Structural changes in helium implanted Zr0.8Y0.2O1.9 single crystals characterized by atomic force microscopy and EXAFS spectroscopy

Available online at www.sciencedirect.com

NIM B Beam Interactions with Materials & Atoms

Nuclear Instruments and Methods in Physics Research B 266 (2008) 1216–1223 www.elsevier.com/locate/nimb

Structural changes in helium implanted Zr0.8Y0.2O1.9 single crystals characterized by atomic force microscopy and EXAFS spectroscopy G. Kuri a,*, D. Gavillet a, M. Do¨beli b, D. Novikov c b

a LWV, NES, Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland Ion Beam Physics, Paul Scherrer Institute and ETH Zurich, CH-8093 Zurich, Switzerland c HASYLAB at Deutsches Elektronen-Synchrotron DESY, D-22603 Hamburg, Germany

Received 19 September 2007; received in revised form 4 February 2008 Available online 21 February 2008

Abstract The present work is devoted to investigate the local atomic environment (of Zr, Y and O) as well as surface modifications associated with excess helium in the cubic phase of (1 0 0)-oriented Zr0.8Y0.2O1.9 single crystal substrates. Commercially available oxide crystals have been implanted at various fluences in the range 0.15–2.0  1016 He-atoms/cm2 using a 2.74 MeV He+ ion beam passing through a 8.0 lm Al foil. The microstructure and surface morphology of the irradiated surface are examined using atomic force microscopy (AFM). The local atomic environments of Zr, Y and O in the implanted layer are studied using synchrotron radiation and by extended X-ray absorption fine structure (EXAFS) measured at glancing angles to probe the implanted layer. From AFM studies it was observed that the surface roughness increases as fluence increases and above a critical fluence stage, small blister-like structures originating from helium bubbles are scattered on the irradiated surface. The radial distribution functions (RDFs), derived from EXAFS data at the Zr K-edge, have been found to evolve continuously as a function of ion fluence describing the atomic scale structural modifications in YSZ by helium implantation. From the pristine data, long range order (beyond the first- and second-shell) is apparent in the RDF spectrum. It shows ˚ . In the implanted specimens, all these peaks are greatly reduced in magseveral nearest neighbour peaks at about 2.1, 3.6, 4.3 and 5.4 A nitude and their average positions are changed, typical of damaged material. A simple model taking into account only the existence of lattice vacancies has been used for the interpretation of measured EXAFS spectra. Ó 2008 Elsevier B.V. All rights reserved. PACS: 61.10.Ht; 61.72.Ji; 61.80.Jh Keywords: Helium implantations; Stabilized zirconia; Point defects

1. Introduction Understanding the effect of helium in stabilized cubic zirconia is crucial for the development of materials resistant to irradiation under fusion and fission conditions. Cubic zirconia is also a potential candidate as inert host matrix fuel (IMF) for transmutation of minor actinides in the management of high-level nuclear waste as well as spe*

Corresponding author. Tel.: +41 56 310 2182; fax: +41 56 310 2203. E-mail address: [email protected] (G. Kuri).

0168-583X/$ - see front matter Ó 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.nimb.2008.02.020

cifically-devoted nuclear reactors. These applications are based on zirconia’s inherent properties in a wide range such as low thermal conductivity and excellent chemical durability, high melting point and toughness, high solubility for actinides and exceptional stability under irradiation conditions. Therefore, the behaviour of helium in zirconia has been extensively studied, both theoretically and experimentally, since many years [1]. A variety of experimental techniques such as transmission electron microscopy (TEM), Rutherford backscattering spectrometry and channeling (RBS/C), positron annihilation spectroscopy, etc. has been

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223

used for understanding the defect structures and materials modifications in helium implanted zirconia. Cationic disorder and oxygen disorder are among the main factors which affect the transport properties of helium in the implanted layer. Some work has also concentrated on the aspects of helium retention and remission [2], helium thermal desorption [3], transport mechanism and related phenomena [4]. It is now established that the degree of radiation swelling and properties of IMF strongly depends on its microstructures and testing scenario [5]. Although the basic helium migration mechanisms are well known, however, due to the complexity of the problem there are still many unknowns regarding He migration in the presence of defects. Thus, an atomic scale description of the lattice disorder in zirconia, which is local in particular and element-specific providing definitive information on the near-neighbour atomic environment, is essential. In this work, we have studied local structural changes of the atomic distributions in helium implanted Zr0.8Y0.2O1.9 single crystals using X-ray absorption fine structure (XAFS) spectroscopy technique, both extended X-ray absorption fine structure (EXAFS) and X-ray absorption near edge structure (XANES). Other experimental techniques mentioned above already employed for the description of lattice disorder in zirconia are not sensitive to any small change in the local atomic distributions. To simulate defects at various levels the oxide crystals were implanted with helium ions at various fluences. Finally, the changes of surface morphology as well as defects at the irradiated surface have been examined using atomic force microscopy (AFM). 2. Experimental and XAFS data analysis Details of the experimental arrangement for helium implantations are described in one of our recent publications [1] and only a short description is provided here. Commercially available single crystals of cubic zirconia stabilized with yttria [Zr0.8Y0.2O1.9 (1 0 0), hereafter YSZ] have been used in the present study. The size of each crystal was 10  10  0.5 mm3 and they were polished to optical finish by the manufacturer. In order to heal the intrinsic (native) point defects as well as near-surface defects induced during polishing, the as-received crystals were preannealed for 1.5 h at about 1250 °C in air and slowly cooled inside the furnace. Thereafter, separate substrates have been implanted at room temperature with a beam of 2.74 MeV He+ ions passing through a 8.0 lm Al foil, covering a fluence range from 1.5  1015 to 2.0  1016 He-atoms/cm2. The use of absorber foil provides a wide depth distribution of incorporated He due to large energy straggling of projectile ions in the Al foil. It also brings the implanted layer closer to the surface which was convenient for later analyses. After the samples preparation the surface morphology of the as-implanted and annealed specimens was observed with a commercial AFM (Nanoscope III from Digital Instruments Ltd.) operated in the tapping mode. The images were treated with a second-order flattening routine

1217

and the root-mean-square (rms) values for the surface roughness were determined. Flattening and rms determination were carried out using the manufacturer’s software. Unless otherwise stated, rms values were calculated based upon 5  5 lm2 scan images, taken at a resolution of 512  512 points. In order to determine the local atomic configuration of constituent elements in irradiated YSZ, i.e. the nearestneighbour atomic distance, the relative coordination number and similar information for some successive shells, we utilized the X-ray absorption spectroscopy (XAS) facility available at the Hamburg synchrotron radiation laboratory (HASYLAB at DESY, Hamburg), on the ROEMO-1 beam line under dedicated ring conditions (4.5 GeV, 140 mA). All measurements have been made in the reflection mode using grazing incidence geometry (of 3.0° providing an X-ray attenuation length 1.2 lm at 18 keV) to enhance the contribution of the implanted layer. The incident X-ray intensities were measured by means of a gasflow proportional counter. The specimens were held at room temperature and the fluorescence excitation spectra at the Zr K-edge were collected with an energy dispersive solid-state silicon detector. Data were also recorded from a pristine YSZ for use as structural comparators and reference spectrum. For calibrating the energy output of cooled Ge (1 1 1) double-crystal Bragg monochromator, the Kedge excitation energy of 17998 eV measured from a high purity zirconium metallic foil, was set equal to the maximum of the first derivative of the XANES data. The experimental station is currently equipped with a diffractometer which gives great freedom to the process of sample alignment, e.g. exposing the ion implanted area with X-ray beam, very precise control of the glancing angle of incidence, adjustment of the specimen height, etc. These were indeed very important in the present work for carrying out the measurements. XAFS data processing and numerical analysis were performed using the computer program IFEFFIT [6]. Considering the measurements geometry, an appropriate correction has been applied to all the measured data for the self-absorption effect [7]. The first step in the data analysis was a background subtraction. We used a linear function in the pre-edge region and a spline fit above the absorption edge. The threshold position E0 was determined from the maximum of the first derivative of an absorption curve. Thus, background removed EXAFS function v(k) was extracted and normalized by the edge jump height. v(k) was then multiplied by k3 to emphasize the higher k region. Thereafter, k3-weighted EXAFS was Fourier transformed (FT) into real space to obtain a radial distribution function (RDF) of the near-neighbours around the absorber atom. Transform termination effects were minimized by choosing an appropriate window range bounded by minima in jv(k)j and using a Hanning window function that smoothed the ends of the region to zero. Phase shift corrected real distances were determined by the Fourier filtering technique for the spectra.

1218

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223

In the next step, the extracted experimental EXAFS signals were best-fitted to the well known EXAFS equation [8] available within the IFEFFIT algorithm. The optimized structural parameters, namely: N – coordination number, R – interatomic distance, r2 – DebyeWaller factor, DE0 (a correction to E0), were derived from the fits, taking into account both the many-electron effects and the vibrations of neighbouring atoms around their equilibrium positions. Quality of the fits were judged by the EXAFS R-factor as defined in the IFEFFIT code, the acceptable values are in the order of a few percent. Theoretical phase shift and backscattering amplitude functions for quantitative EXAFS fitting were generated using the modern relativistic code FEFF-8.2 [9]. This theoretical approach describes the photoelectron final states by an ab-initio curved-wave multiple-scattering calculation using an energy-dependent exchange-correlation for self-energy within a muffin-tin potential. In our analysis, final-state potentials at Zr Kedge were calculated using atomic clusters derived from the known atomic coordinates of cubic phase YSZ structures as referred in [10,11]. 3. Results and discussion 3.1. Surface characterization by AFM In Fig. 1 we have shown the AFM surface images of the as-received (AR) YSZ crystal, the pristine YSZ (obtained by thermal annealing of the AR-one) and the YSZ samples irradiated to different fluences in the range 0.15–4.5  1015 He-atoms/cm2. Each scan shown in the figure represents a 0.5 lm  0.5 lm lateral area. From Fig. 1(a) and (b), it is clear that the surface of two substrates are distinctly different from each other. The surface of the AR specimen shows a topographical variation which contains an intimate mixture of some linear as well as hillock structures with a distribution of sizes. We believe that these atomic scale irregularities are generated in connection with the crystal habit and during the mechanical polishing process commonly used by the crystal manufacturers to obtain a flat YSZ (1 0 0) surface. In agreement with the literature data [12], the measured rms roughness on this surface is about 0.35 nm. However, the subsequent annealing at 1250 °C diminished most of these structures. The corrugation changes significantly with respect to the AR substrate and leads to dramatic smoothing of the AR surface. The surface roughness is much better after annealing and has a rms value of around 0.17 nm from our own measurement. In this case, the YSZ surface also reorganizes to a periodical terraced and stepped structure. The terraces are atomically flat. The origin of these surface features has been described in one of our recent publications [13] and also by other researchers [14]. After irradiation the pristine YSZ shows a noticeable change in the surface steps as depicted in Fig. 1(c)–(e). Each surface segment suffers multiple ion impact and the whole surface is damaged. The remarkable fact about the

Fig. 1. Surface morphologies of the YSZ substrate surfaces, (a) asreceived; (b) pristine; (c)–(e) after helium ion irradiation with different ion fluences of 1.5, 3.0 and 4.5  1015 He/cm2, respectively. The area coverage is 0.5 lm  0.5 lm for all the images and the height scales are 7, 3, 3, 3 and 5 nm for (a)–(e), respectively.

AFM images of the irradiated samples is the disappearance of step patterns on the surface. The line profiles obtained from AFM (not shown here) show the presence of nanoscale hillock-like defects, each created by the impact of helium ions. The average rms surface roughness measured on these samples are 0.30, 0.39 and 0.47 nm for the helium fluences 1.5  1015, 3.0  1015 and 4.5  1015 atoms/cm2, respectively. The magnitudes of this roughening are small, but measurable with the use of surface topographic probes such as AFM. It is evident that the rms roughness increases gradually with ion fluence. The difference in height as well as the distribution of the protrusions seems to be a major factor in describing surface roughness. The sets of micrographs presented in Fig. 1(c–e) exhibit a gradual accumulation of disorder with fluence and resemble the case of a low damage level [15], where the individual damage cascade do not overlap. Consequently, phenomena such as local amorphization and/or formation of any overlapping defects are not expected, although these processes definitely play an

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223

3.2. Damage analysis by XAS The XAS results are shown in Fig. 2. The upper panel in the figure reports representative Zr K-edge XAS spectrum (Fig. 2(a)) obtained from the reference (pristine) YSZ sample. The edge position, taken as equal to the maximum value of the differential coefficients of the spectrum near the edge, reveals an E0 value of 18001 eV which is 3 eV higher than that of the Zr-metal foil. This implies chemically induced shift in excitation energy when going from metallic Zr to Zr4+-oxidation state in YSZ. The lower panel (Fig. 2(b)) plots Fourier transformed EXAFS spectra for the three He implanted samples (presented in Fig. 1) as well as that of a pristine YSZ. For a quantitative analysis of these data, the fitting procedure was adopted using the standard methods as summarized in the experimental section. Table 1 contains the analyzed numerical results for all these specimens. An example of the resulting RDFs

a Pristine YSZ

1.00

Experimental Simulated

8

0.75 FT Magnitude

Normalized Intensity (a.u.)

1.25

0.50

0.25

6

4

2

0 1

2

3

0.00

4

5

6

7

8

18.7

18.8

Radial Distance (Å) 17.9

18.0

18.1

18.2

18.3

18.4

18.5

18.6

Energy (KeV)

b

10 Zr-Zr Zr-O

8

FT Magnitude (a.u.)

important role in high fluence implantations (P1016 atoms/cm2 [1,15]). Therefore, the damaged YSZ surface can be characterized by mainly point defects and/ or tiny isolated defect clusters [16]. The mechanism involves the formation of these defects by helium irradiation, which eventually result in preferential desorption (or diffusion) of atoms from step edges induced by relatively high electronic stopping power during the irradiation process. The certain functional dependence of the hillock characteristics on energy deposition seems to be very important, because significant changes of the radii and the heights with the electronic energy loss have been detected on other targets, like MgO, Al2O3, MgAl2O3, etc. [17], the materials with similar collisional kinematics as YSZ. Importantly, the radiation induced matrix ionization can be assumed to play an eminent role. An assumption that the hillock structures result from the equilibrium population of created point and multiple defects at the surface, can be made on the basis of transport of ions in matter (TRIM) calculations [18]. In view of the present experimental condition for 2.74 MeV He+ ions after passing through 8.0 lm Al foil, the implants reach the YSZ surface with energies in the range 250–450 keV due to energy straggling. The projected range of He ions is about 1.1 lm. These data give average electronic and nuclear stopping powers of 0.5 and 0.001 keV/nm, respectively, at the YSZ surface. Additionally, the critical dose for amorphisation by elastic collisions of YSZ with 250–450 keV helium ions can be roughly estimated to be about 4  1016 ions cm2 [19], which is much higher than the fluence we have used. This suggests that the morphological protrusions at the implanted YSZ surface in the present case do not result from He ion induced amorphisation of the crystalline surface structure. The expansion of the irradiated volume (i.e. protrusion) due to inclusion of helium should be also ruled out, because the maximum height of the surface protrusions ( a few nm) is far smaller than the buried depth of implanted helium atoms.

1219

Pristine 1.5 × 1015 cm-2 3.0 × 1015 cm-2

6

4.5 × 1015 cm-2

4

2

0 1

2

3

4

5

6

7

8

Radial distance (Å)

Fig. 2. (a) Normalized and background removed Zr K-edge absorption spectrum of pristine YSZ specimen in the as-prepared state and before helium implantation. In the inset, magnitude of the k3-weighted Fourier transformed EXAFS signal (dot line) and the best-fitted curve (solid line) are shown. (b) Comparison of the Fourier transformed Zr K-edge EXAFS for pristine YSZ and the three He implanted samples with different fluences as reported in Fig. 1.

for the pristine sample is shown in Fig. 2(a) as an inset. The spectra drawn with points and solid curves are the observed and corresponding best-fitted one, respectively. The EXAFS measurement from pristine YSZ clearly reveals an excellent crystallinity around Zr atoms. The FT data presented in Fig. 2(b) illustrates Fourier features which are characteristics of the fluorite-sites environments. The RDF shows that there are three strong amplitude peaks between 0.15 and 0.60 nm whose distance R is around 0.21, 0.36 and 0.54 nm, respectively. A small structure peaking at about 0.43 nm can be also seen in the spectrum. The first strong peak at about 0.21 nm stems from the first-neighbour (oxygen atoms) of the YSZ lattice. The magnitude intensity of the second peak near 0.36 nm is sensitive to the coordination number as well as geometry

1220

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223

Table 1 Zr EXAFS fit parameters Specimen identification

SC

CN (±10.0%)

RD (nm) (±0.002)

˚ 2) (± 15%) r2 (A

Quality factor (R)

Pristine YSZ

Zr–O Zr–Zr Zr–O Zr–Zr Zr–O Zr–Zr Zr–O Zr–Zr

7.1 11.9 4.8 11.3 4.4 10.6 4.2 10.8

0.215 0.358 0.213 0.364 0.212 0.362 0.210 0.366

0.025 0.021 0.039 0.028 0.031 0.027 0.032 0.035

0.019

1.5  1015 cm2 3.0  1015 cm2 4.5  1015 cm2

0.028 0.037 0.026

SC: scattering configuration; CN: coordination number; RD: radial distance.

of second neighbours (zirconium). The peaks around 0.43 and 0.54 nm contain contributions from long range order (beyond the first- and second-shell). The starting models employed to analyze these data were those already developed and verified in the extremely thorough EXAFS studies presented in some past reports [10,11]. In our case, this was necessary to obtain the structural parameters of Zr in pristine YSZ and for using them in analyzing the data from helium implanted specimens. The refined values obtained from the curve fit (Table 1) give realistic DW factors, correct order for the bonds, an appropriate 7-fold and 12-fold coordination for the first- and second-neighbour shells, respectively. However, dealing with implanted samples, the RDFs exhibit remarkable differences in many respects. For all the spectra shown in Fig. 2(b) quantifiable differences in bond lengths, coordination numbers and radial distribution functions can be distinguished. They clearly display a progressive intensity reduction and asymmetric broadening of the Zr–O peak, which reflects the loss of crystallinity in the implanted lattice. Due to the loss of long range order, only the nearest neighbours peaks (originating from first and second-shell) are observed for the 4.5  1015 He-atoms/cm2 sample. The coordination number for the first nearest neighbour of Zr is decreased greatly on implantation, that is from 7.1 to 4.2 in this sample. Another remarkable point in the table is that the Zr–O bond results shorter and that the distance to the second nearest neighbour to Zr is increased from 0.358 to 0.366 nm. However, the changes in the Zr–Zr second-shell coordination number do not occur to a large extent. At this point let us mention that Zr and Y in YSZ are equivalent backscatterers and indistinguishable from the XAS point of view. Therefore, in the present analysis, the contribution from Zr–Zr correlation in the FT peak implies the result of Zr–Zr(Y) atoms-shell in practice. Considering the information derived from EXAFS analyses several conclusions can be drawn from Table 1. For all fluences, the shift of the Zr–O peak implies the existence of lattice distortion and subsequent relaxation due to the rearrangement of point defects. The progressive decrease of the first-neighbour peak amplitude is indicative of an increase in the number of displaced oxygen atoms in terms of point defects created at the YSZ lattice. To relax the anisotropy of the lattice distortion short

range ordering of oxygen ion vacancies can occur. These would in turn result in the decrease in the interatomic distance from a Zr ion to the first nearest neighbour, or the decrease in volume for the surroundings of Zr, which contributes to the relaxation of the lattice distortion derived from the local strain field. As discussed before following a TRIM calculation, the amount of energy deposited in collision cascades under the implantation conditions used here, is rather low, about 0.5 keV/nm on the average over Zr and O atoms. Most of the energy is deposited in the form of electronic losses inducing ionization or deposited as phonons. Therefore, it is reasonable to consider that the lattice distortion in the first neighbour shell is caused by simple point defects associated with the oxygen sublattice. During implantation the creation of these primary defects is extremely high in the region close to the peak of the implanted profile. This means that the primary defects have a much higher probability of reacting with themselves, which in the case of an interstitial and a vacancy reaction simply results in annihilation but in the case of vacancies can result in the production of the divacancy or larger vacancy agglomerates. Helium irradiation also produces both vacancies and interstitial pairs around Zr ions in YSZ and disturbs the cation network. An apparent second-shell expansion of the Zr–Zr pair seen for the implanted samples may be attributed to the close interaction of the distortion in the direction of the first nearest oxygen neighbours, or [1 1 1], with that of the second nearest Zr neighbours along [1 1 0]; the resultant bond distribution is then weighted in favor of the longer distances. This is consistent with an expected larger value of r2 observed for this shell and there is also TEM evidence showing close correlation between distortions in the first and the second nearest neighbour atoms in YSZ subjected to aging at high temperatures [20]. Another feature of the results in the Table is that the Zr–Zr coordination in the implanted samples decreases to 11% (the maximum value), while the corresponding change in Zr–O shell is 40%. It is plausible that the efficiency of the annihilation and mobility of point defects depend on the sample crystallographic orientation. Previous work on orientation dependence of radiation damage in ion implanted YSZ single crystals has shown that [1 1 0] is the most damage resistant orientation relative to

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223

a He-implanted YSZ

Normalized Intensity (a.u.)

1.25

1.00

0.75

0.50

0.25

0.00 17.8

17.9 18.0

18.1

18.2 18.3

18.4 18.5

18.6

18.7 18.8

Energy (KeV)

b

6.0

4.5

FT Magnitude (a.u.)

both [1 1 1] and [1 0 0], in which the dynamic annealing rate is maximal [21]. In relation of the above discussion on the structural evolution of YSZ material, one additional fact needs to be mentioned. The presence of both He atoms and accumulated vacancies lead to the formation of a free volume within the host lattice in the surface region of the implanted samples, hence to the generation of a high compressive stress. Since the crystal structure cannot account continuously increasing stresses, at a given irradiation fluence, a structural phase transformation may occur to relax the stress. It was previously observed by Hasegawa et al. [22] after N2 ion implantation in zirconia. In YSZ subjected to 120 keV helium irradiations at high fluences (P1016 cm2), a structural change of cubic-to-rhombohedral phase transformation and the associated rhombohedral deformation of the cubic cell along the (1 1 1) directions were also measured [23]. Therefore, we also have carried out X-ray diffraction (XRD) in Bragg geometry on our samples (data in Fig. 2) with a D500 SIEMENS diffractometer using Cu Ka1 radiation and in the angular range from 25° to 80°. The characteristic diffraction peaks of YSZ (0 0 2) and YSZ (0 0 4) have been used to monitor the diffraction patterns, which are the only dominant peaks for the YSZ (1 0 0) single crystals and in the angular range measured. Our experimental results show no change in either the integral intensity of the XRD signal and/or any apparent angular-shift of the peak position due to He+ implantation. This indicates no residual stress in YSZ (1 0 0) within the probing depth of implanted layer. In our opinion, the total helium fluence in the range 1.5–4.5  1015 atoms/cm2 used for the investigation is indeed low and the bulk part of the crystal contributing dominantly, to observe any change in XRD pattern under Bragg condition. In contrast, EXAFS being an atom specific local structure probe, yields information including average interatomic distances and the number and chemical identities of Zr neighbours in these samples. In a forth-coming paper we will discuss the application of grazing incidence XRD as well as EXAFS methods to study the behaviour of helium in a set of polycrystalline YSZ specimens subjected to helium irradiations at higher fluences.

1221

3.0

1.5

0.0 0

1

2

3

4

5

6

7

8

Radial distance (Å)

Fig. 3. (a) The EXAFS data at the Zr K-edge after background subtraction, from a 2.0  1016 He-atom/cm2 YSZ sample. The inset shows top-view AFM picture from a selected area (1 lm  1 lm) of the surface of this specimen, with a height scale of 7 nm. (b) Corresponding radial distribution function (Phase shift corrected) around Zr atoms as Fourier transform of the k-spectrum. See the text for details.

3.3. Blisters on the YSZ surface In the following we present the XAS results obtained for a typical high fluence (2.0  1016 He/cm2) YSZ specimen. The Zr K-edge EXAFS spectrum and the experimental RDF spectrum after irradiation are illustrated in Fig. 3. The AFM surface morphology of the sample is shown in Fig. 3(a) as an inset. The surface of this specimen showed that blister-like defects had been generated during irradiation, apart from many spiky hillocks observed on the surface. The blisters have almost round shapes and a size distribution with a typical diameter in the range of 30– 80 nm. The heights varies between 2 and 10 nm. The rms roughness, which includes the contributions from these blisters, is estimated as 1.12 nm. To connect the blisters

appearance with dense implantation induced damage in the bulk, we also need to assume the creation of small helium bubbles as a common initial structure from which the blisters observed at a higher helium fluence have developed. Although the blisters size is not comparable with the helium ion range (1.1 lm), some bubbles or their agglomerates can be close to the sample surface causing its swelling and appear as blisters on the surface. From the qualitative comparison of the RDF spectrum (Fig. 3(b)) of this specimen relative to that in the pristine state (Fig. 2(b)), we find (i) a distinct splitting of sevenfold coordinated Zr–O first-shell neighbour peak; (ii) a strong reduction of the second-shell Zr–Zr contribution to the

1222

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223

FT data; (iii) a noticeable change in the relative intensity of Zr–O and Zr–Zr amplitude peaks where the cation–cation second-shell peak have a lower height. A further feature is that the evolution toward a medium/long range disordered structure at distances beyond the zirconium second neighbours; note in fact the bump-like structure (centered 0.5 nm) in the apparent overlap region of third, fourth and fifth shells. Since the EXAFS amplitude for a given shell is proportional to both N and exp(2r2k2), damping due to disorder results from the disorder-induced increase of the DW factor and the decrease of coordination numbers. Although the origin of the bimodal distribution feature observed in the Zr–O first coordinational shell is not very clear, it seems to reflect large local environmental distortion of F-type centers and V-type centers and the presence of multiple color centers, associated with oxygen vacancies with trapped electrons nearest to zirconium cations and hole trapped oxygen ions adjacent to Zr [24,25]. Taking into account an attractive interaction of helium atoms and vacancies, the helium inclusions should provoke the segregation of helium rich vacancy complexes in the form of HemVn clusters [1]. Therefore, the oxygen ions neighbouring the vacancy would be expected to show some deviation from the perfect lattice positions due to disturbance of the charge balance by displacement of the interlayer atoms and/or relaxation in response to the electrostatic and elastic strains introduced by the anion vacancies and their complexes. A detail quantitative refinement of the EXAFS spectrum in this kind of sample, where the YSZ lattice contains both point and extended type of defects including gas bubbles, is not a simple task. It requires to develop a structural model with known atomic coordinates of YSZ following the evolution of disorders produced during irradiation in the presence of He, the initial stages of nucleation of He-vacancy clusters, pre-cursors of bubbles and voids and experimental identification of these defects by TEM work in detail. A systematic study to these points is underway. 4. Conclusions Yttria-stabilized zirconia (YSZ) single crystals were implanted with helium ions to the fluences from 1.5  1015 to 2.0  1016 atoms/cm2 and the effects of irradiation were studied using AFM and XAS measurements. AFM gave detailed information on the protrusions of the ion beam treated areas on the nanometer scale. For samples where the helium content was low (1.5–4.5  1015 Heatoms/cm2), a progressive increase in surface roughness with increasing ion fluence has been observed. The results are attributed to the involvement of point defects and their equilibrium population at the YSZ surface. XAS measurements revealed the structural changes like local (dis)order, interatomic distances and number and type of atoms around Zr in the implanted layer. From EXAFS analyses, vacancy-like defects are clearly seen in the decrease of the number of nearest neighbours for both

first (Zr–O) and second (Zr–Zr) coordination shells in the YSZ lattice when the corresponding values are compared with that for the pristine material. After irradiation, a reverse shift of the Zr–O and Zr–Zr bond-length is also observed, i.e. the Zr–O bond becomes shorter and the average distance to the second nearest neighbour to Zr is increased. These results can be related to the relaxation of lattice distortions resulting from short range ordering of oxygen ion vacancies due to the rearrangement of point defects. The increase of the helium fluence to a level of 2.0  1016 cm2 leads to formation of blisters on the YSZ surface driven by the growth of tiny helium bubbles in the implanted layer. Using EXAFS, structural disorder was readily apparent through a strong reduction of both Zr–O and Zr–Zr amplitude peaks. In addition to that, regions of first-shell oxygen neighbours around Zr atoms show highly defective atomic configurations, which cause a bimodal bond length distribution behaviour in the first coordination shell. These interesting results warrant further experimental work as well as theoretical model calculations for the lattice damage to clarify their possible influence on technological applications of helium implantation into YSZ. References [1] See for example: G. Kuri, M. Do¨beli, D. Gavillet, Nucl. Instr. and Meth. Phys. Res. B 245 (2006) 445, and references therein. [2] P.M.G. Damen, A. van Veen, F. Labohm, H. Schut, M.A. van Huis, Nucl. Instr. and Meth. Phys. Res. B 319 (2003) 65. [3] P.M.G. Damen, H. Matzke, C. Ronchi, J.P. Hiernaut, T. Wiss, R. Fromknecht, A. van Veen, F. Labohm, Nucl. Instr. and Meth. Phys. Res. B 191 (2002) 571. [4] P. Trocellier, D. Gosset, D. Simeone, J.M. Costantini, X. Deschanels, D. Roudil, Y. Serruys, R. Grynszpan, S. Saude, M. Beauvy, Nucl. Instr. and Meth. Phys. Res. B 210 (2003) 507. [5] T.A.G. Wiss, P.M.G. Damen, J.P. Hiernaut, C. Ronchi, J. Nucl. Mater. 334 (2004) 47. [6] . [7] C.H. Booth, F. Bridges, Phys. Scripta T115 (2005) 202. [8] P.A. Lee, P.H. Citrin, P. Eisenberger, B.M. Kincaid, Rev. Mod. Phys. 53 (1981) 769. [9] J.J. Rehr, J.M. d. Leon, S.I. Zabinsky, R.C. Albers, J. Am. Chem. Soc. 113 (1991) 5135, we have used the version 8.20. [10] N. Ishizawa, Y. Matsushima, M. Hayashi, M. Ueki, Acta Cryst. B 55 (1999) 726. [11] G.E. Rush, A.V. Chadwick, I. Kosacki, H.U. Anderson, J. Phys. Chem. B 104 (2000) 9597, and references therein. [12] T. Thome, L.P. Van, J. Cousty, J. Euro. Ceramic Soc. 24 (2004) 841. [13] G. Kuri, M. Gupta, R. Schelldorfer, D. gavillet, Appl. Surf. Sci. 253 (2006) 1071. [14] R.G. Green, L. Barre, J.B. Giorgi, Surf. Sci. 601 (2007) 792. [15] L. Thome, J. Fradin, J. Jagielski, A. Gentils, S.E. Enescu, F. Garrido, Euro. Phys. J. Appl. Phys. 24 (2003) 37, and references therein. [16] K. Yasuda, M. Nastasi, K.E. Sickafus, C.J. Maggiore, N. Yu, Nucl. Instr. and Meth. Phys. Res. B 136 (1998) 499. [17] V.A. Skuratov, A.E. Efimov, K. Havancsak, Nucl. Instr. and Meth. Phys. Res. B 250 (2006) 245, and references therein. [18] J.P. Biersack, L.G. Haggmark, Nucl. Instr. and Meth. 174 (1980) 257, we have used the version SRIM-2006. [19] S. Furukawa, Jpn. J. Appl. Phys. 11 (1972) 102. [20] J. Kondoh, S. Kikuchi, Y. Tomii, Y. Ito, Physica B 262 (1999) 177.

G. Kuri et al. / Nucl. Instr. and Meth. in Phys. Res. B 266 (2008) 1216–1223 [21] I.O. Usov, P.N. Arendt, J.R. Groves, L. Stan, R.F. DePaula, Nucl. Instr. and Meth. Phys. Res. B 240 (2005) 661. [22] H. Hasegawa, T. Hioki, O. Kamigaito, J. Mater. Sci. Lett. 4 (1985) 1092. [23] G. Sattonnay, L. Thome, J. Nucl. Mater. 348 (2006) 223.

1223

[24] V.M. Orera, R.I. Merino, Y. Chen, R. Cases, P.J. Alonso, Phys. Rev. B 42 (1990) 9782. [25] J.M. Costantini, F. Beuneu, D. Gourier, C. Trautmann, G. Calas, M. Toulemonde, J. Phys.: Condens. Matter 16 (2004) 3957.