Annealing behaviour of ion tracks in olivine, apatite and britholite

Nuclear Instruments and Methods in Physics Research B 326 (2014) 126–130

Contents lists available at ScienceDirect

Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb

Annealing behaviour of ion tracks in olivine, apatite and britholite B. Afra a, M. Lang b, T. Bierschenk a, M.D. Rodriguez a, W.J. Weber c,d, C. Trautmann e,f, R.C. Ewing b, N. Kirby g, P. Kluth a,⇑ a

Department of Electronic Materials Engineering, Research School of Physics and Engineering, The Australian National University, Canberra, ACT 0200, Australia Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109-1005, USA c Department of Materials Science and Engineering, University of Tennessee, Knoxville, TN, USA d Materials Science and Technology Division, Oak Ridge National Laboratory, TN, USA e GSI Helmholtzzentrum für Schwerionenforschung, 64291 Darmstadt, Germany f Technische Universität Darmstadt, 64289 Darmstadt, Germany g Australian Synchrotron, 800 Blackburn Road, Clayton, VIC 3168, Australia b

a r t i c l e

i n f o

Article history: Received 1 July 2013 Received in revised form 31 October 2013 Accepted 31 October 2013 Available online 6 February 2014 Keywords: SAXS Ion track Apatite Olivine Britholite Activation energy

a b s t r a c t Ion tracks were created in olivine from San Carlos, Arizona (95% Mg2SiO4), apatite (Ca5(PO4)3(F,Cl,O)) from Durango, Mexico, and synthetic silicates with the apatite structure: Nd8Sr2(SiO4)6O2 and Nd8Ca2(SiO4)6O2 using 1.6 and 2.2 GeV Au ions. The morphology and annealing behaviour of the tracks were investigated by means of synchrotron based small angle X-ray scattering in combination with ex situ annealing. Tracks in olivine annealed above 400 °C undergo a significant change in track radius due to recrystallisation of the damage tracks. At temperatures higher than 620 °C, the scattering images indicate fragmentation of the track cylinders into smaller subsections. Ion tracks were annealed at elevated temperatures up to 400 °C in the Durango and Ca-britholite, and up to 560 °C in Sr-britholite. While there was a significant change in the track radii in the Durango apatite, tracks in the two synthetic samples remained almost unchanged. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction When heavy ions with energies in the range of tens of MeV to GeV penetrate a solid, they lose their energy predominantly through inelastic interactions with the electrons of the target material and leave a narrow, cylindrical trail of damage called an ion track [1,2]. Ion tracks are typically a few nanometers in diameter and can be up to tens of micrometers long. When these tracks are annealed, they shrink in size until eventually the damaged material recovers. Ion tracks in olivine provide a record of cosmic rays that have passed through the minerals in meteorites [3]. Chondritic meteorites typically contain olivine ((Fe,Mg)SiO4), which is susceptible to ion track formation from high energetic cosmic particles. Information on the annealing behaviour of ion tracks in olivine can be used to estimate the temperature before and after track formation. Previous investigations of the annealing behaviour of irradiation damage in olivine have been mainly conducted to investigate the charge spectrum of ancient cosmic rays [4,5]. Discrepancies in results on the variation of the track etch velocity with the primary ⇑ Corresponding author. Tel.: +61 2 6125 0358. E-mail address: [email protected] (P. Kluth). http://dx.doi.org/10.1016/j.nimb.2013.10.072 0168-583X/Ó 2014 Elsevier B.V. All rights reserved.

ionization rate or the variation of etchable track length with the atomic number of the incident ion has been attributed to partial annealing of the cosmic ray tracks over time-scales of 107 years [6–9]. In apatite, track formation results from the spontaneous fission of naturally incorporated traces of uranium. Etched tracks of fission fragments are used for dating and constraining the thermal history of geological samples [10–13]. Likewise, fission tracks have been observed from spontaneous fission of 244Cm in Nd8Ca2(SiO4)6O2, a potential host phase for immobilization of minor actinides [14]. Since the discovery of fission tracks, considerable research effort has focused on investigating the annealing behaviour of natural and induced tracks in minerals with a wide range of beam energies [11,12,15,16]. Studies of etched tracks have shown that the annealing rates depend on the chemical composition of the apatite [17–20]. Recently, we have shown that small angle X-ray scattering (SAXS) is capable of resolving radii of un-etched (latent) tracks with high precision and is well suited to study their recovery kinetics using in situ and ex situ annealing experiments [21,22]. Preliminary SAXS measurements of ion tracks in the San Carlos olivine combined with in situ annealing revealed that their track morphology is very similar to that of natural apatite with

127

B. Afra et al. / Nuclear Instruments and Methods in Physics Research B 326 (2014) 126–130

approximately the same density change of 1% within the tracks compared to the matrix [21,22]. However, unlike in apatite, tracks in olivine do not show any substantial change after annealing at 350 °C for approximately 6 h [21]. In this work, we combined synchrotron based SAXS with ex situ annealing in order to study the annealing behaviour of unetched swift heavy ion tracks in two minerals: San Carlos olivine and Durango apatite, as well as two synthetic silicates with apatite structure, Nd8Sr2(SiO4)6O2 and Nd8Ca2(SiO4)6O2. The synthetic silicates (britholite) allow for the systematic study of the effect of composition on track annealing behaviour as well as their potential use for nuclear waste immobilization [23–25]. Tracks are generated with ion energies in the range of GeV, as the focus of this report is to elucidate differences in annealing behaviour and compare the effects of different composition and material structure. While these energies may be different to those of fission tracks and cosmic rays, the energy loss mechanisms (electronic) are the same and thus the damage structure can be expected to be similar.

tion, the temperature of the furnace only dropped by a maximum of 50 °C and equilibrated in less than a minute. After each 30 min annealing step, SAXS measurements were performed. Each given sample was sequentially annealed from room temperature up to a maximum of 950 °C for the San Carlos olivine, 560 °C for Sr-britholite, and 400 °C for Ca-britholite and Durango apatite. A second Durango apatite was annealed only at 380 and 400 °C in order to investigate the effect of thermal history on the annealing behaviour of tracks. Additionally an unirradiated San Carlos olivine was annealed at 1050 °C and measured for comparison. 3. Results and discussion 3.1. Tracks in olivine SAXS images from ion tracks in olivine at room temperature and after annealing are shown in Fig. 1(1–6). The tracks are tilted by 10° with respect to the X-ray beam. The anisotropy in the scattering signal is caused by the high aspect ratio of the tracks that are only a few nanometers in diameter and up to tens of micrometers in length. The presence of strong oscillations in the un-annealed sample is consistent with monodisperse track radii and a sharp density change between the track and the matrix material. With increasing annealing temperature, the spacing of the oscillations gets larger as compared with the room temperature image; thus, the track is expected to be reduced in radius [22]. The image at 650 °C reveals the appearance of a new feature in the scattering signal in the form of a ‘‘bulging’’ of the scattering signal. This effect starts at 620 °C, increases in size with increasing annealing temperature, and is most likely a consequence of fragmentation of the tracks. Fragmentation into smaller subsections of various lengths can occur due to Rayleigh instabilities of the narrow, initially cylindrical tracks [29,30]. These fragments naturally reduce the aspect ratios of the now separated track features, washing out the oscillations and resulting in scattering contributions normal to the originally confined streaks. Additionally, at such high temperatures oxidation of the olivine that occurs in air above 600 °C [31] may also affect the track structure and the observed scattering signal. While we cannot quantify the contribution of oxidation or separate this effect from the scattering signals of the tracks, we have annealed an unirradiated olivine sample at 1050 °C for 30 min and the SAXS image from this sample also shows an anisotropic scattering signal as shown in Fig. 1(7), which further complicates the SAXS analysis at these temperatures and needs to be investigated in the future. For SAXS data analysis below 530 °C, the scattering intensities are extracted by masking and radially integrating the intensity from the streaks of the anisotropic images. The X-ray scattering intensities of radial sectors perpendicular to the streaks (e.g. small arc sectors in the area between the two streaks that do not overlap with either of the streaks) are used as the background intensity as

2. Experiment Olivine (95% Mg2SiO4) from San Carlos, Arizona, apatite (Ca5(PO4)3(F,Cl,O)) from Durango, Mexico and synthetic apatites: Nd8Sr2(SiO4)6O2 and Nd8Ca2(SiO4)6O2 were polished to thicknesses of 40–60 lm and irradiated with 197Au ions to a fluence of 5  1010 ions/cm2 at the UNILAC accelerator at the GSI Helmholtz Centre in Germany. The low irradiation fluence yields well-separated tracks. The olivine was irradiated with a beam energy of 2.2 GeV while the apatite samples were irradiated with 1.6 GeV ions. In both minerals the projected ion range exceeds the sample thickness as estimated by the SRIM-2008 [26]. Irradiations were performed at room temperature under normal incidence. The olivine and the natural apatite were annealed at 450 °C for 24 h before irradiation to remove any natural tracks pre-existing in the minerals. Sample specifications and irradiation parameters are listed in Table 1. Transmission SAXS measurements were performed at the SAXS/ WAXS beam line of the Australian Synchrotron with X-ray energies of 12 and 20 keV and camera lengths of approximately 1600 and 2000 mm for measuring ion tracks in olivine and the apatites, respectively. Samples were mounted on a three-axis goniometer and tilted such that the axis of the tracks had an angle between 5° and 10° to the incoming X-ray beam. The spectra were collected with a Pilatus 1 M detector with exposure times of 5 and 10 s for the X-ray energy of 12 keV and 20 and 30 s for 20 keV, the latter in order to compensate the lower flux delivered at higher energies. Scattering from un-irradiated samples was measured as a reference for background removal. In order to study the annealing behaviour of the ion tracks, the irradiated samples were annealed ex situ in a conventional furnace under ambient atmosphere. The furnace was heated to the desired temperature and the sample was then inserted. During the inser-

Table 1 Sample specifications and irradiation parameters including the surface and average electronic energy loss calculated using SRIM-2008. The average energy loss gives the mean energy loss over the length of the sample or the projected range of the ion, whichever is smaller. We note that SRIM calculations include at least 10% uncertainty. Track radius is from SAXS measurements. Material

Density (g/cm3)

Sample thickness (lm)

Au ions energy (GeV)

Fluence (ions/cm2)

Ion beam projected range (lm)

Surface energy loss (keV/nm)

Average energy loss (keV/nm)

Track radius (nm) at room temperature

San Carlos olivine (95% Mg2SiO4) Durango apatite (Ca5(PO4)3(F,Cl,O)) Synthetic britholite Nd8Sr2(SiO4)6O2 Synthetic britholite Nd8Ca2(SiO4)6O2

3.32

40

2.2

5  1010

90

26.2

30.3

4.6 ± 0.1

1.6

5  10

10

63

28.8

29

5.1 ± 0.1

10

53

36.8

29

5.5 ± 0.2

54

36.5

29

5.1 ± 0.1

3.19

50

5.62 [27]

60

1.6

5  10

5.47 [28]

60

1.6

5  1010

128

B. Afra et al. / Nuclear Instruments and Methods in Physics Research B 326 (2014) 126–130

Fig. 1. (1–6) SAXS images of ion tracks in olivine (1) at room temperature and (2–6) after ex situ annealing (30 min) at selected temperatures. Ion tracks are tilted by 10° with respect to the incoming X-ray beam. At higher temperatures (above 620 °C), the bulging of the scattering signals indicates track fragmentation. (7) SAXS image from an un-irradiated San Carlos olivine sample annealed at 1050 °C for 30 min.

Fig. 2. Scattering intensities as a function of the scattering vector q for ion tracks in the olivine sample irradiated with 2.2 GeV Au ions (5  1010 cm2) before and after ex situ annealing (30 min) at different temperatures up to 530 °C. The solid lines correspond to the fits using the hard cylinder model. The spectra are offset for clarity.

it resembles the scattering signal of an un-irradiated sample. The scattering intensities from the ion tracks up to the annealing temperature where track fragmentation begins are shown in Fig. 2 after background subtraction. More detailed information on SAXS data analysis and modelling can be found in Refs. [22,32]. In order to extract quantitative results from the annealing data at higher temperatures, more sophisticated models need to be developed to account for the observed track fragmentation. The spectra were fitted using appropriate theoretical models by nonlinear least squares fitting. The ion tracks are modelled as parallel, almost identical cylindrical scattering objects [22,33]. The solid lines in Fig. 2 show the fits using the hard cylinder model which assumes a constant electron density of the track and a sharp boundary between the electron density of the track and the matrix material. The form factor for this model is given by f(q) = (2pLRq0/q)J1(R  q), where L is the track length (equivalent to the sample thickness), R the track radius, q0 the density differ-

Fig. 3. Ion track radii in olivine as a function of the annealing temperature. The inset shows an Arrhenius plot of ion track radii for extraction of an activation energy. The solid line is the linear fit to the data points from annealing at temperatures P 360 °C which leads to an activation energy of 0.41 ± 0.05 eV without considering the dependency on the annealing time. The dashed line indicates that the recovery process may have an initial stage, however, there are insufficient data points to calculate an activation energy.

ence between track and matrix material, and J1 the Bessel function of first order. The model also accounts for a variation of the track radius over the depth of the tracks due to changes in the stopping power, and deviations of perfectly sharp boundaries between track and matrix by assuming a narrow Gaussian distribution of the radius and a narrow angular distribution, respectively. For all samples the angular spread was small and the width of fitted angular distribution is below 0.3°. The average width of the fitted radius distribution is 0.6 nm. This corresponds to the fact that the energy loss of the projectile ion only changes by 5% along the length of the ion track (in 40 lm sample). Fitting of both the distribution in radii and the angular spread had no influence on the fitted track radii. This model has been previously used to fit the SAXS data of ion tracks in San Carlos olivine [21], apatite [22], and quartz [32]. The track radii extracted from the fits to the SAXS spectra up to 530 °C are plotted as a function of the annealing temperature in Fig. 3. The track radius decreases with increasing annealing with a significant size reduction starting at about 400 °C. Similar to the annealing of ion tracks in apatite reported in Ref. [22], we approximate the spectrum of activation energies for track recovery

B. Afra et al. / Nuclear Instruments and Methods in Physics Research B 326 (2014) 126–130

in olivine by dividing it into two regions: the recovery process in olivine appears to proceed faster after an initial stage of slow change. Below 360 °C, the number of data points is not sufficient to reliably deduce an activation energy. For temperatures above 360 °C (solid line in the inset of Fig. 3), by linear fit in the corresponding Arrhenius plot [22,34], the activation energy of 0.41 ± 0.05 eV can be calculated. Damage recovery of ion tracks in Durango apatite in a similar approach was previously reported at 300 °C with an activation energy of 0.72 ± 0.02 eV [22]. Since the annealing temperature for olivine in the present paper is higher than for the Durango apatite [22], a lower activation energy for olivine may seem unexpected. Moreover, the prior work of Virk and co-workers [34–36] report an activation energy for swift heavy ion track annealing in natural apatite of about 0.7 eV and much higher values for most silicate minerals. However, similar to zircon structures, orthosilicates and orthophosphates, where much higher critical temperature of orthosilicate phase is attributed to the lower connectivity of the PO4 tetrahedra [37,38], in the case of olivine, easier rotation and reorganization of the structural units may lead to a lower activation energy compared to apatite structure. The estimated activation energy of damage recovery in olivine may be affected by the method of analysis employed that does not take into account the dependency of the annealing behaviour on time. By assuming a characteristic jump frequency or time constant, it is possible to estimate an activation energy that takes into account both the temperature and the time factors [39]:

129

Fig. 4. Scattering intensities as a function of the scattering vector q for ion tracks in Durango, Ca-, and Sr-britholite. SAXS spectra are chosen for measurements of unannealed samples and after annealing at 400 °C. The solid lines correspond to the fits using the hard cylinder model. The spectra are offset for clarity.

Ea ¼ kTðln B tÞ where k is the Boltzmann constant, T is the characteristic recovery temperature (maximum rate or slope of recovery), B is the time frequency factor, and t is the annealing time [40,41]. By assuming a characteristic recovery temperature of 360 °C and a jump frequency of 1  1014 s1 (characteristic of recrystallisation in glass), an activation energy of 2.2 eV can be calculated [39,42]; assuming a jump frequency of 1  1010 s1 (characteristic of point defect diffusion processes [43]), results in an activation energy of 1.7 eV. This illustrates the dependence of the activation energy on a jump frequency or characteristic time constant, and this is largely a first order process. However, measurement of the annealing jump frequency for olivine requires further information which is out of the scope of this work. In order to fully understand the damage recovery of ion tracks in olivine isothermal in situ annealing experiments are needed at a variety of temperatures. In addition, to avoid oxidation, annealing experiments should preferably be carried out in reducing atmospheres.

Fig. 5. Ion track radii in Durango apatite and synthetic Ca- and Sr-britholite as a function of the annealing temperature. The solid lines are to guide the eye.

3.2. Tracks in apatite structures The scattering intensities of ion tracks in Durango apatite, and two synthetic britholite: Nd8Sr2(SiO4)6O2 and Nd8Ca2(SiO4)6O2 at room temperature and at 400 °C are shown in Fig. 4. The solid lines are the fits to the hard cylinder model that provides best fits to the experimental data for all compositions and annealing temperatures. After annealing at 400 °C, the position of the first minima shifts to larger q values (corresponding to smaller track radii); the effect is significantly smaller for the Sr- and Ca-britholite than for the Durango apatite. This indicates that ion track recovery in both synthetic apatite structures (britholite) may start at higher temperatures than in Durango apatite. The extracted track radii in the Durango, Ca- and Sr-britholite samples are plotted in Fig. 5 as a function of the annealing temperature. Prior to annealing, the track radii of Ca-britholite and Durango apatite are similar, while the radius in Sr-britholite is slightly higher. The track radii in the three apatite structures at room temperature are reported in Table 1. The track radius from the Durango apatite agrees with the previous SAXS measurements of similar samples irradiated

with different ion and energy combinations [22]. Extrapolation of the change in the track radii as a function of the energy loss measured in Ca9Nd(SiO4)5(SiO4)F2 [44] indicates that our measurement of radius in Nd8Ca2(SiO4)6O2 is well aligned with the previously reported XRD measurements of un-etched ion tracks in monosilicated fluorapatite [44]. As expected from the scattering intensities in Fig. 4, track radii in the Durango apatite change significantly upon annealing, while track radii in the synthetic apatites do not show substantial reduction. These results confirm that chemical composition is an important cause of variation in annealing behaviour of ion tracks in apatite structures and agree with the previously reported findings that annealing rates may be slower in specimens with appreciable substitution of Ca by other cations [20]. The results are also consistent with the annealing of radiation damage in Ca9Nd(PO4)5(SiO4)F2 britholite that showed no noticeable change when annealed at 300 °C for 4700 min [45]. Furthermore, while Durango apatite is a phosphate, the other apatites, as well as the olivine, are silicates

130

B. Afra et al. / Nuclear Instruments and Methods in Physics Research B 326 (2014) 126–130

indicating a higher temperature resistance of ion tracks in the latter. This is also known for fission tracks in zircon (ZrSiO4) [46], another silicate with significance in thermochronology. A detailed study of the radiation damage in a variety of phosphates has previously shown that materials with zircon structure are amorphized at higher temperatures than those with monazite structure [47]. Phosphates and silicates with the same crystal structure show large differences in their response to irradiation. Phosphates have lower critical temperatures of amorphization and activation energies for dynamic diffusion-driven irradiationenhanced epitaxial recrystallization [37,47,48]. The results of the present work are well aligned with the previously reported comparison between silicates and phosphates and provide a direct comparison of the change in the damage track radii in apatite and the two silicates. SAXS radii of tracks in Durango apatite annealed only at 380 and 400 °C are also plotted in Fig. 5. Comparison with the Durango apatite which was annealed at all temperature steps below 400 °C, reveals that the track radii are slightly larger, however, remain within the uncertainty. This result demonstrates that the annealing history in a timeframe of several minutes may not have a significant effect on the damage recovery of ion tracks. Therefore, the approach of sequential annealing of the same sample yields meaningful results.

4. Conclusions We have investigated the annealing behaviour of ion tracks in San Carlos olivine, Durango apatite, and synthetic Nd8Sr2(SiO4)6O2 and Nd8Ca2(SiO4)6O2 britholite (with the apatite structure) by combining synchrotron based SAXS with ex situ annealing. Results indicate fragmentation of the tracks in olivine above 620 °C and appearance of an additional signal in the SAXS images possibly related to oxidation of the material at higher temperatures. Using an Arrhenius plot, an activation energy of 0.41 ± 0.05 eV was calculated for the track recovery due to recrystallisation of the damage in olivine. Unlike in Durango apatite, track radii in both silicates with apatite structure did not show substantial change upon annealing which confirms the important role of chemical composition for the recovery of track damage. Although the set of experiments performed for this work cannot provide a full explanation of the track recovery process, our SAXS results provide new insight into the annealing behaviour of the tracks and are useful for designing appropriate experiments in order to further quantify details of track recovery.

Acknowledgements This research was undertaken on the SAXS/WAXS beam line at the Australian Synchrotron, Victoria, Australia. PK acknowledges the Australian Research Council for financial support. The work at the University of Michigan was supported by the Office of Basic Energy Sciences of the USDOE (DE-FG02-97ER45656).

References [1] D.A. Young, Nature 182 (1958) 375. [2] E.C.H. Silk, R.S. Barnes, Phil. Mag. 4 (1959) 970. [3] V.L. Ginzburg, E.L. Feinberg, N.G. Polukhina, N.I. Starkov, V.A. Tsarev, Dokl. Phys. 50 (2005) 283. [4] R.L. Fleischer, P.B. Price, R.M. Walker, M. Maurette, J. Geophys. Res. 72 (1967) 331. [5] T. Plieninger, W. Kratschmer, W. Gentner, Proc. 3rd Lun. Sci. Conf. 2933–2939 (1972). [6] P.B. Price, D. Lal, A.S. Tamhane, V.P. Perelygin, Earth Planet. Sci. Lett. 19 (1973) 377–395. [7] P.F. Green, R.K. Bull, S.A. Durrani, Nucl. Instrum. Methods 157 (1978) 185–193. [8] R.K. Bull, S.A. Durrani, Earth Planet. Sci. Lett. 32 (1976) 35–39. [9] R.K. Bull, S.A. Durrani, Nucl. Track Detect. 1 (1977) 75–80. [10] R.L. Fleischer, P.B. Price, R.M. Walker, Nuclear Tracks in Solids: Principles and Applications, University of California Press, Berkeley, 1975. [11] K. Gallagher, R. Brown, C. Johnson, Annu. Rev. Earth Planet Sci. 26 (1998) 519. [12] A.J.W. Gleadow et al., Rev. Miner. Geochem. 48 (2002) 579. [13] T. Tagami, P.B. O’Sullivan, Rev. Miner. Geochem. 58 (2005) 19. [14] W.J. Weber, H. Matzke, Mater. Lett. 5 (1986) 9. [15] W. Li, L. Wang, K. Sun, M. Lang, C. Trautmann, R.C. Ewing, Phys. Rev. B 82 (2010). [16] D. Schauries et al., J. Appl. Cryst. 46 (2013), http://dx.doi.org/10.1107/ S0021889813022802. [17] K.D. Crowley, M. Cameron, B.J. McPherson, Nucl. Tracks Radiat. Meas. 17 (1990) 409–410. [18] J. Barbarand, A. Carter, I. Wood, T. Hurford, Chem. Geol. 198 (2003) 107–137. [19] P.F. Green, I.R. Duddy, A.J.W. Gleadow, P.R. Tingate, G.M. Laslett, Nucl. Tracks Radiat. Meas. 10 (1985) 323–328. [20] W.D. Carlson, R.A. Donelick, R.A. Ketcham, Am. Mineral. 84 (1999) 1213–1223. [21] B. Afra, M.D. Rodriguez, M. Lang, R.C. Ewing, N. Kirby, C. Trautmann, P. Kluth, Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact. Mater. Atoms 286 (2011) 243. [22] B. Afra et al., Phys. Rev. B 83 (2011) 64116. [23] W.J. Weber, Y. Zhang, H. Xiao, L. Wang, RSC Adv. 2 (2012) 595–604. [24] W.J. Weber et al., J. Mater. Res. 13 (1998) 6. [25] I. Bae, Y. Zhang, W.J. Weber, J. Mater. Res. 23 (2008) 4. [26] J. F. Ziegler, The Stopping and Range of Ions in Matter, 2008. Available from: . [27] Y. Masubuchi, M. Higuchi, H. Katase, T. Takeda, S. Kikkawa, K. Kodaira, S. Nakayama, Solid State Ionics 166 (2004) 213. [28] J.A. Fahey, W.J. Weber, F.J. Rotella, J. Solid State Chem. 60 (1985) 145. [29] A.M. Glaeser, Interface Sci. 9 (2001) 65. [30] F.A. Nichols, W.W. Mullins, Trans. Metall. Soc. Aime 233 (1965) 1840. [31] S.E. Haggerty, I. Baker, Contrib. Miner. Petrol. 16 (1967) 233–257. [32] B. Afra et al., J. Phys.: Condens. Matter 25 (2013) 045006. [33] P. Kluth et al., Phys. Rev. Lett. 101 (2008) 175503. [34] H.S. Virk, Radiat. Eff. Defects Solids 133 (1995) 87. [35] A.S. Sandhu, L. Singh, R.C. Ramola, S. Singh, H.S. Virk, Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact. Mater. Atoms 46 (1990) 122. [36] S.K. Modgil, H.S. Virk, Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact. Mater. Atoms 12 (1985) 212. [37] A. Meldrum, L.A. Boatner, R.C. Ewing, Mineral. Mag. 64 (2000) 183. [38] R.C. Ewing, A. Meldrum, L.M. Wang, W.J. Weber, L.R. Corrales, Rev Mineral. Geochem. 53 (2003) 387. [39] W.J. Weber, Radiat. Eff. Defects Solids 77 (1983) 295. [40] W. Primak, Phys. Rev. 100 (1955) 1677. [41] W. Primak, J. Appl. Phys. 31 (1960) 1524. [42] K.G. Cheng, J. Phys. Chem. B 103 (1999) 8272. [43] P. Peter Pichler, Intrinsic Point Defects, Impurities, and Their Diffusion in Silicon, Springer, 2004. [44] R. Tisserand, M. Rebetez, M. Grivet, S. Bouffard, A. Benyagoub, F. Levesque, J. Carpena, Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact. Mater. Atoms 215 (2004) 129. [45] R. Tisserand, M. Grivet, M. Rebetez, B. Ravat, J. Carpena, S. Bouffard, Nucl. Instrum. Methods Phys. Res. Sect. B-Beam Interact. Mater. Atoms 239 (2005) 164. [46] W. Li, L. Wang, M. Lang, C. Trautmann, R.C. Ewing, Earth Planet. Sci. Lett. 302 (2011) 227–235. [47] A. Meldrum, L.A. Boatner, R.C. Ewing, Phys. Rev. B 56 (1997) 13805. [48] W.J. Weber, R.C. Ewing, A. Meldrum, J. Nucl. Mater. 250 (1997) 147.