Abnormal photoluminescence of ZnO thin film on ITO glass - CiteSeerX

Materials Letters 61 (2007) 201 – 204 www.elsevier.com/locate/matlet

Abnormal photoluminescence of ZnO thin film on ITO glass Xuemei Teng, Hongtao Fan, Shusheng Pan, Cong Ye, Guanghai Li ⁎ Key Laboratory of Materials Physics, Anhui Key Laboratory of Nanomaterials and Nanotechnology, Institute of Solid State Physics, Chinese Academy of Sciences, Hefei 230031, PR China Received 19 December 2005; accepted 6 April 2006 Available online 5 May 2006

Abstract The photoluminescence (PL) properties of ZnO thin films on ITO glass substrate deposited by rf magnetron sputtering with different oxygen partial pressures were studied. It was found that the exciton related emission of ZnO thin films depends on oxygen partial pressure, and that the visible emission related to intrinsic defects has no obvious change with various oxygen partial pressures. Abnormal UV-PL characteristics were observed, and its intensity was obviously enhanced. The emission position has a strong red-shift with increasing excitation intensity, and the emission intensity increases notably with increasing excitation cycle. © 2006 Elsevier B.V. All rights reserved. Keywords: ZnO thin film; Photoluminescence; ITO glass; Sputtering

1. Introduction Due to its wide band gap (3.37 eV) and large exciton binding energy (60 meV), ZnO is suitable for the production of light emitting devices and a promising candidate for the next generation of electronic devices [1,2]. ZnO thin films play an important role in solid-state display devices, solar cells and exciting acoustic waves at microwave frequencies [3,4]. Therefore, an accurate knowledge of the structural and optical properties of ZnO is indispensable for the design and analysis of various optical and optoelectronic devices. The structural and optical properties of ZnO thin films depend on the preparation methods, substrate temperature, substrate material and subsequent annealing treatment [5–9]. Above all, the selection of substrate is very important for the growth of thin film because the matching in lattice parameters and crystal structure between the film and substrate strongly affects the crystal growth behavior of the film [10]. Indium tin oxide (ITO) glass substrate has many excellent properties, such as electrically conductive and optically transparent, high Vis-Nir light transmission, uniform transmission homogeneity and reflection in the infrared range. Since there is a small lattice mismatch (3%) between the neighboring oxygen–oxygen (O–O) distance on the closet-packed ITO (111) and ZnO (0001) planes [11], and O dangling bonds on the ⁎ Corresponding author. E-mail address: [email protected] (G. Li). 0167-577X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.matlet.2006.04.031

ITO layer surface, thus it benefits the initial nucleation and the subsequent growth of high quality ZnO films on ITO substrate. Up to date, a few studies have reported on the growth of the photoluminescent ZnO thin film on ITO glass substrate. In this paper, the influence of the ITO buffer layer and oxygen partial pressure on the structural and photoluminescence properties of ZnO thin films were reported, and abnormal PL features were observed and discussed. 2. Experimental ZnO thin films with thickness of about 200 nm were deposited on ITO glass substrate at a constant substrate temperature of 400 °C by rf magnetron sputtering from high purity ZnO target (99.999%). The chamber was firstly pumped to a base vacuum better than 5× 10− 6 Pa, and then filled with a mixture of Ar and O2 through independent mass flow controllers, in which the oxygen partial pressure, PO2 [=O2 /(O2 +Ar)], changed from 0% to 100%. The sputtering was carried out at a constant gas working pressure of 1 Pa, rf power of 80 W and bias voltage of 50 V. Since the growth rate of the ZnO thin film decreases with increasing PO2, and thus the proper selection of the growth time for different oxygen partial pressures is needed in order to obtain ZnO films with approximately the same thickness. To study the crystalline structure of ZnO thin films, X-ray diffraction (XRD) was carried out on Philips X'Pert PRO

202

X. Teng et al. / Materials Letters 61 (2007) 201–204

diffractometer (40 KV, 40 mA) using Cu Kα line (λ =0.15406 nm). Field-emission scanning electron microscopy (FE-SEM, FEI Sirion 200) was used to study the morphology of ZnO thin films. The optical properties of the films were investigated by the use of photoluminescence (PL). PL spectra were measured on a LABRAM-HR spectrometer (He–Cd laser) with an excitation wavelength of 325 nm and an excitation intensity of 2000 W/cm2 at room temperature. 3. Results and discussion To optimize the deposition conditions, the influence of PO2 on the structure of deposited ZnO thin films was studied firstly. Fig. 1 shows the XRD patterns of ZnO thin films deposited on ITO glass with different PO2. One can see that only one strong (0002) peak of hexagonal wurtzite ZnO appears (JCPDS NO.36-1451), which suggests that the obtained ZnO films exhibit (0002) preferential orientation with the C-axis perpendicular to the substrate surface. In addition, the peak intensity firstly increases and then decreases with gradually increasing PO2. The initial increase may result from the improved stoichiometry of the films, associated with the incorporation of oxygen at oxygen vacancies. However, exorbitant PO2 might worsen the stoichiometry of the films by introducing interstitial oxygen or zinc vacancies. It has been reported that the decreased peak intensity with increase in the oxygen partial pressure might be due to the formation of new ZnO type of crystals, but in this study there is no evidence for the existence of new ZnO type of crystals. The previous report has indicated that the position of ZnO (0002) peak in XRD is related to stoichiometric condition, and the deviations from ideal stoichiometry might be expected to induce lattice defects [12]. The angular peak positions of ZnO films are less than the bulk value (2θ=34.43°) [13], indicating a uniform state of stress with tensile components parallel to the C-axis. The strain along the C-axis, εzz is given by the following equation [14]: ezz ¼ ðC−C0 Þ=C0 100%

ð1Þ

where C is the lattice parameter of the strained ZnO films calculated from XRD and C0 (5.2066) is the unstrained lattice parameter of ZnO [14]. At the same time, the strain can be positive (tensile) or negative (compressive). Fig. 2 shows the lattice parameter C, grain size d, and strain εzz of ZnO thin films on ITO glass with different PO2. The lattice parameter firstly decreases with increasing PO2, after reaching the minimum value at 60% PO2 and then slightly increases with further increasing PO2. This may be attributed to the improvement in stoichiometry of the film with increasing

Fig. 1. XRD patterns of ZnO thin films on ITO glass with different oxygen partial pressures.

Fig. 2. Lattice parameter, C, grain size, d, and strain, εzz, of ZnO thin films on ITO glass with different oxygen partial pressures.

PO2. The strain value of ZnO thin films on ITO glass substrates is tensile and reaches the minimum value at 60% PO2, differing from the compressive strain of ZnO films on Si substrate [10]. The grain size evaluated from the well-known Scherrer's relation decreases with increasing strain, which is in agreement with the results reported by Ghosh et al. [10]. This may come from the retarded crystal growth due to the stretched lattice that can increase the lattice energy and diminish the driving force of growth. The FE-SEM observations have also shown that the ZnO grain size increases with increasing PO2, see Fig. 3. Fig. 4 shows the PL spectra of ZnO thin films on ITO glass with different PO2 (with the same excitation intensity P=2000 W/cm2). One can see that the PL spectra have a strong UVemission and a weak broad band in the visible region (green emission), being similar to the result by oxidizing the metallic Zn [15]. However, it is different from the PL spectrum of ZnO deposited on ITO coated glass substrate by electrophoretic deposition method, which only gave a strong emission at 390 nm [16]. The intensity of the UV peak situated at about 395 nm firstly increases with increasing PO2 due to the increased oxygen stoichiometric, and reaches the maximum at 60% PO2 and then decreases. Based on the relationship between the grain size and PO2 (see Fig. 2), one can see that the PL intensity of exciton related states was noticeably enhanced with increasing grain size, which is consistent with the results observed in ZnO films grown by molecular beam epitaxy (MBE) [5]. This trend can be understood that surface-to-volume ratio becomes smaller with increasing grain size, and the larger grains have smaller nonradiative relaxation rates over the surface states resulting in the enhancement of PL intensity. The influence of excitation intensities on UV-PL of ZnO thin films on ITO glass (60% PO2) is shown in Fig. 5(a) (P=2000, 500 and 200 W/cm2).

Fig. 3. FE-SEM images of ZnO thin films on ITO glass with oxygen partial pressure of (a) 20% and (b) 60%.


Fig. 4. PL spectra of ZnO thin films on ITO glass with different oxygen partial pressures.

It can be observed that UVemission intensity increases with the increase of the excitation intensity, which is in agreement with the previously reported result for ZnO nanowires [17], while the peak position of the emission band shifts from 383 (3.24 eV) to 395 nm (3.14 eV) by 100 meV, and this redshift is different from the results observed in semiconductor quantum dots and quantum well. Further PL measurement indicates that the emission intensity of ITO glass increases slightly and the peak position has no shift with the increased excitation intensity (P=2000, 500 and 200 W/cm2). Since the exciton density increases with excitation intensity, different exciton related emissions might occur. In the intermediate density regime, biexcitonic, exciton–exciton, and exciton–carrier emissions may be observed. Biexcitonic emissions are only observed at cryogenic temperatures. At higher temperatures exciton ionization increases the free carrier densities, and thus an increased probability of exciton–carrier emission. At high excitation intensities, the dominated peak can be attributed to electron hole plasma recombination shifted by band gap renormalization [18]. Since the energy shift with excitation intensity is too high to attribute this transition completely to exciton emission, the transition may also come from the recombination located at the interface between ZnO and ITO glass substrate.

203

The Gaussian fitting of the UV-PL band (P= 2000 W/cm2) gives three peaks situated respectively at 380, 395 and 410 nm, as shown in Fig. 5(b). It is evident that at high excitation intensity, the spectrum is dominated by the low-energy component, i.e. the peak at 395 nm, but low excitation intensity is favorable for the high-energy component, i.e. the peak at 380 nm. It is well-known that the UV emission peak at 380 nm originates from the recombination of free excitons through an exciton–exciton collision process corresponding to near band edge (NBE) emission of wide band gap ZnO [1]. The emission at 3.1 eV (400 nm) has been frequently observed in natural silica and oxygen-deficient silica glass [19,20]. In the present study the PL peak from the ITO layer is situated at 395 nm with a wider full width at half maximum (FWHM), as shown in Fig. 5(c) (P=2000 W/cm2). Sekiguchi et al. [21] have observed the blue emission at about 413 nm in ZnO thin films grown under oxygen rich condition by the chemical vapor deposition method, and attributed the PL to the possible existence of cubic ZnO, which might exist near the substrate interface in ZnO thin films [22]. Because there is no evidence for the existence of cubic ZnO from the XRD analysis in our samples, the O dangling bonds on ITO surface layer or the interface between substrate and ITO might contribute to the anomalous emission (410 nm, 3.0 eV). This result indicates that the PL peak position of ZnO thin film on ITO glass depends on the excitation intensity, and could be adjusted by the PL from ITO buffer layer. Fig. 5(d) presents the PL spectra of ZnO thin film on ITO glass substrate (60% PO2) measured at different cycles with the same excitation intensity (P= 2000 W/cm2). One can see that the intensity of the UV-PL peak increases notably with the excitation cycles and nearly reaches the saturated level after about eight cycles, while that in the visible region almost does not change. It was found that the PL intensity of ITO glass has no change with increasing excitation cycle under the same excitation intensity (P=2000 W/ cm2), and thus the increased UV-PL peak intensity of ZnO thin film on ITO glass substrate originates from the ZnO itself. Laser annealing has been frequently used in the semiconductor industry for dopant activation and defects elimination. In our case, the laser irradiation during the PL measurement can be regarded as a laser annealing process, which can decrease the nonradiative centers and extended defects in ZnO thin film, which thus increases the luminescence efficiency of the films at the following cycles. Note that such laser annealing effect can enhance both the peak intensity and integrated photoluminescence efficiency at the UVregion, up to a factor

Fig. 5. PL spectra of ZnO thin films on ITO glass (60% PO2): (a) excited with different excitation intensities (W/cm2); (b) Gaussian fitting of the top PL band in (a) (P = 2000 W/cm2); (c) normalized PL spectra of curve (1) ZnO film on Si substrate, curve (2) ZnO film on ITO substrate and (3) ITO substrate (P = 2000 W/cm2); (d) excited at different excitation cycles.

204


of 2–3 compared to the as-grown thin films, without any spectral shift of the peak position. It is noticeable that the PL characteristics of ZnO thin films on ITO substrate are strongly dependent on the film stoichiometry, which is consistent with the previous studies [23,24]. The emission intensity is determined by the radiative and nonradiative transition. In the case of ZnO, the nonradiative transition is induced by crystal imperfections, such as point defects, dislocation and grain boundaries. As well-known, an important characteristic of ZnO materials is nonstoichiometry due to native point defects like interstitial zinc or oxygen vacancies, which are generated by thermal variations in the oxygen (or zinc) pressure and temperature during deposition [25]. Besides ZnO has a tendency to lose its oxygen and become nonstoichiometric [25,26], the chemical component of ZnO films prepared by rf sputtering is nonstoichiometric, and usually consists of oxygen vacancies. Therefore, there are many lattice defects and surface defects in ZnO thin films, and these defects produce various nonradiative centers and reduce the emission intensity from ZnO thin films. It is reasonable that the number of oxygen vacancies and Zn interstitials is reduced and the stoichiometry of ZnO films is improved with the proper PO2, which is proved by the enhanced PL intensity with proper PO2 and increasing excitation cycle.

4. Conclusion In summary, the C-crystal axis orientated ZnO thin films on ITO glass substrate have been grown by rf magnetron sputtering. PL characteristics of ZnO thin films depend on the oxygen partial pressure, and ZnO film grown at 60% PO2 has the maximum PL intensity. The UV-PL peak intensity increases and the peak position shifts to long wavelength with the increase of excitation intensity. The PL intensity and integrated PL efficiency of ZnO thin films on ITO glass increase with increasing excitation cycle due to the decrease of nonradiative centers and extended defects in the films upon laser annealing. Acknowledgment This work was supported by the National Natural Science Foundation of China (No. 10474098).

References [1] Y.C. Kong, D.P. Yu, B. Zhang, W. Fang, S.Q. Feng, Appl. Phys. Lett. 78 (2004) 407. [2] Z.F. Liu, Z.G. Jin, W. Li, X.X. Liu, J.J. Qiu, W.B. Wu, Mater. Lett. 60 (2006) 810. [3] G.C. Yi, C.R. Wang, W.I. Park, Semicond. Sci. Technol. 20 (2005) S22. [4] R. Triboulet, J. Perriere, Prog. Cryst. Growth Charact. 47 (2003) 65. [5] T. Matsumoto, H. Kato, K. Miyamoto, M. Sano, T. Yao, Appl. Phys. Lett. 81 (2002) 1231. [6] M. Zerdali, S. Hamzaoui, F.H. Teherani, D. Rogers, Mater. Lett. 60 (2006) 504. [7] T.W. Kim, K.D. Kwack, H.K. Kim, Y.S. Yoon, J.H. Bahang, H.L. Park, Solid State Commun. 127 (2003) 635. [8] I.S. Kim, S.H. Jeong, S.S. Kim, B.T. Lee, Semicond. Sci. Technol. 19 (2004) L29. [9] K. Keem, H. Kim, G.T. Kim, J.S. Lee, B. Min, K. Cho, M.Y. Sung, S. Kim, Appl. Phys. Lett. 84 (2004) 4376. [10] R. Ghosh, D. Basak, S. Fujihara, J. Appl. Phys. 96 (2004) 2689. [11] C.H. Yi, I. Yasui, Y. Shigesato, Jpn. J. Appl. Phys. 34 (1995) 1639. [12] S. King, J.G.E. Gardeniers, I.W. Boyd, Appl. Surf. Sci. 96–98 (1996) 811. [13] American Standard for Testing of Materials (ASTM) 36-1451. [14] H.C. Ong, A.X.E. Zhu, G.T. Du, Appl. Phys. Lett. 80 (2002) 941. [15] S.L. Cho, J. Ma, Y.K. Kim, Y. Sun, G.K.L. Wong, J.B. Ketterson, Appl. Phys. Lett. 75 (1999) 2761. [16] Z.J. Wang, Z.J. Wang, S.H. Li, Z.H. Wang, Y.M. Lv, J.S. Yuan, Chin. Phys. 13 (2004) 750. [17] S.C. Lyu, Y. Zhang, H. Ruh, H.J. Lee, H.W. Shim, E.K. Suh, C.J. Lee, Chem. Phys. Lett. 363 (2002) 134. [18] D.M. Bagnall, Y.F. Chen, Z. Zhu, T. Yao, M.Y. Shen, T. Goto, Appl. Phys. Lett. 73 (1998) 1038. [19] M. Cannas, M. Barbera, R. Boscaino, A. Collura, F.M. Gelardi, S. Varisco, J. Non-Cryst. Solids 245 (1999) 190. [20] Y. Sakurai, J. Non-Cryst. Solids 271 (2000) 218. [21] T. Sekiguchi, K. Haga, H. Inaba, J. Cryst. Growth 214/215 (2000) 68. [22] Y. Xin, P.D. Brown, C.B. Boothroyd, A.R. Preston, C.J. Humphreys, T.S. Cheng, C.T. Foxon, A.V. Andrianov, J.W. Orton, Mater. Res. Soc. Symp. Proc. 423 (1996) 311. [23] Z.K. Tang, G.K.L. Wong, P. Yu, Appl. Phys. Lett. 72 (1998) 3270. [24] G.B. Zhang, C.S. Shi, Z.F. Han, J.Y. Shi, Z.X. Fu, M. Kirm, G. Zimmerer, Chin. Phys. Lett. 18 (2001) 441. [25] D.H. Zhang, PhD Thesis University of Waterloo (1993). [26] D.H. Zhang, Mater. Chem. Phys. 45 (1996) 248.