Nanotribological behavior of ZnO films prepared by atomic layer deposition

(1)

Nanotribological behavior of ZnO

ﬁlms prepared by atomic

layer deposition

Wun-Kai Wang

a,1

, Hua-Chiang Wen

b,1

, Chun-Hu Cheng

c

, Wu-Ching Chou

b

,

Wei-Hung Yau

d,n

_{, Ching-Hua Hung}

a

_{, Chang-Pin Chou}

a

Department of Mechanical Engineering, National Chiao Tung University, Hsinchu 300, Taiwan, ROC

b

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

c

Department of Mechantronic Technology, National Normal University, Taipei 106, Taiwan, ROC

d

Department of Mechanical Engineering, Chin-Yi University of Technology, Taichung 400, Taiwan, ROC

a r t i c l e i n f o

Article history:

Received 20 February 2013 Received in revised form 14 August 2013

Accepted 18 September 2013 Available online 12 October 2013 Keywords:

A. Thinﬁlms B. Vapor deposition C. Mechanical properties

a b s t r a c t

We used atomic layer deposition to form ZnO thin-film coatings on Si substrates and then evaluate the effect of pile-up using the nanoscratch technique under a ramped mode. The wear volume decreased with increasing annealing temperature from room temperature to 4001C for a given load. Elastic-to-plastic deformation occurred during sliding scratch processing between the groove andfilm for loading penetration of 30 nm. The onset of non-elastic behavior and greater contact pressure were evident for loading penetration of 150 nm; thus, full plastic deformation occurred as a result of a substrate effect. We suspect that elastic–plastic failure events were related to edge bulging between the groove and film, with elastic–plastic deformation attributable to adhesion discontinuities and/or cohesion failure of the ZnOfilms.

1. Introduction

Zinc oxide (ZnO), a metal oxide that form transparentfilms, is a wide-bandgap semiconductor with high thermal and chemical stability and large exciton binding energy (60 meV); it is one of the main materials used in blue and UV optical devices [1–3]. ZnO can be deposited as a thin film using several techniques, including pulsed laser deposition, spray pyrolysis, and aqueous solution methods[4–6]. In recent years, atomic layer deposition (ALD) has become an attractive method for the formation of smooth films, with the additional benefits of high conformality and precise control over thickness and composition. ALD forma-tion of ZnO through cyclic self-limiting deposiforma-tion can yield conformal and uniform thinfilms at the atomic level[7,8]. The electrical, optical, and structural properties of such ZnO materials have been investigated in detail for various processing para-meters and conditions[9].

ZnOﬁlms have recently been used in acoustic wave biosensing and microﬂuidic applications; their potential utility in the fabrica-tion of lap-on-chip (LOC) systems may affect the development of the biotechnology industry. Kim et al. reported that a passivated

ALD aluminum oxide surface could be used in an acoustic wave sensor for early biofilm detection[10]. Most acoustic wave devices need to be sensitive to input signals, such as mechanical stress, chemical exchange, and electrical perturbations. Thus, a strong texture, a low density of defect traps, a smooth surface, and good stoichiometry are basic requirements for good piezoelectric prop-erties. As mentioned above, ALD with precise compositional control using layer-by-layer deposition can meet these require-ments at a low thermal budget of 2001C; the ability to integrate ALD with CMOS processing is useful in the preparation of LOC systems. However, the acoustic velocity of a ZnO film strongly depends on substrate effects due to lattice mismatch, which limits the piezoelectric properties. The crystallinity of a ZnO film is responsible for its piezoelectric response. Although the crystal-linity can be improved via high-temperature annealing, the inter-face quality will suffer from adhesion issues as a result of local strain or stress originating from de-bonding of Zn atoms[11].

Nanoindentation can provide a good understanding of the mechanical properties of a thin ﬁlm interface, primarily its elastic–plastic properties, friction, and adhesion. These parameters characterize the mechanical stability of a working device and can provide useful predictions regarding durability. Nanoindentation is performed using an instrument that measures force and displace-ment continuously during indentation[13–18]. Slip band move-ment and dislocation nucleation mechanisms have been proposed to explain pop-in events within the epilayers of ZnO_ﬁlms[14_–17]. Contents lists available atScienceDirect

journal homepage:www.elsevier.com/locate/jpcs

Journal of Physics and Chemistry of Solids

http://dx.doi.org/10.1016/j.jpcs.2013.09.016

n_{Corresponding author. Tel.:}_{þ886 423924505; fax: þ886 423930681.}

E-mail address:ywhung1@gmail.com (W.-H. Yau).

(2)

Annealing treatment is commonly applied to minimize defects and improve the quality of as-grown thinfilms. Kim et al. found that annealing significantly improved the optical properties and sur-face morphology of ALD ZnO thin films compared to as-grown samples[12].

In this study we investigated the nanotribological behavior of ALD ZnO ﬁlms and the temperature-dependent crystallization effects of different annealing conditions. To gain an insight into the failure mechanisms involved in nanotribological responses, we analyzed the surface morphology, _{ﬁlm crystallinity,} microstruc-ture, and adhesion and cohesion using atomic force microscopy (AFM), X-ray diffraction (XRD), transmission electron microscopy (TEM), and scanning probe microscopy (SPM). We also examined the quality of the resulting ZnO layers in terms of pile-up events under sliding scratch-induced deformation.

2. Experimental

Water and diethylzinc were used as precursors and nitrogen as the purge gas. Thin ZnO films were deposited on Si (1 0 0) substrates at 2001C in an ALD system operated using the flow-rate interruption method[8,19]. The number of deposition cycles wasfixed at 150 in each experiment and the deposition rate was controlled at 0.2 Å per cycle. The thickness of each as-deposited ZnO_{film was approximately 30 nm. As-deposited ZnO films were} subjected to rapid thermal annealing at 300 or 4001C for 1 h under N2. The heating rate was set at 201C s1; the furnace

required approximately 30 min to cool to RT after annealing. The ZnO crystallinity was analyzed using XRD (PANalytical X'Pert Pro, MRD) with Cu K_α(

λ

¼0.154 nm) radiation. Scans were conducted over the 2

θ

range 20–801 (scan rate 21 min1_{) at a grazing angle of}

0.51 under 30 kV and 30 mA. XRD data conformed to the Joint Committee on Powder Diffraction Standards card (JPCDS). Cross-sectional TEM (JEOL JEM-2010F, operating voltage 200 kV) was used to examine the ZnO/Si microstructure.

The nanotribological properties of ZnOﬁlms were determined using an SPM system in which AFM (Digital Instruments Nano-scope III) was combined with a nanoscratch measurement system. The scanning probe microscope (Hysitron Corporation) was oper-ated at a constant scan rate of 2

μ

m s1. The adhesion properties of as-deposited samples and samples annealed at 300 and 4001C were evaluated using a nanoindentation system. The constant depth mode was used and penetration depths of 30 and 150 nm were applied for each ZnOﬁlm. The maximum load was main-tained while forming 10-

μ

m-long scratches. Surface proﬁles before and after scratching were obtained by scanning the tip at a normal load of 0.02 mN (i.e., a load sufﬁciently small to produce any measurable displacement). After scratching, the wear tracks were imaged using AFM.

3. Results and discussion

Fig. 1shows XRD patterns for as-deposited ZnO and samples annealed at 300 and 4001C. The diffraction peaks for ZnO (1 0 0), (0 0 2), (1 0 1), (1 0 2), (1 1 0), (1 0 3), and (1 1 2) evident in the XRD patterns confirm that all the films were polycrystalline with a hexagonal wurtzite structure. The intensity of the ZnO (0 0 2) peak was higher for the sample annealed at 4001C than for the film annealed at 3001C, indicating that the higher annealing tempera-ture increased the crystallinity. The results suggest that the ZnO layer remained largely polycrystalline with excellent crystalline quality.

We obtained more detailed structural information using TEM. Fig. 2shows a cross-sectional TEM image of the ZnO/Si system,

which features a layer of polycrystalline ZnO. Randomly oriented ZnO nanocrystals within the 30-nm-thick ZnO layers appear to have very good crystallinity. A nanometer-thick interface region is evident between Si and the ZnO layer. This interface appears blurred and the ZnO grains it contains have relatively poor crystallinity. As described below, this interface has a signiﬁcant effect on the ZnO/Si adhesion properties.

Fig. 3presents typical AFM profiles of ZnO films deposited on Si before and after nanoscratch tests at a penetration of 30 or 150 nm. The scratch images reveal that an increase in load led to an increase in wear volume (total volume swiped by the indenter tip) for each ZnO coating; the wear volume is consistent with the projected area of the indenter tip. Using such a scratch system to determine elastic_{–plastic contact, Pelletier et al. found that the} shape of the residual groove was related to the plastic strainfield [20]. The profile in Fig. 3a indicates that the surface material underwent an elastic reaction as a result of elastic deformation between the groove and thefilm when the penetration depth was 30 nm. The indenter resulted in greater contact pressure for the sample annealed at 3001C, leading to the onset of non-elastic behavior (Fig. 3b). For the sample annealed at 4001C, large areas of the coating were removed from the scratch track as a result of elastic–plastic deformation at 30 nm (Fig. 3c). The penetration depth traces reveal a slight increase for the sample treated at 4001C compared to the sample treated at 300 1C. This SPM trend matches the lateral force curves inFig. 4a, from which it is evident that an elastic reaction between the groove and film dominates. A similar response is evident inFig. 3d–f: an increase in lateral

Fig. 1. XRD patterns for ALD ZnOfilms. (a) As-deposited film and films annealed at (b) 3001C and (c) 400 1C.

(3)

force is evident for a penetration depth of 150 nm, without marked oscillations. However, the substrate effect may dominate as a result of the increased lateral force.

The annealing temperature was sufﬁcient to drive transforma-tion to a denser crystal structure so that ductile deformatransforma-tion was inhibited. The critical load was 29.8 and 50.6

μ

N for the samples annealed at 300 and 4001C, respectively. No signal was detected for the as-deposited sample. Sample annealing at 4001C led to relaxed crystallization of the ZnO coating, as observed inFig. 4a,b. The changes in scratch groove morphology also indicate an improvement in the nanoscale wear resistance. At the beginning of each scan, all of theﬁlms exhibited similar behavior, but the effect of the annealing temperature became apparent thereafter. This behavior may be due in part to shrinkage events, attributable to a very thin coating that responds to a susceptible defect density for scratch tests at a penetration depth of 30 nm. Schreiber and Vasnyov used dynamic-mode cathodoluminescence to observe local plastic deformation during in situ scratching of a (0001) ZnO sample[22]. They found evidence of the propagation of highly

mobile dislocations in various prismatic slip systems for their as-deposited samples. Thus, we suspect that the features of the nanoscratch zone in the vicinity of the interface might not be merely accidental artifacts resulting from sample preparation[18]. Additional evidence will be required because this phenomenon controls oscillations under friction, thereby affecting the adhesion and/or cohesion failure of ZnOfilms. Fig. 5a shows the friction coefficient recorded at a penetration depth of 30 nm. Oscillations under the friction force are evident, particularly for an annealing temperature of 4001C. Owing to thin film elastic deformation, the friction coefficient decreased, gradually approaching the true value; elastic–plastic deformation of the friction coefficient was slightly disrupted in the interval 15–25 s. At 22.5 s, film dislocation caused oscillation of the friction coefficient, which then stabilized at 25 s. The friction coefficient recorded at a penetration depth of 150 nm (Fig. 5b) also exhibited oscillation under the friction force. Similar analysis of scratchedfilms has been reported previously [19,20]. Notably, oscillation trends for each sample were similar in both on- and off-load scans [21]. The oscillations in the plots

Fig. 3. AFM images combined with nanoscratch trace measurements for nanoscratched samples obtained using a ramp force at a depth of (a–c) 30 nm and (d–f) 150 nm. (a,d) As-deposited sample; (b,e) sample annealed at 3001C; and (c,f) sample annealed at 400 1C.

(4)

recorded at a penetration depth of 150 nm are more stable than those recorded at 30 nm. We suspect that the lower degree of adhesion reflects interlinks and rearrangements that tend to result influctuations of the friction coefficient (

μ

) at a penetration depth of 150 nm. Instability in the initial

μ

proﬁle, so-called pile-up, is evident inFig. 5a for the sample annealed at 3001C. We calculated values of

μ

for the ZnO ﬁlms by averaging results measured at scratch durations ranging from 8 to 38 s (Fig. 6). From the trends observed in our nanotribological investigations, changes in the value of

μ

appear to signal the onset of adhesive failure, with cracking or delamination resulting between the interface and the sliding nanoscratch track[15,23].

The discordant curves and irregularities observed during elastic and plastic deformation can be attributed to adhesion disconti-nuities and/or cohesion failure, as determined from the nano-scratch traces (induction of pile-up) [24,25]. The profile for the sample annealed at 4001C was softer than for the as-deposited sample owing to shrinkage of the nanoscratch traces. Thus, the ZnO film structure changed when the thermal treatment tem-perature was increased from 300 to 4001C, presumably via alteration of the composition of the intrinsic structure. Hence, edge bulging occurred between the groove and film and the material was crushed as a result of elastic failure. In summary, the friction traces were fairly reproducible, but varied greatly for the different samples.

4. Conclusion

We used AFM and nanoscratch techniques to characterize high-quality ALD ZnOﬁlms on Si substrates. We performed nanoscratch tests to determine the adhesion/cohesion failure mechanism of

Fig. 4. Normal displacement plotted with respect to scratch duration for ZnOﬁlms annealed at 300 and 4001C. Each sample was subjected to penetration at (a) 30 nm and (b) 150 nm.

Fig. 5. Coefficient of friction (μ) plotted with respect to scratch duration for as-deposited ZnOfilm and films annealed at 300 and 400 1C for penetration depths of (a) 30 nm and (b) 150 nm. Inset to (a): 15–25-s interval.

Fig. 6. Distinct critical coefﬁcient of friction for ZnO ﬁlms annealed at various temperatures and then subjected to penetration at depths of 30 and 150 nm.

(5)

ZnO-coated Si. An increase in annealing temperature was suf ﬁ-cient to drive transformation to a denser crystal structure so that ductile deformation was inhibited. Pile-up occurred at a penetra-tion force of 29.8 and 50.6

μ

N for samples annealed at 300 and 4001C, respectively. The systems exhibited significant suppression of lateral indented deformation, which induced a degree of pile-up. In addition, tip drop and lift phenomena can be explained in terms offilm stiffness as a function of interatomic distances during scratch tests. For ZnO films compressed by an indenter, more

μ

oscillations were observed for a penetration depth of 30 nm than for 150 nm. It is reasonable to suspect that edge bulging occurred between the groove andﬁlm, resulting in elastic failure.

Acknowledgment

This study was supported by the National Science Council of Taiwan (grant nos. NSC 101–2221-E-167-004 and NSC 102–2221-E-167-007) and the National Chin-Yi University of Technology (NCUT11-REM-004: 2012.1.1–2012.10.31 and NCUT 12-REM-003: 2013.1.1–2013.10.31). We thank Dr. H.C. Wen and Prof. W.C. Chou for technical suggestions, and Prof. Y.R. Jeng for the SPM equip-ment used in this study.

References

[1]A.P. Alivisatos, Science 271 (1996) 933–937.

[2]K.C. Aw, Z. Tsakadze, A. Lohani, S. Mhaisalkar, Scr. Mater 60 (2009) 48–51. [3]Y.R. Ryu, T.S. Lee, J.A. Lubguban, H.W. White, B.J. Kim, Y.S. Park, C.J. Youn, Appl.

Phys. Lett. 88 (2006) 241108.

[4]S. Christoulakis, M. Suchea, E. Koudoumas, M. Katharakis, N. Katsarakis, G. Kiriakidis, Appl. Surf. Sci. 252 (2006) 5351–5354.

[5]I. Kortidis, K. Moschovis, F.A. Mahmoud, G. Kiriakidis, Thin Solid Films 518 (2009) 1208–1213.

[6]X.X. Liu, Z.G. Jin, S.J. Bu, J. Zhao, Z.F. Liu, J. Am. Ceram. Soc. 89 (2006) 1226–1231.

[7]Y.H. Lin, Y.C. Lai, C.L. Lu, W.K. Hsu, J. Mater. Chem. 21 (2011) 12485–12488. [8]H.C. Wen, C.I. Hung, H.J. Tsai, C.K. Lu, Y.C. Lai, W.K. Hsu, J. Mater. Chem. 22

(2012) 13747–13750.

[9]Y.W. Zhu, H.I. Elim, Y.L. Foo, T. Yu, Y.J. Liu, W. Ji, J.Y. Lee, Z.X. Shen, A.T.S. Wee, J.T.L. Thong, C.H. Sow, Adv. Mater. 18 (2006) 587–592.

[10]Y.W. Kim, S.E. Sardari, M.T. Meyer, A.A. Iliadis, H.C. Wu, W.E. Bentley, R. Ghodssi, Sens. Actuators B 163 (2012) 136–145.

[11]T. Riemann, J. Christen, G. Kaczmarczyk, A. Kaschner, A. Hoffmann, A. Zeuner, D. Hofmann, B.K. Meyer, Phys. Status Solidi b 229 (2002) 891–895. [12]C.R. Kim, C.M. Shin, J.Y. Lee, J.H. Heo, T.M. Lee, J.H. Park, H. Ryu, C.S. Son,

J.H. Chang, Curr. Appl. Phys. 10 (2010) S294–S297.

[13]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) 7900–7905. [14]W.H. Yau, P.C. Tseng, H.C. Wen, C.H. Tsai, W.C. Chou, Microelectron. Reliab. 51

(2011) 931–935.

[15]M.H. Lin, H.C. Wen, Y.R. Jeng, C.P. Chou, Nanoscale Res. Lett. 5 (2010) 1812–1816.

[16]M.H. Lin, H.C. Wen, C.Y. Huang, Y.R. Jeng, W.H. Yau, W.F. Wu, C.P. Chou, Appl. Surf. Sci. 256 (2010) 3464–3467.

[17]H.C. Wen, C.S. Yang, W.C. Chou, Appl. Surf. Sci. 256 (2010) 2128–2131. [18]S.R. Jian, I.J. Teng, P.F. Yang, Y.S. Lai, J.M. Lu, J.G. Chang, S.P. Ju, Nanoscale Res.

Lett. 3 (2008) 186–193.

[19]C.J. Hu, Y.H. Lin, C.W. Tang, M.Y. Tsai, W.K. Hsu, H.F. Kuo, Adv. Mater. 23 (2011) 2941–2945.

[20]H. Pelletier, A.L. Durier, C. Gauthier, R. Schirrer, Tribol. Int. 41 (2008) 975–984. [21]J. Zhang, J. Zhang, Tribol. Lett. 49 (2013) 77–83.

[22]J. Schreiber, S. Vasnyov, J. Phys. Condens. Matter 16 (2004) S75–S84. [23]R. Saha, W.D. Nix, Acta Mater. 50 (2002) 23–38.

[24]Y.R. Jeng, P.C. Tsai, Y.H. Liu, Mater. Res. Bull. 44 (2009) 1995–1999. [25]K. Shinozaki, T. Honma, T. Komatsu, Mater. Res. Bull. 46 (2011) 922–928.