Abstract
Zinc oxide (ZnO) thin films were grown by nonreactive RF sputtering at room temperature under varying argon pressures (). Their optical band gap was found to increase from 3.58 to 4.34 eV when the argon pressure increases from 2.67 to 10.66 Pa. After annealing at 200°C and 500°C, optical band gaps decrease considerably. The observed widening of the band gap with increasing can be understood as being a consequence of the poorer crystallinity of films grown at higher pressures. Measurements of morphological and electrical properties of these films correlate well with this picture. Our main aim is to understand the effects of on several physical properties of the films, and most importantly on its optical band gap.
1. Introduction
ZnO is a wide-band-gap semiconductor with a high transparency in the whole visible range and electrical properties that can be tailored from insulating to semimetallic by doping. This has attracted strong interest in this material because of its potential for applications, which include transparent conducting electrodes, gas sensors, light emitting devices, laser diodes, and optical waveguides [1, 2].
ZnO thin films have been obtained by techniques such as sol-gel [3, 4], metal organic chemical vapor deposition [5, 6], pulsed laser deposition [7, 8], and sputtering [9, 10]. As compared to other deposition methods, sputtering has several advantages. Uniformity of film thickness over large areas, a high degree of film adhesion, and relatively simple scalability properties are some of the most important advantages. Several recent works [11–14] discuss the mechanisms involved in ZnO thin film growth by magnetron sputtering in a mixture of argon and oxygen, that is, by reactive sputtering, at varying Ar/O2 ratios. This technique produces films with a fixed optical band gap, in the absence of doping.
Most applications of ZnO as a semiconductor, on the other hand, require the tailoring of the band gap. Band gap modulation of ZnO is usually obtained by doping with Cd and Mg ions [15]. A significant widening of the band gap, reaching values as high as 5.23 eV, has been reported for ZnMgO films [16, 17]. However, doping may sometimes be undesirable, for example, because this may compromise the crystallographic properties of the film [17]. Therefore, there is a strong interest in exploring different approaches to control the band gap in ZnO films.
An approach to control band gap that has not been widely explored up to now consists in depositing ZnO from a ceramic target, without the addition of O2, under varying Ar pressures. This procedure, as we report here, allows one to modify the band gap without doping. In this work, we show that the optical band gap of ZnO films can be tuned from 3.58 to 4.34 eV, when depositing by the above mentioned procedure with Ar pressures is in the range from 2.67 to 10.66 Pa.
As-deposited films were polycrystalline with low intensity diffraction peaks corresponding to the wurtzite hexagonal phase of ZnO. After annealing at temperatures of 200°C and 500°C in air, the polycrystalline films became c-orientated with high intensity (002) diffraction peak. The surfaces of as-grown films were smooth, with an average roughness of 12 Å. Their morphology changed drastically after crystallization, showing average roughness ranging from 40 to 60 Å. All films had transmittance values above 90% in the whole visible range. While resistivities of as-grown films were larger than 107 Ω cm at 302 K, these were found to range from 103 Ω cm ( = 2.67 Pa) to 106 Ω cm ( = 10.66 Pa) after annealing at 200°C and from 10−1 Ω cm to 101 Ω cm after annealing at 500°C.
2. Experimental Details
Zinc oxide thin films were grown on Corning glass 2947 substrates by RF magnetron sputtering technique. All films were prepared using an INTERCOVAMEX-V3 sputtering system with a Lesker’s Torus magnetron gun connected to a 600 watt advanced energy RF power. All films were grown at room temperature, using a 3 in. diameter ZnO target (Kurt J. Lesker, 99.9% purity). An on-axis configuration was used for the sputtering process. The distance between the target and the substrate was kept at 2.5 cm. The vacuum chamber was evacuated by a turbo molecular pump to a base pressure lower than 10−4 Pa before deposition. The target was then presputtered during 10 minutes to remove any contamination from its surface. Argon pressures ranged from 2.67 to 10.66 Pa, in steps of 2.67 Pa. The RF power was 100 W and growth time was 10 min in all cases.
Film morphology was determined with the help of a Nanosurf atomic force microscopy (AFM) and using a NanoWorld silicon probe in contact mode (constant force 0.18 N/m, resonance frequency 14 kHz). The mean roughness values reported in this work are arithmetic averages of the absolute surface height measured from the mean plane. These were provided by the Easy-Scan AFM analysis software. Structural properties of films were obtained by X-ray diffraction (XRD), using a Siemens D500 diffractometer equipped with Cu Kα radiation (λ = 0.154 nm). Film thicknesses were measured using a profiler Veeco Dektak 8 Stylus. Transmittance spectra were recorded with the help of an Agilent 8453 UV-Vis spectrophotometer. Resistivity values were obtained with an Ecopia HMS-5000 Van der Pauw measurement system at 302 K.
3. Results and Discussion
3.1. Structure and Morphology
Figure 1 shows XRD spectra for as-grown films and after thermal annealing. As-grown films were polycrystalline with low intensity (100), (101), and (110) diffraction peaks corresponding to hexagonal wurtzite structure of ZnO (JCPDS 36-1451) [18]. After annealing at 200°C and 500°C during 2 h in air, a preferred orientation along (002) plane was evident, indicating an improved crystallinity orientation of the ZnO films along the c-axis. The intensity of the (002) peak increased, and its 2θ value shifted toward a higher angle, with increasing annealing temperature. As shown in Table 1, the 2θ values of (002) peaks for films annealed at 200°C are slightly smaller than 2θ = 34.4220, as reported for JCPDs 36-1451 [18]. This suggests the existence of compressive strain in c-axis direction. This residual stress was released after annealing at 500°C, as evidenced by measured values of 2θ of around 34.42. During the annealing process, the atoms of ZnO films gain more kinetic energy and their vibration amplitude grows. This kinetic energy of the atoms allows them to rearrange in the crystal lattice releasing the compressive stress. Similar results have been reported in the literature previously [19, 20].
Crystallite sizes were estimated using Scherrer’s equation [21], considering the FWHM and 2θ values of (002) peaks, see Table 1. As seen in Table 1, crystallite sizes were independent of at a fixed annealing temperature. However, the average crystallite size increased from 12 nm for films annealed at 200°C to 14.2 nm after annealing at 500°C, as a consequence of the Ostwald ripening process where atoms migrate forming larger crystallites.
Table 1 shows that deposition rates decrease when increases from 2.67 Pa to 10.66 Pa. These variations can be attributed to changes in the mean-free path of the sputtered species in the background argon gas. As increases, the sputtered species undergoes a higher number of collisions; hence, fewer particles arrive at the substrate with the subsequent decrease in deposition rates [22, 23].
As-grown films had smooth surfaces, with an average roughness of 12 Å. Figure 2(a) shows an AFM image of a typical as-deposited film. A dense film structure, without pores or defects, can be observed. After annealing at 200°C and 500°C in air, sample morphology changed significantly due to crystallization; see Figures 2(b) and 2(c). Average roughness increased to 38 Å after annealing at 200°C and to 59 Å after annealing at 500°C; see Figure 3. AFM images showed that these increases are to be correlated with the appearance of what resembled a hill-valley structure throughout the entire film surface. The observed surface roughening after the annealing process can be ascribed to the Ostwald ripening process, where atoms migrate to reduce surface energy forming larger grains, as can be seen in AFM images.
(a)
(b)
(c)
3.2. Optical and Electrical Properties
Figure 4 shows the wavelength dependence of transmittance for films grown with ranging from 2.67 to 10.66 Pa. All transmittances were larger than 90% in the visible range. Notice that the absorption edge shifts towards shorter wavelengths with increasing .
In Figure 4, an increasing tail is observed for λ < 400 nm when increases. Similar tails in the transmittance spectra of ZnO films have been observed previously [6, 24, 25]. These tails have been correlated with poor crystallinity of films [26]. A possible explanation for our observation of increasing tail regions for increasing (see Figure 4) is as follows. Our experiment was conducted in a high (>2 Pa) regime, where transport is mostly diffusive [27]. Hence, sputtered species were thermalized by collision before reaching the substrate. Therefore, incident atoms had little energy available to diffuse across the surface, leading to poor crystallinity in the growing film.
After annealing, the average transmittance of ZnO films did not change significantly, but the optical absorption edge shifted to longer wavelengths (see Figure 5). It is known [28, 29] that oxygen vacancies modify the absorption edge for ZnO films. In our study, it is reasonable to assume that the density of oxygen vacancies increased for increasing annealing temperature. This provides a possible explanation for the observed shifts.
Optical band gaps were determined by extrapolation from a linear fit of absorption coefficients in the high-energy region. Assuming parabolic valence and conduction bands, the absorption coefficient of a semiconductor near the absorption edge is given by [30] as follows: , where is a constant, is the optical band gap, and is a coefficient that equals 1/2 for allowed direct transitions. Figure 6 shows the optical band gap energy as a function of , before and after annealing. These values increase from 3.58 eV for = 2.67 Pa to 4.34 eV for = 10.66. Our optical gap values are larger than those reported for undoped ZnO thin films [2]. Similar optical band gap shifts to values of around 4 eV have been reported by Tan et al. [31]. These authors propose that the amorphous phase in these films is the main reason for the increase in optical band gaps. This explanation is consistent with the observed increases in transmittance tails in Figure 4, which have been correlated with poor crystallinity [26, 31].
Relying on the XRD results, it seems reasonable to assume that the widening of the optical band gap is due to the reduced crystallinity of our films. As the increases sputtered atoms lost kinetic energy, so the fraction of amorphous ZnO increases. In fact there are only low intensity (100), (101), and (110) diffraction peaks in the XRD pattern for as-grown films (Figure 1). The physical model about the change of optical band gap induced by increasing can be viewed as crystalline regions embedded in a matrix of amorphous ZnO. As a result, the absorption of photon is mainly contributed by amorphous ZnO. On the other hand, after annealing, the crystalline quality of ZnO films improves, and consequently the optical band gap decreases considerably. Amorphous phase as a possible cause of the widening of the optical band gap in polycrystalline ZnO films has also been reported by Mera et al. [32].
After annealing, band gap values decreased considerably. In the case of = 2.67 Pa, for example, the as-grown was 3.37 eV, but it decreased to 3.24 eV after annealing at 500°C. This is consistent with our earlier discussion of transmittance results and its relation with oxygen vacancies and crystallization process. Thermal annealing favors oxygen desorption, thus leading to a decrease in . Li et al. reported that substrate temperatures have a similar effect on the optical band gap values of ZnO films [33]. Hiramatsu et al. have found that the intensity of O2 desorption from ZnO films increases rapidly for temperatures larger than 300°C [34].
ZnO is an important semiconductor with a wide range of applications in electronic devices. Therefore, the investigation of its electrical properties is important. The resistivities of our films at room temperature, before annealing, were higher than 107 Ω cm, which is the experimental detection limit of our Hall system. Resistivities after annealing at 200°C and 500°C are shown in Table 2. The observed decrease in resistivity with increasing annealing temperature can be ascribed to oxygen desorption and to crystallization, in accordance with our discussion above. After annealing, the crystalline quality increased, which also will cause the decrease of resistivity due to the weakening of the grain boundary scattering. The large resistivity of as-grown films might make them apt for buffer layers in flexible solar cells [35].
4. Conclusions
The present work investigated the effect of varying the argon pressure on the morphological, optical, and electrical properties of zinc oxide films grown by RF sputtering from a ZnO target. Additionally, the influence of the annealing, temperature was also studied. Relying on XRD measurements the films were polycrystalline and became c-orientated after annealing at 200°C and 500°C. The main result of this investigation is that optical band gap values increase from 3.58 eV to 4.34 eV when the argon pressure increases from 2.67 to 10.66 Pa. The observed widening of the optical band gap can be understood as being due to changes in the mean-free path of the sputtered species in the background argon gas, which in turn influences the degree of crystallinity of the films. Optical band gaps and resistivities were found to decrease after annealing, which is consistent with oxygen desorption and the improvement of the crystalline quality of films during this process. Our results demonstrate that the optical properties of ZnO films grown under the conditions of our experiment can be tuned by controlling the argon pressure.
Acknowledgments
The authors would like to thank Daniel Aguilar and Patricia Quintana for XRD analysis and Willian Cauich for film thickness measurements. The financial support from FORDECYT-116157 is gratefully acknowledged.