Rapid thermal annealing effects on the structural and nanomechanical properties of Ga-doped ZnO thin films

(1)

Rapid thermal annealing effects on the structural and nanomechanical properties of

Ga-doped ZnO thin

ﬁlms

S.-R. Jian

a,

⁎

, G.-J. Chen

a

, S.-K. Wang

a

, T.-C. Lin

a

, J.S.-C. Jang

b,c

, J.-Y. Juang

d

, Y.-S. Lai

e

, J.-Y. Tseng

f a

Department of Materials Science and Engineering, I-Shou University, Kaohsiung 840, Taiwan

b_{Institute of Materials Science and Engineering, National Central University, Chung-Li 320, Taiwan} c

Department of Mechanical Engineering, National Central University, Chung-Li 320, Taiwan

d

Department of Electrophysics, National Chiao Tung University, Hsinchu 300, Taiwan

e

Central Product Solutions, Advanced Semiconductor Engineering, Nantze Export Processing Zone, Kaohsiung 811, Taiwan

f

Institute of Physics, Academia Sinica, Taipei 11529, Taiwan

a b s t r a c t

a r t i c l e i n f o

Available online 27 June 2012 Keywords:

Ga-doped ZnO thinﬁlms XRD

AFM

Nanoindentation Hardness

In this study, the structural and nanomechanical properties of Ga-doped ZnO (GZO) thinfilms on glass sub-strates followed by rapid thermal annealing (RTA) process were investigated by X-ray diffraction (XRD), atomic force microscopy (AFM) and nanoindentation techniques. The XRD results indicated that the annealed GZO thinfilms are textured, having a preferential crystallographic orientation along the hexagonal wurtzite (002) axis. Both the grain size and surface roughness of the annealed GZO thinfilms exhibit an in-creasing trend after RTA treatment. The hardness and Young's modulus of the annealed GZO thinfilms were measured by a Berkovich nanoindenter operated with the continuous contact stiffness measurements (CSM) option. Furthermore, the hardness and Young's modulus were found to increase with increasing grain size when the RTA time was prolonged from 0.5 to 3 min. The deformation behavior is referred to the inverse Hall–Petch effect commonly observed in systems deformed primarily via grain boundary sliding. The sup-pression of dislocation movement-associated deformation mechanism might be arisen from strong pinning effects introduced by Ga-doping.

1. Introduction

ZnO has been considered a material of high potential in optoelec-tronics and spinoptoelec-tronics applications[1–3]. Although recently doping ZnO into p-type has been the subject that has attracted much more at-tention in ZnO researches, the n-type materials with high quality crys-tallinity and controllable electron carrier concentration are also indispensable toward the applications. For example, being able to con-trol the electron carrier concentration is important for making the ZnO-based dilute magnetic semiconductors because their magnetic properties can be thereby modulated[4]. Moreover, group-III-doped ZnO with large electron carrier concentration is a potential candidate for replacing conventional transparent conducting oxides such as in-dium tin oxide. Among group-III element, Ga is an excellent n-type dopant for ZnO owing to the more compatible covalent bond length between Ga-O and ZnO (1.92 Å for Ga-O and 1.97 Å for Zn-O) than that of Al-O (2.7 Å)and In-O (2.1 Å).

Ga-doped ZnO (GZO) thinﬁlms have been widely investigated and used in optoelectronic as well as electronic devices [5,6].

However, their mechanical properties are largely ignored. Since for most device fabrication processes contact-induced damage may sig-niﬁcantly affect the ultimate optical and electronic properties of the device, thus a quantitative assessment of the material's mechanical properties is of crucial importance. Nanoindentation has been widely used for charactering the mechanical properties of solid surface and thin_ﬁlms[7,8]. Among the mechanical properties of interest, hard-ness, Young's modulus, elastic/plastic deformation behavior can be obtained from nanoindentation testing[9–11].

It is well known that the properties of ZnO thin_{films are strongly} affected not only by the deposition conditions[12,13]but also by the post-annealing conditions[14]. Post-annealing has a large effect on the crystallinity of ZnO thinfilms, such as grain size, residual strain and so on, thus, it is also an important method for manipulating the mechanical characteristics of the materials [15]. In this study, we aimed to investigate the nanomechanical properties of GZO thin films by means of nanoindentation. The GZO films were deposited on glass substrates using a radio frequency (rf) magnetron sputtering system. Thefilm microstructures were characterized by X-ray diffrac-tion (XRD) and the surface morphology was examined by atomic force microscopy (AFM). Furthermore, the influences of the rapid thermal annealing (RTA) on the properties of the GZO thinfilms are also presented.

Surface & Coatings Technology 231 (2013) 176–179

⁎ Corresponding author. Tel.: +886 7 6577711x3130; fax: +886 7 6578444. E-mail address:srjian@gmail.com(S.-R. Jian).

Contents lists available atScienceDirect

Surface & Coatings Technology

(2)

2. Experimental details

Experimentally, the 3 at% Ga-doped ZnO (GZO) thin_{ﬁlms were} deposited on Corning 1737 F glass substrates at room temperature by using rf-magnetron sputtering. The detailed growth procedures and operation parameters can be found elsewhere[16]. Subsequently, the as-deposited GZO thinﬁlms were post-annealed at 500 °C for 0.5, 1, 2 and 3 min in oxygen environment by rapid thermal annealing (RTA). During the annealing process, the rising or cooling rate of the temperature was kept at 30 °C/s.

The crystal structure of GZO thinﬁlms were analyzed by X-ray dif-fraction (Panalytical X'Pert XRD, CuKα, λ=1.5406 Å). The Scherrer's formula shown below was employed to estimate the mean grain size of GZO thinﬁlms[17]:

D¼_{B cos}0:9λ_θ ð1Þ

Hereλ, B and θ denote the X-ray wavelength, the FWHM of (002) peak, and the corresponding Bragg diffraction angle, respectively.

In addition, the examinations of surface features were carried out by using AFM (Topometrix-Accures-II). The average surface rough-ness, RRMSof the surface was calculated by the following equation[18]:

RRMS¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 N XN n¼1 r2 n v u u t _ð2Þ

Here N is the number of data and rnis the surface height of the nth

datum.

Nanoindentation experiments were preformed on a MTS Nano In-denter® XP system with a three-sided pyramidal Berkovich indenter tip by using the continuous stiffness measurement (CSM) technique [19]. This technique is accomplished by imposing a small, sinusoidal varying force on top of the applied linear force that drives the motion of the indenter. The displacement response of the indenter at the exci-tation frequency and the phase angle between the force and displace-ment were measured continuously as a function of the penetration depth. Solving for the in-phase and out-of-phase portions of the dis-placement response gives rise to the determination of the contact stiffness as a continuous function of depth[19]. As such, the mechan-ical properties changing with respect to the indentation depth can be obtained. The nanoindentation measurements were carried out as fol-lows. First, prior to applying loading on ZnO sample, nanoindentation was conducted on a standard fused silica sample to obtain the reason-able range (the Young's modulus of fused silica is 68 ~ 72 GPa). Then, a constant strain rate of 0.05 s−1was maintained during the increment of load until the indenter reached a depth of 100 nm into the surface. The load was then held at the maximum value of loading for 10 s in order to avoid the creep which might signiﬁcantly affect the unloading behavior. The indenter was then withdrawn from the surface at the same rate until the loading has reduced to 10% of the maximum load. Then, the indenter was completely removed from the material. In this work, constant strain rate was chosen in order to minimize the strain-hardening effects. At least 20 indentations were performed on each sample and the distance between the adjacent indents was kept at least 50μm apart to avoid interaction.

The hardness, H, and elastic modulus, E, were calculated from the load–displacement data following the work reported by Oliver and Pharr[20].

H¼Pmax

Ap ð3Þ

where Pmaxis the load measured at a maximum penetration depth (h)

and Apis the projected contact area between the indenter and the

sample at Pmax.

The reduced elastic modulus (Er), which is the combined elastic

modulus of both the tested sample and the indenter, is calculated as follows: Er¼ ffiffiffi π p 2 Sffiffiffiffiffiffi_A p q ð4Þ

where S = dP/dh (stiffness) is the slope of the upper portion of the unloading curve in the load–displacement curve.

The elastic modulus of thinﬁlms, Ef, is then calculated as follows:

Ef ¼ 1−v 2 f ₁ Er− 1−v2 i Ei !₋₁ ð5Þ Here v is the Poisson's ratio and the subscripts i and f denote the parameters for the indenter and the annealed ZnO thinﬁlms, respec-tively. For diamond indenter tip, Ei= 1141 GPa, vi= 0.07 and, vf=

0.25 are chosen for ZnO sample[21]. 3. Results and discussion

Fig. 1presents the XRD patterns of GZO thinfilms with RTA time varying from 0.5 to 3 min. It can be found that the preferred orienta-tion of thefilms has been modified with the variation of RTA time du-rations. With the increase in the RTA time, peak intensities are seen to increase. Note that there is no evidence of the existence of Ga2O3

phase discernible from the XRD results, implying that Ga atoms either have completely substituted for Zn atoms in the ZnO lattice or they might segregate into the amorphous regions in the vicinity of grain boundaries. In addition, an intense (101) peak is observed in all of the annealed GZO_{films, indicating that the films were largely} grow-ing with the (101) plane parallel to the surface of the substrate. On the other hand, (002) is commonly regarded as the preferred orienta-tion forfilms with hexagonal wurtzite structure, because thermody-namically the (001) plane has a lower surface energy than all the other planes in this particular structure. Therefore, increasing the RTA time properly may favor the atoms to diffuse to the equilibrium positions, hence the preferred orientation becomes more of the (002) characteristic and thefilm crystallinity was improved signifi-cantly at the same time. The mean grain sizes of GZO thinfilms were estimated to be 38.4, 42.3, 44.2 and 45.8 nm (compared with 23.5 nm for as-deposited GZOfilm) for films respectively annealed at RTA time of 0.5, 1, 2 and 3 min.

Fig. 1. XRD patterns of (a) as-deposited GZO thinﬁlm, and GZO thin ﬁlms annealed at different RTA time durations: (b) 0.5 min, (c) 1 min, (d) 2 min, and (e) 3 min.

177 S.-R. Jian et al. / Surface & Coatings Technology 231 (2013) 176–179

(3)

Fig. 2illustrates the surface morphologies of the annealed GZO thinﬁlms obtained by AFM. It is evident that, as the RTA time in-creases from 0.5 to 3 min, the average surface roughness is increased from 7.2 to 9.3 nm. It is because the longer RTA time provides more activation energy to stimulate the migration of grain boundaries and coalescence of adjacent grains, leading to larger grains and rougher surface [15,22]. Therefore, it can be observed that fewer grain boundaries and larger grains are evidently displayed inﬁlms with longer RTA time.

The effects of RTA on the mechanical performance of GZO thinfilms were evaluated directly by using nanoindentation measurements. The typical load–displacement curves for the as-deposited and annealed GZO thinfilms are plotted inFig. 3. The load–displacement responses obtained by nanoindentation contain information about the elastic and plastic deformation behaviors and can be regarded as the “finger-prints_{” of the film's mechanical properties. It is evident that, for the} RTA-treatedfilms, both the film hardness and the elasticity recovery are significantly enhanced, as compared to the as-deposited ones. The hardness and Young's modulus of as-deposited and annealed GZO thinfilms, extracted from the load–displacement curves as a function of penetration depth, are displayed inFig. 4(a)–(b). As a function of the penetration depth, both the hardness and Young's modulus reach a maximum value and then start to decrease untilfinally reaching a con-stant value. Comparing the hardness curve of annealed GZO thinfilms with that of the as-depositedfilms, an enhancement of the hardness from 8.5 to 10.8 GPa was observed, while the Young's modulus is en-hanced from 120 to 155 GPa when the indenter penetration depth is greater than 30 nm. The effects of RTA time variation on the hardness and Young's modulus of the annealed GZO thinfilms are summarized inTable 1. Furthermore, Wang and Li[23]proposed a similar size de-pendence of elastic modulus for ZnO nanowires. Because of the in flu-ences of surface stress effect, while the diameter size is larger than 20 nm, the Young's modulus increases with the diameter size. Similar

size effect might work in the GZO thin_{ﬁlms. In overall, the values of} Young's modulus of GZO thinﬁlms are lower than those of the bulk GZO, which is consistent with the previous studies[24,25].

Also shown inTable 1are the ratios H/Efand H3/Ef2for all GZO

samples. The H/Efvalue characterizes the susceptibility of materials

to the elastic strain. Therefore, increasing the H/Efvalue can be

trans-lated into a correspondent enhancement in the wear resistance[26]. On the other hand, the parameter, H3_/E

f

2_{, describes the ability of a}

ma-terial in resisting against the plastic deformation and, thus, character-izes its toughness and resistance to crack propagations [27]. The larger values of H/Efand H3/Ef2often represent as a reliable indicator

of good wear resistance. The current results indicated that GZO ﬁlms subjected to 3 min-RTA treatment exhibited the best combina-tions in terms of the primary mechanical property parameters examined.

According to the above-mentioned results, it is evident that the primary effects of RTA, represented mainly by the duration of the RTA time here, were to increase the grain size of GZOﬁlms, which

Fig. 2. AFM images of annealed GZO thinﬁlms at the different RTA time durations: (a) 0.5 min, (b) 1 min, (c) 2 min, and (d) 3 min.

Fig. 3. Typical load–displacement curve for the as-deposited and annealed GZO thin ﬁlms.

(4)

in turn led to significant enhancements in various aspects of the film's mechanical properties. Since the variation of hardness with the grain size appears to follow the inverse Hall–Petch effect[28], that is the hardness increases with grain size, it is suggestive that the most prominent deformation mechanism prevailing in the current GZO films could be arisen from grain boundary sliding instead of from dis-location movements. We believe that the doping of Ga may have in-duced significant dislocation pinning effects, which in turn hindered the dislocation activities and activated the secondary deformation mechanism as an alternative. It should be interesting to carry out fur-ther prolonged RTA treatments to see what will happen when the film grain size is further increased.

4. Conclusion

A combination of XRD, AFM and nanoindentation techniques has been carried out to investigate the structural, surface morphological features, and nanomechanical properties of GZO_{films subjected to} RTA treatment with various time durations. The XRD results indicated that the annealed GZO thinfilms had the hexagonal wurtzite struc-ture with a predominant (002) growth orientation. Prolonged RTA duration appeared to facilitate the grain growth, leading to increased film surface roughness, as evident from AFM observations. The hard-ness and Young's modulus of GZOfilms were respectively ranging from 7.8 to 10.8 GPa and from 105.4 to 130.6 GPa when the RTA time was increased from 0.5 to 3 min. The fact that the hardness of the GZOfilms behaves in accordance with the inverse Hall–Petch ef-fect indicates that Ga doping might have changed the prevailing de-formation mechanism from dislocation movement for pristine ZnO to grain boundary sliding by introducing substantial pinning effects.

Acknowledgements

This work was partially supported by the National Science Council of Taiwan, under Grant No.: NSC100-2221-E-214-024. JYJ is partially supported by the NSC of Taiwan and the MOE-ATU program operated at NCTU. The author likes to thank Dr. P.-F. Yang and Dr. Y.-T. Chen for their technical supports.

References

[1] T. Aoki, Y. Hatanaka, D.C. Look, Appl. Phys. Lett. 76 (2000) 3257. [2] C. Klingshim, Phys. Status Solidi B 244 (2007) 3027.

[3] F. Pan, C. Song, X.J. Liu, Y.C. Yang, F. Zeng, Mater. Sci. Eng., R 62 (2008) 1. [4] Z. Yang, J.L. Liu, M. Biasini, W.P. Beyermann, Appl. Phys. Lett. 92 (2008) 042111. [5] H. Makino, T. Yamada, N. Yamamoto, T. Yamamoto, Thin Solid Films 519 (2010)

1521.

[6] Y.Y. Liu, S.Y. Yang, G.X. Wei, H.S. Song, C.F. Cheng, C.S. Xue, Y.Z. Yuan, Surf. Coat. Technol. 205 (2011) 3530.

[7] S.R. Jian, J.S.C. Jang, Y.S. Lai, P.F. Yang, C.S. Yang, H.C. Wen, C.H. Tsai, Mater. Chem. Phys. 109 (2008) 360.

[8] S.R. Jian, H.G. Chen, G.J. Chen, J.S.C. Jang, J.Y. Juang, Curr. Appl. Phys. 12 (2012) 849.

[9] S.R. Jian, G.J. Chen, T.C. Lin, Nanoscale Res. Lett. 5 (2010) 935.

[10] S.R. Jian, J.Y. Juang, N.C. Chen, J.S.C. Jang, J.C. Huang, Y.S. Lai, Nanosci. Nanotechnol. Lett. 2 (2010) 315.

[11] S.R. Jian, J.Y. Juang, J. Nanomater. 2012 (2012) 914184.

[12] P.F. Yang, H.C. Wen, S.R. Jian, Y.S. Lai, S. Wu, R.S. Chen, Microelectron. Reliab. 48 (2008) 389.

[13] H. Chen, J. Ding, S. Ma, Phys. E 42 (2010) 1487.

[14] J.H. Han, Y.S. No, T.W. Kim, J.Y. Lee, J.Y. Kim, W.K. Choi, Appl. Surf. Sci. 256 (2010) 1920.

[15] L.Y. Lin, D.E. Kim, Thin Solid Films 517 (2009) 1690.

[16] J.Y. Tseng, Y.T. Chen, M.Y. Yang, C.Y. Wang, P.C. Li, W.C. Yu, Y.F. Hsu, S.F. Wang, Thin Solid Films 517 (2009) 6310.

[17] B.D. Cullity, S.R. Stock, In: Element of X-ray Diffraction, Prentice Hall, New Jersey, 2001, p. 170.

[18] K. Miyoshi, Y.W. Chung, Surface Diagnostics in Tribology: Fundamental Principles and Applications, World Scientiﬁc Publishing, Singapore, 1993..

[19] X.D. Li, B. Bhushan, Mater. Charact. 48 (2002) 11. [20] W.C. Oliver, G.M. Pharr, J. Mater. Res. 7 (1992) 1564.

[21] C.Y. Yen, S.R. Jian, G.J. Chen, C.M. Lin, H.Y. Lee, W.C. Ke, Y.Y. Liao, P.F. Yang, C.T. Wang, Y.S. Lai, J.S.C. Jang, J.Y. Juang, Appl. Surf. Sci. 257 (2011) 7900.

[22] Y. Lin, J. Xie, H. Wang, Y. Li, C. Chavez, S. Lee, S.R. Foltyn, S.A. Crooker, A.K. Burrell, T.M. McCleskey, Q.X. Jia, Thin Solid Films 101 (2005) 492.

[23] G.F. Wang, X.D. Li, Appl. Phys. Lett. 91 (2007) 231912. [24] X.Y. Tao, X.N. Wang, X.D. Li, Nano Lett. 7 (2007) 3172. [25] H. Ni, X.D. Li, Nanotechnology 17 (2006) 3591. [26] A. Leyland, A. Mattews, Wear 246 (2000) 1.

[27] D. Galvan, Y.T. Pei, J.T.M. De Hosson, Surf. Coat. Technol. 200 (2006) 6718. [28] J. Schiotz, T. Vegge, F.D. Di Tolla, K.W. Jacobsen, Phys. Rev. B 60 (1999) 11971. Fig. 4. (a) Hardness and (b) Young's modulus as a function of the penetration depth for

as-deposited and annealed GZO thinﬁlms.

Table 1

The data evaluated from XRD patterns, AFM images and nanoindentation results of measured GZO thinﬁlms in this study.

GZO D (nm) RRMS(nm) H (GPa) Ef(GPa) H/Ef H3/Ef2

As-deposited 23.5 5.5 8.5 113.4 0.075 0.047 RTA-0.5 min 38.4 7.2 7.8 105.4 0.074 0.043 RTA-1 min 42.3 8.4 9.5 115.2 0.081 0.064 RTA-2 min 44.2 8.7 10.2 126.8 0.080 0.066 RTA-3 min 45.8 9.3 10.8 130.6 0.083 0.074 179 S.-R. Jian et al. / Surface & Coatings Technology 231 (2013) 176–179