is the extinction coefficient.

The refractive index of the substrate can be expressed as n s = 1 / T s + ( 1 / T s 2 – 1 ) 1 / 2 {n}_{s}=1/{T}_{s}+{(1/{T}_{s}^{2}\mbox{--}1)}^{1/2} , where T _s is the optical transmission of the substrate. The best model that describes the refractive index is the simple classical dispersion relation for a single oscillator centered at a wavelength λ _o [27]. The refractive index in this model can be written as n 2 = 1 + ( n 0 2 – 1 ) λ 2 / ( λ 2 – λ o 2 ) {n}^{2}=1+({n}_{0}^{2}\hspace{.5em}\mbox{--}\hspace{.5em}1){\lambda }^{2}/({\lambda }^{2}\hspace{.5em}\mbox{--}\hspace{.5em}{\lambda }_{o}^{2}) , where n _o is the refractive index at an infinite wavelength. It should be mentioned that this model is suitable for most II–VI semiconductor films. The dependence of the absorption coefficient on wavelength involves many complicated scattering and absorption processes. The absorption coefficients in the transparent and medium absorption regions are very low and can be approximated by a second-order polynomial function of 1/λ. On the other hand, the absorption coefficient near the absorption edge can be expressed by Urbach’s relation [16]. Furthermore, α can be expressed as α = c + g/λ + f/λ ² + Exp(m + h/λ), where c, g, f, m, and h are the fitting parameters. Inserting the expressions for n and α in equation (1) enables one to fit experimental transmittance data to equation (1) in order to reveal the values of d, σ, n _o , and λ _o as shown in Figure 7. The obtained values were used to accurately calculate the absorption coefficient near the absorption edge as α = −ln(x)/d, where the value of x can be calculated as x = { ( C 1 + A / T ) x=\{({C}_{1}+A/T) − [ ( C cos ( φ ) + A / T ) 2 − 4 B D ] 1 / 2 } / ( 2 D ) -{{[}{(C\cos (\varphi )+A/T)}^{2}-4BD]}^{1/2}\}/(2D) .

Figure 7

Transmittance data along with the fitting curve of Zb4.

In the case of an allowed direct transition, α varies with the photon energy (hν) as αhν = A′(hν – E _g)^1/2. This relation is known as Tauc’s direct optical transition model, where A is a constant that depends on the optical transition probability [5] and E _g is the optical energy gap. E _g can be extracted from (αhν) ² against the photon energy by extrapolating the linear portion of (αhν)² versus hν to intercept the x-axis at α = 0, as shown in Figure 8.

Figure 8

A plot of (αhν)² against photon energy (hν).

The film thickness along with the root mean square surface roughness and refractive index as a function of the wavelength and optical energy gap is listed in Table 3. The results showed that at a substrate temperature of 250°C, the film thickness increased with the antimony evaporation rate. This is because the Zn to Te vapor flux ratio is almost 2, which allows the Sb atoms to substitute for the Te deficiency. Furthermore, excess Zn will not be re-evaporated, as in the case of the pure ZnTe films. However, a higher substrate temperature (350°C) caused Sb atoms to re-evaporate faster, leading to a much smaller amount of Sb in the prepared films. This is because there was no observable change in the film thickness and surface roughness, although there was a relatively small change in the refractive indices and optical energy gaps of the films. This increase in the refractive index and decrease in the optical energy gap were mostly due to the doping process, as in the case of silver doping [16] and copper doping [28], the refractive index increases with increases with doping concentration. The conductivity was observed to increase by several orders of magnitude, making the material suitable for applications that require medium conductivity and good optical quality, such as window layers for solar cells and other optical devices. Moreover, films deposited at 250°C showed a significant increase in the refractive index accompanied by a clear narrowing of the optical energy gap, an increase in the absorption region, and a large improvement in the electrical conductivity of the ZnTe films. This makes the material suitable for use as a back contact for CdTe solar cells.