This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright Author's personal copy Vacuum 85 (2010) 307e311 Contents lists available at ScienceDirect Vacuum journal homepage: www.elsevier.com/locate/vacuum Grain growth and structural transformation in ZnS nanocrystalline thin films Shiv P. Patel a, *, J.C. Pivin b, V.V. Siva Kumar c, A. Tripathi c, D. Kanjilal c, Lokendra Kumar a a Physics Department, University of Allahabad, Allahabad-211002, India CSNSM, IN2P3-CNRS, Batiment 108, F-91405 Orsay Campus, France c Inter-University Accelerator Centre (IUAC), Aruna Asaf Ali Marg, New Delhi 110067, India b a r t i c l e i n f o a b s t r a c t Article history: Received 26 April 2010 Received in revised form 28 June 2010 Accepted 30 June 2010 Fabrication of ZnS thin films having similar stoichiometry at different substrate temperatures (TS) e.g. 200  C, 300  C and 400  C by means of RF magnetron sputtering method is presented. The films grown at TS of 200  C are in cubic zinc-blende phase and textured along (111) plane. The films deposited at TS of 300  C and 400  C are in hexagonal wurtzite phase. The surface roughness and grain size of the films increase with increasing TS. The ultra-violet and visible absorption studies show that the bandgap of films can be tailored by varying TS, taking advantage of the structural transformation. Ó 2010 Elsevier Ltd. All rights reserved. Keywords: Thin films Surface roughness Structural transformation X-ray diffraction Atomic force microscopy Bandgap 1. Introduction Zinc sulfide (ZnS), a IIeVI compound semiconductor, is an important material having various potential applications [1e5] particularly as laser [6] and light emitting diode [7]. ZnS is a suitable material as infrared (IR) window and solar dome of missiles [8]. An emerging field of applications of ZnS are data storage devices, data transfer, optical interfacing because electronics require semiconducting nanocrystalline thin film, which are sensitive in ultra-violet part of electromagnetic spectrum [9]. The exploration of properties of thin films as function of growth conditions like substrate temperature (TS), working pressure and fabrication method is a topic attracting intensive activities in materials science. Among these, substrate temperature greatly influences the atomic motion and grain growth kinetics of films, and is also an essential parameter for the development of crystalline films [10]. The substrate temperature also determines the crystal structure and stoichiometry of compound semiconductors thin films which affects quite a few properties [11e14]. Several techniques like sputtering [1,2], pulsed laser deposition (PLD) [12,13], thermal evaporation [15,16], chemical bath deposition (CBD) [17,18] and molecular beam epitaxy (MBE) [19,20] have been * Corresponding author. Tel./fax: þ91 532 2460993. E-mail addresses: shivpoojanbhola@gmail.com (S.P. Patel), lkumarau@gmail. com (L. Kumar). 0042-207X/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.vacuum.2010.06.011 used by various researchers for the development of ZnS thin film. Each fabrication method offers some advantages as well as drawbacks from the viewpoint of adhesion, film uniformity, stoichiometry and effective cost. However, RF magnetron sputtering method overcomes above drawbacks and is widely used in numerous applications [1,2,11]. The magnetic field applied behind the cathode improves the films quality because the probability of ionization of gases in the sputtering chamber increases due to increase in path length of electrons in the presence of magnetic and electric field. ZnS has two polytypes with zinc-blende (ZB) cubic and wurtzite (W) hexagonal crystal structures, out of which the ZB phase is more stable at ambient temperature and pressure. The bandgap of bulk ZnS is 3.67 eV for ZB phase and 3.90 eV for W phase, and the ZB phase transforms into W phase at 1020  C [21]. In the case of nanoparticles, the transition temperature depends on their size and is for instance of 400  C for 3 nm [22]. There are other experimental and theoretical studies which deal with stability of these phases [23,24], but only one at our knowledge which discussed the transformation in ZnS thin films with substrate temperature and the increase in bandgap was correlated with the volume fraction of W phase [25]. The structural transformation has been observed between 300  C and 400  C. As for any IIeVI compound semiconductor, the bandgap of ZnS is a direct consequence of the crystal structure and stoichiometry [12e14,26]. Therefore, it is interesting to study variation in the bandgap and crystal structure of ZnS thin films with substrate temperature. In the present paper, the RF magnetron sputtering is used to grow crystalline and textured ZnS Author's personal copy 308 S.P. Patel et al. / Vacuum 85 (2010) 307e311 thin films at different substrate temperatures (TS). The correlation between the grain growth, surface roughness, structural transformation from ZB to W phase, stoichiometry and bandgap of films with TS will be discussed in detail. 2. Experimental RF magnetron sputtering (13.56 MHz) method was utilized to fabricate ZnS thin films on (100)-oriented silicon (Si) and fused silica substrates. The films were deposited at 200  C, 300  C and 400  C substrate temperatures. The vacuum chamber was evacuated up to 1.2  10 5 mbar base pressure using a turbo molecular pump. A commercially available high purity (99.99%) ZnS target of 51 mm in diameter was used for the sputtering. The sputtering was carried out with Argon-plasma at gas pressure of the order of 2.0  10 2 mbar using 150 W RF power. A 200-Gauss magnet was kept behind the target to apply the magnetic field in the sputtering chamber. The distance from the ZnS target to substrates was fixed to about 40 mm for the deposition. The Rutherford backscattering spectrometry (RBS) measurements were performed to determine the composition and thickness of films on Si substrates. The RBS analyses were carried out using 1.8 MeV Heþ1 ions provided by the ARAMIS accelerator of CSNSM at Orsay campus, France, at normal incidence to the sample surface with collection of the backscattered ions at an angle of 165 . The crystal structures and grain size of the film were analyzed by X-ray diffraction (XRD). The XRD patterns were recorded with Bruker D8 Advance diffractometer at glancing A) source. The surface topogangle of 1, using Cu Ka (l ¼ 1.5406  raphy, roughness and grain size of films were analyzed by means of atomic force microscopy (AFM) in taping mode, using a Digital Instruments Nanoscope IIIa. Ultra-Violet and Visible (UVeVis) absorption spectra were recorded on dual beam Hitachi U-3300 spectrophotometer for electronic bandgap studies of films. 3. Results and discussion Fig. 1 displays the RBS spectrum of ZnS nanocrystalline thin film deposited on Si at substrate temperature 400  C. There is no detectable contamination of carbon, oxygen or any other impurity Fig. 1. RBS spectrum of ZnS thin film deposited at substrate temperature (TS) 400  C. The star symbols show the experimental data and solid line shows the simulated result carried out by using XRUMP code. The film thickness is around 800 nm. The content of Zn is 50.5 atomic % and S is 49.5 atomic % in the film. in the film. The contents of Zn and S in the film are found to be 50.5 at % and 49.5 at %, respectively as analyzed by XRUMP simulations code with an accuracy better than 1 atomic%. The thickness of the film is around 800 nm as determined by simulation using RBS data. The RBS analyses of films deposited at 200 and 300  C gave similar results of stoichiometry and film thickness. The XRD patterns of films deposited on silicon at different TS are shown in Fig. 2. The film deposited at 200  C shows an intense diffraction peak centered at 28.54 corresponding to the (111) lattice plane of zinc-blende phase (Joint Committee on Powder Diffraction Standard, JCPDS Card No. 77-2100), marked by dotted line and labeled ZB in the pattern. No other diffraction peaks is observed in the pattern which confirms that the film deposited at 200  C TS is textured along (111) plane. The films deposited at 300  C TS contain a mixture of W and ZB phases. The films deposited at 400  C exhibit diffraction peaks at 26.84 , 28.34 , 47.36 and 56.18 which are characteristic positions of (001) (002), (110) and (112) lattice planes for the wurtzite-ZnS (JCPDS Card No. 75-1547), as labeled W in the pattern. Inset of Fig. 2 shows more clearly the shifting of (111) peak of ZB phase to (002) peak of W phase containing structural transformation of ZnS from ZB to W phase. Similar XRD pattern and structural transformation have been observed on fused silica substrate as well. It is found that the growth of textured ZnS film is independent of underlying periodicity of the substrate. The texturing of films has been reported in some of the oxides materials [27,28] and ZnS [29] with different film deposition method, in which films orientation is independent of periodicity of substrates. The fused silica has no long range periodicity that can help the oriented/epitaxial growth of crystalline films as in case of single crystal (100)-oriented Si substrate but we have observed texturing on fused silica as well [29]. The minimization of surface energy, which is driving force for the film growth, is responsible for the texturing of ZnS film on fused silica and silicon. In growth process, the orientation of films preferred in such a crystallographic plan which have highest atomic density lies parallel to the substrate surface [27,28]. The DebyeScherrer formula (D ¼ kl/bCosq) is used to determine the grain size, where D is the mean grain size, k is a geometric factor (¼0.9), l is the wavelength of the X-ray wavelength, b is the full width at half Fig. 2. XRD patterns of ZnS thin films grown at different TS. The film deposited at TS of 200  C is extremely textured along (111) plane of ZB-ZnS while those deposited at TS of 300 and 400  C are in W-ZnS and lose their textured nature. The inset shows the extended part of Fig. 2 which shows the shifting of (111) plane of ZB phase to (002) plane of W phase. Author's personal copy S.P. Patel et al. / Vacuum 85 (2010) 307e311 maximum (FWHM) of diffraction peaks and q is the diffraction angle [30]. The grain size of films at different TS of 200  C, 300  C, and 400  C are found to be 19 nm, 23 nm, and 25 nm, respectively. Fig. 3(a), (b) and (c) shows the AFM images of ZnS thin films deposited on Si at TS of 200  C, 300  C, and 400  C, respectively. The surface roughness and grain size of the films at various TS are shown in Fig. 3(d). As it is clear from the topographic images that the higher TS lead to bigger grains at the surface which are in agreement with XRD results. It should be pointed out here that the XRD gives the in plane grains size while AFM gives the out of plane grain size. The lower surface roughness and grain size at TS of 200  C is due to the extremely textured nature of the films. On the contrary, the film which is grown at TS of 300 and 400  C grains has more random orientation instead of columnar shape. At higher TS, the enhancement of surface mobility of adatom also promotes the grain growth at the expense of smaller grains [31,32]. Both factors, loss of texture and grain growth explain the higher surface roughness. The final grain size in the film is a function of the respective crystal nucleation and growth rates, and the latter is affected by the stress developed in the film [31]. Fig. 4 shows the UVeVisible absorption spectra of ZnS thin film deposited at different TS. The bandgap of films is calculated by using 309 Fig. 4. UVeVis absorption spectra of ZnS thin films deposited at different TS. The inset shows the plot of (ahn)2 versus hn for determination of bandgap. The bandgap of films grown at TS of 200, 300 and 400  C are found to be 3.38, 3.44 and 3.49 eV, respectively. Fig. 3. AFM images of ZnS thin films deposited at (a) TS ¼ 200  C (b) TS ¼ 300  C (c) TS ¼ 400  C and (d) variation in roughness and grain size with TS. Author's personal copy 310 S.P. Patel et al. / Vacuum 85 (2010) 307e311 Tauc’s plots [12e14] as shown in inset of Fig. 4. The bandgaps are found to be 3.38 eV, 3.44 eV, and 3.49 eV at TS of 200  C, 300  C, and 400  C respectively. The observed bandgap in present study are less as compared to that of bulk ZnS of 3.67 eV for ZB and 3.90 eV for W phase [21]. A smaller value of bandgap for all the films compared to bulk compounds as observed here is usually due to defect states which widen the tail in band edge. Thin film may contain a few hundreds ppm of S vacancies or Zn interstitials as supported by RBS results, explaining the observation of lower values of bandgap than that for bulk stoichiometric phases [12]. One interesting observation is that the bandgap of films increases with increasing TS. Theoretically, the bandgap of nanocrystalline ZnS thin films can be affected by following factors: i) the small grain size which tends to increase the gap of any semiconductor, ii) the fraction of W and ZB phases [25], iii) the Zn to S stoichiometry ratio. Recently, Luo et al. have reported an increase in bandgap of ZnS thin films with increasing deposition temperature and explained their results on the basis of Zn and S stoichiometry [13]. They proposed that at low temperatures, the films are deficient in S, which causes the reduction in ZnS bandgap. With the increase in deposition temperature, the composition ratio of Zn and S become close to 1, so that the bandgap increases up to 3.7 eV. In the present work, RBS analyses indicates that the films deposited at different TS have similar composition, and the stoichiometry does not play any role in increasing the bandgap. The observed growth of grains in the films with increasing TS should lead to a narrowing of the gap by suppression of the confinement effect, so that the actual increase in bandgap of films with TS is surprising. Therefore remain only the structural phase change as origin of increase of the bandgap with TS. The similar explanation about increase in bandgap of ZnS thin film with increasing TS has been proposed by Yeung et al. [25]. In their study the bandgaps were 3.28 eV, 3.44 eV, 3.47 eV and 3.5 eV for the film deposited at 150  C, 250  C, 350  C and 450  C, respectively. Qadri et al. [22] have reported that ZnS nanoparticles of sizes 2.7 nm change its structure at 400  C during annealing. They observed no W phase up to 350  C but above this annealing temperature, W phase started to appear and the proportion of W phase was of 28% at 400  C. The proportion of W phase is already preponderant in our samples deposited at 300  C and films deposited at 400  C are made of W only, probably because of the different preparation technique than in Qadri et al. study [22]. As far as size dependent properties of low-dimensional materials are concerned, the present study relates to nanostructured thin films, whereas Qadri et al. [22] studied structural transformations in nanoparticles. Nanostructured materials (NSs) and nanoparticles (NPs) may have different properties, such as melting temperature, lattice expansion, and cohesive energy [33,34]. The differences of cohesive energy of NSs in comparison with NPs reflect that the thermal properties of NSs are more stable than the NPs. Clearly, the differences in properties is caused by the coordination number imperfection at the surface related to the presence of broken bonds in NPs while in NSs is attributed to the bond elongation at the grain boundaries [33]. Thermodynamically, a chemical process always develops towards the lower Gibbs free energy. The ZnS films minimize their Gibbs free energy by taking the W structure at higher TS of 300 and 400  C [35,36]. The coefficient of volume thermal expansion, defined by aV ¼ (dV(T)/dT)/V(T), of W-ZnS is lower than that of ZB-ZnS [36]. The Gibbs free energy variation DG ¼ DF þ PDV, and the smaller volume expansion coefficient of W phase make it stable at higher TS [35,36]. The change in Helmholtz free energy DF is always negative as temperature increases, and W phase of ZnS has larger DF than the ZB phase up to 1200 K [36]. It may be noted here that the temperature is not the only factor which decides the structural phase stability and phase transition. There are several other parameters like, pressure and thickness of the film. Experimentally, it was observed that when ZnS nanoparticles of 3 nm sizes were heated in vacuum, transformed into W phase from ZB phase, slightly at lower temperature [23]. Le et al. have developed a model and included many parameters in variation of Gibbs free energy to explain the phase selection phenomena of ZnS thin films and nanoparticles. The variation of Gibbs free energy is given as DG (T,P,D) ¼ DGv (T,0,N) þ DGs (D) þ DGe (P,D), where DGv (T,0,N) is variation of Gibbs free energy in bulk crystal, DGs (D) is a Gibbs surface free energy as function of size, DGe (P,D) is Gibbs elastic energy dependent on size and pressure [35]. In the present case, the nanocrystalline nature of film, substrate temperature and very low deposition pressure makes the formation of stable W phase of ZnS at 400  C. 4. Conclusion Textured and well crystallized ZnS thin films have been synthesized at substrate temperature (TS) of 200  C. The films grown at TS of 300  C and 400  C lose their textured nature. The films deposited at different TS have similar stoichiometry as shown by RBS analyses. Grain growth in the film is observed with increasing TS which leads to the increase in surface roughness. A structural transition of ZnS from ZB to W phase is observed which causes the increase in bandgap of films with increasing TS. In conclusion, by controlling the substrate temperature (TS), one can engineer several properties like grain size, surface roughness, crystal structure and bandgap of the ZnS films. Acknowledgment The author (SPP) would like to express his sincere thanks to Council of Scientific and Industrial Research (CSIR), New Delhi for providing financial assistance through junior research fellowship (JRF) to carry out this work. The authors are thankful to Mr. P. K. Kulariya for XRD measurements. We are also thankful to Mr. S. A. Khan for kind help. The help received from Mr. Jai Prakash is greatly acknowledged. References [1] Nitta A, Tanaka K, Maekawa Y, Kusabiraki M, Aozasa M. Thin Solid Films 2001;384:261. [2] He Q, Guo H-bin, Wei J-jun, Askari SJ, Wang H-bin, Zhang S-yu, et al. Thin Solid Films 2008;516:4695. [3] Leftheriotis G, Yianoulis P, Patrikios D. Thin Solid Films 2000;306:92. [4] Kashani H. Thin Solid Films 1996;288:50. [5] Goudarzi A, Aval GM, Sehraei R, Ahmadpoor H. Thin Solid Films 2008; 516:4953. [6] Zapien JA, Jiang Y, Meng XM, Chen W, Au FCK, Lifshitz Y, et al. Appl Phys Lett 2004;84:1189. [7] Katayama N, Oda S, Kukimoto H. Appl Phys Lett 1975;27:697. [8]. Deom AA, Luc A, Amara S, Balageas DL. Wear 1999;233-235:13. [9] Yano S, Schroeder R, Ullrich B, Sakai H. Thin Solid Films 2003;423:473. [10] Afifi HH, Mahmoud SA, Ashour A. Thin Solid Films 1995;263:284. [11] Mandal SK, Chaudhuri S, Pal AK. Thin Solid Films 1999;350:209. [12] Patel Shiv P, Chawla AK, Chandra R, Prakash Jai, Kulriya PK, Pivin JC, et al. Solid State Comm 2010;158:1158. [13] Luo P-f, Jiang G-S, Zhu C-f. Chin J Chem Phys 2009;22:97. [14] Daranfed W, Aida MS, Hafdallah A, Lekiket H. Thin Solid Films 2009;518:1082. [15] Wu X, Lai F, Lin L, Lv J, Zhuang B, Yan Q, et al. Appl Surf Sci 2008;254:6455. [16] Krasnov AN, Hofstra PG, McCullough MT. J Vac Sci Technol A 2000;18:671. [17] Roy P, Ota JR, Srivastav SK. Thin Solid Films 2006;515:1912. [18] Koutsogeorgis DC, Mastio EA, Cranton WM, Thomas CB. Thin Solid Films 2000;365:134. [19] Valkonen MP, Lindroos S, Resch R, Leskela M, Friedbacher G, Grasserbauer M. Appl Surf Sci 1998;136:131. [20] Tran TK, Park W, Tong W, Kyi MM, Wagner BK, Summers CJ. J Appl Phys 1997;81:2803. [21] Gilbert B, Frazer BH, Zhang H, Haung F, Banfield JF, Haskel D, et al. Phys Rev B 2002;66:245205. Author's personal copy S.P. Patel et al. / Vacuum 85 (2010) 307e311 [22] Qadri SB, Skelton EF, Dinsmore AD, Hu JZ, Kim WJ, Nelson C, et al. J Appl Phys 2001;89:115. [23] Zhang H, Huang F, Gilbert B, Banfield JF. J Phys Chem B 2003;107:1305. [24] Wright K, Gale JD. Phys Rev B 2004;70:035211. [25] Yeung KM, Tsang WS, Mak CL, Wong KH. J Appl Phys 2002;92:3636. [26] Kale RB, Lokhande CD. Semicond Sci Technol 2005;20:1. [27] Yoon J-G, Oh HK, Lee SJ. Phys Rev B 1999;60:2839. [28] Singh MP, Thakur CS, Shalini K, Banerjee S, Bhat N, Shivashankar SA. J Appl Phys 2004;96:5631. 311 [29] Wei M, Choy K-L. Chem Vap Deposition 2002;8:15. [30] Cullity BD. Element of X-ray diffraction. 2nd ed. Notre Dame: Addison-Wesley Publishing Company, Inc.; 1977. [31] Thomson CV. J Appl Phys 1985;58:763. [32] Palmer JE, Thomson CV, Smith HI. J Appl Phys 1987;62:2492. [33] Zhu YF, Zheng WT, Jiang Q. Appl Phys Lett 2009;95:083110. [34] Xiao S, Hu W, Yang JJ. Phys Chem B 2005;109:20339. [35] Li S, Lian JS, Jiang Q. Chem Phys Lett 2008;455:202. [36] Wang SQ. Appl Phys Lett 2006;88:061902.