007) 8718–8722www.elsevier.com/locate/tsf

Thin Solid Films 515 (2

Effect of substrate temperature on the structure and optical properties of ZnOthin films deposited by reactive rf magnetron sputtering

Sukhvinder Singh a, R.S. Srinivasa b, S.S. Major a,⁎

a Department of Physics, Indian Institute of Technology Bombay, Mumbai-400076, Indiab Department of Metallurgical Engineering and Materials Science, Indian Institute of Technology Bombay, Mumbai-400076, India

Available online 6 April 2007

Abstract

Zinc Oxide films were deposited on quartz substrates by reactive rf magnetron sputtering of zinc target. The effect of substrate temperature onthe crystallinity and band edge luminescence has been studied. The films deposited at 300 °C exhibited the strongest c-axis orientation. AFM andRaman studies indicated that the films deposited at 600 °C possess better overall crystallinity with reduction of optically active defects, leading tostrong and narrow PL emission.© 2007 Elsevier B.V. All rights reserved.

Keywords: ZnO; rf magnetron sputtering; Microstructure; Optical properties

1. Introduction

ZnO is a versatile, wide band gap semiconductor with largeexciton binding energy (60 meV) and interesting piezoelectricand ferroelectric properties. Its high chemical and thermalstability and abundance make it an attractive material for a widevariety of applications, such as, UV emitters and detectors,SAW devices, gas sensors and transparent conducting electro-des [1]. Several growth techniques, such as, spray pyrolysis,sputtering, pulsed laser deposition, MBE and MOCVD havebeen extensively used for the deposition of un-doped and dopedZnO films, among which, sputtering has been the most widelyused. Most of this work has been reviewed in Refs. [1] and [2].High quality epitaxial films of un-doped ZnO have usually beendeposited on single crystalline substrates, such as sapphire byMBE [3,4], MOCVD [5,6], PLD [7,8] and sputtering [9–11].However, reports of sputtering of ZnO films on amorphoussubstrates such as quartz and glass with properties comparableto those of epitaxial films are limited [12–14]. In this work, wereport the sputtering of un-doped ZnO films of high opticalquality deposited on fused quartz substrates. In particular, the

⁎ Corresponding author. Tel.: +91 22 25767567; fax: +91 22 25767552.E-mail address: syed@phy.iitb.ac.in (S.S. Major).

0040-6090/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.tsf.2007.03.168

effect of substrate temperature on the crystallinity and bandedge luminescence has been studied.

2. Experimental details

ZnO films were deposited on quartz substrates by reactive rfmagnetron sputtering.A 99.9%pure Zn target of 3-in diameterwasused. The target to substrate distance was 55 mm. The basepressure was 1×10−5 mbar. The flow rates of argon (24 sccm) andoxygen (6 sccm) were controlled by mass flow controllers.Deposition was carried out at a working pressure of 10−2 mbar,after pre-sputtering with argon for 10 min. The sputtering powerwas maintained at 400W during deposition. The depositions werecarried out in the temperature range of room temperature to 600 °C.

X-ray diffraction (XRD) studies were performed with aPANalytical X'Pert PRO powder diffractometer using Cu Kα

radiation. Specular transmittance (T) and reflectance (R)measurements at normal incidence were carried out with aPerkinElmer Lambda-950 UV/Visible spectrophotometer.Photoluminescence (PL) measurements were carried out atroom temperature using a 325 nm He–Cd laser and a JOBINYVON HR-460 monochromator. Raman spectra of the filmswere recorded at room temperature in back scattering geometry,using JOBIN YVON HORIBA HR-800 instrument equippedwith a 20 mW Ar+ laser (514.5 nm). Ambios XP-2 surface

8719S. Singh et al. / Thin Solid Films 515 (2007) 8718–8722

profilometer was used for thickness measurements. AFM studieswere carried out in contact mode with silicon nitride probesusing Digital Instruments Nanoscope IV Multimode SPM.

3. Results and discussion

The growth rate of films increased from ∼1.0 μm/h at roomtemperature, to ∼1.4 μm/h at 100 °C and remained unchangedat higher substrate temperatures. All the films were approxi-mately of the same thickness (600±50 nm). XRD patterns oftypical ZnO films are shown in Fig. 1. The films deposited atroom temperature and 100 °C showed a strong peak at 2θ ∼34°and a weak peak ∼72°, which were identified respectively as(0002) and (0004) reflections of hexagonal ZnO, indicating, astrong c-axis orientation of crystallites. However, the filmsdeposited in the temperature range of 150–250 °C (a typicalcase of 200 °C is shown in Fig. 1) exhibited multiple reflections(as indexed in Fig. 1), though with a dominant (0002) peak.Interestingly, the films deposited at 300 °C and highertemperatures, again showed peaks corresponding to only(0002) and (0004) reflections. The intensity variation of

Fig. 1. XRD patterns of ZnO films deposited at different substrate temperaturesas indicated in the figure.

Fig. 2. (a) Strain and (b) crystallite size along c-axis plotted against substratetemperature.

(0002) peak with substrate temperature shown in Fig. 1 alsoreveals an interesting pattern. The film deposited at 300 °Cexhibited (0002) peak intensity, nearly 40–60 times that fromthe films deposited at lower temperatures. At higher tempera-tures, the intensity decreases gradually but remains within thesame order of magnitude.

The enhancement of c-axis orientation with increasingsubstrate temperature above 100 °C is attributed to increase insurface diffusion of the adsorbed species and is in tune withmost of the earlier studies [1,2]. The unusual reduction in theextent of c-axis orientation at intermediate temperatures and itssubsequent enhancement at higher temperatures seen in thepresent work, has been reported earlier in sputtered ZnO films[15]. There are also a few reports on ZnO films deposited bysputtering [12,16] and MOCVD [17], showing a strong c-axispreferred orientation near room temperature followed by itsreduction with increase in substrate temperature to the range150–300 °C. The present results also show that the filmdeposited at 300 °C is of the highest quality in terms of c-axisorientation of crystallites. The marginal decrease in intensity of(0002) peak seen at higher substrate temperatures has not beenreported earlier and is attributed to mis-orientation of (0002)planes with substrate surface.

A comprehensive understanding of the processes leading toc-axis orientation over the complete range of substratetemperatures, which is required to explain all the abovefeatures, seems to be lacking. Recently, Kajikawa [18] hasextensively reviewed and analyzed the complex dependence ofc-axis preferred orientation in sputtered ZnO films on processessuch as, preferential nucleation, preferential crystallization,

Fig. 4. Raman spectra of ZnO films deposited at different substrate temperaturesalong with the spectrum of the quartz substrate.

Fig. 3. AFM images of ZnO films deposited at (a) room temperature, (b) 200 °C, (c) 300 °C and (d) 600 °C.

8720 S. Singh et al. / Thin Solid Films 515 (2007) 8718–8722

sticking, surface diffusion and grain growth and concluded thatthe validity of each process under given experimental condi-tions cannot be ensured, due to lack of reliable data on stickingcoefficients and surface diffusivities.

Fig. 1 also reveals shifts in the positions of (0002) and (0004)peaks, with substrate temperature. The position of (0002) peakshifts monotonically from 34.18° for the film deposited at roomtemperature to 34.44° for the film deposited at 400 °C, the laterbeing quite close to the corresponding value of 34.467° for bulkZnO (JCPDS file No. 75-1526). The position of the (0002) peakremains unchanged for the films deposited at higher substratetemperatures. The lattice constant ‘c’ was obtained from theposition of the (0002) peak for all the films, and was used toestimate the strain (ε) in the films along c-axis given by, ε=(cfilm−cbulk) /cbulk. Fig. 2(a) shows the variation of strain withsubstrate temperature. The film deposited at room temperatureshowed the maximum compressive strain (∼8×10−3), whichdecreased monotonically to nearly zero at ∼400 °C. Theexpected thermal strain introduced by differences in thermalexpansion coefficients of the film (αZnO=4×10

−6 K−1) andquartz substrate (αquartz = 0.5×10

− 6 K− 1) at a substratetemperature of 300 °C is ∼10−3 and is expected to be lesserat lower substrate temperatures. The higher values of strain (2–8×10−3) in the films deposited below 400 °C and its decreasewith increasing substrate temperature indicate that thermalstress does not contribute significantly to the observed strain.Further, the amorphous nature of quartz substrate rules outstrain due to lattice mismatch. Hence the compressive strain inthe films deposited below 400 °C is not attributed to extrinsicstress but is considered intrinsic to the growth process. It hasbeen reported in the case of ZnO films that O− ions formed at

the target have sufficient energies to bombard the growing filmand cause implantation or displacement of surface atoms deeperinto the film, resulting in compressive strain [19]. The decreasein strain at higher substrate temperatures is attributed toannealing effects and consequent reduction in defects.

The FWHM of (0002) peak was used to estimate thecrystallite size along c-axis (growth direction) by Scherrer'srelation. The variation of crystallite size with substratetemperature is shown in Fig. 2(b). It is seen that in the filmsdeposited up to 100 °C, the crystallite size along the growthdirection is ∼100 nm. A noticeable decrease is seen incrystallite size for films deposited in the temperature range of150–250 °C, which had earlier exhibited reduced c-axis

Fig. 6. PL spectra of ZnO films deposited at different substrate temperatures.

Fig. 5. (a) Reflectance (R) and Transmittance (T) spectra and (b) α2 vs. hν plotsfor ZnO films deposited at different substrate temperatures.

8721S. Singh et al. / Thin Solid Films 515 (2007) 8718–8722

preferred orientation. However, at substrate temperature of300 °C, the crystallite size increases to ∼80 nm and remainspractically unchanged till ∼600 °C, whence a slight increase to∼110 nm is noticed.

Typical AFM images of the films are shown in Fig. 3. Fig. 3(a)shows surface features of uniform lateral size of 50–70 nmfor room temperature deposited film. In contrast, the filmdeposited at 200 °C shows a mixture of small (40–80 nm) andlarge (150–200 nm) surface features (Fig. 3(b)). However, thefilm deposited at 300 °C again shows uniform features similar tothe room temperature deposited films, but with larger size of60–100 nm (Fig. 3(c)). The film deposited at 600 °C, showsdistinctly different surface morphology, as large features of∼250 nm along with smaller features of 60–100 nm are seen.The increase in the lateral size of morphological features withsubstrate temperature is attributed to the increase in lateral sizeof crystallites.

Fig. 4 shows the Raman spectra of ZnO films along with thatof quartz substrate. The ZnO film deposited at room temper-ature shows two very broad peaks centered ∼398 cm−1 and∼569 cm−1. The broad peak ∼398 cm−1 is attributed to acombination A1(TO) and E1(TO) modes and the small shoulder∼440 cm−1 is assigned to E2-high mode [20,21]. These featuresare rather unexpected, since in a highly c-axis oriented ZnOfilm in backscattering geometry, only strong E2 modes andweak A1(LO) mode are expected to be present [1,21]. The broad

peak ∼569 cm−1 is assigned to a combination of A1(LO) andE1(LO) modes [20,21]. The appearance of LO modes has beenattributed to the presence of oxygen defects and zinc interstitials[22]. These observations, including the weak E2-high mode asa shoulder, indicate the presence of structural defects, mis-orientation of (0002) planes and disorder in the film [22–24], inspite of its c-axis preferred orientation, seen in the XRD pattern.The film deposited at 300 °C shows an enhancement of E2-highintensity, though a broad band of TO modes is still seen. Theincrease in the prominence of E2-high peak corroborates wellwith the significant improvement in c-axis orientation of the film.Interestingly, the film deposited at 600 °C shows only an intenseE2-high peak along with weak A1(LO) mode∼575 cm−1, whichis characteristic of strong c-axis orientation [1,21]. Thereduction in intensity as well as shift of LO modes to higherfrequencies seen in this film, are attributed to decrease in pointdefects [21]. The decrease in defects correlates well with theincrease in crystallite size and decrease in stress, leading tooverall improvement in crystallinity of the film deposited at600 °C. It is interesting to note that such features are seen in thefilms deposited at 600 °C, which shows a lesser degree of c-axisorientation, as compared to that deposited at 300 °C.

Typical specular R and T spectra of the films are shown inFig. 5(a). All the films exhibit high transmittance in the visibleregion, limited by 10–15% reflectance. The sharp fall intransmittance below 400 nm is due to the onset of fundamental

8722 S. Singh et al. / Thin Solid Films 515 (2007) 8718–8722

absorption of ZnO. Absorption coefficient (α) was obtainedfrom specular R and T data and thickness (d) of the films, usingthe expression,

TðkÞcð1� RðkÞÞ2e�aðkÞd

for a self supporting film, with high absorption and lowreflectance. The corresponding plots of α2 vs. hν are shown inFig. 5(b), from which, the direct band gap of the films has beenestimated. The band gaps for all the films are ∼3.30 eV, whichagrees well with bulk band gap of ZnO. It is however observedthat with the increase in substrate temperature, the α2 vs. hνplots become sharper, indicating a reduction of defects in thefilms. The presence of optically active defects in films depositedat room temperature and 300 °C is evident from the relativelyenhanced sub-band gap absorption in the films.

Fig. 6 shows room temperature PL spectra of the films. Allthe films show strong emission ∼377 nm, attributed to nearband edge UV luminescence of ZnO [25]. Weak and broadbands are seen in green–yellow region, which are known toarise from oxygen defects and zinc interstitials [1,26]. Theintensity of band edge luminescence is found to increase withsubstrate temperature, accompanied by narrowing of the peak,as reported earlier for sputter deposited ZnO films on Si [27].The PL spectrum of the ZnO film deposited at 600 °C showed ahigh intensity band edge emission with FWHM of 102 meV.This value of PL peak width is comparable to reported valuesfor ZnO films deposited by MBE [3], MOVPE [25] and PLD[26] on single crystalline substrates. The PL results are in linewith optical absorption, which indicate reduction of opticallyactive defects in the films deposited at 600 °C. It may be notedthat the most intense and narrow PL emission does not comefrom the ZnO film deposited at 300 °C, which exhibitedstrongest c-axis orientation. The improved band edge lumines-cence of the film deposited at 600 °C is attributed to its betteroverall crystallinity resulting from reduction of strain anddefects as well as improvement in crystallite size (both lateraland along growth direction) as indicated by the structuralstudies presented above and supported by Raman spectra.

4. Conclusions

ZnO thin films of high optical quality have been depositedon quartz substrates by reactive sputtering. Though XRDstudies showed the strongest c-axis orientation for the filmsdeposited at 300 °C, AFM and Raman studies indicated that thefilms deposited at 600 °C possess better overall crystallinityresulting from reduction of strain and defects and improvementin crystallite size. As a consequence of reduction in opticallyactive defects, the ZnO film deposited at 600 °C showed a highintensity band edge luminescence with FWHM of ∼102 meV,which is comparable to the reported values for films depositedon single crystalline substrates.

Acknowledgements

The financial support from MHRD (Govt. of India) for thiswork is gratefully acknowledged. Sukhvinder Singh is thankfulto CSIR, New Delhi (India) for Senior Research Fellowship.FIST (Physics)-IRCC central SPM Facility and CRNTS of IITBombay are respectively acknowledged, for providing AFMand Raman facilities.

References

[1] Ü. Özgür, Ya.I. Alivov, C. Liu, A. Teke, M.A. Reshchikov, S. Doğan,V. Avrutin, S.-J. Cho, H. Morkoç, J. Appl. Phys. 98 (2005) 041301(and references therein).

[2] K. Ellmer, J. Phys. D: Appl. Phys. 33 (2000) R17 (and references therein).[3] Y. Chen, D.M. Bagnall, H.J. Koh, K.T. Park, K. Hiraga, Z. Zhu, T. Yao,

J. Appl. Phys. 84 (1998) 3912.[4] A. Bakin, A. El-Shaer, A. Che Mofor, M. Kreye, A. Waag, F. Bertram,

J. Christen, J. Stoimenos, J. Cryst. Growth 287 (2006) 7.[5] Y. Wang, S. Wang, S. Zhou, J. Xu, J. Ye, S. Gu, R. Zhang, Q. Ren, Appl.

Surf. Sci. 253 (2006) 1745.[6] J. Dai, H. Su, L. Wang, Y. Pu, W. Fang, F. Jiang, J. Cryst. Growth 290

(2006) 426.[7] Y.R. Ryu, S. Zhu, J.M. Wrobel, H.M. Jeong, P.F. Miceli, H.W. White,

J. Cryst. Growth 216 (2000) 326.[8] J.A. Sans, A. Segura, M. Mollar, B. Mar, Thin Solid Films 453–454 (2004)

251.[9] I.-S. Kim, S.-H. Jeong, S.S. Kim, B.-T. Lee, Semicond. Sci. Technol. 19

(2004) L29.[10] K.-K. Kim, J.-H. Song, H.-J. Jung, W.-K. Choi, S.-J. Park, J.-H. Song,

J. Appl. Phys. 87 (2000) 3573.[11] K.C. Ruthe, D.J. Cohen, S.A. Barnett, J. Vac. Sci. Technol., A, Vac. Surf.

Films 2 (2004) 2446.[12] R. Ondo-Ndong, F. Pascal-Delannoy, A. Boyer, A. Giani, A. Foucaran,

Mater. Sci. Eng., B, Solid-State Mater. Adv. Technol. 97 (2003) 68.[13] T. Inukai, M. Matsuoka, K. Ono, Thin Solid Films 257 (1995) 22.[14] S. Zhu, C.-H. Su, S.L. Lehoczky, P. Peters, M.A. George, J. Cryst. Growth

211 (2000) 106.[15] L.-J. Meng, M. Andritschky, M.P. dos Santos, Vacuum 45 (1994) 19.[16] K. Wasa, S. Hayakawa, Thin Solid Films 7 (1971) 135.[17] X.L. Chen, X.H. Geng, J.M. Xue, D.K. Zhang, G.F. Hou, Y. Zhao, J. Cryst.

Growth 296 (2006) 43.[18] Y. Kajikawa, J. Cryst. Growth 289 (2006) 387.[19] O. Kappertz, R. Drese, M. Wuttig, J. Vac. Sci. Technol., A, Vac. Surf.

Films 20 (2002) 2084.[20] T.C. Damen, S.P.S. Porto, B. Tell, Phys. Rev. 142 (1966) 570.[21] N. Ashkenov, B.N. Mbenkum, C. Bundesmann, V. Riede, M. Lorenz,

D. Spemann, E.M. Kaidashev, A. Kasic, M. Schubert, M. Grundmann,J. Appl. Phys. 93 (2003) 126.

[22] Y.W. Chen, Y.C. Liu, S.X. Lu, C.S. Xu, C.L. Shao, C. Wang, J.Y. Zhang,Y.M. Lu, D.Z. Shen, X.W. Fan, J. Chem. Phys. 123 (2005) 134701.

[23] J.Z. Wang, M. Peres, J. Soares, O. Gorochov, N.P. Barradas, E. Alves, J.E.Lewis, E. Fortunato, A. Neves, T. Monteiro, J. Phys., Condens. Matter 17(2005) 1719.

[24] M. Gomi, N. Oohira, K. Ozaki, M. Kayano, Jpn. J. Appl. Phys. 42 (2003)481.

[25] Y. Ma, G.T. Du, T.P. Yang, D.L. Qiu, X. Zhang, H.J. Yang, Y.T. Zhang,B.J. Zhao, X.T. Yang, D.L. Liu, J. Cryst. Growth 255 (2003) 303.

[26] X.M. Fan, J.S. Lian, Z.X. Guo, H.J. Lu, Appl. Surf. Sci. 239 (2005) 176.[27] S.-H. Jeong, J.-K. Kim, B.-T. Lee, J. Phys. D: Appl. Phys. 36 (2003) 2017.