Surface and Coatings Technology 174 175 (2003) 187 192 Deposition of aluminum-doped zinc oxide films by RF magnetron sputtering and study of their surface characteristics a b b a a, S.H. Jeong, S. Kho, D. Jung, S.B. Lee, J.-H. Boo * a Department of Chemistry and Institute of Basic Science, Sungkyunkwan University, Suwon 440-746, South Korea b Department of Physics and Institute of Basic Science, Sungkyunkwan University, Suwon 440-746, South Korea Abstract Transparent conductive, undoped and aluminum-doped ZnO (AZO) thin films were prepared on the glass substrates at deposition temperature in the range of room temperature (R.T.) ;300 8C by RF magnetron sputtering. Highly oriented AZO films in the w002x direction were obtained with specifically designed ZnO targets. A systematic study on the dependence of deposition parameters on the structural, optical and electrical properties of the as-grown AZO films was mainly investigated in this work. The AZO film prepared at R.T. with 4 wt.% Al(OH) 3 doped a ZnO target under a target-to-substrate distance (D ts) of y2 45 mm, has not only a high transmittance of 85% at the visible region but has also a resistivity of 9.8=10 VØcm. In addition, the resistivity of AZO films increases with increasing T sub. We investigated that this tendency was altered after 4 wt.% doping. 2003 Elsevier Science B.V. All rights reserved. Keywords: AZO film; RF magnetron sputtering; Surface characteristic 1. Introduction Zinc oxide (ZnO) films have been investigated in recent years as transparent conducting oxide layers, because of their good electrical and optical properties in combination with large band gap, abundance in nature, and absence of toxicity. In addition, the ZnO films could be deposited at relatively low deposition temperature w1x and good stability (in H2 plasma) w2x. The electrical conductivity of ZnO is controlled by intrinsic defect, i.e. oxygen vacancies andyor zinc interstitials w3x, which act as n-type donors. The resistivity is lowered further by extrinsic doping with group III elements such as B, Al, Ga or In. Amongst them, Al-doped ZnO thin films are considered as candidate materials for organic electroluminescence displays w4x. The physical properties of ZnO films are generally dependent on deposition parameters w5x and post-deposition conditions such as postannealing, etc. w6x. However, there are not many reports on the systematic study of deposition parameter effect and film properties. Therefore, in this report we describe the effect of sputtering parameters on the structural, electrical, and *Corresponding author. Tel.: q82-31-290-7072; fax: q82-31-290-7075. E-mail address: jhboo@chem.skku.ac.kr (J.-H. Boo). optical properties of un-doped ZnO and Al doped ZnO thin films. 2. Experimental AZO thin films were deposited on glass substrates in an RF magnetron sputtering system, and the targets used in the experiment were specifically designed using high purity zinc oxide (99.99%) and aluminum hydroxide (99.99%) powders. The sputtering system was pumped y6 down to 3=10 Torr using a turbo molecular pump. The working pressure which mainly consisted of a highpurity Ar(99.99%) gas was 32 mtorr. The glass substrates were ultrasonically cleaned in sequentially distilled water, acetone, alcohol, alkaline solution, distilled water and finally dried with nitrogen gas. AZO films were deposited on the substrate at temperature range of R.T. to 300 8C (especially at R.T. and 250 8C) with RF power of 150 W after cleaning the targets with Ar plasma for 5 min. Due to the practical applicability of AZO as a transparent conductive film, 150 200-nmthick samples were typically prepared for optical measurements. The crystal structure and the surface morphologies were observed using X-ray diffraction (XRD) and atomic force microscope (AFM) under contact mode. The growth rate and the thickness of the 0257-8972/03/$- see front matter 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/s0257-8972(03)00600-5
188 S.H. Jeong et al. / Surface and Coatings Technology 174 175 (2003) 187 192 Fig. 1. Typical XRD pattern of AZO film deposited on glass substrate at room temperature (a) and high resolution XRD pattern (b) as a function of wt.% of Al. The insets of each figures shows XRD data of targets withywithout dopant of Al(OH) 3 (a) and of different substrate temperature (b). samples were determined by cross-sectional scanning electron microscope (SEM) and X-ray photoelectron spectroscopy (XPS) was also utilized to analyze the chemical state of AZO films. Optical transmittance measurements were performed with a UV visible spectrophotometer. The resistivities were measured by a four-point probe method. 3. Results and discussion 3.1. Structural characterization Fig. 1a shows the typical XRD pattern of a ZnO thin film and in the inset of Fig. 1a is also shown the XRD patterns obtained from the specially designed un-doped ZnO and Al-doped ZnO targets for comparison. Only the (002) peak in Fig. 1a is observed at 2us34.308 indicating that all of the obtained films had a preferred orientation with the c-axis perpendicular to the substrate. A similar XRD pattern to the un-doped ZnO target was obtained from a target with below 2 wt.% Al(OH). 3 Above 2 wt.% Al(OH) 3 doped ZnO target, however, the shoulder peaks attributed to aluminum oxide diffraction appeared (see inset of Fig. 1a). This resulted from a target fabricated at high temperature (1200 8C) under high Ar atmosphere pressure. However, in the case of AZO film deposition, neither metallic Zn or Al characteristic peaks nor aluminum oxide peak are observed. As shown in Fig. 1a, there are two diffraction peaks due to AZO(002) and AZO(004) phases suggesting that the AZO films were grown with a highly oriented w002x direction. The obtained 2u values are slightly less than that of the standard ZnO crystal (34.458). The reason for the occurrence of different diffraction angles between bulk ZnO crystal and ZnO films could be explained by the following. Kroynagi w7x observed that the diffraction angle of the (002) plane of AZO film shifted towards lower angle due to ionization of ZnO rather than Al doping. Fig. 1b shows high-resolution XRD patterns of AZO (002) plane. From this figure, the peak position of the (002) plan is shifted to higher 2u values with increasing variation of Al content to 4 wt.%. This means that the lattice parameter of ZnO was decreased in the c-axis with Al doping. Above 4 wt.%, the peak remains constant up to 8 wt.% Al content. However, there was a drastic change of the AZO(002) peak intensity with increasing Al content. This phenomenon can be explained in two ways. The first reason is that the decrease of the XRD peak with changing 2u values is reflected from single crystal to amorphous crystal. The second reason is that the decreasing tendency of the XRD peak without changes in 2u values is attributed to small crystal formation. This indicates that below 4 wt.%, the Al-doping mechanism of Si or other materials is quite different compared to that beyond 4 wt.%. Park et al. w8x reported that they reasoned that the ionic radii 2q 3q of Zn and Al are 72 and 53 pm, so the length of the c-axis is expected to be shorter, respectively, and that Al atoms are substituted into the Zn site in the crystal. In our case, this phenomenon was apparent up to 4 wt.% Al doped ZnO targe. Moreover, the intensity of (002) peaks gradually decreases and broadens with increasing dopant. It reasoned that the crystallinity of the Al atom substituted into the Zn site decreased with increasing dopant. The inset of Fig. 1b shows the XRD spectra for ZAO films prepared on glass with 4 wt.% AZO target at R.T. and 250 8C, respectively. Only the (002) diffraction peak is observed. The full width half maximums (FWHMs) of the peak also changed from 0.329 to 0.3068 with increasing substrate temperature (T ). It is sub worth noting that with increasing substrate temperature, the FWHM as well as location of the measured diffrac-
S.H. Jeong et al. / Surface and Coatings Technology 174 175 (2003) 187 192 189 smaller with increasing Al contents. This result agrees well with XRD data shown in Fig. 1b. 3.2. Optical properties For most optical applications a high transmittance in the visible range is very important. We measured that the average transmittance in the visible range was over 85% for all samples regardless of the Al(OH) content 3 in the target. As Al(OH) 3 content increases, the absorption edge shifted to a shorter wavelength region. The variation of optical band gap (E ) for the as-grown 2 films can thus be obtained by plotting a vs. hn (a is the absorption coefficient and hn is the photon energy) from UV transmittance data (not shown). Eg is then determined by the extrapolation method. Fig. 3a shows the variation of optical band gap as a function of Al contents. It shows that the band gap widens with increasing Al contents up to 4 wt.% dopant. The reason g Fig. 2. Variations of growth rate as a function of target-to-substrate distance (D ts)(a) and Al(OH) 3 content as well as substrate temper- ature (b); (c) and (d) show the AFM images of undoped ZnO film (c) and 4 wt.% Al-doped ZnO film (d). tion peaks are not significantly shifted compared with those of Al content. This is caused by the crystallinity of the resulting films that have improved and formed a more stoichiometric oxide with increasing T. sub Fig. 2a shows the growth rate of AZO films as a function of the target-to-substrate distance (D ts). Since the best quality of as-grown AZO was obtained under the condition of Dts of 45 mm at 4 wt.%, we fixed the distance to be 45 mm. However, we found that there is no significant influence of growth rate on Al(OH) 3 content and T sub (see Fig. 2b). Fig. 2c,d shows AFM images of the ZAO films prepared with un-doped and 4 wt.% AZO targets, respectively. It shows the large variation of surface roughness and grain size. The grain size of the un-doped ZnO film (see Fig. 2c) is larger than that of 4 wt.% Al-doped ZnO film (Fig. 2d), suggesting that the particles in ZnO films become Fig. 3. (a) Change of optical band gap of AZO films with different Al(OH) 3 content. (b) XP survey spectra of 4 wt.% AZO film before (under curve) and after Ar sputtering for 10 min (upper curve).
190 S.H. Jeong et al. / Surface and Coatings Technology 174 175 (2003) 187 192 Fig. 4. High resolution XP spectra of AZO films prepared with 4 wt.% Al content at R.T. and 250 8C: (a) Zn 3p 3p3y2 250 8C; and (d) O 1s, 250 8C. 3y2, R.T.; (b) O 1s, R.T.; (c) Zn is that the density of electrons decreased after the Al 3q 2q ion was substituted into the Zn ion site in the films. This means that the valence level was lowered. However, beyond 4 wt.% dopant the energy band gap becomes narrow because the excess Al atoms are segregated into the grain boundaries. These segregated Al atoms do not act as dopant. This broadening effect of optical band gap width can be understood based on the Burstein effects w9x. 3.3. Chemical state on the surface Fig. 3b shows XP survey spectra obtained beforey after Ar sputtering for a AZO film prepared at R.T. with 4 wt.% Al(OH) 3 doped ZnO target. Only zinc and oxygen peaks are observed, before Ar sputtering, and there is no contamination except for carbon. However, the intensity of C 1s component is found to decrease rapidly with Ar sputtering (see also Fig. 3b) with XPS study only, an attempt to detect aluminum in the prepared AZO films and thereby predicting its chemical state are not successful because of (i) the low concentration of aluminum in the matrix and (ii) the low values of the ionization cross-sections of aluminum. Additional difficulties are caused by the presence of X-ray satellites on the low binding energy side of the Zn 3p photoelectron peak w10x. Therefore, in this study we also used the EDX method to detect the Al-content in the films and the EDX results revealed that Al-content in the film can be detected above 2 wt.%. Fig. 4 shows high-resolution XP spectra for Zn 2p 3y2 (Fig. 4a,c) and O 1s (Fig. 4b,d). We mainly studied the core line of Zn 2p3y2 of Zn 2p3y2 which appeared due to the film s surface area, exhibits high symmetry. In the case of Fig. 4a,c, the binding energy of Zn 2p 3y2 obtained from the AZO films prepared, regardless of Tsub is observed as 1022.20"0.10 ev. This value is larger than that of Zn 2p3y2 in the bulk ZnO, and the metallic Zn peak (1021.50 ev). It also confirms that the vast majority of Zn atoms remain in all probed films 2q in the same formal valence states of Zn within an oxygen deficient ZnO1yx matrix w11x. In contrast with Zn 2p 3y2, the O 1s spectra for both specimens were found to consist of two components as shown in Fig. 4b,d. Firstly, the higher binding energy component (i.e. shoulder peak) at 532.30"0.15 ev is
S.H. Jeong et al. / Surface and Coatings Technology 174 175 (2003) 187 192 191 temperature may have high oxygen vacancy rather than that grown at high temperature, suggesting high conductivity (i.e. low resistivity). Therefore, change in the intensity of the higher binding energy component may only be related to variation of the concentration of stoichiometric oxygen in ZnO. Consequently, the oxygen deficiency was decreased with increasing T sub. It is known that more oxygen-deficient film commonly exhibits a lower resistivity because of the carrier increasing with increasing oxygen vacancies w11x. 3.4. Electrical properties In the experimental data of Fig. 5b, with increasing Al content from 0 to 4 wt.%, the resistivity was y1 decreased from 5.0=10 VØcm to a minimum value y2 of 9.8=10 VØcm. However, the resistivity of the AZO film prepared with )6 wt.% Al(OH) 3 doped ZnO target has been increased up to 12 VØcm. We already mentioned this reason in Section 3.3. In addition, the resistivity of AZO films increased with increasing T sub. It was pointed out in the XRD data, the crystallinity of the resulting films is improved and the films contained a more stoichiometric oxide with increasing T sub. 4. Conclusions Fig. 5. (a) Depth profiling of atomic ratio of Zn and O for AZP films grown at R.T. and 250 8C. (b) Plot of the resistivity of the AZO films grown on different temperature as a function of Al(OH) content. 3 attributed to the presence of loosely bound oxygen or non-stoichiometry oxygen on the AZO film surface and partially belongs to a specific species, e.g. CO3 or adsorbed H2O, and O 2. Secondly, the lower binding energy component (i.e. main peak) of the O 1s spectra at 531.25"0.15 ev is associated with stoichiometric 2y oxygen bonded as O ions within the matrix of ZnO. After 10 min Ar sputtering, however, the intensity of higher binding energy component is more rapidly decreased in the case before Ar sputtering. In detail, the intensity of higher binding energy component in Fig. 4d (at 250 8C) is higher than that in Fig. 4b (at R.T.) after Ar sputtering for 10 min. This means that the shoulder O 1s peak appeared due mainly to the Al-doping to the ZnO crystals since we used Al(OH) 3 as dopant with annealing at 1200 8C. With these spectra of Zn 2p 3y2 and O 1s, considering their atomic sensitivity, one can get the atomic ratio of ZnyO to be 1.12 at R.T. and 1.02 at 250 8C after Ar sputtering for 5 min (see Fig. 5a). This indicates that the AZO film grown at low Highly preferred (002) orientation AZO films were prepared by RF magnetron sputtering with home-made ZnO targets containing different amounts of Al(OH) 3 powder as doping source. As deposited films with 4 wt.% Al(OH) doped ZnO target at R.T. under D of 3 ts y2 45 mm have resistivities as low as 9.8=10 VØcm with 85% transmittance in the visible region. This deposited film had a low valence level after the Al atoms were substituted into the Zn site in the films prepared with up to 4 wt.% dopant. Excess Al atoms are segregated into the grain boundaries beyond 4 wt.% dopant. This effectively lowers the optical band gap again. We investigated such a tendency was changed before and after 4 wt.% doping. In addition, the resistivity of AZO films increases with increasing T sub. With decreasing ZnyO ratio in AZO films, crystallity increased with increasing T sub. Acknowledgments This work was supported by the BK21 project of the Ministry of Education, Korea and by the Center for Advanced Plasma Surface Technology at the Sungkyunkwan University. References w1x D.H. Zhang, T.L. Yang, J. Ma, Appl. Surf. Sci. 158 (2000) 43.
192 S.H. Jeong et al. / Surface and Coatings Technology 174 175 (2003) 187 192 w2x W.S. Lan, S.J. Fonash, J. Electron. Mater. 16 (1987) 141. w3x S. Major, K.L. Chopra, Sol. Energy Mater. 17 (1998) 319. w4x J. Zhao, S. Xie, S. Han, Z. Yang, L. Ye, T. Yang, Synth. Met 114 (2000) 251. w5x K. Ellmer, J. Phys. D 33 (2000) R17. w6x A. Malik, A. Seco, R. Nunes, E. Fortunato, R. Martins, Mater. Res. Soc. Symp. 471 (1997) 47. w7x A. Kuroyanagi, J. Appl. Phys. 66 (1989) 5492. w8x K.C. Park, D.Y. Ma, K.H. Kim, Thin Solid Films 305 (1997) 201. w9x E. Burstein, Phys. Rev. 93 (1954) 632. w10x M.N. Islam, T.B. Ghosh, K.L. Chopra, H.N. Acharya, Thin Solid Films 280 (1996) 20. w11x M. Chen, Z.L Pei, C. Sun, L.S. Wen, X. Wang, Mater. Lett. 48 (2001) 194.