Preparation and Characterization of Coevaporated Cd_1−𝑥Zn_𝑥S Alloy Thin Films

Abstract

Cd_1−𝑥Zn_𝑥S thin films have been prepared by the vacuum coevaporation method. The structural, compositional, and optical properties of Cd_1−𝑥Zn_𝑥S thin films have been investigated using X-ray diffraction, X-ray fluorescence, and optical transmittance spectra. As-deposited Cd_1−𝑥Zn_𝑥S thin films are polycrystalline and show the cubic structure for 𝑥=1 and hexagonal one for 𝑥<1 with the highly preferential orientation. The composition of Cd_1−𝑥Zn_𝑥S thin films determined from Vegard's law and quartz thickness monitors agrees with that determined from the X-ray fluorescence spectra. Optical absorption edge of optical transmittance for Cd_1−𝑥Zn_𝑥S thin films shows a blue shift with the increase of the zinc content. The band gap for Cd_1−𝑥Zn_𝑥S thin films can be tuned nonlinearly with 𝑥 from about 2.38 eV for CdS to 3.74 eV for ZnS. A novel structure for CuInS₂-based solar cells with a Cd_0.4Zn_0.6S layer is proposed in this paper.

1. Introduction

ZnS-based II-VI materials can form ternary alloys, such as Cd_1−𝑥Zn_𝑥S, with a direct fundamental band-gap range from 2.4 to 3.7 eV at room temperature. The energy position of the conduction band is changed primarily by varying the Zn content. Therefore, these materials have potential applications for efficient electron-confined structures, photovoltaic heterojunction, and electro-optical devices.

In photovoltaic system, the replacement of CdS with the higher band-gap Cd_1−𝑥Zn_𝑥S has led to a decrease in window absorption losses and has resulted in an increase in the short-circuit current [1]. This Cd_1−𝑥Zn_𝑥S ternary compound is also a useful buffer material for fabrication of p-n junctions without lattice mismatch [2].

Cd_1−𝑥Zn_𝑥S thin films have been prepared by different techniques, which include evaporation [2–5], chemical bath deposition [1, 6, 7], and organic vapor deposition [8]. However, very little information is available on the ternary Cd_1−𝑥Zn_𝑥S alloy films with the desired composition prepared by the vacuum evaporation technique. Cd_1−𝑥Zn_𝑥S thin films, to our best knowledge, have been prepared by vacuum evaporation involved with three methods, one of which is the thin films being prepared by using a mechanically alloyed mixture of CdS and ZnS [3]. It is obvious that a homogenous mixture of Cd_1−𝑥Zn_𝑥S is difficult to obtain. So Kumar et al. [4] synthesized the fine powders of Cd_1−𝑥Zn_𝑥S by solid-state reaction and then deposited Cd_1−𝑥Zn_𝑥S thin films. Considering the large difference in the vapor pressures of CdS and ZnS, Torres and Gordillo [5] prepared Cd_1−𝑥Zn_𝑥S thin films by evaporation, varying the diameter of the openings of the coaxial chambers which contain the CdS and ZnS precursors. However, homogenous Cd_1−𝑥Zn_𝑥S thin films with desired content are not readily deposited by the evaporation mentioned above. In this work, Cd_1−𝑥Zn_𝑥S(0≤𝑥≤1) thin films have been prepared by using a coevaporation of CdS and ZnS from two horizontal sources, which allows independent control of the temperature and consequently control of the partial vapor pressures of both precursors. The structural, compositional, optical properties have been investigated.

2. Experimental Details

Cd_1−𝑥Zn_𝑥S thin films were deposited by the vacuum coevaporation technique, as described elsewhere [9], to a thickness of ~300 nm at room temperature. The vacuum system had a base pressure of 10⁻⁴ Pa and was partitioned into two parts, one for a CdS powder source (5 N purity) and the other for a ZnS powder source (5 N purity). The CdS and ZnS deposition rates were measured by separate LHC-2 quartz monitors. The deposition rate of ZnS was adjusted to obtain different Zn concentrations in CdS thin films, varying between 0 and 1. The thickness was confirmed precisely by a surface profilometer. X-ray fluorescence (XRF) spectrometry was subsequently used to determine the composition of each sample. The crystallographic phase and lattice constants of each film were obtained by X-ray diffraction (XRD). Measurements of the optical transmittance were made over the wavelength range 200∼800 nm with a data interval of 0.5 nm. Transmittance measurements were also made with the beam at near-normal incidence to the sample to prevent reflected light being returned to the monochromator.

3. Results and Discussion

To accurately determine the composition of the Cd_1−𝑥Zn_𝑥S alloys, X-ray fluorescence measurements of Cd_1−𝑥Zn_𝑥S thin films were also carried out (Figure 1). Table 1 shows composition of the Cd_1−𝑥Zn_𝑥S thin films from XRF using the fundamental parameter (FP) method.

Table 1 

Composition of Cd_1−𝑥Zn_𝑥S thin films.

(a)

(b)

(c)

Figure 1 

Typical XRF of Cd1−𝑥Zn𝑥S(𝑥=0.61) thin films. (a) Cd-𝐾𝛼, (b) S-𝐾𝛼, and (c) Zn-𝐾𝛼.

Figure 2 demonstrates X-ray diffraction spectra of Cd_1−𝑥Zn_𝑥S(0⩽𝑥⩽1) thin films obtained by using Cu 𝐾_𝛼 radiation, where only one diffraction peak with a 2𝜃 value between 26.38 and 28.37⁰  is present. The absence of diffraction peaks associated with CdS and ZnS suggests that those thin films are single-phase materials. The presence of sharp peaks in XRD patterns demonstrates the polycrystalline nature of these Cd_1−𝑥Zn_𝑥S(0⩽𝑥⩽1) thin films, which have a strong preferential orientation. From XRD patterns, one can see that the diffraction angle (2𝜃) shifts towards a slightly higher angle with increasing 𝑥 (Zn content) in the thin films. This is due to the substitution of Zn²⁺ for Cd²⁺, in which different ionic radius of Cd²⁺(0.095nm) and Zn²⁺(0.074nm) [10] leads to the increase of lattice constants of Cd_1−𝑥Zn_𝑥S thin films. Figure 2 also shows that the intensity of XRD peaks decreases at higher concentration of Zn. The structure of Cd_1−𝑥Zn_𝑥S thin films varies continuously from being cubic zincblende material with a preferred orientation of (111) plane at the ZnS end to being hexagonal wurtzite with a preferential orientation of (002) plane when zinc content decreased.

Figure 2 

XRD patterns of as-deposited polycrystalline Cd1−𝑥Zn𝑥S thin films.

Since the lattice constant of Cd_1−𝑥Zn_𝑥S thin films follows Vegard’s law, the composition of Cd_1−𝑥Zn_𝑥S thin films can be determined by𝑥_{Cd1−𝑥Zn𝑥S}=𝑎_{Cd1−𝑥Zn𝑥S}−𝑎_CdS𝑎_ZnS−𝑎_CdS×100%, (1) where 𝑎_{Cd1−𝑥Zn𝑥S},  𝑎_CdS, and 𝑎_ZnS are, respectively, the lattice constants of Cd_1−𝑥Zn_𝑥S,CdS, and ZnS. From the 𝑑 values of the X-ray diffraction spectra shown in Figure 2 the lattice constants can be calculated as follows:𝑎𝑑=(4/3)ℎ²+ℎ𝑘+𝑘²+(𝑎/𝑐)²𝑙². (2)

Therefore, from (1) and (2), one can obtain the composition of Cd_1−𝑥Zn_𝑥S thin films as shown in Table 1. It can be seen that there is a good agreement among the three sets of 𝑥 values. As for the low angle diffraction line the variation of the diffraction angle with the composition is small and the determined film composition is not so accurate. The error is estimated to be about 0.05 in 𝑥. The precision suggests that we can control the composition of Cd_1−𝑥Zn_𝑥S thin films, which are prepared by coevaporation of CdS and ZnS. In addition, the comparison of calculated 𝑑 values from the diffraction spectra (Figure 2) with the standard 𝑑^∗ values from the JCPDS XRD spectra data for Cd_1−𝑥Zn_𝑥S is shown in Table 2.

Table 2 

XRD results for Cd_1−𝑥Zn_𝑥S thin films.

The optical transmittance spectra of the Cd_1−𝑥Zn_𝑥S films with different 𝑥 values were obtained, and steep absorption edges were observed, demonstrating that the thin films are homogeneous (Figure 3). The optical transmittance of the Cd_1−𝑥Zn_𝑥S films is typically 60% at wavelengths beyond the absorption edge. It can be seen that Cd_1−𝑥Zn_𝑥S films show the blue shift in the absorption edge with the increase of the zinc content.

Figure 3 

UV-visible transmission spectra of Cd1−𝑥Zn𝑥S films.

From the transmittance curves of Figure 3, the absorption coefficient 𝛼 of the thin films is estimated. The absorption coefficient 𝛼 with the energy of the photon ℎ𝜐 could be expressed by the following equation: 𝛼ℎ𝜐∝ℎ𝜐−𝐸_𝑔^𝑚, (3) where 𝐸_𝑔 is the optical band gap of the thin films, and the exponent 𝑚 may take values, 1,2,3,1/2,3/2, depending on the electronic transitions in 𝑘-space. A linear fit is achieved for the Cd_1−𝑥Zn_𝑥S films with 𝑚=1/2, indicating that this material is a direct gap semiconductor. So (𝛼ℎ𝜐)² is plotted as a function of ℎ𝜐 for various zinc content (not shown there). The optical band energy gap can be determined from the intercept of the extrapolation of the straight portion to the ℎ𝜐 axis.

The band-gap dependence on composition is shown in Figure 4. It can be seen that the band gap varies with Zn content in a nonlinear way. This dependence can be determined by fitting the band-gap values to a parabolic form on the method of least square [11] and described as 𝐸_𝑔(𝑥)=2.38+0.69𝑥+0.67𝑥². (4)

Figure 4 

The variation of energy gap of the Cd1−𝑥Zn𝑥S as a function of 𝑥.

It gives the bowing parameter (𝑏) of 0.67 eV, which is much smaller than that reported on another Cd_1−𝑥Zn_𝑥S alloy system in thin films by Borse et al. [6] (𝑏=0.85eV) and that in single crystal by Muhamad [12] (𝑏=0.82eV), but agrees well with that in thin films by Yamaguchi et al. [1] (𝑏=0.724 eV). In the present case of Cd_1−𝑥Zn_𝑥S thin films, 𝐸_𝑔 values vary from 2.38 to 3.74 eV for 𝑥=0 to 1 in a nonlinear way. This change in the band gap of CdS by addition of Zn shows formation of a continuous series of solid solutions. Therefore, the band energy gap of Cd_1−𝑥Zn_𝑥S can be adjusted in the range of the binary band gaps. And the energy position of the conduction band, that is, electron affinity 𝜒, can be also adjusted by varying the zinc content [2].

For the photovoltaic solar cells, CuInS₂ has a desirable direct band gap of 1.5 eV, which is also well-matched with the solar spectrum. Efficiency of 11.4% for CuInS₂-based solar cells has been successfully made [13]. However, further improvements on performance of CuInS₂-based solar cells have been restricted by the conventional structure due to the lattice mismatch between ZnO and CuInS₂. Therefore, an approach to overcome this problem is to introduce an effective buffer layer, Cd_1−𝑥Zn_𝑥S, in CuInS₂-based solar cells.

Furthermore, according to Anderson’s model, the band alignments with 𝜒_absorber⩽𝜒_buffer⩽𝜒_window are desired to prevent the formation of a conduction band spike in the CuInS₂-based solar cells. For Cu-rich films, electron affinity for CuInS₂ thin films is about 4.1 eV (𝐸_𝑔∼1.53 eV) [14, 15], and for ZnO thin films is about 4.2 eV (𝐸_𝑔∼3.2 eV), so the electron affinity of 4.1 eV for Cd_1−𝑥Zn_𝑥S thin films (i.e., 𝑥=0.6) [2] is taken into account. Based on these mentioned above, we propose a modified structure of CuInS₂-based solar cells. The energy band diagram of CuInS₂-based solar cells is shown in Figure 5. Cd_0.4Zn_0.6S thin films (𝐸_𝑔∼3.03 eV, from (4)) provide lattice constant matching to the absorber and reduce the interface state density between the 𝑝- and 𝑛-type materials, meanwhile, the conduction band and valence band spikes are eliminated too.

Figure 5 

Flat band diagram of a CuInS2/ZnO heterojunction with a Cd0.4Zn0.6S buffer layer.

4. Conclusions

Cd_1−𝑥Zn_𝑥S thin films with the required composition have been successfully deposited on glass substrates by the vacuum coevaporation method. The thin films are found to be cubic at the ZnS end and hexagonal for 𝑥<1, which show the highly preferred orientation. The measurements of optical transmittance for Cd_1−𝑥Zn_𝑥S films demonstrate the blue shift in the absorption edge, and the optical band gap increases from 2.38 to 3.74 eV in a nonlinear way as composition varies from 𝑥=0 to 1. And the energy position of the conduction band can be tuned by varying the zinc content. Therefore, a buffer layer Cd_1−𝑥Zn_𝑥S for 𝑥=0.6 is introduced in the CuInS₂-based solar cells to eliminate band spikes and provide lattice matching to the absorber.

Acknowledgments

Financial support from the National Basic Research Program of China (Grant no. 2011CBA007008), National Natural Science Foundation of China (Grant no. 61076058), and the Science and Technology Program of Sichuan Province, China (Grant no. 2008GZ0027) is gratefully acknowledged. The authors thank Ms. Huiqin Sun, Analytical and Testing Center, Sichuan University, for optical transmittance spectra measurements.