Otwarty dostęp

Annealing effect on the structural and optoelectronic properties of Cu-Cr-O thin films deposited by reactive magnetron sputtering using a single CuCr target


Zacytuj

Introduction

Thin films with visible-light transmittance above 80% and electrical resistivity below 1 × 10−3 Ω-cm are called transparent conducting oxides (TCOs). The representative TCO thin-film materials include SnO2, In2O3, and ZnO [1]. However, the optoelectronic properties of these TCO thin films are not stable, especially at high ambient temperature. Impurity-doped TCOs, such as Sb2O3-doped SnO2, Al2O3-doped ZnO, and SnO2-doped In2O3, are usually preferred to solve this problem [24]. Proper dopant content could greatly increase free-electron concentration without affecting visible-light transmittance, resulting in extensive applications in optoelectronic devices. However, these aforementioned TCOs with excellent optoelectronic properties are n-type; p-type TCOs with high visible-light transmittance and low electrical resistivity are also required to develop transparent p–n diode elements in the optoelectronic industry. As a p-type TCO material, CuCrO2 has the advantages of abundant mineral content, low cost, and nontoxic constituent elements, thus showing the greatest potential in optoelectronic applications [5].

Many methods could be used to prepare Cu-Cr-O thin films, including pulsed laser deposition [6], chemical solution deposition [7], solid-state reaction method [8], hydrothermal method [9], chemical vapor deposition [10], and reactive sputter deposition [1114]. Among them, reactive sputter deposition is one of the most frequently used methods for growing thin films because it can provide highly uniform thickness in large-area depositions. The sputtering deposition techniques for Cu-Cr-O thin films often utilize a sintered ceramic target as the source [11]. Ceramic targets, however, require a more complicated manufacturing process than metal targets, and thus, their cost is extremely high overall. In addition, they are prone to cracking due to the high temperature generated by the plasma during the sputtering process. By using a dual-target sputtering system in an oxygenincorporation atmosphere, Cu and Cr metal targets could be used to directly deposit Cu-Cr-O thin film [14]. This method eliminates the need to prepare Cu-Cr-O powder, which is time and resource intensive. In addition, the conductivity of the metal target is higher than that of the Cu-Cr-O target; this facilitates the use of DC sputtering techniques. Nevertheless, the compositional uniformity of dualtarget sputtering is significantly poor under large-area deposition due to the difference in the positions of both targets. A single CuCr alloy target could be used for sputtering in an oxygen atmosphere to resolve the problems mentioned above. Thus, this method was adopted in the present work to synthesize Cu-Cr-O thin film. The actual optoelectronic performance of this method, however, remains to be verified in this study.

On the other hand, the fabrication of as-deposited Cu-Cr-O thin films is substantially dependent on good control of the post-annealing temperature as the as-sputtered thin films generally reveal poor crystallinity. Hence, this study focused on optimizing the quality of thin films by altering the post-annealing temperature. The influence of phase composition and microstructural variation on the optoelectronic characterization of the Cu-Cr-O thin film was also explored. Finally, the feasibility of the method was evaluated.

Materials and experimental methods

High-vacuum DC magnetron sputter deposition equipment was employed to grow Cu-Cr-O thin films on fused silica substrates. The target used was a CuCr alloy target with diameters of 4 in, and its atomic ratio was 1:1. By using a turbo molecular pump with a rotary pump, the system achieved an ultimate vacuum of 3.0 × 10−6 Torr. The target and the substrate had a working distance of about 90 mm. Before deposition, the substrates were sequentially cleaned and rinsed with ethanol and distilled water in an ultrasonic bath. Film deposition was carried out at a CuCr-target power of 200 W, a substrate temperature of room temperature, and an O2/(O2 + Ar) flow rate ratio of 10%. The working pressure was 5.0 × 10−3 Torr at this moment. The sputtering time was set to 15 min to obtain a film thickness of 330 nm. Following deposition, the specimens were post-annealed in a tube furnace under an Ar flow rate of 400 sccm (standard cubic centimeter per minute). The oxygen partial pressure in the tube furnace was kept below 7.0 × 10−5 atm. The annealing temperature was ramped from room temperature to a temperature range of 500°C–800°C (equivalent to 773 K—1073 K) at a rate of 5°C/min for 2 hours, and then cooled down to room temperature at the same rate.

Electron spectroscopy for chemical analysis (ESCA, PHI 5000 VersaProbe) was used to determine the chemical composition of Cu-Cr-O films. X-ray diffraction (XRD, Mac-Science MXP3) equipped with Cu Kα radiation was employed to examine the phase structure of these films, with a scanning speed of 1°/min. Field-emission scanning electron microscopy (FESEM, JEOL JSM-6700F) and field-emission transmission electron microscopy (FETEM, JEOL JEM-1200EXII) were utilized to observe the microstructure of Cu-Cr-O films. The atomic force microscope (AFM, Seiko SPA400) was utilized to examine the surface topography of Cu-Cr-O films. Ultraviolet/visible spectrophotometry (Hitachi U3010) was applied to gauge the transmittance of Cu-Cr-O films. The Hall effect measurement system (Accent HL5500PC) was operated to measure the electrical property of Cu-Cr-O films.

Results and discussion
Chemical composition

The variation in the chemical composition of Cu-Cr-O films with annealing temperature is revealed in Table 1, which shows two significant findings. First, the ratio of Cu to Cr in these films was measured to be about 2, although an equimolar CuCr target was used in this study. The higher Cu content in the films may be due to the difference in sputtering yields between Cu and Cr [15]. Second, only marginal variation was observed in the chemical composition with an increment in annealing temperature. The evaporation of these constituent elements did not occur in the films even when annealed at temperatures up to 800°C.

The elemental composition of Cu-Cr-O films as a function of annealing temperature

Annealing temperature (°C) Cu (at.%) Cr (at.%) O(at.%) Cu/Cr ratio
As-deposited 28.41 14.05 57.54 2.02
500 31.69 15.95 52.36 1.99
600 31.20 16.24 52.55 1.92
700 31.23 15.68 53.09 1.99
800 31.81 15.68 52.51 2.03
Phase transformation

Figure 1 shows the XRD patterns of the Cu-Cr-O films with different post-annealing temperatures. No visible diffraction peaks appeared in the as-deposited film, indicating that it was amorphous. As the annealing temperature was heightened to 500°C, the film began to crystallize, mainly forming spinel CuCr2O4 and monoclinic CuO phases. From the perspective of elemental composition, the higher the copper content, the more preferentially the copper-rich phase is formed. However, in terms of thermodynamics, the Gibbs free energy of the CuCr2O4 phase is relatively negative at lower temperatures, indicating that the generation of CuCr2O4 phase is thermodynamically favorable [16]. When the annealing temperature attained 600°C, the spinel CuCr2O4 phase disappeared, and it was replaced by the delafossite CuCrO2 phase, because increasing the annealing temperature rapidly reduces the Gibbs free energy of the CuCrO2 phase even lower than that of the CuCr2O4 phase, thereby promoting the reaction of CuCr2O4 with CuO to form CuCrO2. These relevant reaction equations and thermodynamic data can be expressed as follows: 1/2Cu2O(s)+1/4O2(g)CuO(s) \[\begin{align}1/2C{{u}_{2}}{{\text{O}}_{(\text{s})}}+1/4{{\text{O}}_{2(\text{g})}}\to \text{Cu}{{\text{O}}_{(\text{s})}} \end{align}\] ΔG1=ΔG1+RTln(pO2)=65222.25+46.955TRTln(pO2) CuO(s)+Cr2O3(s)CuCr2O4(s) \[\begin{align}Cu{{O}_{(\text{s})}}+C{{r}_{2}}{{\text{O}}_{3(\text{s})}}\to CuC{{r}_{2}}{{\text{O}}_{4(\text{s})}} \end{align}\] ΔG2=ΔG2+RTln(pO2)=366715.85T \[\begin{align}\Delta\!\!\text{ }{{\text{G}}_{2}}=\text{ }\!\!\Delta\!\!\text{ G}_{2}^{\circ }+RT\ln (\text{p}{{\text{O}}_{2}})=3667-15.85\text{T} \end{align}\] CuCr2O4(s)+CuO(s)2CuCrO2(s)+1/2O2(g) \[\begin{align} CuC{{r}_{2}}{{\text{O}}_{4(\text{s})}}+Cu{{\text{O}}_{(\text{s})}}\to 2CuCr{{\text{O}}_{2(\text{s})}} \\ +1/2{{\text{O}}_{2(\text{g})}} \\ \end{align}\] ΔG3=ΔG3+RTln(pO2)=79512.568.15TRTln(pO2) \[\begin{align} \text{ }\!\!\Delta\!\!\text{ }{{\text{G}}_{3}} & =\text{ }\!\!\Delta\!\!\text{ G}_{3}^{\circ }+RT\ln (\text{p}{{\text{O}}_{2}})=79512.5 \\ {} & -68.15\text{T}-RT\ln (\text{p}{{\text{O}}_{2}}) \\ \end{align}\] where ΔG and ΔG° are the variations of the Gibbs free energy and the standard Gibbs free energy, respectively. R is the gas constant, T is the absolute temperature, and pO2 is the oxygen partial pressure. In these equations, the subscript “(s)” is utilized to indicate substances in the solid state, while the subscript “(g)” represents substances in the gaseous state. From the above equations, when the pO2 is 7.0 × 10−5 atm, the ΔG can be calculated, and the variation of ΔG with temperature can be drawn, as shown in Figure 2. When the annealing temperature exceeds 600°C, it is found that ΔG3<ΔG2<ΔG1, which means that the CuCrO2 phase becomes a thermodynamically favorable phase. As the annealing temperature was heightened to 800°C, the diffraction peaks became stronger and sharper, implying grain growth in the CuCrO2 and CuO phases.

Fig. 1.

XRD patterns of Cu-Cr-O films with different post-annealing temperatures

Fig. 2.

The ΔG of reactions 2, 4, and 6 with different annealing temperatures

Microstructure

Figure 3 shows the SEM image of the Cu-Cr-O films with different post-annealing temperatures. The as-deposited film was composed of nanoparticles with size ranging from approximately 4 nm to 10 nm (Figure 3a). An annealing procedure at 500°C generated larger particles, about 30–40 nm in size, with an agglomerate surface characteristic that left voids inside the film (Figure 3b). With the further increase in annealing temperature to 800°C, the film grains grew significantly, and the particles size increased to hundreds of nanometers (Figures 3c3e). While higher temperature increases particle size by enhancing atomic diffusion, it also enlarges the void by causing atoms to coalesce near the void edges due to capillary forces.

Fig. 3.

Cross-section and plane-view SEM images of Cu-Cr-O films with different post-annealing temperatures: (a) as-deposited, (b) 500°C, (c) 600°C, (d) 700°C, and (e) 800°C

Thus, whether the film properties obtained by high-temperature annealing are better is worth exploring. Figure 4 shows the AFM images of the Cu-Cr-O films with different post-annealing temperatures. These results were in agreement with the FESEM observations. The high heat induces coalescence of adjacent grains, resulting in a rapid increase in RMS roughness from 0.75 nm to 20.01 nm with increasing annealing temperature. Figure 5 shows TEM images with the selected area diffraction (SAD) patterns of the Cu-Cr-O films with different post-annealing temperatures. The 500°C-annealed film consisted of equiaxed CuO and CuCr2O4 grains (Figure 5a). The grains and void size increased along the film growth direction. The thickness-dependent change in grain size is a common phenomenon for annealed growth. When the annealing temperature was increased to 600°C, the film exhibited increased void density and grain growth (Figure 5b). In this situation, the CuCr2O4 reacted with CuO to form CuCrO2 until CuCr2O4 was depleted. However, no distinct two-phase structure in this film could be observed, probably due to insufficient atom diffusion at the present annealing temperature. When the annealing temperature was continued to increase to 800°C, these grains and voids also continued to grow (Figure 5c). Visible phase segregation was present in the microstructure, as confirmed by EDS analysis (not shown). The above findings were in good agreement with the XRD analysis.

Fig. 4.

AFM images of Cu-Cr-O films with different post-annealing temperatures: (a) as-deposited, (b) 500°C, (c) 600°C, (d) 700°C, and (e) 800°C

Fig. 5.

TEM images with corresponding SAD patterns (insets) of Cu-Cr-O films with different post-annealing temperatures: (a) 500°C, (b) 600°C, and (c) 800°C

Optical measurements

Figure 6a shows the light transmittance of the Cu-Cr-O films with different post-annealing temperatures. The as-deposited film only exhibited a very low transmittance of 7.9% in a wavelength region of 500–900 nm. Increasing the annealing temperature to 500°C resulted in sharper fundamental absorption edges and increased transmittance up to 31.4%, which was caused by improved crystal quality. These CuO and CuCr2O4 with narrow band gap and high photo response in the film were responsible for this result [17, 18]. At an annealing temperature of 600°C, the film achieved the highest transmittance of 49.5%. The increase in grain size and the replacement of the CuCr2O4 phase by the CuCrO2 phase with a wide band gap were the major factors. When the annealing temperature was 800 °C, however, the transmission of the film decreased to 36.7%, which was associated with higher roughness, increased void defects, and larger CuO grains. The part of incident light was redirected to diffuse scattering instead of specular scattering because of the increment in surface roughness and void defects [19]. The as-deposited film was characterized by a gentle absorption edge, implying its high defect density in film. As annealing temperature increased, the absorption edge became sharper, which was attributed to crystallization and grain growth. The band gap of the Cu-Cr-O films was measured by a linear extrapolation on the x-axis of plot of the (αhv)2 vs. photon energy (hv), as shown in Figure 6b. The band gap energy values of the as-deposited, 500°C-annealed, 600°C-annealed, 700°C-annealed, and 800°C-annealed films were found to be 1.25, 1.78, 1.98, 1.96, and 1.84 eV, respectively. The low band gap of the deposited films was also associated with a high defect density. The initial increase of band gap with was dependent on the formation and growth of these crystal phases. This subsequent decrease of band gap may be related to a reduction in carrier concentration, known as the Burstein–Moss effect [20, 21].

Fig. 6.

(a) Light transmittance of Cu-Cr-O films with different post-annealing temperatures. (b) variation of (αhv)2 versus hv

Electrical measurements

Table 2 shows the Hall effect measurement results for the Cu-Cr-O films with different postannealing temperatures. Due to the positive Hall coefficients measured in the Cu-Cr-O films, the holes were confirmed to be the major carriers of these films, indicating a p-type semiconductor behavior. A considerable variation was found in the electrical property of CuO, CuCr2O4, and CuCrO2 films studied by different authors, likely due to differences in the films’ composition and structure that contributed to this inability to fully explain electrical behavior. The as-deposited film had a very high electrical resistivity that could not be measured by a Hall system, hence it was omitted from Table 2. After annealing was conducted at 500°C, the films exhibited measurable electrical resistivity. When the annealing temperature attained 600°C, the carrier concentration and Hall mobility of the film increased. CuCrO2 is generally known to have lower electrical resistivity than CuCr2O4 [2225]. Coupled with the larger grain size, the substitution of CuCr2O4phase with the CuCrO2 phase was thought to be responsible for the superior electrical performance. As the annealing temperature was further increased, however, the electrical property of the film deteriorated, mainly due to the decrease in carrier concentration. Yoshida et al. have confirmed that the major source of conductive carriers in Cu-Cr-O p-type semiconductors was interstitial O and Cu vacancies [26, 27]. In addition, Banerjee et al. found that electrical resistivity was reduced by extra O atoms present at interstitial sites and grain boundaries [28]. In the present study, while annealing could increase the grain size, it also reduced some point defects that contributed to hole concentration, such as such as interstitial O and Cu vacancies, leading to increased electrical resistivity. The values of Hall mobility showed scattering variations with increasing annealing temperature from 600°C to 800°C. The increase in Hall mobility originated from the decrease in grain boundary scattering and point defect scattering due to the increase in grain size [29]. However, the larger surface roughness and void density at higher annealing temperature facilitated the chemisorption of nitrogen, oxygen, and water vapor into the film, thus increasing barrier height and degrading Hall mobility [30]. Accordingly, the film annealed at 600°C had the lowest electrical resistivity of 41 Ω-cm.

The carrier concentration, Hall mobility, and electrical resistivity of Cu-Cr-O films as a function of annealing temperature

Annealing temperature (°C) Carrier concentration (×1017 cm−3) Hall Mobility (cm2/V-s) Electrical resistivity (Ω-cm)
500 2.81 0.121 183
600 4.05 0.373 41
700 3.25 0.223 86
800 0.43 0.268 540

Based on the above results, although 600°C-annealed film has the optimal optoelectronic performance, its phase composition contains a large amount of CuO phase with narrow energy gap, resulting in poor light transmittance compared with other published papers. The ESCA result shows that the Cu content in film is about twice the Cr content. The composition difference between the film and the target caused by the magnetron sputtering method is too large to be ignored, which will be an unfavorable factor that must be considered and eliminated by adjusting the target composition in the further study of single-target magnetron sputtering in the future.

Conclusion

This study investigated the fabrication of transparent conductive Cu-Cr-O films on fused silica substrates through DC reactive magnetron sputtering using a single equimolar CuCr alloy target in conjunction with a post-annealing process. Annealing at 500°C transformed the phase structure of the film from amorphous to monoclinic CuO and spinel CuCr2O4. Further annealing at 600°C depleted the CuCr2O4 phase to generate the delafossite CuCrO2 phase. As the annealing temperature increased, the microstructure of the films converted from a compact nanograins feature to large agglomerated grains with numerous voids, thereby roughening the surface. Accordingly, grain growth combined with the formation of the CuCrO2 phase led to the optimized optoelectronic performance of the 600°C-annealed films. However, some subsequent effects associated with grain growth adversely affected the optoelectronic properties. As the annealing temperature continued to increase, the grains also grew continuously, along with the surface roughness and void density. The increased grain size resulted in the reduction in the defects contribution to the hole concentration. The great surface roughness and void density provided electrical barriers for hole transport, thereby lowering the Hall mobility. They also redistributed some of the incident light into diffuse scattering, resulting in significant light-scattering losses.

eISSN:
2083-134X
Język:
Angielski
Częstotliwość wydawania:
4 razy w roku
Dziedziny czasopisma:
Materials Sciences, other, Nanomaterials, Functional and Smart Materials, Materials Characterization and Properties