Available online at www.sciencedirect.com
Applied Surface Science 254 (2008) 4179–4185 www.elsevier.com/locate/apsusc
Structural and electrical properties of zinc oxides thin films prepared by thermal oxidation Mihaela Girtan a,b,*, G.G. Rusu b, Sylvie Dabos-Seignon a, Mihaela Rusu b b
a Physics Department, Angers University, 2, Boulevard Lavoisier, 49045 Angers, France Faculty of Physics, Al.I. Cuza University, Boulevard Carol I, No. 11, Iasi 700506, Romania
Received 25 September 2007; received in revised form 29 December 2007; accepted 31 December 2007 Available online 9 January 2008
Abstract We report on zinc oxide (ZnO) thin films (d = 55–120 nm) prepared by thermal oxidation, at 623 K, of metallic zinc films, using a flash-heating method. Zinc films were deposited in vacuum by quasi-closed volume technique onto unheated glass substrates in two arrangements: horizontal and vertical positions relative to incident vapour. Depending on the preparation conditions, both quasi-amorphous and (0 0 2) textured polycrystalline ZnO films were obtained. The surface morphologies were characterized by atomic force microscopy and scanning electron microscopy. By in situ electrical measurements during two heating–cooling cycles up to a temperature of 673 K, an irreversible decrease of electrical conductivity of as flash-oxidized Zn films was revealed. The influence of deposition arrangement and oxidation conditions on the structural, morphological and electrical properties of the ZnO films is discussed. # 2008 Elsevier B.V. All rights reserved. PACS : 61.10.Nz; 68.37.Ef; 68.37.Ps; 72.80. r Keywords: Vacuum deposition; ZnO; Thin films; XRD; Electrical conductivity
1. Introduction ZnO films belong to transparent conducting oxides (TCO) class. Other transparent conducting oxides are: In2O3, SnO2, CdO, as non-doped or doped materials. The most common use of such materials is as electrodes for solar cells and other electronic devices [1–4]. For this type of application, best performances are given for instance, by ITO (Tin doped Indium Oxide). However, the production cost of ITO remains high, this being one of the factors which limit the development of solar cells market at large scale. These facts lead scientific world attention to more abundant and lower cost materials as ZnO. Generally, lower price depositions techniques of ZnO films are chemical methods and in this way interesting structures were already obtained [5].
* Corresponding author at: Physics Department of Angers University, POMA, Laboratory 2, Boulevard Lavoisier, 49045 Angers, France. Tel.: +33 241 735359; fax: +33 241 735216. E-mail address:
[email protected] (M. Girtan). 0169-4332/$ – see front matter # 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.apsusc.2007.12.055
Other simple method to prepare ZnO thin films, but for instance more less used, is thermal oxidation of metallic Zn films. Studies on other metallic oxides thin films prepared by this technique showed that the structure and morphology of the oxidized films strongly depend on the heating rate during oxidation process [6,7,8]. We can distinguish two manners to proceed: one’s corresponds to a gradually heating oxidation process (low heating) and the other to a directly subjection of films at high temperature: ‘‘flash-oxidation’’ (very fast heating). In our previous studies on ZnO films [9,10] were analysed the properties of Zn films oxidized by heating with slow rates (5 K/min). In this paper, we investigate the structural, morphological and electrical properties of ZnO films obtained by flash-oxidation of evaporated Zn thin films. 2. Experimental Preparation method of ZnO films consisted in: firstly, Zinc metallic films deposition onto glass substrates by thermal evaporation in vacuum, followed secondly, by thermal oxidation in an open atmosphere. Thermal evaporation was performed using a small resistance-heated tungsten boat in a
4180
M. Girtan et al. / Applied Surface Science 254 (2008) 4179–4185
chamber evacuated to a pressure of 5 10 5 Torr. The source material was zinc metal with a purity of 99.99% in the form of pellets. The substrates were soda lime glass plates of size 1.1 cm 1.1 cm 0.1 cm cleaned using chromic acid and washed with distilled water. The metallic zinc was evaporated by quasi-closed volume technique onto unheated substrates, simultaneously in two arrangements: horizontally and vertically relative to the vapour flux. Consequently, taking into account the geometry of our system, the angle, b, between the normal to the substrate and the direction of the incident vapours was of b = 08 for horizontal arrangement and about b = 608 for vertical arrangement. Source evaporation temperature was kept constant at 723 K. The as-deposited zinc thin films had a grey metallic lustre. More details on the deposition procedure are described in [11]. The thickness, d, of thin films was determined by Fizeau’s method for fringes of equal thickness [12], using an interferential microscope. The obtained values for the investigated samples were ranged between 55 and 120 nm. After deposition, Zn films were directly subjected to a temperature of 623 K (flash-oxidation) in an open resistive furnace and were maintained here at constant temperature for 20 min. After oxidation, films become transparent, the transmission coefficient being of about 70% at 900 nm. Samples structures were investigated by X-ray diffraction (XRD) using an D8 Advance Brucker diffractometer (Cu Ka 1,2), equipped with a linear Vantec super speed detector. The morphology of studied films was visualised by contact mode atomic force microscopy (AFM) and by scanning electron microscopy (SEM). The temperature dependence of electrical conductivity during some heating–cooling cycles was studied in the temperature range 300–673 K, with increasing (decreasing) temperature rate of 20 K/min. In order to avoid supplementary effects due to the photoconductivity of ZnO samples, the electrical resistivity was measured in dark at atmospheric air pressure, using surface type cells. Some deposition and characteristic parameters for typical studied samples are listed in Table 1.
Fig. 1. XRD patterns for ZnO films prepared by flash-oxidation at 623 K by direct subjection at this temperature of Zn films deposited on: (a) horizontal arranged substrates, and (b) vertical arranged substrates.
films. Samples ZnO-H10 and ZnO-V10 were obtained by flashoxidation of zinc films Zn-H10 and Zn-V10, respectively, for 20 min, at a temperature of 623 K. By H were denoted the films deposited in horizontal arrangement and by V—those deposited in the vertical ones. From Fig. 1b one can observe that the sample ZnO-V10 has a well-defined polycrystalline structure, whereas the XRD pattern of ZnO-H10 sample presented in Fig. 1a, indicates only the start of the crystallization process. The diffraction pattern for the ZnO-V10 sample fits correctly with the ZnO bulk wu¨rtzite crystallographic phase structure according to the ASTM standard [13]. The sharp diffraction peak at 2u = 34.48 (Fig. 1b) corresponds to the reflection on (002) plane of ZnO and indicates that in respective film, the ZnO microcrystallites grow preponderantly with their c-axis orientated perpendicular to the substrate surface. The additional weak peak at 2u = 43.28 corresponds to the Zn (1 0 1) reflection plane and indicates that an amount of zinc atoms was not completely reacted during the flash-oxidation process. These results relating to the structural characteristics of flash-oxidized Zn films are somehow different from those obtained in [10] for slow oxidized Zn films. In that
3. Results and discussion 3.1. Structural characteristics Fig. 1 shows the typical X-ray diffraction patterns for zinc oxide films prepared by flash-heating of the evaporated Zn Table 1 Sample
Zn-H10 Zn-V10
As-deposited
Flash-oxidized
Ts (K)
Tv (K)
d (nm)
l (cm)
b (degree)
r (nm/s)
Tox (K)
Ra (nm)
sc (V
300 300
723 723
58 93
7 5
0 60
0.64 1.03
623 623
40 90
46.7 45.6
1
m 1)
Ts, substrate temperature; Tv, temperature of Zn evaporation source; d, film thickness; l, source–substrate distance; b, incidence angle; r, growth rate; Tox, flashoxidized temperature; Ra, average roughness; sc, electrical conductivity before heating–cooling cycles.
M. Girtan et al. / Applied Surface Science 254 (2008) 4179–4185
4181
Fig. 2. Atomic force microscopy images obtained for as-deposited Zn films prepared in: (a) horizontal arrangement, and (b) vertical arrangement.
case, both typical ZnO films had a (1 0 1) textured structures. We should mention that both previously and actually studied Zn films exhibited the same crystalline structure, namely, Zn films deposited onto normally arranged substrates present (0 0 2) texture, whereas, those deposited onto vertical arranged substrates show a (1 0 1) texture. The different structures observed for the actually studied ZnO films could be correlated both with structural characteristics of the virgin Zn films and with the high heating rate used for their flash-oxidation. Aida et al. [14] revealed that the oxidation mechanism in the case of Zn films is achieved by the ionization of oxygen atoms at the films surface and subsequently followed by the diffusion of the produced ions in the film lattice. Such an oxidation mechanism could take place also in heated Zn films, but this process could be influenced by the specific crystalline structures of respective films. It is known that the thermal oxidation rate of the metallic films depends on the oxidation time, oxidation temperature and also on the crystallographic orientation [15]. In the case of virgin Zn-H10 sample, its (0 0 2) texture can be responsible for the quasi-amorphous structure of obtained ZnO-H10 sample. Indeed, for the Zn hexagonal structure, the (0 0 2) plane provides the highest atomic packing density [16]. This fact can determine a decreasing of the oxygen diffusion rate in the respective films, hence a lower thermal oxidation rate. In the same time, the higher treatment temperature (623 K) leads to the partial destroy of the Zn crystalline lattice (the Zn melting point is about of 693 K [17]). Owing to the high density of the crystallization centers determined by both the structural defects and different impurities, a quasi-amorphous structure for ZnOH10 sample will result (Fig. 1a). The lower sizes of the grains in Zn-H10 sample by comparison with those of Zn-V10 sample (see Fig. 2) could also contribute to the specific structure of ZnO-H10 sample. In the case of the ZnO-V10 sample, the preferred (1 0 1) orientation of the crystallites in virgin Zn-V10 film can favors the diffusion with a higher rate of the oxygen atoms through Zn lattice. Also, the larger thickness of the sample Zn-V10 (greater Zn amount), compared with those of the Zn-H10 sample, could limit the Zn melting effect in the respective films. This assumption is confirmed by the presence of the Zn (1 0 1) peak at 2u = 43.28 in the XRD pattern from Fig. 1b, which proves
that the un-oxidized Zn amount is still crystalline. All these factors determine the nanocrystalline c-axis textured structure for the ZnO-V10 sample. The presence of the (0 0 2) ZnO peaks in the both XRD patterns from Fig. 1 shows that the flash-oxidation technique favors the c-axis orientation growth of the ZnO crystallites in both typical evaporated Zn films. This assumption is confirmed by the XRD patterns from Fig. 3, recorded for the same ZnOH10 and ZnO-V10 samples after having been subjected to the heating–cooling cycles, during the electrical measurements. The intense (0 0 2) and the much weak (004) peaks of the hexagonal ZnO structure in the respective patterns, indicate that both samples present, finally, a c-axis oriented structure. Therefore, one can conclude that the flash-heating (oxidation) of the evaporated Zn films determines the start of the c-axis oriented grow of the ZnO crystallites. This orientation can be
Fig. 3. XRD patterns for the same samples from Fig. 1 recorded after them were subjected to heating–cooling cycles during electrical measurements.
4182
M. Girtan et al. / Applied Surface Science 254 (2008) 4179–4185
improved by the subsequent heating of the resulting ZnO films. Similar strongly c-axis crystal orientation was reported by other researches for ZnO thin films prepared by r.f. sputtering [18,19]. Using the well-known Debye–Schrerrer formula [20], the grain size values, relative to (0 0 2) peaks (Fig. 3), for finally oxidized ZnO samples, were calculated. The obtained values were of 13.2 nm for ZnO-H10 sample and 16.7 nm for ZnOV10 sample, respectively. The lower value obtained for ZnOH10 sample could be attributed to their initial quasi-amorphous structure after flash-heating oxidation. The different crystallization modes of two typical Zn samples (normally and obliquely deposited, respectively) during their thermal flash-oxidation were confirmed by the AFM and SEM studies.
The AFM images for typical samples, before and after their flash-oxidation, are presented in Figs. 2 and 4. As one could observe, after oxidation, the ZnO-H10 sample is characterized by hillock-like grains with nanometric dimensions, whereas the ZnO-V10 sample shows stacks of rectangular shaped pellets, arranged one over the other. The lower thickness of the successive pellets in comparison with their surface can be related with both direction dependent oxidation rate and different growth rate of various crystal faces bounding the microcrystallites [21,22]. The different surface morphologies of the virgin Zn samples (Fig. 2) can also contribute to their specific morphologies after oxidation. The film average roughness determined by AFM for oxidized samples were of about 40 nm for ZnO-H10 sample and about 90 nm for ZnO-V10 sample.
Fig. 4. Atomic force microscopy images obtained for ZnO films prepared by flash-oxidation of Zn films deposited in: (a) horizontal arrangement, and (b) vertical arrangement.
M. Girtan et al. / Applied Surface Science 254 (2008) 4179–4185
4183
Fig. 5. Scanning electron microscopy images of ZnO films prepared by flash-oxidation of Zn films deposited in: (a) horizontal arrangement, and (b) vertical arrangement.
At other scale, the SEM images (Fig. 5) have revealed that grown conditions favor the formation of needle-shaped grains in random orientations. We think that this kind of structure could be a promising alternative for rough transparent conducting electrode for solar cells. 3.2. Electrical behavior It is known [23,24] that between the electronic transport properties of polycrystalline semiconducting thin films and their structural characteristics there is a strong correlation. Particularly, both the values and the variation of the electrical conductivity of such films are in connection with their structure and its changes. On the other hand, thermal treatments of the respective films will modify their structural characteristics and consequently, their electrical properties. On this basis, the study of the temperature dependence of the electrical properties of thin films, may offer useful information about the possible changes of the structural characteristics of the films. Moreover, when this study is carried out in situ during successive heating and cooling cycles, such structural changes can be revealed [6,7,25–27]. In the case of transparent conducting oxides with applications in optoelectronics (as contact electrodes), the measurements of the electrical resistivity during many successive heating and cooling cycles may provide useful information on the thermal stability of the electrode and temperature limitations.
In the following, the variation of electrical resistivity during two successive heating–cooling cycles within temperature range 300–673 K for the two typical flash-oxidized ZnO samples are presented. The temperature dependencies of the electrical conductivity obtained for both ZnO-H10 and ZnOV10 samples are plotted in Figs. 6 and 7, respectively. From respective figures some conclusions can be deduced. Firstly, both samples present an irreversible variation of the
Fig. 6. The temperature dependence of electrical conductivity during heating– cooling cycles for ZnO films prepared by flash-oxidation of Zn films deposited in horizontal arrangement.
4184
M. Girtan et al. / Applied Surface Science 254 (2008) 4179–4185
possible dissimilarity between the values of heating (cooling) rate and oxidation rate. 4. Conclusions
Fig. 7. The temperature dependence of electrical conductivity during heating– cooling cycles for ZnO films prepared by flash-oxidation of Zn films deposited in vertical arrangement.
electrical conductivity during heating–cooling cycles. This reveals that the structures of as flash-oxidized ZnO films are still not stabilized and that an increase of heating temperature determines subsequent structural changes. From the same figures, one can remark that at a temperature of about 580 K begins the irreversible decrease of the electrical conductivity, mainly in the first heating step. We consider that the main factor that determines these behaviors of the electrical conductivity is the oxidation process that continues to take place during electrical measurements. Namely, due to the considerable interaction at higher temperature between the remaining Zn amount and atmospheric oxygen, a sharp decrease of the oxygen vacancies which act as shallow donors [28,29] takes place. The loss of the free charges being irreversible, the electrical conductivity will decrease continuously until the complete oxidation of ZnO films. Also, the evaporation of interstitial Zn atoms due to the higher vapour pressure of zinc [30] can reduce the carrier concentration and hence the electrical conductivity [31]. Certainly, the structural changes during ZnO film annealing can also influence the carrier mobility, m, hence the electrical conductivity. But this influence is less pronounced in comparison with that of changes in the stoichiometry of ZnO films [32–36]. On the other hand, the non-linear dependence of the oxygen diffusion rate on the concentration gradient, time and temperature [15] can play an important role in the variation of the electrical conductivity and their irreversibility during the annealing process. In the case of ZnO-H10 sample, the effect of these factors can be greater due to the transition from amorphous to crystalline structure of respective film. This can explain the sharp decrease of their electrical conductivity at higher temperature during first heating step. In the same time, from Figs. 6 and 7 one can observe that for both two ZnO samples the curves plotted in the cooling step do not match with those for next heating step within the same temperature range. Such behavior can be attributed to the
Polycrystalline Zn films were deposited in vacuum by quasiclosed volume technique onto horizontal and vertical arranged glass substrates. The obtained results indicate that the position of Zn substrates during their preparation, influence the structure and morphology of ZnO films subsequently obtained by ‘‘flash’’ thermal oxidation. The ZnO films originate in normally deposited Zn films present a quasi-amorphous structure whereas those originate in obliquely deposited Zn films are polycrystalline and c-axis oriented. By in situ electrical measurements during heating–cooling cycles, an irreversible decrease of electrical conductivity for both typical samples was revealed. This behavior is attributed mainly to the oxidation of remaining Zn amount in the films due to the higher heating temperature during electrical measurements. This assumption is sustained by the strong c-axis orientate polycrystalline structures received for both typical two ZnO films after electrical measurements. Acknowledgements Authors are grateful to N. Mercier from CIMMA— Laboratory for providing the necessary facilities for XRD studies and to R. Filmon from SCIAM—Angers for SEM scans. References [1] K.L. Chopra, S. Major, D.K. Pandya, Thin Solid Films 102 (1983) 1. [2] K.L. Chopra, S.R. Das, Thin Film Solar Cells, Plenum Press, New York, 1983. [3] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, IOP Publishing Ltd., 1995. [4] T. Soga, Nanostructured Materials for Solar Energy Conversion, Elsevier, 2006. [5] L. Vayssieres, Appl. Phys. A 89 (2007) 1. [6] M. Girtan, G.I. Rusu, G.G. Rusu, Mater. Sci. Eng. B 76 (2000) 156. [7] M. Girtan, G.I. Rusu, G.G. Rusu, S. Gurlui, Appl. Surf. Sci. 162–163 (2000) 490. [8] L. Leontie, M. Caraman, M. Alexe, C. Harnagea, Surf. Sci. 507–510 (2002) 480. [9] G.G. Rusu, M. Girtan, M. Rusu, Superlattice Microst. 42 (2007) 116. [10] G.G. Rusu, M. Rusu, N. Apetroaei, Thin Solid Films 515 (2007) 8699. [11] M. Rusu, I.I. Nicolaescu, G.G. Rusu, Appl. Phys. A 70 (5) (2000) 565. [12] K.L. Chopra, Thin Film Phenomena, McGraw-Hill, New York, 1969, p. 102. [13] ASTM X-ray Powder Diffraction Data File, Card 5-0664. [14] M.S. Aida, E. Tomasella, J. Cellier, M. Jacquet, N. Bouhssira, S. Abed, A. Mosbah, Thin Solid Films 515 (4) (2006) 1494. [15] H.F. Wolf, Semiconductors, Wiley-Interscience, New York, 1971. [16] J.Q. Hu, Q. Li, N.B. Wong, C.S. Lee, S.T. Lee, Chem. Mater. 14 (2002) 1216. [17] L.I. Maissel, R. Glang (Eds.), Handbook of Thin Film Technology, McGraw Hill Hook Company, 1970. [18] M. Gioffre, M. Angeloni, M. Gagliardi, M. Iodice, G. Coppola, C. Aruta, F.G. Della Corte, Superlattice Microst. 42 (2007) 85. [19] G. Rolo, J. Ayres de Campos, T. Viseu, T. de Lacerda–Aroso, M.F. Cequeira, Superlattice Microst. 42 (2007) 265.
M. Girtan et al. / Applied Surface Science 254 (2008) 4179–4185 [20] B.D. Cullity, R.S. Stock, Elements of X-Ray Diffraction, third ed., Prentice Hall, 2001. [21] R.Q. Guo, J. Nishimura, M. Ueda, M. Higashata, D. Nakamura, T. Okada, Appl. Phys. A 89 (2007) 141. [22] W.J. Li, E.W. Shi, W.Z. Zhong, Z.W. Yin, J. Cryst. Growth 203 (1999) 186. [23] L.L. Kazmerski (Ed.), Polycrystalline and Amorphous Thin Films and Devices, Academic Press, New York, 1980. [24] G. Harbeke (Ed.), Polycrystalline Semiconductors: Physical Properties and Applications, Spring, Berlin, 1985. [25] G.G. Rusu, J. Optoelectron. Adv. Mater. 8 (3) (2006) 931. [26] G.G. Rusu, M. Rusu, Solid State Commun. 116 (7) (2000) 363. [27] M. Girtan, Surface Coatings Technol. 184 (2004) 219. [28] Z.L. Wang, J. Phys. Condens. Matter 16 (2004) R829. [29] S. Kohiki, M. Nishitani, T. Wanda, J. Appl. Phys. 75 (1994) 2069. [30] Y. Igasaki, H. Saito, J. Appl. Phys. 69 (1991) 2190.
4185
[31] A. El Manouni, F.J. Manjon, M. Perales, M. Mollar, B. Mari, M.C. Lopez, J.R. Ramos Barrado, Superlattice Microst. 42 (2007) 134. [32] G. Epurescu, G. Dinescu, A. Moldovan, R. Birjega, F. Dipietrantonio, E. Verona, P. Verardi, L.C. Nistor, C. Ghica, G. Van Tendeloo, M. Dinescu, Superlattice Microst. 42 (2007) 79. [33] D. Dimova-Malinovska, H. Nichev, O. Angelov, V. Grigorov, M. Kamenova, Superlattice Microst. 42 (2007) 1233. [34] B.N. Mukashev, A.B. Aimagambetov, D.M. Mukhamedshina, N.B. Beisenkhanov, K.A. Mit’, I.V. Valitova, E.A. Dmitrieva, Superlattice Microst. 42 (2007) 103. [35] B.N. Mukashev, S.Zh. Tokmoldin, N.B. Beisenkhanov, S.M. Kikkarin, I.V. Valitova, V.B. Glazman, A.B. Aimagambetov, E.A. Dmitrieva, B.M. Veremenithev, Mater. Sci. Eng. B 118 (1–3) (2005) 164. [36] Ya.I. Alivov, A.V. Chernykh, M.V. Chukichev, R.Y. Korotkov, Thin Solid Films 473 (2) (2005) 241.