Structural effects due to the incorporation of Ar atoms in the lattice of ZrO2 thin films prepared by ion beam assisted deposition

Structural effects due to the incorporation of Ar atoms in the lattice of ZrO2thin films prepared by ion beam assisted deposition J.P. Holgado a , R. Escobar Galindo b , A. van Veen b , H. Schut b , J.Th.M. de Hosson c , A.R. Gonz aalez-Elipe a,* a Instituto de Ciencia de Materiales de Sevilla (CSIC/Universidad de Sevilla) and Dpto. de Qu ıımica, Inorg aanica Univ. Sevilla., Av. Americo Vespucio S/N, E-41092 Sevilla, Spain b Interfaculty Reactor Institute, Defects in Materials, Delft University of Technology, Mekelweg 15, 2629 JB Delft, The Netherlands c Materials Science Center, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands Received 4 October 2001; received in revised form 27 January 2002 Abstract Two sets of ZrO2thin films have been prepared at room temperature by ion beam induced chemical vapour deposition and subsequently annealed up to 1323 K. The two sets of samples have been prepared by using either Oþ 2or mixtures of (Oþ 2þArþ) ions for the decomposition of a volatile metallorganic precursor of zirconium. The structure and microstructure of these two sets of samples have been determined by means of X-ray diffraction, Fourier transform infrared spectroscopy and positron beam analysis (PBA). The samples were very compact and dense and had a very low-surface roughness. After annealing in air at TP573 K both sets of films were transparent and showed similar refraction indexes. For the (Oþ 2þArþ)-ZrO2thin films it is shown by X-ray photoelectron spectroscopy and Rutherford back scattering that a certain amount of incorporated Ar (5–6 at.%) remains incorporated within the oxide lattice. No changes were detected in the amount of incorporated Ar even after annealing at T¼773 K. For higher annealing temperatures (T>1073 K), the amount of Ar starts to decrease, and at T¼1223 K only residual amounts of Ar (<0.4%) remain within the lattice. It has been found that as far as Ar atoms remain incorporated within the ZrO2network, the (Oþ 2 –Arþ)-ZrO2films present a cubic/tetragonal phase. When the amount of ‘‘embedded’’ Ar decreases, the crystalline phase reverts to monoclinic, the majority phase observed for the (Oþ 2)-ZrO2films after any annealing treatments. The microstructure of the films after different annealing treatments has been investigated by PBA. The presence of Ar ions and the initial amorphous state of the layers were detected by this technique. An increase of the open volume was observed after annealing up to 773 K in both sets of samples. For higher annealing temperatures the samples showed a progressive crystallisation resulting in a decrease of the open volume. During this sintering the samples without embedded Ar present a higher concentration of open volume defects. After the release of Ar occurs (TP1223 K) both samples approach to a similar defect free state. Nuclear Instruments and Methods in Physics Research B 194 (2002) 333–345 www.elsevier.com/locate/nimb * Corresponding author. Fax: +34-954-460-665. E-mail address: [email protected] (A.R. Gonz aalez-Elipe). 0168-583X/02/$ - see front matter Ó2002 Elsevier Science B.V. All rights reserved. PII: S0168-583X(02)00695-X The incorporation of Ar within the ZrO2thin film structure, is proposed as the main factor contributing to the stabilisation of the cubic/tetragonal phase of ZrO2at room temperature. Ó2002 Elsevier Science B.V. All rights reserved. Keywords: Chemical vapour deposition; Ion bombardment; Thin film; Positron annihilation; Phase transformation; Ion assisted deposition; ZrO2 1. Introduction Among many other applications, such as optical coatings [1], buffer layers for growing superconductors [2], heat resistant coatings [3], oxygen sensors [5], etc. ... , ZrO2thin films have attracted great attention because of their properties as ion conductors [4], to be used in the so-called solid oxide fuel cells (SOFCs) [5]. For all these applications, and specially in the case of SOFCs, it is generally desirable to obtain the tetragonal or cubic phase of zirconia. It is known that although the tetragonal phase of ZrO2is unstable at room temperature, it can be stabilised by doping with cations such as Y3þ,Ca 2þ, etc. [6], or by decreasing the particle size of the crystalline domains [7]. The synthesis of ZrO2or ZrO2doped thin films (i.e. ZrO2-M) has been attempted by many different methods including spray deposition [8], chemical vapour deposition (CVD) [9], electron evaporation [10], and other methods [11]. Very compact ZrO2ðMÞthin films can be prepared by ion beam assisted deposition procedures (IBAD) [12] that involve the bombardment of the growing film with low-energy ion beams (i.e. generally EP1000 eV), usually Oþ 2,Ar þor mixtures of both. By these methods, it is very common to observe the appearance of different crystallographic effects, such as amorphisation or the preferential growth of some crystallographic planes [13,14] that are attributed to the influence of the bombardment with low-energy ions [15]. Martin [16] has shown that monoclinic to tetragonal phase transformations can be induced in ZrO2thin films by bombardment with Arþions of relatively low energy. For bulk ZrO2the effect of high-energy Arþbombardment in inducing crystallographic transformations and the stabilisation of the cubic/ tetragonal phase is well known [17]. In most cases an implicit explanation of this type of monoclinic to cubic/tetragonal transformation relies on the assumption that ballistic interactions between the energetic ions and the target atoms are responsible for the phase change. The results presented in the present paper do not exactly support such an explanation based on energetic effects, but rather emphasise the role of Ar embedded within the ZrO2thin film in producing the stabilisation of the cubic/tetragonal phase. We compare the structural evolution of ZrO2thin films prepared by ion beam induced chemical vapour deposition (IBICVD) by using either Oþ 2or mixed Oþ 2þArþ ion beams. In the latter case the cubic/tetragonal phase is stabilised upon annealing, as long as Ar atoms remain incorporated within the thin films. In a previous paper [18] we have studied in detail how the texture, morphology and optical properties of the resulting films are affected by the use of Arþions. The first motivation for the present paper is then to highlight how the presence of incorporated Ar may stabilise the cubic/tetragonal structure of zirconia acting in a similar way than ‘‘doping’’ cations such as Y3þ,Ca 2þ, etc. and to provide an analysis of the evolution of structural defects in the samples upon annealing, as determined by positron beam analysis (PBA). 2. Experimental Thin films have been prepared by IBICVD [19]. This method consists of the decomposition of a volatile metallic precursor by means of an ion beam. Further details about the method as applied to the synthesis of oxides and nitrides can be found in previous publications [19–21]. Zirconium tetra ter-butoxide, Zr(Obut)4has been used as precursor. Typical partial pressure 334 J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 during deposition was 4 105mbar. Oþ 2or mixtures of Oþ 2þArþions, accelerated to 400 eV, were supplied by an ‘‘IQP’’ broad ion source. O2þAr mixtures with a molar ratio 10:1 were used for preparation of Ar-incorporated samples. A beam current of around 40 lAcm 2was measured at the sample holder during deposition. The total pressure in the chamber during deposition was 5 104mbar. Under these conditions a growth rate of approximately 2.0 nm min1was measured with a quartz crystal monitor at the sample position. Silicon (1 0 0) wafers and fused quartz have been used as substrates for preparation of the samples. X-ray diffraction (XRD) spectra were recorded by using the Cu Karadiation with a Siemens D5000 diffractometer. An incidence angle of 0.5° was chosen to increase the sensitivity. IR spectra were obtained with a Nicolet spectrometer at 8 cm1resolution for samples prepared on Si wafers. The silicon substrate was partially transparent to infrared radiation showing a small feature at 609 cm1. Annealing of the samples in air was carried out in a conventional atmospheric furnace, using a linear heating ramp of 10 K min1. At the end of the ramp the selected final temperature was kept for 3 h followed by cooling down slowly. Rutherford back scattering (RBS) spectra were recorded with a tandem ion accelerator (Pelletron 9SH-2 from NEC) at the Centro Nacional de Aceleradores (CNA) in Seville. Alpha particles of 2.021 MeV and usual settings of scattering angle and detector geometry were used for recording of spectra. The data acquisition was done with a total sample charge of 5 lC. Simulations of composition profiles have been done with the RUMP program code [22]. X-ray photoelectron spectroscopy (XPS) was carried out using an ESCALAB 210 spectrometer working in pass energy constant mode with a value of 50 eV. The surface concentration of the elements was determined by measuring the peak areas and correcting by the appropriate sensitivity values. The PBA was performed using the Delft variable energy positron beam (VEP) [23]. The positrons were injected in the samples with energies between 100 eV and 25 keV. Annihilation of positrons with electrons results, in the center-of-mass system, in the emision of two collinear photons of 511 keV each. In the laboratory frame the nonzero momentum of the positron-electron pair causes an energy shift of the photons. This results in a broader energy distribution, called Doppler broadening of the annihilation radiation (DBAR) around the 511 keV peak. This broadened energy spectrum is characterised by two parameters: S and W. The S(shape) parameter is defined as the ratio between the central area and the total area of the annihilation photopeak. This parameter reflects the annihilation of positrons with valence electrons (low-momentum electrons). In general, a high value of Sindicates positron annihilation in open volume defects. If the material allows the formation of a Positronium state (Ps) this will contribute to the DBAR with a narrow component because of the low-intrinsic moment of the parapositronium. A second useful parameter for the analysis of DBAR is the W(wing) parameter, which reflects the positron annihilation with highmomentum electrons (core electrons). The W- parameter is sensitive to changes in the chemical environment. Both parameters can be combined in S–W maps with a third variable (e.g. the implantation energy, temperature or strain) as a running parameter. These maps are useful to trace positron trapping in, e.g., layers as in the samples described in this work. The data was analysed with the variable energy positron fit (VEPFIT) program [24]. The samples studied here are modelled by a number of stacked slabs. Each of these slabs is characterised by a thickness (di), Siand Wiparameters, and a positron diffusion length (Li). The measured SðEÞand WðEÞas a function of the implantation energy E, are a sum of the characteristic Siand Wivalues of the trapping layers (cluster points) weighted by the fraction of positrons trapped in each layer fiðEÞ. The spectra were recorded with a single Ge solid-state detector. All experiments were carried out at room temperature under a vacuum of about 106Pa. The values of S and Wparameters are normalised with respect to the Sand Wvalues for a crystal Si substrate (SSi ¼0:58, WSi ¼0:031). J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 335 3. Results 3.1. Chemical composition The determination of the chemical composition of the zirconia films was done by XPS and RBS. The photoelectron spectra of both set of samples, prepared by using Oþ 2or (Oþ 2þArþ) ions, were typical for a ZrO2stoichiometry. Besides a carbon peak, which is easily removed by bombardment with Arþor Oþ 2(and thus proving that it is due to contamination by exposure to air), no other contaminant species were detected at the surface of samples. In the case of (Oþ 2þArþ)-ZrO2samples, Ar2p and Ar2s peaks could be detected in the photoelectron spectra of the films. Their intensities were equivalent to an atomic percentage of approximately 5–6%, a value that did not change after cleaning by sputtering with Oþ 2ions of 3 keV energy. The samples were also analysed by RBS to obtain a more precise quantitative determination of the amount of Ar present in the samples, as well as an evaluation of the chemical composition profile for the whole film thickness. Fig. 1 shows experimental and theoretically calculated spectra for the as prepared (Oþ 2þArþ)-ZrO2and Oþ 2- ZrO2samples. The right of this figure shows, in an enlarged scale, the Ar region of some spectra corresponding to (Oþ 2þArþ)-ZrO2samples annealed at increasing temperatures up to 1423 K. In the (Oþ 2)-ZrO2and (Oþ 2 –Arþ)-ZrO2samples the O/ Zr ratio was close to 2.0, although a slightly smaller ratio (i.e. ffi1.95–2.00) was obtained for the as prepared (Oþ 2 –Arþ)-ZrO2thin film. For these samples, Ar (5–6 at.%) is found to be distributed homogeneously over the whole film thickness even after their annealing at temperatures as high as 873 K. After annealing of the samples at T>773 K the concentration of Ar within the film decreases. It is found that at 1073 K only 3% of Ar is still retained by the thin film, and that a smaller concentration is observed at its surface. After annealing at 1323 K, only a 0.5 at.% of Ar is present, with an even lower concentration at the surface of the thin film. These results indicate that annealing treatments above 1073 K provoke a progressive loss of the Ar incorporated within the structure of zirconia films. This loss is more pronounced from the topmost surface layers of the ZrO2thin films. Fig. 1. Experimental and simulated RBS spectra of the (Oþ 2þArþ)-ZrO2sample. The inset shows, in an enlarged scale, the region corresponding to embedded Ar after annealing treatments at indicated temperatures. 336 J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 3.2. Crystallographic structure Fig. 2 shows XRD patterns for (Oþ 2)-ZrO2thin films prepared at 298 K and after annealing at increasingly higher temperatures up to 1323 K. For comparison, the reference patterns of the most intense peaks of the different phases of zirconia are included in the top part of the figure. The spectrum of the as prepared (Oþ 2)-ZrO2sample reveal the existence of at least two phases: monoclinic and cubic and/or tetragonal. It should be remarked here that the diffractograms corresponding to these two phases present a series of peaks with very similar intensity and with very close positions. Due to the large width of the peaks obtained in the diagram of the thin film and the possible displacement of the positions of the peaks respect to those of the pure phases due to stress and strain effects, even a careful analysis of the positions of the peaks do not permit to discriminate between the cubic and/or the tetragonal phase of zirconia in the thin film. Thus, in the following discussion a mention to cubic/tetragonal phase should always be taken as a possible existence of any of those phases or a mixture of both. Annealing at T<773 K leads to a relative increase of the monoclinic phase, although only a small change in the width of the peaks is observed in this range of temperatures. By contrast, when the samples are annealed at higher temperatures (T>873 K), a drastic decrease is observed in the intensity of the peaks associated with the cubic/ tetragonal phase. These peaks practically disappear from the spectrum when the film was annealed at TP1223 K. This annealing treatment induces an increase in the height and a decrease in the width of the peaks corresponding to the monoclinic phase (Table 1). For the whole temperature range of annealing, the relative intensity of the peaks of the monoclinic phase was far from the ratio observed in a single crystal, thus indicating the preferential growth in the films of some crystallographic planes (cf. Table 1). XRD patterns for (Oþ 2þArþ)-ZrO2films prepared at 298K and subsequently annealed at increasingly higher temperatures up to 1323 K are presented in Fig. 3. The XRD pattern of the ‘‘as deposited’’ (Oþ 2þArþ)-ZrO2films depicts only two broad features at diffraction angles which suggest the existence of a badly ordered cubic or tetragonal structure of ZrO2. Again, the width of the peaks makes it very difficult to discriminate between these two phases, and in the following discussion it should always be kept in mind the possible existence of either cubic/tetragonal phases or even a mixture of both. Although the position of the diffraction lines is consistent with those of the cubic/ tetragonal phases, the observed intensities are far Fig. 2. XRD patterns of Oþ 2-ZrO2thin film for the as prepared samples and after annealing treatments at indicated temperatures. The patterns of the most intense peaks of monoclinic (bottom), cubic (top, full line) and tetragonal (top, dashed line) phases of zirconia are included for comparison. J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 337 from reference patterns. In particular, the high intensity of the peak at 2hffi59:4°, indicates a preferential growth of the film according to the corresponding (3 1 1) planes of the cubic phase and/ or the (2 1 1) plane of the tetragonal phase. The preferential growth of planes with high Miller indices has been recently reported by us for Fe2O3 thin films prepared also by IBICVD [25]. Such a preferential growth was related with the presence of Ar embedded in the structure. By annealing these zirconia samples at TP573 K the diffraction pattern changed. The maximum of the peak at 2h¼59:4°becomes slightly broader with a small asymmetry at higher diffraction angles. Simultaneously, a broad feature appears at 2h¼33–35°. The comparison of these structures with the lines corresponding to the reference pattern of the bulk material, suggests the existence of a badly ordered cubic/tetragonal phase of zirconia in the as prepared samples and their progressive ordering or restructuring with the annealing treatments. A significant change in the diffraction diagram of the (O2þArþ)-ZrO2samples is observed after annealing at T¼1073 K. In this case the peak corresponding to planes (3 1 1) is still the most intense, but becomes narrower (cf. Table 2). In this diagram peaks at 33–35°are now clearly visible, probably corresponding to some diffraction planes of the cubic/tetragonal phase of zirconia and/or (2 0 0) of the monoclinic phase. The peaks of the tetragonal/cubic phases of the zirconia are no longer visible after annealing at T¼1223 K, being substituted by a clear pattern corresponding to the monoclinic phase. The diffraction peaks of this phase are also much narrower now (cf. Table 2), having a full width at half maximum (FWHM) similar to that of the peaks of the (Oþ 2)-ZrO2thin films reported in Fig. 2 (cf. Table 1). After annealing at 1323 K, a diffraction diagram similar to that depicted by the Oþ 2-ZrO2sample annealed at the same temperature was obtained. Fourier transform infrared spectroscopy (FTIR) spectroscopy can be used as a complement of XRD for a more accurate description of the structure of ZrO2thin films. Although FTIR is normally used on a ‘‘fingerprint’’ basis, a systematic analysis of IR spectra may also provide insight into the ZrO2thin film structure [25–29]. Fig. 4 shows FTIR spectra for the (Oþ 2)-ZrO2 sample as prepared and after different annealing treatments up to 1323 K. Initially, the FTIR spectra of the as prepared Oþ 2-ZrO2film is characterised by a structure where some broad bands are defined at 340, 407 and 490 cm1. These bands can be attributed to some of the 15 vibrational modes of the monoclinic phase of ZrO2[30]. By annealing at T6773 K, the bands become slightly sharper and increase in intensity. This tendency is even clearer after annealing TP1223 K, while some new bands centred at 258, 460 and 570 cm1develop in the spectrum. These bands can be also assigned to some of the 15 vibrational modes of the monoclinic phase of ZrO2[30]. The increase in intensity at TP1073K, the development of new bands and the smaller width of the observed bands above this temperature are facts which are consistent with the observed increase in the crystallinity of the monoclinic phase. This tendency can be also deduced from the decrease of Table 1 Normalised intensity, referred to most intense peak of each phase and FWHM (in brackets) of the most significant peaks appearing in XRD diagrams of Fig. 2 for (Oþ 2)-ZrO2samples after indicated annealing treatments Temperature Monoclinic Cubic/tetragonal (1 1 0) (1 1 1) (1 1 1) (1 1 1) (2 2 0) (3 1 1) 300 K 18 (0.98) 18 (0.90) 100 (0.82) 100 (0.92) – – 573 K 19 (0.85) 45 (0.83) 100 (0.82) 100 (0.88) – – 773 K 21 (0.75) 58 (0.75) 100 (0.62) 100 (0.81) – – 1073 K 14 (0.72) 58 (0.57) 100 (0.39) 100 (0.40) – – 1223 K 18 (0.32) 21 (0.32) 100 (0.26) – – – 1323 K 25 (0.25) 63 (0.29) 100 (0.23) – – – Single crystal 25 100 70 100 50 20 338 J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 the FWHM of the peaks of the XRD diagrams presented in Fig. 2 (cf. Table 1). Fig. 5 shows the FTIR spectra corresponding to a(O þ 2þArþ)-ZrO2sample subjected to different annealing treatments. The as prepared sample presents only a wide structure with a maximum at 399 cm1. After annealing at T6773 K, this structure becomes slightly sharper and two maxima can be observed at 343 and 434 cm1. To our knowledge typical frequencies of the transverse vibrational modes of the cubic or tetragonal phases of pure ZrO2are not reported in literature. Therefore, comparison is made with the vibrational frequencies of the cubic and tetragonal phases of Yttria-stabilised ZrO2[31] reported in Table 3. Assuming that the presence of yttria might produce some shifts in the position of bands, the spectra in Fig. 5 after annealing at TP573 K could be consistent with the formation of the tetragonal phase. By contrast, the broad band at 399 cm1found in the as prepared sample, could be attributed to a badly ordered cubic and/or tetragonal phase. The broad shapes of all these spectra make it difficult to discard some minority contribution of a cubic phase after annealing at 573 or 773 K. The FTIR spectra of the (Oþ 2þArþ)-ZrO2samples annealed at TP1073 K undergo drastic changes. The shape of spectra is better defined, and shows bands centred at 330, 405 and 490 cm1. The position of these new bands is similar to those that were ascribed to the monoclinic phase of zirconia observed in the (Oþ 2)- ZrO2samples. However, the superposition of the broad IR structure of the tetragonal phase cannot be discarded in these spectra. It is interesting that no clear evidence of the presence of the monoclinic phase can be deduced from the XRD pattern only of the (Oþ 2þArþ)-ZrO2sample annealed at this temperature (cf. Fig. 2). The different long range order sensitivity of XRD and IR is likely the reason of the different information provided by these two techniques and indicate the advantage of using both methods for a proper evaluation of the structure of some thin films. After annealing at T¼1323 K, the bands increase in intensity and become better defined, while some other bands at 258, 340, 407, 460, 490 and 570 cm1become also visible. At this final temperature, the spectrum is almost identical to that obtained for the (Oþ 2)- ZrO2sample annealed at the same temperature. 3.3. Positron annihilation experiments The S-parameter versus implantation energy in the as prepared samples of (Oþ 2þArþ)-ZrO2and (Oþ 2)-ZrO2series is plotted in Fig. 6(a). Both samples present almost identical variation of this Fig. 3. XRD patterns of (Oþ 2þArþ)-ZrO2thin film for the as prepared samples and after annealing treatments at indicated temperatures. The patterns of the most intense peaks of monoclinic (bottom), cubic (top, full line) and tetragonal (top, dashed line) phases of zirconia are included for comparison. J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 339 parameter, indicating that the fraction of open volume in the samples is very similar. By contrast, the variation of the W-parameter, presented in Fig. 6(b), shows some differences between the as prepared samples of (Oþ 2þArþ)-ZrO2and (Oþ 2)- ZrO2. In principle, this change can be ascribed to the positron annihilation in the vicinity of argon ions, as they are heavier than oxygen ions and, accordingly, positron annihilation with its core electrons would be enhanced. Fig. 4. FTIR spectra of the (Oþ 2)-ZrO2thin film for the as prepared sample and those obtained after annealing treatments at indicated temperatures. Fig. 5. FTIR spectra of the (ArþþOþ 2)-ZrO2thin film for the as prepared sample and those obtained after annealing treatments at indicated temperatures. Table 2 Normalised intensity, referred to most intense peak of each phase and FWHM (in brackets) of the most significant peaks appearing in XRD diagrams of Fig. 3 for (ArþþOþ 2)-ZrO2samples after indicated annealing treatments Temperature Monoclinic Cubic/tetragonal (1 1 0) (1 1 1) (1 1 1) (1 1 1) (2 2 0) (3 1 1) 300K–––––100(2.56) 573K–––––100(2.56) 773K–––––100(2.10) 1073 K –––––100(1.22) 1223 K 18 (0.29) 21 (0.32) 100 (0.26) – – – 1323 K 25 (0.23) 64 (0.29) 100 (0.23) – – – Single crystal 25 100 70 100 50 20 340 J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 Although the variation of Sand Wparameters reflects the properties of the samples, a simultaneous representation of these two parameters in S–W plot is usually preferred for a clearer description of the evolution of the thin film properties. Fig. 7 presents the S–W map for the as prepared (Oþ 2þArþ)-ZrO2and (Oþ 2)-ZrO2samples. In this kind of plots, information is obtained for the different layers of which the sample is composed. In the current study a three-layer system (surface/ zirconia layer/silicon substrate) was applied. As expected, the S–W cluster points of the surface and the silicon substrate are identical in both samples but different in those of the zirconia layer. Thus, changes in the cluster points of the zirconia layer will reflect changes in the structure of the film. The variation of the fitted S–W points corresponding to the zirconia layer for samples of (Oþ 2þArþ)- ZrO2and (Oþ 2)-ZrO2after annealing treatments at different temperatures is shown in Fig. 8. In Fig. 9 the fitted Sand Wparameters of the zirconia layer are plotted independently versus the annealing temperature. In both figures for the different sets of samples two annealing stages can be considered: annealing from room temperature to 773 K and annealing above that temperature. In the first case, the S–W characteristic points move to higher S values while the Wparameter decreases. Changes in the positron diffusion length were also observed. In fact, for the (Oþ 2)-ZrO2samples, the positron diffusion length increases from 7.5 nm in the as prepared sample, to 50 nm after annealing at 773 K. For the as prepared (Oþ 2þArþ)-ZrO2sample a shorter positron diffusion length was measured (3 nm) and it remains shorter after annealing at Table 3 Some of the IR vibrational modes for tetragonal and Yttria stabilised cubic ZrO2(taken from [30]) Phase mT(cm1) Tetragonal 467 339 Yttria-stabilised cubic 358 Fig. 6. Variation versus energy of implantation in the as prepared sample of (Oþ 2þArþ)-ZrO2and (Oþ 2)-ZrO2thin films (a) of the S-parameter (b) of the W-parameter. The lines indicate the fitted values obtained from VEPFIT. Fig. 7. S–W plot for the as prepared (Oþ 2þArþ)-ZrO2and (Oþ 2)-ZrO2as prepared, indicating the three-layer system (surface/zirconia layer/silicon substrate). The arrows indicate the positron implantation depth. J.P. Holgado et al. / Nucl. Instr. and Meth. in Phys. Res. B 194 (2002) 333–345 341