Variation of electrical properties in thickening Al-doped ZnO films: role of defect chemistry

Abstract

This study addresses the variation in electrical properties in a thickening Al-doped ZnO (AZO) film up to 348 nm and correlates this with its defect chemistry. To this end, AZO films were deposited on soda lime glass substrates at varying deposition times, which ranged from 15 to 120 min at an interval of 15 min, by RF magnetron sputtering. A range of experimental techniques, such as Hall effect measurement, PL measurement, AFM, XPS, XRD, FE-SEM and UV-vis-NIR spectrophotometry, were used to understand the microstructure and optoelectronic properties of the films. It was determined that all the films grew as ZnO hexagonal wurtzite structures with the strong (002) orientation of the crystallites, with an average transmittance of 87–93% in the visible range. However, the electrical properties of these films showed strong dependence on film thickness because of the associated defect chemistry resulting from the film microstructure. A sharp increase in the film carrier concentration is strongly influenced by the presence of shallow donor level defects caused by Zn and extended Zn interstitials, whereas deep donor level defects caused by the occurrence of the oxygen vacancies and excess surface oxygen do not contribute to the carrier generation. The combination of an increased concentration of shallow donor level defects and absence of deep donor level defects correspond with the best values of carrier concentration and carrier mobility leading to the lowest electrical resistivity of 0.913 × 10⁻³ Ω cm and highest observed film quality at 242 nm thickness, which was obtained at 60 min deposition time. Moreover, this study reveals the necessity of using an appropriately thick AZO film as the TCO. Finally, a narrow PL spectrum, with strong violet and/or blue emissions, is an indication of a high quality TCO. This study also sheds light on the use of PL as an alternate tool to understand the microstructure and electrical properties of a TCO.

---

1. Introduction

Due to the simultaneous occurrence of low electrical resistivity and high optical transmittance, Al-doped zinc oxide (AZO) films have a wide range of applications in technologically relevant areas, such as thin film solar cells, light emitting diodes (LEDs), liquid crystal displays and electro chromic devices, thus leading to a significant amount of research to understand and optimize the optoelectronic properties of this material.^1–5 Designing a transparent conducting oxide (TCO) with the optimal combination of microstructure and optoelectronic properties requires optimization of the deposition conditions, which improves its microstructure and subsequent optoelectronic properties.^6–14

As observed in this study, it is interesting to note the substantial variation in the electrical properties of the film by only changing film thickness (unlike the optical properties), even if all the other deposition conditions (using radio frequency (RF) magnetron sputtering), such as substrate temperature, power to the target, substrate-to-target distance, argon flow rate, target type, and chamber pressure, were maintained constant. Understanding the origin of these variations is extremely crucial because this would lead to the fabrication of an appropriately thick AZO film for any application. Although similar observations have been reported in the literature, their mechanisms are yet to be understood.^15,16

Therefore, in this study, the variation of the electrical properties (namely carrier concentration and carrier mobility) of a thickening AZO thin film is interpreted in terms of its defect chemistry and microstructure. To this end, AZO films were deposited on soda lime glass substrates by varying the deposition time from 15 to 120 min at an interval of 15 min via RF magnetron sputtering. A range of experimental techniques, such as atomic force microscopy (AFM), X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), X-ray photoelectron spectroscopy (XPS), Hall effect measurements, UV-vis-NIR spectrophotometry and photoluminescence (PL) measurements, were used to understand the overall microstructure and optoelectronic properties of the films. While AFM, XRD, FESEM and XPS obtain overall information about the film microstructure, Hall effect measurements and UV-vis-NIR spectrophotometry obtain information about its optoelectronic properties. In PL, excited electrons from the conduction band undergoes relaxation and thus emit various wavelengths depending on the defects states present within the band gap.¹⁷ This could then be correlated with the microstructure and subsequent electrical properties because the position and concentration of these defect states are influenced by the film microstructure and strongly contribute to the film carrier concentration and therefore giving information on the film quality.

2. Experimental procedure

2.1 Fabrication of AZO films

A series of Al-doped ZnO (AZO) films with thicknesses varying from 115 to 348 nm were deposited on ultrasonically cleaned 1 cm × 1 cm soda lime glass (SLG) substrates by varying the deposition time (t) from 15 to 120 min via RF magnetron sputtering. An AZO ceramic target (of 99.99% purity with a composition of 98 wt% ZnO and 2 wt% Al₂O₃) was used for the deposition. The chamber pressure (P_chamber) of 7.0 × 10⁻³ mbar was maintained throughout the deposition process by providing 40 SCCM argon (of purity 99.99%) to the chamber with the help of a mass flow controller. Optimized deposition conditions of 623 K substrate temperature (T_s) with 15 rpm substrate rotation, 7.0 cm substrate-to-target distance and an RF power of 50 W to the target were used.¹⁸ Before every deposition, the target was pre-sputtered for 5 min to remove the presence of any impurities.

2.2 Microstructural and optoelectronic characterization

After the AZO thin films were fabricated, their overall structure and surface morphology were analyzed using a grazing incident X-ray diffractometer (GIXRD; Model: D8 Discover, Supplier: Bruker Corporation) and atomic force microscope (AFM; Model: NanoScope Multimode 8.0, Supplier: Bruker Corporation), respectively. A field emission scanning electron microscope (FESEM; Model: JSM 7600 F, Supplier: JEOL Ltd.) was used to determine film thickness and uniformity. For Hall effect measurements (Model: 8400 series HMS, Supplier: Lakeshore), indium was used to contact four corners of the AZO coated SLG and electric and magnetic fields were applied in the van der Pauw configuration to measure sheet resistance, carrier concentration and carrier mobility. Photoluminescence (PL, Model: DM 500i, Supplier: Dongwoo Optron Co. Ltd) spectra were obtained on a spectrometer excited with a 325 nm He–Cd laser. Transmission spectra and optical bandgap of the AZO thin films were studied using an UV-vis-NIR spectrophotometer (Cary 5000, Agilent Technologies). X-ray photoelectron spectroscopy (XPS; Model: PHI 5000, Versa Prob II, Supplier: FEI Inc. spectrometer) was employed to probe the surface chemical states of the AZO thin films.

3. Results and discussions

The AZO films deposited on SLG substrates by varying t from 15 to 120 min (at an interval of 15 min) at T_s = 623 K via RF magnetron sputtering were found to have grown in the (002) oriented ZnO hexagonal wurtzite structure with an average transmittance of 87–93% in the visible range (see Fig. S1 and S2(a)†). The corresponding optical band gaps (E_g) of these films were found to increase from 3.67 to 3.73 eV with an increase in thicknesses from 115 to 272 nm and then decrease with further increase in the film thickness to 3.69 eV at 348 nm. This behaviour is predominantly attributed to the Burstein–Moss shift that results from the carrier concentrations of the films (n_e), which is given by the following relation (see Fig. S2(b)† and 1).^18–20where ΔE_g is the change in optical band gap, m* is the effective mass of an electron and h is Planck's constant.

Fig. 1 Variation in film thickness and crystallite size of Al-doped ZnO films deposited on soda lime glass substrates by RF magnetron sputtering at t varying in the range of 15–120 min.

Fig. 2 shows the variation in the electrical properties (i.e., electrical resistivity (ρ), carrier concentration (n_e) and carrier mobility (μ_e)) of the thickening AZO films with respect to deposition time. It is found that with an increase in the film thickness from 115 nm to 348 nm because of an increase in t, ρ decreases from 5.58 × 10⁻³ Ω cm at t = 15 min to 0.913 × 10⁻³ Ω cm at t = 60 min, whereas with further increase in the film thickness, ρ increases up to 2.32 × 10⁻³ Ω cm at t = 120 min. The trend of ρ was found because of the observed trends in n_e and μ_e, with the maximum values of 5.17 × 10²⁰ cm⁻³ and 16.32 cm² V⁻¹ s⁻¹ being obtained at t = 75 and 60 min, respectively (see Fig. 2). The lowest ρ and highest μ_e values observed at t = 60 min in the present case suggest the growth of the optimum quality AZO film on the SLG substrate at 242 nm film thickness.

Fig. 2 Variation in electrical resistivity (ρ), carrier concentration (n_e) and carrier mobility (μ_e) of Al-doped ZnO films deposited on soda lime glass substrates by RF magnetron sputtering at t varying in the range of 15–120 min.

To understand the variation of these electrical properties in a thickening AZO film deposited under optimized experimental conditions,¹⁸ the defect densities of all the films (obtained from the PL spectra) are presented in Fig. 3. The films deposited at a lower t of 15, 30 and 45 min and higher t of 105 and 120 min showed relatively broader spectra than those deposited for 60, 75 and 90 min. For quantification, the PL spectra were deconvoluted with Gaussian fitting of the peaks (goodness of the fits in all the PL spectra are within 0.98–0.99). This resulted in five emission peaks for the broader PL spectra (obtained for both lower and higher t) and three or four emission peaks for the narrower PL spectra obtained from the AZO films deposited for 60 to 90 min. Among the peaks, an ultraviolet (UV) emission peak centred at around 376 ± 1 nm, which indicates the near band edge emission that is typical for the free exciton peak of ZnO, was present for all the AZO films.²¹ Apart from this characteristic UV emission, the PL spectra of the films also showed the presence of a broad visible emission bands.²² They are at 419 nm, 521 nm, 602 nm and 723 nm for the AZO film grown for 15 min duration, which thus emits violet, green, orange and red colors, respectively. It can be noted that in a PL spectrum, violet, blue, green, yellow, orange and red emissions occur in the wavelength ranges of 380–449 nm, 450–497 nm, 498–569 nm, 570–589 nm, 590–619 nm and 620–750 nm, respectively.^23–26 Similar emission spectra were also observed for the AZO films deposited for 30 and 45 min with the emissions corresponding to the longer wavelengths of the visible band (i.e., emissions except violet) being shifted towards the shorter wavelengths with an increase in t (see Table S1 of the ESI† for details). With a further increase in the film thickness to 242 nm (i.e., at t = 60 min), the emissions corresponding to green, yellow, orange and red peaks were found to completely disappear with the emergence of a new emission peak at 465 nm (thus emitting blue color) along with a violet peak at 426 nm. With a further increase in film thicknesses (due to the increase in t from 60 to 90 min), the green emission peak that appeared at 509 nm for t = 75 min shifted to 549 nm at t = 90 min, thereby increasing the broadness of the spectra (see Fig. 3(e) and (f)). However, upon further thickening of the film (i.e., films formed at the larger t of 105 and 120 min), the PL spectra became very broad with five emission peaks analogous to those observed at the lower film thicknesses (compare Fig. 3(g) and (h) and (a)–(c)). The emissions in all these AZO films are because of the presence of defects within the band gap of the material. There may be many defects such as zinc interstitials (Zn_i), extended zinc interstitials (ex-Zn_i), single charged oxygen vacancies (Vo⁺), double charged oxygen vacancies (Vo⁺⁺) and excess oxygen present on the surface of the AZO films.^27–34 Zn_i can be generated by the Frenkel reaction using the following equation:^32,34

Zn_zn ↔ Zn_i + V^*_Zn (2)

whereas ex-Zn_i (denoted as both the Zn_i⁺ and Zn_i⁺⁺) form because of the further ionization reactions of the already formed Zn_i as per the following reactions:

Zn_i ↔ Zn_i⁺ + e⁻ (3)

Zn_i ↔ Zn_i⁺⁺ + e⁻ (4)

Fig. 3 Photoluminescence spectra of Al-doped ZnO films deposited on soda lime glass substrates by RF magnetron sputtering at t varying in the range of 15–120 min (a)–(h).

Zeng et al. introduced a higher concentration of defects (in the form of Zn_i or ex-Zn_i) into nanoscale ZnO under non-equilibrium processes (i.e., laser ablation-induced extreme conditions and zinc-rich annealing) and observed their transition to the valence band, which they attributed to violet and blue emissions.³² In an another study, Chen et al. deposited Al doped ZnO films at different RF powers via RF sputtering and then investigated the mechanism of blue and green emissions in these films, wherein they concluded that the blue emissions originated from zinc interstitial related defects.³³

It should be noted herein that although an extrinsic donor such as substitutional aluminium (Al³⁺) is one of the major type of defect responsible for the increase in n_e in the film, the defect state is not discernible from the PL spectra because it forms very close to the conduction band and because of its low ionization energy in ZnO (i.e., 0.053 eV) it is easily donated to this band at room temperature.² Similar observations have also been reported in the literature, wherein the identification of defect states formed due to Al doping in ZnO was not possible.^35,36 Gunjan et al. synthesized ZnO nanoparticles by varying the Al doping concentration and concluded that Al-doping was responsible for the blue shift in the near band edge emission, which clearly indicates that these defect states are merged with the conduction band.³⁵ In another study, Mahmood et al. observed an increase in the intensities of the near band edge emissions for Al doped ZnO films as compared to those ZnO with no additional defect states due to Al doping observable.³⁶

The origin of violet and blue emissions is known to be due to the transition of the electrons present in the shallow donor level defects at the top of the valence band. Similarly, emissions corresponding to the visible light of longer wavelengths (i.e., the observed green, yellow, orange and red peaks) are known to be due to the transition of the electrons present in the deep donor level defects in the valence band.^32,33,37,38 A comparison between the emission characteristics and carrier concentrations of these films resulted in the co-existence of a larger n_e in the AZO film and narrower emission spectra (i.e., the films obtained at t = 60 and 75 min; see Fig. 2 and 3) comprising only strong violet and blue peaks (i.e., absence of all the deep donor level emissions). For further quantification, the intensity ratios of violet to UV (I_V/I_UV), blue to UV (I_b/I_UV) and those of the deep-level emissions (i.e., combined green, yellow, orange and red emissions occurring from the deep donor level defects) to UV (I_DLE/I_UV) for all the deposited AZO films were plotted along with their n_e in Fig. 4. Herein, the highest value of n_e obtained in the AZO film (deposited at t = 75 min) was found to coincide with the highest (I_V/I_UV) and (I_b/I_UV) ratios, which evidence the strong contribution of the shallow donor level defects towards the n_e of the films. The absence of (I_DLE/I_UV) in this film further supports the abovementioned finding of the deep donor level defects in the AZO films not contributing to the carrier generation in the films. Accordingly, an increase in this ratio coincides with a decrease in n_e for all the AZO films studied. The shallow donor level defects occur due to the presence of Zn_i (emitting violet light) and ex-Zn_i (emitting blue light), whereas deep donor level defects occur due to the presence of single and double charged oxygen vacancies (Vo⁺ and Vo⁺⁺ emitting green and yellow light, respectively) and adsorbed oxygen present on the surface of the AZO films (emitting orange-red emissions; see Fig. 5).^{17,32–34,39} It could be interpreted that the electrons present in these shallow donor level defects can easily jump into the conduction band and thus increasing the number of n_e in these films. On the other hand, electrons present in the deep donor level defects are difficult to ionize and thus do not contribute to the carrier generation in the conduction band and therefore resulting in a lower value of n_e.

Fig. 4 Variation in the intensity ratio of the violet to UV (I_V/I_UV), blue to UV (I_b/I_UV) and deep-level emission to UV (I_DLE/I_UV) emission peaks in Fig. 3 and the carrier concentration (n_e) of the Al-doped ZnO films as a function of deposition time (t).

Fig. 5 Schematic of Al-doped ZnO thin film based on photoluminescence data. The indicated optical band gap here is from the measured UV emission in the PL spectra for all the AZO films studied. The data for the electron affinity (energy difference between the top of the conduction band and the vacuum level) has been obtained from the literature.³⁹

To understand the transition of these defect states in the thickening AZO film, the surface morphologies of all the deposited films (obtained using AFM) at a regular interval of t along with the surface chemical constitutions of the selected samples (obtained using XPS for t = 15, 60 and 120 min) are presented in Fig. 6(a)–(h) and 7(a)–(c). In addition, the AFM line section analysis for the three selected samples is presented in Fig. S3(a)–(c) of the ESI.† Film formation starts with the nucleation and growth of the (incoming) chemisorbed sputtered species on the SLG substrate. Therefore, for a very low deposition time (i.e., at t = 15 min), a discontinuous (as apparent from the spread in the thickness value; see Fig. 1), a porous and loose film with hillocks structure and smaller crystallites was observed (see Fig. 1), which increased the possibility of adsorbed oxygen on the surface.^31,38,40 The presence of this excess surface oxygen was also confirmed from the XPS spectra, wherein the maximum amount of chemisorbed oxygen was found for this film out of the three samples analyzed (comparing the intensities of the O_c peak at the binding energy value of 532.2 eV; see Fig. 7(a)–(c) and also the ESI† for details).^41,42 The intensity ratio of the emissions from the excess adsorbed oxygen on the film surface (i.e., the combined orange and red emissions) to UV (I_OR/I_UV) was also found to be the highest among all the AZO films studied (see Fig. 8). This defective film can trap charge carriers, thereby reducing μ_e substantially (see Fig. 2). With an increase in t, crystallite size increased, followed by the coalescence of islands covering the substrate surface, which resulted in the growth of increasingly uniform films of higher thicknesses (see the spread in the average thickness values of Fig. 1). The AZO film deposited at t = 60 min was found to have the most oriented crystallite growth (see Fig. 6(d)). It may be stated that at around this film-thickness, the rate of arrival of the incoming sputtered species on the substrate surface and their temperature induced surface diffusion into the equilibrium sites followed by subsequent island coalescence and growth reach a balanced state, thereby leading to this type of crystallite growth with compact film. Subsequently, this resulted in the smallest spread in its average thickness value as compared to the other reported results (see Fig. 1). The amount of chemisorbed oxygen was found to be minimum from XPS (see Fig. 7(a)–(c) and the ESI† for details), which is on par with the PL finding (i.e., no orange and red emissions were found; see Fig. 3 and 4). The development of relatively bigger crystallites and oriented growth in this film led to the highest μ_e (see Fig. 2). With a further increase in the film-thickness at higher t > 60 min, local heating (because of the longer deposition time at a particular temperature) and lateral surface diffusion led to the formation of bigger aggregates and therefore the formation of a very rough film (with the rms roughness value of 13.63 nm deposited at t = 120 min; see Fig. 6(h)). This rough surface increases the quantity of chemisorbed oxygen (see Fig. 7(a)–(c) and the ESI† for details), thus increasing the orange and red emissions in the PL spectra (see Fig. 8). The defective film then becomes responsible for the reduction in μ_e (see Fig. 2). In addition, the intensities of the shallow donor level defects, such as Zn_i and ex-Zn_i, were found to increase with an increase in t from 15 to 75 min and therefore increasing the violet and blue emissions (see Fig. 4), beyond which they decrease until the studied t of 120 min. This behaviour can be explained through the following hypothesis.

Fig. 6 AFM micrograph (1 μm × 1 μm) of Al-doped ZnO films deposited on soda lime glass substrates by RF magnetron sputtering at t varying in the range of 15–120 min (a)–(h).

Fig. 7 Deconvoluted XPS spectra for O (1s) of Al-doped ZnO films deposited on soda lime glass substrates by RF magnetron sputtering at t of (a) 15 min, (b) 60 min and (c) 120 min.

Fig. 8 Variation of the intensity ratios of the emissions from that of the excess adsorbed oxygen on the film surface (i.e., the combined orange and red emissions) to UV (I_OR/I_UV) of the Al-doped ZnO films as a function of deposition time (t).

With an increase in t (up to 75 min in this case) and at the relatively higher T_s of 623 K, due to the longer heating period, the thermally induced kinetic energy of the Al dopants increases, thus leading to a higher substitution of Al atoms in the Zn lattice sites. This increase in Al doping at Zn sites may tend to drive Zn away from their substitutional positions and eventually Zn_i and ex-Zn_i may form because of the Frenkel and further ionization reactions of the already formed Zn_i. This phenomenon can be observed from the gradual shift in the Zn peak in the XPS spectra towards a lower binding energy with an increase in t, as observed in this case with the shift in the Zn 2p_3/2 peak from 1024.05 to 1023.57 eV with an increase in t from 15 to 60 min (see Fig. 9(a) and (b)). This is because of the weakening of the binding energy of the Zn ions due to Zn–Al–O coordination as compared to the case when they are coordinated only with O²⁻ (i.e., Zn–O) in ZnO. This behaviour has also been observed for ZnO in the literature when it is doped separately by In³⁺ and Mn²⁺.^43,44 However, heating even longer (as observed for t ≥ 90 min) may lead to the formation of a higher amount of Al₂O₃ in the film (because of the higher Al–O bond strength), thus leading to less Al³⁺ substitution at Zn sites, which reduces the concentration of Al doping as compared to that at a t of 60 min. Subsequently, the amount of Al₂O₃ formed was found to be maximum in the XPS spectra for the film deposited at the maximum time of 120 min (see Fig. 7(a)–(c) and the ESI† for details). This reduces the formation of both Zn_i and ex-Zn_i defects, thus quenching the shallow donor level defects to some extent and therefore decreasing the intensities of the violet and blue emissions. Consequently, the Zn 2p_3/2 peak shifts back to a slightly higher binding energy (i.e., to 1023.77 eV at t = 120 min; see Fig. 9(c)).

Fig. 9 XPS spectra for the Zn 2p_3/2 peak for the Al-doped ZnO films deposited on soda lime glass substrates by RF magnetron sputtering at t of (a) 15 min, (b) 60 min and (c) 120 min.

4. Conclusions

In this study, an understanding of the origin and variation in the electrical properties of a thickening Al-doped ZnO thin film prepared using RF magnetron sputtering was devised. To this end, Hall effect measurements, PL measurements, AFM, XPS, XRD, FE-SEM and UV-vis-NIR spectrophotometry were used. The overall film morphology and its chemical constitutions were found to strongly influence its defect chemistry, which was responsible for determining the carrier concentration and carrier mobility values. The presence of shallow donor level defects caused by Zn and extended Zn interstitials were found to directly contribute to the carrier concentrations of the film, whereas deep donor level defects caused by oxygen vacancies and excess oxygen on the film surface did not add to the carrier generation. The combination of an increased concentration of shallow donor level defects and the absence of deep donor level defects corresponded with the best values of the carrier concentration and carrier mobility, thus leading to the lowest electrical resistivity and high material quality of the film. Two important conclusions were drawn herein: (1) strong dependency of the electrical properties on the defect densities (2) significant variation in the electrical properties of the film with thickness because of the change in the defect chemistry. This study reveals the necessity of using an appropriately thick AZO film as a TCO. Finally, a narrow PL spectrum with strong violet and/or blue emissions is noted to be an indication of a high quality TCO. This study also sheds light on the use of PL as an alternate tool to understand microstructure and electrical properties.

Acknowledgements

We gratefully acknowledge financial support from Science and Engineering Research Board (SERB), Department of Science and Technology, Government of India (Project No. SR/FTP/PS-088/2012). Furthermore, we acknowledge Professor S. S. Major (HMS Central Facility of IIT Bombay, India) for providing access to the Hall Measurement facility, Dr V. Singh (Sophisticated Instrument Centre (SIC), IIT Indore, India) for Photoluminescence measurement, Dr Gauthama (Advanced Centre for Materials Science (ACMS), IIT Kanpur, India) for X-ray Photoelectron Spectroscopy measurement and Mr S. Mughal (Shah-Schulman centre for Surface Science and Nanotechnology, Nadiad, India) for UV-vis-NIR spectrophotometer measurement.

References

1. F. Matino, L. Persano, V. Arima, D. Pisignano, R. Blyth, R. Cingolani and R. Rinaldi, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 72, 085437.
2. A. Luque and S. Hegedus, Handbook of Photovoltaic Science and Engineering, A John Wiley and Sons, 2011, pp. 717–799.
3. T. Singh, T. Lehnen, T. Leuning, D. Sahu and S. Mathur, Appl. Surf. Sci., 2014, 289, 27.
4. T. Minami, Semicond. Sci. Technol., 2005, 20, S35.
5. K. Ellmer, Nat. Photonics, 2012, 6, 809.
6. K. Mahmood, B. S. Swain and H. S. Jung, Nanoscale, 2014, 6, 9127.
7. H. Lai, T. Basheer, V. L. Kuznetsov, R. G. Egdell, R. M. J. Jacobs, M. Pepper and P. P. Edwards, J. Appl. Phys., 2012, 112, 083708.
8. S. Venkatachalam, Y. Iida and Y. Kanno, Superlattices Microstruct., 2008, 44, 127.
9. J. P. Lin and J. M. Wu, Appl. Phys. Lett., 2008, 92, 134103.
10. G. J. Fang, D. Li and B. L. Yao, Thin Solid Films, 2002, 418, 156.
11. M. D. Barankin, E. Gonzalez II, M. Ladwig and R. F. Hicks, Sol. Energy Mater. Sol. Cells, 2007, 91, 924.
12. E. G. Fu, D. M. Zhuang, G. Zhang, W. F. Yang and M. Zhao, Appl. Surf. Sci., 2003, 217, 88.
13. J. F. Chang and M. H. Hon, Thin Solid Films, 2001, 386, 79.
14. W. Yang, Z. Wu, Z. Liu, A. Pang, Y. L. Tu and Z. C. Feng, Thin Solid Films, 2010, 519, 31.
15. Z. L. Tseng, P. C. Kao, Y. C. Chen, Y. D. Juang, Y. M. Kao and S. Y. Chu, J. Electrochem. Soc., 2011, 158, J310.
16. S. S. Lin and J. L. Huang, Ceram. Int., 2004, 30, 497.
17. B. Panigrahy, M. Aslam, D. S. Misra, M. Ghosh and D. Bahadur, Adv. Funct. Mater., 2010, 20, 1161.
18. C. C. Singh, T. A. Patel and E. Panda, J. Appl. Phys., 2015, 117, 245312.
19. M. J. Lee, J. Lim, J. Bang, W. Lee and J. M. Myoung, Appl. Surf. Sci., 2008, 255, 3195.
20. T. S. Kang, T. Y. Kim, K. S. Yoon, J. M. Kim, H. S. Im and J. P. Hong, J. Mater. Chem. C, 2014, 2, 10209.
21. N. M. Flores, U. Pal, R. Galeazzi and A. Sandoval, RSC Adv., 2014, 4, 41099.
22. K. H. Tam, C. K. Cheung, Y. H. Leung, A. B. Djurisic, C. C. Ling, C. D. Beling, S. Fung, W. M. Kwok, W. K. Chan, D. L. Phillips, L. Ding and W. K. Ge, J. Phys. Chem. B, 2006, 110, 20865.
23. E. Hecht, Hecht Optics, Addison-Wesley Publishing Company, 2011, p. 72.
24. A. B. Djurisic, Y. H. Leung, K. H. Tam, Y. F. Hsu, L. Ding, W. K. Ge, Y. C. Zhong, K. S. Wong, W. K. Chan, H. L. Tam, K. W. Cheah, W. M. Kwok and D. L. Phillips, Nanotechnology, 2007, 18, 095702.
25. A. Van Dijken, E. A. Meulenkamp, D. Vanmaekelbergh and A. Meijerink, J. Lumin., 2000, 454, 87.
26. K. Vanheusden, C. H. Seager, W. L. Warren, D. R. Tallant and J. A. Voigt, Appl. Phys. Lett., 1996, 68, 403.
27. T. Yatsui, M. Ohtsu, S. J. An and J. Yoo, Opt. Rev., 2006, 13, 218.
28. J. Xu, W. Ji, X. B. Wang, H. Shu, Z. X. Shen and S. H. Tang, J. Raman Spectrosc., 1998, 615, 1997.
29. Y. Liu and J. Lian, Appl. Surf. Sci., 2007, 253, 3727.
30. C. Bundesmann, N. Ashkenov, M. Schubert, D. Spemann, T. Butz, E. M. Kaidashev, M. Lorenz and M. Grundmann, Appl. Phys. Lett., 2003, 83, 1974.
31. J. D. Ye, S. L. Gu, F. Qin, S. M. Zhu, S. M. Liu, X. Zhou, W. Liu, L. Q. Hu, R. Zhang, Y. Shi and Y. D. Zheng, Appl. Phys. A, 2004, 81, 759.
32. H. Zeng, G. Duan, Y. Li, S. Yang, X. Xu and W. Cai, Adv. Funct. Mater., 2010, 20, 561.
33. H. Chen, J. Ding, W. Guo, G. Chen and S. Ma, RSC Adv., 2013, 1, 12327.
34. F. Kayaci, S. Vempati, I. Donmez, N. Biyikli and T. Uyar, Nanoscale, 2014, 6, 10224.
35. G. Srinet, R. Kumar and V. Sajal, Ceram. Int., 2014, 40, 4025.
36. K. Mahmood, R. Munir, B. Swain, G. Han, B. Kim and H. S. Jung, RSC Adv., 2014, 4, 9072.
37. A. Kushwaha, H. Tyagi and M. Islam, AIP Adv., 2010, 3, 042110.
38. T. Chen, G. Z. Xing, Z. Zhang, H. Y. Chen and T. Wu, Nanotechnology, 2008, 19, 435711.
39. L. Li, Z. Yang, J. Y. Kong and J. L. Liu, Appl. Phys. Lett., 2009, 95, 232117.
40. Y. Zuo, S. Ge, Z. Chen, L. Zhang, X. Zhou and S. Yan, J. Alloys Compd., 2009, 47, 470.
41. M. Chen, X. Wang, Y. Yu, Z. Pei, X. Bai, C. Sun, R. Huang and L. Wen, Appl. Surf. Sci., 2000, 158, 134.
42. K. H. Ri, Y. Wang, W. L. Zhou, J. X. Gao, X. J. Wang and J. Yu, Appl. Surf. Sci., 2011, 258, 1283.
43. S. Ghosh, M. Saha and S. K. De, Nanoscale, 2014, 6, 7039.
44. U. Ilyas, R. S. Rawat, T. L. Tan, P. Lee, R. Chen, H. D. Sun, L. Fengji and S. Zhang, J. Appl. Phys., 2012, 111, 033503.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra06513a

---