Effect of H and He irradiation on cavity formation and blistering in ceramics

Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

Contents lists available at SciVerse ScienceDirect

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

Effect of H and He irradiation on cavity formation and blistering in ceramics S.J. Zinkle ⇑ Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, TN 37831, USA

a r t i c l e

i n f o

Article history: Received 9 September 2011 Received in revised form 9 March 2012 Available online 25 April 2012 Keywords: Ion irradiation Ceramic oxides Silicon carbide Silicon nitride Alumina Spinel Magnesia Transmission electron microscopy Bubble Void Proton irradiation Ion cutting

a b s t r a c t Single- or poly-crystalline specimens of SiC, Si3N4, MgO, Al2O3 and MgAl2O4 were implanted with 0.4– 1 MeV H+ or He+ ion beams at room temperature and 650 °C up to fluences of 1  1022/m2. This produced peak implanted gas and displacement damage levels as high as 50 at.% and 34 displacements per atom (dpa). The specimens were subsequently examined optically, and in cross-section using transmission electron microscopy. Subsurface blistering occurred for specimens irradiated to H or He fluences greater than about 3  1021/m2 (15 at.% peak implanted gas concentration), and surface exfoliation occurred for fluences above 1  1022/m2 (40 at.% implanted gas). Both helium and hydrogen had comparable effectiveness for inducing blistering and exfoliation on an atomic basis. The threshold blistering and exfoliation fluences for both ions were weakly dependent on temperature between 25 and 650 °C. Both H and He were found to be very effective in inducing matrix cavity formation, due to their low solubility in these ceramics. The implanted gas concentrations that resulted in visible cavity formation generally ranged from 1 to 5 at.%. Visible cavity formation was readily induced during room temperature irradiation despite the limited vacancy mobility in these ceramics at room temperature. Three general types of cavity morphologies were observed: isolated cavities, clusters of small cavities (typically associated with dislocation loops), and two-dimensional platelets. Cavity formation was observed to initiate at the periphery of dislocation loops in some cases. During elevated temperature irradiation, cavity formation was often observed to be preferentially associated with certain low-index habit planes, particularly if  0 0} for Al2O3, the habit plane was oriented nearly parallel to the irradiated surface: (0 0 0 1) and {1 1 (0 0 0 1) for a–SiC, {0 0 1} and {1 1 0} for MgO, and {1 1 0} and {1 1 1} for MgAl2O4. The bubble formation and blistering behavior of the ceramics was similar to that observed in other studies of metals irradiated at comparable homologous temperatures. Ionization-induced diffusion effects associated with dual-beam light ion irradiation appeared to exert only a weak effect on cavity and dislocation loop growth compared to the single ion irradiation conditions. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Proposed fusion energy systems will utilize a variety of inorganic insulators and ceramic composites for plasma diagnostics, plasma heating, and structural applications [1–3]. Ceramics such as Al2O3, MgAl2O4 and SiC produce high concentrations of gaseous (H, He) transmutation products when exposed to high-energy neutrons. Neutronics calculations indicate the hydrogen and helium production rates in Al2O3 and SiC exposed to 14 MeV fusion neutrons is about 60–150 atomic parts per million (appm) for a damage level of one displacement per atom (dpa), whereas the corresponding transmutation rate in a fission reactor is only 2–5 appm/dpa [4,5]. This would produce H and He impurity concentrations exceeding 1 at.% for ceramic materials exposed to anticipated lifetime doses of 200 displacements per atom (dpa)

⇑ Tel.: +1 8655765785. E-mail address: [email protected] 0168-583X/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nimb.2012.03.030

in the first wall region of proposed fusion reactors. It is widely recognized that helium can enhance cavity formation in irradiated materials due to its limited solubility. For example, MgAl2O4 exhibits negligible cavity swelling after irradiation to high damage levels by ions or fission neutrons [6,7], whereas pronounced cavity formation in the matrix and at grain boundaries occurs after dual beam irradiation at conditions relevant for fusion energy [8]. Hydrogen has limited solubility in most ceramics and therefore might be trapped at cavities (stimulating cavity nucleation and growth). In addition, hydrogen may react chemically with some ceramics [9]. The plasma-facing regions of fusion reactors would also be exposed to intense fluxes of low energy hydrogen and helium ions. Previous work has shown that blistering and eventual exfoliation of the blister cap occurs in many materials during prolonged exposure to keV to MeV H and He ion beams [10,11]. A variety of materials are being considered as potential plasma-facing armor, and therefore further information on the effects of high fluxes of H and He ions on the threshold doses for blister formation and exfo-

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

5

liation is useful for determining the material parameters which provide the best radiation resistance. Fundamental studies on blistering and exfoliation following H and He implantation are also relevant for evaluating the suitability of ion implantation methods for precision fabrication of complex microelectronic and optoelectronic systems such as semiconductor on insulator devices [12–14]. Previous studies on Si have observed that H irradiation tends to form planar cavities, whereas inert gas (e.g., He) irradiation typically forms three-dimensional cavities [15]. This paper investigates the surface and bulk microstructural changes observed in H and He implanted MgO, Al2O3, MgAl2O4, Si3N4, and SiC, including the threshold fluences for blistering and exfoliation. The selected irradiation temperatures of room temperature and 650 °C cover a wide range of vacancy mobilities [16]. Long-range vacancy migration should not occur in any of these ceramics at room temperature. At 650oC, vacancies are mobile in the oxide ceramics [16,17] but are still not mobile in SiC [18].

2. Experimental Procedure Polycrystalline blocks of Al2O3 (GE Lucalox or Wesgo Al995), MgO (Ube UMP-0505), MgAl2O4 (Ceradyne) ß-SiC (Cercom) and Si3N4 (Kyocera SN733) and (0 0 0 1)-oriented single crystal wafers of a-SiC (Cree) were machined into 3 mm diameter  0.5 mm thick samples using a combination of diamond sawing and ultrasonic cutting. The grain sizes of the polycrystalline specimens ranged from 1 lm for the hot isostatically pressed Si3N4 to 30 lm for the sintered oxide specimens. The polycrystalline samples were mechanically polished using 0.05 lm diamond paste, whereas the SiC single crystals were polished by the manufacturer. The specimens were bombarded in a 3  3 target array at room temperature or 650 °C in the triple ion beam accelerator facility at Oak Ridge National Laboratory [19]. These temperatures correspond to 0.15 and 0.4 TM for all of the materials except MgO, where TM is the melting temperature (sublimation temperature for SiC and Si3N4). The irradiation temperatures corresponded to 0.1 and 0.3 TM for MgO. For the oxides, these two temperatures are above and below the temperature for long-range vacancy migration [16,17]. For SiC, vacancies should be immobile and C interstitials should be mobile at both temperatures, whereas Si interstitial atoms should be immobile at room temperature and mobile at 650 °C [16,18,20]. The specimens were exposed to H or He ion beams with energies ranging from 0.4 to 1 MeV. The central 2 mm diameter surface region of the polished specimens was exposed to the ion beam, whereas the outer 0.5 mm rim was shielded by an overlying metal mask that securely held the specimens in place during the irradiation. Nine specimens arranged in a 3  3 array were simultaneous irradiated in a given irradiation campaign; the irradiation matrix typically consisted of duplicate or triplicate specimens of three or four materials. The beam was electronically wobbled to provide uniform exposure to all specimens [19]. Average specimen temperatures were continuously monitored using thermocouples in the base of the specimen holder and attached to the surface of a dummy stainless steel specimen occupying one of the 9 irradiation positions. Typical particle fluxes for the irradiation ranged from 0.8–6  1017 H/m2-s and 0.4–4  1017 He/m2-s. Some specimens were also exposed to simultaneous dual beams of H and He ions with H/He flux ratios ranging from 3 to 10. Fig. 1 shows the calculated (SRIM 2008 [21]) depth-dependent displacement damage and implanted gas ion profiles for 1 MeV H and He irradiated Al2O3 for ion fluences of 1  1022/m2. The maximum exposure fluences in this study were 1.7  1022 H/m2 and 1  1022 He/m2, which correspond to calculated [21] peak damage

Fig. 1. Calculated displacement damage and implanted gas ion profiles for 1 MeV H and He irradiated Al2O3 for ion fluences of 1  1022/m2. The arrows mark the peak dpa and implanted gas concentrations.

and implanted gas ion concentrations of 3 dpa and 30 at.% H (1  105 appm H/dpa) and 34 dpa and 54 at.% He (1.6  104 appm He/dpa) for 1 MeV H and 1 MeV He ions, respectively, assuming an average displacement energy of 40 eV for all of the elements in the ceramics in this study. The minimum investigated H and He fluences in this study were 1  1021/m2 (2 at.% H and 5 at.% He at peak implantation region). The implanted ion depths were 10 lm and 2.2 lm for 1 MeV H and 1 MeV He ions, respectively. The calculated implanted gas concentration per dpa decreased rapidly outside the peak implantation region. The calculated SRIM-2008 [21] full widths at half maximum for the implanted ion distributions were 0.57 lm and 0.15 lm for 1 MeV H and He ions, respectively. At depths midway between the surface and implanted ion region, the calculated implanted H or He concentrations were negligible (<1 appm) and the calculated displacement damage was 3.5% of the peak damage levels. Following irradiation, the samples were examined with an optical microscope to detect blistering and surface exfoliation, and further characterization of exfoliated specimens was performed using scanning electron microscopy. Selected samples were prepared for cross-section transmission electron microscopy (TEM) using techniques that are described elsewhere [22]. The TEM specimens were examined in a Philips CM-12 or CM-30 electron microscope, operating at 120 and 300 keV, respectively.

3. Results and discussion For most of the investigated ceramics, the H and He ion irradiation produced a mixture of dislocation loops and cavities at both room temperature and 650 °C. In general, the loop and cavity formation was localized within a subsurface band corresponding to the peak damage and implanted ion region (with the exception of He-irradiated Al2O3 where loops occurred throughout the irradiated region and He-irradiated MgAl2O4 at 650 °C where cavities were observed throughout the irradiated region). The dislocation loops are a result of displacement damage associated with the energetic ion irradiation that transferred energies to the host lattice atoms in excess of the values needed to create vacancies and self-interstitial atoms [16]. Since the interstitial atoms are mobile in all of the investigated materials at room temperature and higher temperatures (except for SiC at room temperature), these point de-

6

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

fects can diffuse and cluster to produce interstitial-type dislocation loops. For SiC, irradiation at room temperature did not produce resolvable dislocation loops due to the immobility of interstitials and vacancies [16,18,20]; instead, amorphization was induced in SiC for damage levels above 0.8 dpa. The dislocation loop microstructure following ion irradiation for many of the investigated materials has been previously summarized elsewhere by the author [16,23–25]. 3.1. Threshold concentrations for blistering and surface exfoliation Fig. 2 shows the low-magnification optical microstructures of several specimens following high-fluence H or He ion irradiation at 650 °C. The circular arc visible in the upper portions of several of the photos marks the boundary between the irradiated and unirradiated region. The dark regions within the irradiated portion of the photos correspond to surface exfoliation. Pronounced surface exfoliation occurred in the proton-irradiated oxide specimens after 1.7  1022/m2 (30 at.% H) with the largest percentage exfoliation occurring in MgAl2O4. Blistering was observed in proton-irradiated Si3N4 at 0.4–1  1022/m2, but surface exfoliation did not occur. Scanning electron microscope observation of the exfoliated regions following high-dose proton and helium ion irradiation revealed relatively smooth surfaces. Surface undulations of typical lateral

wavelength 0.5 to 1 lm and depth variability of <50 nm were observed in 1 MeV H irradiated Al2O3 after a fluence of 1.7  1022/m2 at 650 °C. Similarly, the surface roughness determined by TEM of blistered regions of MgAl2O4 irradiated with 1 MeV He or 0.4 MeV H at room temperature or 650 °C to fluences of 5 to 10  1021/m2 was 0.3–0.5 lm in the lateral direction and <50 nm depth variability (see, e.g., Figs. 3a and 7b). Surface exfoliation occurred in all of the oxide specimens irradiated with 1 MeV He ions to 1  1022/m2 (50 at.% He) as well as in b-SiC irradiated to 0.4  1022/m2 (20 at.% He). Table 1 summarizes the threshold H and He ion fluences (/t) and implanted gas concentrations that produced blistering and surface exfoliation at room temperature and 650 °C, based on optical microscopy. Blister formation and surface exfoliation generally occurred in specimens irradiated to fluences above 3  1021/m2 and 1  1022/m2, respectively. The corresponding peak implanted gas concentrations are 10–20 at.% and 20–50 at.%, respectively. Hydrogen implantation generally required a higher fluence than the helium implantations to produce blistering or surface exfoliation. However, it is worth noting that the calculated peak H concentrations that induced blistering and amorphization are comparable or lower than the corresponding He concentrations (Table 1) due to more pronounced range straggling for the relatively light and more deeply penetrating H ions. As will be

Fig. 2. Plan view optical micrographs of polycrystalline Al2O3, MgO, MgAl2O4, Si3N4 and ß-SiC implanted at 650 °C with 1 MeV H (1.7  1022/m2), 0.4 MeV H (1.0  1022/m2), or 1 MeV He (1.0  1022/m2 for the oxides and 0.4  1022/m2 for SiC).

7

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

Fig. 3. Cross-section microstructure of Al2O3 following 1 MeV He ion irradiation at 650 °C to a fluence of 1  1022 m 2 (bottom figure). The top figure shows a higher magnification image of typical dislocation loops formed in the midrange irradiated regions.

discussed in Sections 3.3 and 3.4, blistering was preceded by the formation of small cavities in the subsurface gas-implanted region. The threshold dose to produce blistering and surface exfoliation decreased slightly with increasing temperature, in agreement with results previously obtained on metals [11]. There was not a pronounced difference in the blistering and exfoliation behavior of the five ceramics. Silicon nitride and single crystal a–SiC had the highest resistance to blistering and exfoliation, although most of the materials exhibited roughly comparable behavior. It is worth noting that the single crystal a–SiC specimens exhibited similar threshold doses for blistering and exfoliation (i.e., exfoliation occurred soon after the presence of visible blisters), whereas the polycrystalline SiC and other ceramics typically blistered at doses about three times lower than the dose required to produce surface exfoliation. The dose for blister formation in SiC for He ions at 650oC was approximately four times larger for a–SiC single crystal specimens than for b–SiC polycrystal specimens. The present observations are in general agreement with previous extensive studies on metals [11,26] and limited data on ceramics [13,27–35] that the threshold fluence for blister formation for monoenergetic H and He ions occurs near 2 to 5  1021/m2 for 1 MeV ions (10–20 at.% implanted gas), with lower fluences required at lower ion energies due to reduced range straggling. Significantly lower blistering thresholds have been previously reported for H- or He-implanted samples subjected to postirradia-

tion annealing. For example, blistering has been observed in MgAl2O4 implanted with 1 at.% He after annealing at 800 °C [36]. Similarly, the threshold gas concentration for blistering and exfoliation in H implanted Si is reduced from 20–30 at.% in as-implanted Si to 5 at.% H in samples that were annealed at 500 °C after implantation [12,15,37], and for SiC implanted with H ions at room temperature the threshold concentration for blistering decreased from >15 at.% in the as-implanted case to 5 at.% H after annealing at 900 °C [15]. In the present study, the size of the blisters was noticeably smaller for 1 MeV He (2 lm implantation depth) compared to 1 MeV H implantation (10 lm implantation depth), in agreement with previous studies on ceramics [28,29,31]. Surface exfoliation was generally initiated within individual grains in the He-implanted specimens, whereas the blister formation and exfoliation extended over several grains in the H-implanted specimens (Fig. 2). These differences may be attributable to the deeper range of the H ions compared to the He ions (10 vs. 2 lm for 1 MeV ions in ceramics). Previous studies on metals have reported that the blister size increases with increasing implantation depth, with a roughly linear relationship [10,26] At energies between 20 and 500 keV an approximately correlation of r/R5 was observed where r is the blister radius and R is the ion range [10,26]. Observations of H blistering in silicon have similarly noted that the blister size increases with implantation depth, although a power law distribution (r  R)n with n  0.35 to 0.5 was used to describe the experimental data [37]. Although numerous helium implantation studies have been performed on metals [38–40], relatively little is known about the microstructural effects of simultaneous high displacement damage and gas concentration in ceramics [8,34,41–46]. Therefore, the microstructure of the H and He ion irradiated ceramic specimens was analyzed by cross-section TEM to determine the physical processes responsible for the blistering and surface exfoliation. The cross-section technique allowed the microstructure of regions containing between zero and the maximum implanted gas concentration to be studied, in order to differentiate between displacement damage and implanted gas atom effects. Comparisons between specimens irradiated to different fluences provided a cross-check on the effect of implanted gas on the cavity evolution.

Table 1 Summary of threshold fluences (/t) and peak implanted gas concentration to produce blistering and surface exfoliation in H and He implanted ceramics. The parentheses denote conditions where blistering or exfoliation was not observed up to the maximum investigated fluence. Ion and Tirr Material Blistering /t Blistering Exfoliation /t Exfoliation (1022 m 2) conc. (at.%) (1022 m 2) conc. (at.%) H, 50 °C ‘‘ ‘‘ ‘‘ H, 650 °C ‘‘ ‘‘ ‘‘ ‘‘ He, 50 °C ‘‘ ‘‘ ‘‘ He, 650 °C ‘‘ ‘‘ ‘‘ ‘‘ ‘‘

a–SiC Si3N4 Al2O3 MgAl2O4 a–SiC Si3N4 MgO Al2O3 MgAl2O4 a–SiC Si3N4 Al2O3 MgAl2O4 b-SiC a–SiC Si3N4 MgO Al2O3 MgAl2O4

(>1) (>0.3) 0.3 0.3 0.8 0.4 0.3 0.3 0.2 (>0.5) 0.3 0.4 0.2 0.2 0.8 (>0.2) 0.3 0.3 0.1

(>20) (>5) 5 5 15 7 5 5 4 (>30) 15 20 10 10 40 (>10) 15 15 5

(>1) (>0.3) (>1) (>1) 0.8 (>1) 1 1 1 (>0.5) (>0.3) (>0.5) (>0.5) 0.4 0.8 (>0.2) 0.8 1 0.4

(>20) (>5) (>20) (>20) 15 (>20) 20 20 20 (>25) (>15) (>25) (>25) 20 40 (>10) 40 50 20

8

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

3.2. Measured ion ranges Table 2 summarizes the measured ion ranges for 1 MeV H and He ions in four of the investigated ceramic materials as determined by cross-section TEM observation. The measured ion ranges are greater than the ranges predicted by SRIM-2008 [21] by 6% to 15%. Similar discrepancies between the experimental and SRIM calculated ion ranges in Al2O3 for a wide range of ions have been previously reported [23]. The most likely cause of the discrepancy is associated with errors in the electronic stopping power (Se) [47,48], which is a particular challenge for conditions intermediate between the low energy behavior defined by the LSS theory of Lindhard, Scharff and Schiøtt (Se  E1/2, where E is the ion energy) and the high energy Bethe-Bloch behavior, Se  (lnE)/E [47]. This crossover in functional dependence occurs for light ion energies near 100 keV. A recent survey of stopping powers reported significant errors in the SRIM electronic stopping powers for slow heavy ions compared to LSS theory and experimental neutron scattering data [48]. The discrepancies between the measured (Table 2) and SRIM calculated ion ranges should not be due to swelling of the irradiated region during the implantation. From previous work on ion irradiated Al2O3 and MgAl2O4 [49] and neutron-irradiated SiC [50,84], the radiation-induced ‘‘point defect’’ volumetric swelling should be less than 1–3%, and the cavity swelling in the peak implanted region was 1% for the lower-fluence samples used for the ion range measurements. 3.3. Effects of He on cavity formation Fig. 3 shows the general cross-section microstructure of Al2O3 following 1 MeV He ion irradiation at 650 °C to a fluence of 1  1022 m 2 (50 at.% He). The calculated displacement damage associated with this irradiation (cf. Fig. 1) ranged from 34 dpa at the peak damage region near 2.2 lm to 1 dpa in the near-surface region (0.5 lm depth). The dark region extending up to a depth of 2.5 lm is associated with a high density of small dislocation loops. These dislocation loops were visible throughout the irradiated region, and are shown in more detail in the top image in Fig. 3. Near the damage peak, the dislocation loop density was 2  1022/m3 and the mean loop diameter was 15 nm. A nearly continuous band of large interconnected cavities was observed at a depth near 2.3 lm (corresponding to the peak He implantation region), which is indicative of the onset of extensive exfoliation. Examination of Al2O3 specimens irradiated to an order of magnitude lower fluence at 650 °C revealed preferential nucleation of cavities on (0 0 0 1) basal plane dislocation loops in the He implanted region, as shown in Fig. 4. In the grain imaged in this figure, the (0 0 0 1) basal plane is oriented approximately parallel to the surface normal and the cavity-loop complexes are viewed edge-on (visible as white lines in Fig. 4). The average size of the dislocation loop-cavity complexes in the peak region of the Al2O3 specimen irradiated with 1 MeV He ions to a fluence of 1  1021 m 2 (3.4 dpa, 5 at.% He at peak region) at 650 °C was 25 nm diameter. The planar loop-cavity complexes consisted of numerous small spherical cavities (3 nm diameter) grouped within the dislocation loop habit plane. Fig. 5 shows another example of small cavities nucleated on (0 0 0 1) dislocation loops in Al2O3 following He ion irradiation at 650 °C. In this grain, the (0 0 0 1) ba-

Table 2 Ion ranges for 1 MeV H and He in ceramics as determined by cross-section TEM. Ion

Al2O3

MgAl2O4

MgO

SiC

1 MeV H 1 MeV He

9.4 lm 2.3 lm

10.7 lm 2.4 lm

11.0 lm 2.45 lm

2.37 lm

Fig. 4. Preferential cavity nucleation on (0 0 0 1) dislocation loops in Al2O3 following 1 MeV He ion irradiation at 650 °C to a fluence of 1  1021 m 2 (5 at.% He at peak). The micrograph was taken in the underfocused condition so that the cavities appear as low-contrast (white) objects. The original irradiated surface is located to the left  0 0] and [1 2  1 0], allowing the of the micrograph. The foil normal is between [1 1 basal plane loop-cavity complexes to be viewed edge-on.

sal plane was oriented about 25 degrees from the surface normal. This demonstrates the cavity alignment is preferentially associated with the crystallographic basal plane in Al2O3 rather than the specimen surface normal, at least for misorientation angles up to 25° between the basal plane and specimen surface. Cavity formation was also observed in the He-implanted regions of Al2O3 irradiated at room temperature. Fig. 6 shows the cavity microstructure of Al2O3 following 0.8 MeV He ion irradiation at room temperature to a fluence of 1.5  1021 m 2 (8 at.% He and 5 dpa in peak implanted region). A high density of planar cavities in the peak He-implanted region is visible with habit planes nearly parallel to the irradiated surface. The cavities appear to reside on at least three different habit planes, each of which is roughly parallel with the irradiated surface; the specific habit planes were not identified for this irradiation condition. As indicated in Fig. 6, the (0 0 0 1) habit plane in this grain was oriented about 40° from the irradiated surface normal and none of the visible cavities appeared to reside on this basal habit plane. It is worth noting that long range vacancy migration would not be expected to occur in Al2O3 irradiated at room temperature, based on the recommended [51] vacancy migration energies of 1.9 eV and 3.7 eV for the Al and O sublattices, respectively. Numerous previous studies have reported either disk-shaped platelets or loop-bubble complexes located on (0 0 0 1) habit planes in Al2O3 following He ion irradiation (with or without postirradiation annealing) [42,52–55]. In the present study, we observed preferential bubble formation on (0 0 0 1) dislocation loops in Al2O3 for an irradiation temperature of 650 °C (e.g., Figs. 4 and 5). However, He ion irradiation of Al2O3 at room temperature produced diskshaped cavities approximately parallel to the irradiated surface,

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

9

Fig. 5. (a) Underfocus and (b) overfocus images showing preferential cavity nucleation on (0 0 0 1) dislocation loops in the peak implantation region of Al2O3 following 1 MeV He ion irradiation at 650 °C to 1  1021 m 2 (5 at.% He at peak). Cavities are visible as white and dark objects for the underfocus and overfocus conditions, respectively. The  0], allowing the basal plane loop-cavity complexes to be viewed edge-on. original irradiated surface is located to the left of the micrographs. The foil normal is near [1 0 1

Fig. 6. Cross-section microstructure (underfocused image) of Al2O3 following 0.8 MeV He ion irradiation at room temperature to a fluence of 1.5  1021 m 2 1  0]. The (5 dpa, 8 at.% He at peak implanted region). The foil normal is near [2 1 original irradiated surface is located to the left of the micrograph.

irrespective of the crystallographic orientation of the polycrystalline grains (cf. Fig. 6). This suggests that irradiation-induced stress effects or another physical mechanism might be providing the dominant influence during room temperature irradiation (particu-

larly if the (0 0 0 1) basal plane has a large misorientation with respect to the irradiated surface, as in the grain shown in Fig. 6). It has long been recognized that volumetric expansion within the ion implanted region leads to significant lateral compressive stress (due to constraint from the nonirradiated substrate material) and an accompanying out-of-plane tensile stress [14,56]. This implantation stress can lead to significant anisotropy in the nucleation of dislocation loops [57], and would also favor the formation of disk-shaped cavities with habit planes parallel to the irradiated surface [58]. Since the volumetric ‘‘point defect’’ swelling of Al2O3 and many other ceramics increases with decreasing temperature between 650 °C and room temperature (by about a factor of two for Al2O3, MgAl2O4, and SiC [49,50]), the magnitude of the ion implantation stress at room temperature may be large enough to overcome the energetically favored (0 0 0 1) orientation for cavities in Al2O3. Cavity formation in H-implanted (0 0 1) Si crystals preferentially occurs as (0 0 1) platelets parallel to the implanted surface due to implantation compressive stress effects, rather than on {1 1 1} habit planes that have lower surface energy [37]. Helium implantation had a profound effect on cavity formation in MgAl2O4, where radiation-induced cavity formation is rare in the absence of co-implanted gases [6–8,43,57,59–62]. According to conventional radiation damage models, helium can be a very potent agent to stabilize cavity formation due to gas pressurization effects [63,64]. The injected helium promoted cavity formation in the grain interior and at grain boundaries in MgAl2O4. Fig. 7 compares the cavity microstructures of MgAl2O4 irradiated with 1 MeV He ions at 650 °C to fluences of 1  1021 and 1  1022 m 2. The latter fluence corresponds to 35 dpa and 50 at.% He in the peak implanted region and 2 dpa and 0.08 at.% He at a depth of 1.5 lm (0.5 lm from the peak implanted region). Small cavities (2.5 nm average diameter) and a few extended microcracks were

10

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

visible within a 200 nm band at a depth near 2 lm (corresponding to the peak damage and injected He region) in the lower-dose specimen. Irradiation to an order of magnitude higher fluence produced only a slight increase in the average cavity diameter to 5 nm within the peak implanted region, indicating that the cavity growth rate was slow (Fig. 7). Extensive cracking occurred within the He implanted region at the higher fluence, producing a nearly continuous interconnected buried cavity layer. Cavities were visible at all depths from the bombarded surface (0.5 dpa) to the implanted ion region (35 dpa) at the higher fluence. The cavity microstructure for the high fluence specimen near a depth of 1 to 1.5 lm is shown in Fig. 7c. The cavity diameter (d) was nearly constant in irradiated regions outside the peak implanted region (d = 3.0 nm at the surface and d = 3.5 nm at a depth of 1.5 lm) in the higher fluence spinel specimen. The corresponding cavity den-

Fig. 7. Comparison of the cross-section microstructures of MgAl2O4 implanted with 1 MeV He ions at 650 °C to fluences of (a) 1  1021 m 2 and (b), (c) 1  1022 m 2. All images taken in underfocus condition. The boxed region in 7b is shown at higher magnification in 7c.

sity increased rapidly from 4  1021/m3 at the surface to 2  1022/ m3 at a depth of 1.5 lm. The presence of cavities in irradiated regions >1 lm from the peak He implanted region suggests significant diffusion of helium and/or He-vacancy complexes, since the calculated as-implanted He concentration at depths shallower than 1 lm is <10 appm He (as noted at the beginning of this paragraph, cavity formation generally has not been observed in irradiated MgAl2O4 in the absence of cavity nucleation stimulants such as H or He). Previous studies on MgAl2O4 implanted at room temperature with H and He did not detect any visible cavities for gas contents up to 2000 appm [43]. Helium ion irradiation of MgAl2O4 at room temperature and 650 °C to fluences of 1  1021 He/m2 produced uniformly distributed matrix cavities. There was also some evidence for enhanced cavity formation associated with {1 1 0} and {1 1 1} dislocation loops, but due to the low number density of dislocation loops in the He-irradiated spinel outside of the peak damage region the predominant midrange cavity microstructure consisted of randomly dispersed individual cavities. A previous simultaneous dual-beam (He plus Al ion) irradiation study on MgAl2O4 at 650 °C found that cavity formation was enhanced at dislocation loops and adjacent to grain boundaries for fusion-relevant He/ dpa ratios of 150 appm He/dpa [8]. At high dual-beam doses (>20 dpa), very large cavities that completely covered the grain boundary region were observed during irradiation at 650 °C [8]. The matrix cavities were preferentially associated with dislocation loops lying on {1 1 1} and {1 1 0} habit planes [8]. Conversely, Yamada et al. did not observe preferential cavity association with {1 1 0} and {1 1 1} loops in spinel preimplanted with H, He and C and then irradiated with Ar ions (cavities were not observed in Ar ion-irradiated spinel that was not preimplanted) [43]. This suggests the method of He gas implantation (preimplantation vs. coimplantation) and perhaps irradiation spectrum [24,65,66] may have a significant influence on the microstructural evolution in MgAl2O4. Fig. 8 shows the microstructure of MgAl2O4 after helium ion irradiation at room temperature to a fluence of 4.7  1021 He/m2 (16 dpa and 25 at.% He in the peak implanted region). Cavity formation is visible within a 250 nm wide band centered near the peak He implantation region. Elongated cavities aligned approximately parallel to the irradiated specimen surface are visible in the peak implanted He region, along with randomly distributed small individual spherical cavities. The wavy orientation of the elongated cavities suggests their formation is not limited to a few low-index habit planes. Unlike the behavior at 650 °C (where cavity formation occurred throughout the irradiated region), visible cavity formation for the room temperature He irradiation was limited to the subsurface He-implanted region even after high fluence irradiation. This observation is in agreement with thermal desorption results by Neeft et al. [67] on room temperature He-implanted MgAl2O4 where it was observed that He migration and release began at temperatures above 525 °C (activation energies of 1.8–2.35 eV depending on dose). Irradiation of MgO with 1 MeV He ions at 650 °C produced cavity formation associated with {001} dislocation loops in the peak implantation region at both investigated fluences of 1  1021 and 1  1022 He/m2 (5 and 50 at.% He peak concentrations). A previous study of 200 keV 3He-implanted MgO did not detect visible cavity formation for irradiation at room temperature or 600 °C for a He ion fluence of 1.8  1020 He/m2 (1.5 at.% He) [42]. Prior work on MgO implanted with He by cyclotron irradiation at room temperature to 0.1–0.24 at.% He followed by postirradiation annealing found that He platelets on {0 0 1} habit planes were produced after annealing at 800–1000 °C [55]. Nanoscale cavities on {0 0 1} habit planes parallel to the irradiated surface have also been observed in MgO following annealing at 1100 °C of specimens irradiated with 30 keV He ions to 1  1020 He/m2 (1 at.% He) [68].

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

11

Fig. 8. Cavity microstructure (underfocused image) of MgAl2O4 after 0.8 MeV He ion irradiation at room temperature to a fluence of 4.7  1021 He/m2 (16 dpa and 25 at.% He in the peak implanted region). The original irradiated surface is located to the left of the micrograph.

A high density of small (<3 nm diameter) cavities were observed in the He implanted region of SiC irradiated at 650 °C to fluences of 0.8 to 10  1021 He/m2. An example of the typical cavity microstructure after a relatively low fluence of 8  1020 He/m2 is shown in Fig. 9, where the cavities are present as randomly distributed individual objects. The cavities were visible only in the He implanted region, consisting of a 130 nm wide band for the fluence of 8  1020 He/m2 (4 at.% He at peak). A grain boundary is present near the top of Fig. 9. There was no evidence for a cavity-free zone next to the grain boundary. At high fluence at 650 °C, the cavities appeared to be preferentially associated with (0 0 0 1) habit planes. Previous studies have found similar results for elevated temperature He ion irradiations regarding preferential cavity orientation along (0 0 0 1) habit planes for a–SiC [69–74]. The He cavities typically appear as closely spaced individual bubbles (1–3 nm diameter) within 10–30 nm diameter disk-shaped regions on (0 0 0 1) habit planes for [0 0 0 1]-oriented a–SiC. Irradiation of SiC with 1 MeV He ions near room temperature resulted in the formation of an amorphous band near the peak damage region for doses greater than 0.6–0.8 dpa (0.6 lm wide amorphous band for 4.7  1021 He/m2; also see Ref. [25]). Fig. 10 shows the representative microstructure in SiC irradiated with 1 MeV He ions at room temperature to a fluence of 4  1021 He/ m2. Cavity formation was visible in the amorphous He-implanted regions of the SiC specimens in a band with a width of 150 nm and 320 nm for fluences of 0.8  1021 He/m2 and 4  1021 He/ m2, respectively (>4 and 20 at.% He peak concentration, respectively). The cavity geometry was indistinct, perhaps due to the amorphous structure of the material, and the cavity size (1– 3 nm) was near the TEM resolution limit. A grain boundary is present in the lower region of Fig. 10. There appeared to be a slightly enhanced cavity width along the residual grain boundary in the peak He implanted region. The residual grain boundary appeared to contain a continuous cavity layer throughout the amorphous irradiated region, which extended outside (closer to the irradiated surface) the implanted He region by several hundred nanometers (cf. Fig. 1 for approximate damage and implanted He profiles). The grain boundary cavity regime also extended about 20 nm deeper than the amorphous region, suggesting enhanced He mobility in the grain boundary compared to the crystalline matrix. A previous study of single crystal 6H a-SiC irradiated with 19 keV He at room temperature reported the formation of large oval cavities after irradiation-induced amorphization had occurred [75]. The

Fig. 9. Cavity microstructure (underfocused image) of the peak damage region of beta-SiC after 1 MeV He ion irradiation at 650 °C to a fluence of 8  1020 He/m2 (2.6 dpa and 4 at.% He in the peak implanted region). The original irradiated surface is located to the left of the micrograph.

Fig. 10. Cavity microstructure (underfocused image) of the peak damage region of beta-SiC after 1 MeV He ion irradiation at room temperature to a fluence of 4  1021 He/m2 (13 dpa and 21 at.% He in the peak implanted region). The irradiation induced amorphization in the midrange and peak damage regions. The original irradiated surface is located to the left of the micrograph. The upper left inset figure shows the electron diffraction pattern for the amorphous region.

critical fluence for He blistering in their study was about 2  1022 He/m2 (>100 at.% He peak concentration)

12

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

3.3.1. Threshold He concentrations for visible cavity formation Table 3 summarizes observations on the threshold H and He implanted ion concentration for visible cavity formation in ceramics from the present study and previous literature results [35,42,44,52,53,55,69,72,76–80]. The threshold fluences for cavity formation in SiC, Al2O3, MgAl2O4, and Si3N4 irradiated with He ions at room temperature and 650 °C were found to be 1.0, 1.0, 1.0, and 1.5  1021/m2 (peak concentration of 5–7.5 at.% He), respectively. Cavity nucleation and growth in irradiated materials typically involves three-dimensional diffusion of vacancies and/or gas-vacancy defect complexes [63,64], and this is presumed to be the operative mechanism for cavity formation at 650 °C in the present study. Although the migration energies of He-vacancy clusters are not as well known as vacancies, in general the addition of He reduces the mobility of vacancies in materials [64,81]. Therefore, cavity formation via long-range migration of vacancies or Hevacancy complexes at room temperature would not be expected to occur in the ceramics investigated in the present study since vacancies are immobile [16–18,20], with the possible exception of amorphous SiC where defect mobilities are not known. The operative mechanism for growth of cavities to TEM-visible sizes near room temperature without involving long-range vacancy (or vacancy–He complex) migration is uncertain. For uniformly dispersed gas concentrations of a few atomic percent, each gas atom is within 3–4 nearest neighbors of another gas atom. The short-range migration and clustering of gas atoms might be enhanced by localized lattice strain in the vicinity of the implanted He. In addition, collisional displacement damage during prolonged irradiation could induce short-range diffusion and thereby facilitate the clustering of He-vacancy complexes into small cavities. Loop punching or other mechanisms such as helium desorption from He-vacancy complexes combined with interstitial He migration [70,82] might also be viable causes. Cavity formation was detected in Al2O3 at the minimum investigated fluences of the present study, i.e., as low as 1  1021 He/m2 (5 at.% He at peak) at room temperature and 650 °C. The width of the TEM-visible cavity regime in 1 MeV He ion irradiated Al2O3 was 210 nm at 650 °C (Figs. 4 and 5) for a fluence of 1  1021 m 2 and 120 nm at room temperature for a fluence of 1.5  1021 m 2 (Fig. 6). From a comparison of the calculated implanted He profile (Fig. 1), this implies the threshold He concentration for cavity formation in Al2O3 is 1.5 at.% at 650 °C and 4.5 at.% at room temperature. Cavity formation in He implanted MgAl2O4 extended throughout the irradiated region at high fluences at 650 °C (cf. Fig. 7b, 7c). At low fluences near 1  1021 He/m2 (5 at.% He at peak), the visible cavity formation was limited to a band 220 nm wide near the He implanted region (Fig. 7a). At room temperature, the visible cavity formation was limited to the implanted He region

and was observed within a band 250 nm wide after irradiation to a fluence of 4.7  1021 He/m2. From a comparison with the calculated He profile (Fig. 1), this implies the threshold He concentration for cavity formation in He-implanted MgAl2O4 is 1.2 at.% at 650 °C and 3.8 at.% at room temperature. Visible cavity formation in MgO irradiated with He ions at 650 °C was restricted to heterogeneous nucleation associated with {001} dislocation loops. The width of the cavity region was 400 nm for a fluence of 1  1022 He/m2 (50 at.% He at peak), which implies that the threshold He concentration for cavity formation at 650 °C is 2 at.% He. The threshold concentration for cavity formation in MgO following room temperature implantation was not measured in this study. Using similar methods, the threshold He concentration for cavity formation in SiC was determined to be 2 at.% at 650 °C and 1.7 at.% at room temperature. In contrast to the results for Al2O3 and MgAl2O4 (Table 3), it is worth noting that the threshold He concentration for visible cavity formation in SiC was lower at room temperature than at 650 °C. The formation of bubbles in He ion irradiated SiC at room temperature was preceded by amorphization in the peak damage region, and it is possible that bubble nucleation and growth to visible sizes might be easier in amorphous SiC compared to crystalline SiC. Muto et al. [75] reported the formation of large He bubbles in SiC that was amorphized by He ion irradiation at room temperature. Zhang et al. [72,73] found that the threshold He concentration for visible cavity formation in SiC that was implanted and then annealed was lower for room temperature irradiation (0.35 at.% He) than for irradiation at 500 K (1.5–2.7 at.% He). These present observations generally agree with previous work on SiC [44,79] and Al2O3 [35] irradiated at room temperature. Conversely, Chen et al. [45,70] observed He platelet formation on (0 0 0 1) habit planes in a–SiC after room temperature implantation of 0.25 at.% He (0.15 dpa), where a degrader wheel was used to produce a uniform depth distribution of implanted He (2.3  1022/m2 total fluence). This is a significantly lower implanted helium concentration for visible bubble formation in SiC compared to other studies that have reported threshold He concentrations of 2.5–8 at.% He [44,78–80]. The lack of amorphization prior to bubble formation in the study by Chen et al. [45,70] might be attributable to a slightly higher irradiation temperature, since the amorphization dose for SiC is very sensitive to temperature near 300–350 K [25,83,84]. In agreement with previous ion irradiation studies on Al2O3 [8,41,42,85] and SiC [69,72,86] performed at temperatures above 500 °C, He implantation in a-SiC and Al2O3 at 650 °C enhanced the formation of cavities along (0 0 0 1) basal plane dislocation loops (cf. Figs. 4 and 5). Irradiation to high fluence (>1  1021/ m2; >5 at.% He) in the current study produced nearly continuous

Table 3 Threshold implanted gas concentrations (CH, CHe) for visible cavity formation in H and He ion-irradiated ceramics as determined by the present study and previous reports [35,42,44,52,53,55,69,72,76–80]. The threshold gas concentrations were determined from a comparison of the visible bubble region width and the calculated (Fig. 1) gas concentrations. Threshold H or He gas concentration for visible cavities (at.%) MgO

MgAl2O4

Al2O3

SiC

Si3N4

He, 25 °C

– >1 [42]

4.5 5–12 [35]

2 0.24–1 [42,55]

1.7 2.5 < CHe < 8 [44,78–80] 2 <3.8 [69,72]

7.5

He, 650 °C

H, 25 °C

–

3.8 1 < CHe < 15 [35,76] 1.0 1 < CHe < 15 [35,76] 13

H, 650 °C

0.5

1

1.5 1 [42,52,53] <10 10 [77] 0.5 1.5 [53,77]

<10 >20 [44] <10

<5

–<15

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

cavity bands and cracking in the helium-implanted region. The results from previous studies suggest the threshold helium concentration for cavity formation in He ion irradiated SiC and Al2O3 at 500–800 °C is 1 at.% (Table 3) [35,42,43,52,53,69,72]. With the notable exception of SiC discussed above, the threshold helium concentration for visible cavity formation in the investigated ceramics at 650 °C is comparable or lower than the concentration needed to produce visible cavities during room temperature irradiation (Table 3). The moderate reduction in He concentration needed to produce visible cavities at higher temperatures might be due to higher mobility of helium-vacancy clusters which would facilitate the nucleation and growth of bubble nuclei. It is interesting to note that the threshold He concentrations for visible bubble formation are not markedly different between room temperature and 650 °C, despite current understanding (as noted in Section 1) that long range vacancy migration should not occur in any of these ceramics at room temperature whereas vacancies are mobile in all of the ceramics but SiC at 650 °C [16–18]. Previous work on He-implanted Al2O3 by Sasajima, Furuno and coworkers [35,53] found that the threshold fluence for visible bubble formation at room temperature was 1.4  1021 He/m2 (12 at.% He, 8 dpa) for a-axis oriented single crystals and 0.6  1021 He/m2 (5 at.% He, 4 dpa) for c-axis oriented crystals. The threshold fluence for visible bubble formation decreased at elevated irradiation temperatures to 1.2  1020 He/m2 (1 at.% He) at 650 °C, and at 800–1000 °C numerous small bubbles were detected inside dislocation loops.

3.3.2. Influence of dual-beam light ion irradiation on microstructure Selected SiC, Al2O3 and MgAl2O4 specimens were exposed at room temperature and 650 °C to simultaneous dual beams of 0.4 MeV H and 0.36–0.8 MeV He ions with H/He particle flux ratios of 3 to 10 in order to investigate possible enhanced diffusion effects [24,65,87] associated with highly ionizing radiation. For these irradiation conditions, the more deeply penetrating H beam provided a high ionizing radiation level within the He-irradiated region while contributing a relatively small amount of atomic displacements. The calculated total ionizing radiation dose rates for the dual beam irradiations near the He-implanted regions were 1 to 10 MGy/s, with 50–80% of the combined ionization associated with the H ion beam and 95–99% of the calculated displacements in the implanted He region due to the He ion beam (electronic to nuclear stopping power ratios >1000) [24]. In general, the highly ionizing dual beam irradiations produced qualitatively similar microstructural behavior compared to the single He ion beam results. A slight increase in He cavity size and decrease in cavity density was observed for the dual beam irradiation conditions compared to the single He ion beam irradiations. Fig. 11 shows the microstructure near the He-implanted region of MgAl2O4 after simultaneous dual beam irradiation with 0.4 MeV H and 0.36 MeV He ions at 650 °C at a H/He flux ratio of 10. The combined damage level in the He implanted region (visible as the vertical band of defect clusters in the right side of Fig. 11a) was about 3.5 dpa, with <5% of the calculated displacement damage due to the H ion beam. At a depth of 0.5 lm midway between the irradiated surface and He peak damage region (left hand portion of Fig. 11a), the calculated combined irradiation parameters were 0.2 dpa (50% due to H beam), 50 appm He and <1 appm H. The dual beam irradiation did not have a pronounced effect on the dislocation loop microstructure compared to single He ion beam irradiation conditions, although a low density of very large (100 nm diameter) dislocation loops were observed in the midrange regions about 0.5 lm from the irradiated surface for the dual beam irradiation (Fig. 11a).

13

Fig. 11. a) Microstructure near the He-irradiated region of MgAl2O4 after simultaneous dual beam irradiation with 0.4 MeV H and 0.36 MeV He ions at 650 °C. The original irradiated surface is located to the left of Fig. 11a. The H/He flux ratio was 10 and the H ion fluence was 1  1022 m 2. Fig. 11b shows a higher magnification image of the cavity microstructure outlined by the inset box in Fig. 11a.

Fig. 12 compares the measured depth-dependent He cavity size and density in MgAl2O4 after single and dual ion beam irradiation at 650 °C. The decrease in the cavity density within 0.2 lm of the peak implanted He depth for both irradiation conditions was due to coalescence of adjoining helium bubbles. Overall, the dual beam irradiation (open and filled triangles in Fig. 12) caused a slight enhancement in the cavity size and a reduction in the visible cavity density over most of the He-irradiated region compared to the 1 MeV He single ion irradiation condition (open and filled squares in Fig. 12). Unlike the single He ion beam case where the cavity size was 3–3.5 nm diameter from the surface to near the peak He implantation region, a significant variation of cavity size with depth was observed in the midrange region for the dual beam irradiated MgAl2O4 (Fig. 11b and Fig. 12). The dual beam irradiation cavity diameter was about 3.5 nm at the irradiated surface and increased to about 5 nm at a depth of 0.4 lm, then decreased to a diameter near 3.5 nm with increasing depths up to 0.7 lm whereupon it began to increase again near the He implanted peak region. The dual beam irradiation visible cavity density also exhibited a local maximum at a depth of 0.4 lm (0.7 lm from the peak He implantation region).

3.4. Effects of H on cavity formation Hydrogen implantation in Al2O3 at 650 °C induced the formation of a moderate density of heterogeneously nucleated cavities. Fig. 13 shows the cross-section microstructure of the peak damage region in Al2O3 following 1 MeV H ion irradiation at 650 °C (3 dpa, 30 at.% implanted hydrogen at peak). The cavities within the 1.2 lm wide H-implanted region were spatially segregated into banded regions containing high and low densities of cavities. The volume-averaged cavity density in the peak damage region was 6  1022/m3 and the mean diameter was 7 nm. The corresponding dislocation loop density near the peak damage region was 4  1020/m3 and the mean loop diameter was 200 nm. Examination of the irradiated regions adjacent to the peak damage region indicated that the heterogeneous grouping of individual cavities was due to their association with dislocation loops, as shown in

14

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

Fig. 12. Comparison of the depth-dependent He cavity size and density in MgAl2O4 after irradiation at 650 °C with single (1 MeV He, 1  1022 He/m2) and simultaneous dual ion beams (0.36 MeV He and 0.4 MeV H; 1  1021 He/m2and 1  1022 H/m2).

Fig. 14. The cavity-loop habit planes included both (0 0 0 1) basal  0 0} prism planes. The two arrows in Fig. 14 point planes and {1 1 to dislocation loops that have 7 to 8 cavities associated with the periphery of the loops. The calculated damage level and implanted hydrogen concentration in this region are 0.3 dpa and 0.1 at.% H, respectively. Other loops depicted in Fig. 14 have cavities distributed nearly uniformly along the entire face of the loop. Similar heterogeneous clustering of individual cavities within dislocation loops was observed in Al2O3 irradiated with 1 MeV H ions at 650 °C to an order of magnitude lower fluence, where the mean cavity diameter was 2.5 nm and the volume-averaged cavity density was 5  1022/m3 (the corresponding dislocation loop size and density in the peak H implanted region were 40 nm and 3  1020/m3, respectively). Preferential nucleation of cavities at  0 0} prism plane dislocation the interior of (0 0 0 1) basal and {1 1 loops has also been observed by Furuno, Sasajima and coworkers [53,77] for 15 keV H2 ion irradiated (0 0 0 1) single crystal Al2O3 at 650 °C. Fig. 15 shows an example of small (5–10 nm diameter by <1 nm thick) platelet cavities on (0 0 0 1) habit planes near the peak

Fig. 14. Cavity formation at dislocation loops in Al2O3 irradiated with 1 MeV H ions at 650 °C to 1.7  1022 H/m2. The image was obtained at an irradiation depth of 9 lm (1 lm from the peak implanted region).

H-implanted region in (0 0 0 1) single crystal SiC irradiated with 0.4 MeV H ions at 650 °C. The original surface spalled off during the irradiation, so the fluence in the region shown in Fig. 15 is less than the nominal specimen exposure fluence of 1  1022 H/m2. The cavity formation was localized within the H-implanted region in a band 125 nm wide. The microstructure in the H-implanted region (gas-filled platelets on (0 0 0 1) basal planes) is qualitatively similar to that observed in He-implanted SiC by other investigators [45,69,86]. Previous work on 4H SiC irradiated with 70 keV H and then annealed at 950 °C reported the threshold fluence for exfoliation decreased slightly with increasing irradiation temperature between room temperature (4.4  1020 H/m2; 5 at.% H at peak) and 600 °C (2.7  1020 H/m2; 3 at.% H) [13], but the as-irradiated cavity microstructure was not characterized in detail. Most of the implanted hydrogen was found to be associated with planar cavities

Fig. 13. Heterogeneous cavity formation in the peak damage region of Al2O3 irradiated with 1 MeV H ions at 650 °C to a fluence of 1.7  1022 H/m2. The original irradiated surface is located to the left of the micrograph.

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

Fig. 15. Over-focused cross-section TEM image of a-SiC implanted with 0.4 MeV H ions at 650 °C showing heterogeneous cavity nucleation on (0 0 0 1) dislocation loops. The cavities appear as dark vertical lines in the micrograph due to their edgeon configuration and overfocus imaging condition. Due to surface exfoliation that occurred sometime during the irradiation, the fluence in the displayed region is less than the nominal specimen exposure fluence of 1  1022 H/m2. The original irradiated surface is located to the left of the micrograph.

in SiC irradiated near room temperature to a peak He concentration of a few at.% and subsequently annealed near 900 °C [15]. An investigation of cavity formation in SiC irradiated at room temperature with hydrogen, deuterium and helium ions (where cavity formation occurred following irradiation-induced amorphization) found that the H and D irradiations required about twice the irradiation fluence to induce cavity formation and blistering compared to He irradiations [75]. Evidence for C–H chemical interactions in the hydrogen isotope irradiated specimens was obtained [75]. Fig. 16 shows the microstructure of MgO near the peak H-implanted region after high-fluence irradiation with 1 MeV H ions at 650 °C. Cavity formation in the implanted hydrogen region was highly heterogeneous and preferentially associated with dislocation loops lying on {0 0 1} and {1 1 0} habit planes (visible as dark lines for the edge-on loops in Fig. 16). The cavity microstructure for 1 MeV H ion irradiated MgO was not examined in detail, but the cavities appeared to occur as platelets coinciding with the loops as opposed to a collection of small spherical cavities. Since hydrogen diffusion in MgO has been reported to be considerably enhanced by the ionizing radiation associated with proton irradiation [88,89], it is possible that some of the implanted hydrogen may have diffused away from the bombarded region during the irradiation. Cavity formation in H-implanted MgAl2O4 was also distinguished by significant spatial heterogeneity. However, the cavity formation did not appear to be associated with dislocation loops. As shown in Fig. 17, the cavities in MgAl2O4 irradiated with H ions at room temperature to a high fluence of 1  1022 H/m2 (2 dpa and 20 at.% H in the peak implant region) exhibited a wavy structure and were aligned approximately parallel to the irradiated sur-

15

Fig. 16. TEM image of MgO near the damage peak after irradiation with 1 MeV H ions at 650 °C to a fluence of 1.7  1022 m 2. The zone axis is near (0 0 1). Large dislocation loops on {0 0 1} and {1 1 0} habit planes are visible.

Fig. 17. Cavity microstructure (underfocused image) of MgAl2O4 after 0.4 MeV H ion irradiation at room temperature to a fluence of 9.5  1020 H/m2 (2 dpa and 20 at.% H in the peak implanted region). The original irradiated surface is located to the left of the micrograph.

face, with no distinct crystallographic relationship with the spinel matrix. The visible cavities were confined to a 250 nm wide band near the peak implantation region. The room temperature cavities were small platelets (5 nm diameter by 1 nm thickness).

16

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

Fig. 18. Large heterogeneous cavities produced in the peak implantation region of MgAl2O4 irradiated with 1 MeV H ions at 650 °C to a fluence of 1  1021 m 2 (0.28 dpa, 1.8 at.% H at peak). The original irradiated surface is located to the left of the micrograph.

As shown in Fig. 18, irradiation of MgAl2O4 with 1 MeV H ions at 650 °C produced a low to moderate density of relatively large cavities that were heterogeneously distributed in the peak implanted region for a fluence of 1  1021 m 2 (0.28 dpa, 1.8 at.% H). The cavities at 650 °C did not exhibit a distinct crystallographic orientation relationship with the spinel matrix, and were much larger and of lower density compared to the room temperature H-irradiated MgAl2O4. The volume-averaged cavity density in the hydrogen-implanted regions of MgAl2O4 at 650 °C (1.5  1021/m3) was more than one order of magnitude lower and the mean cavity size (10 nm) was higher than that in Al2O3 irradiated at comparable conditions. The cavities were slightly elongated in directions parallel to the irradiated surface. Irradiation of MgAl2O4 with 1 MeV H ions at 650 °C to an order of magnitude higher fluence produced matrix cavity growth and extensive cavity formation along grain boundaries. The large three-dimensional matrix cavities were faceted along {1 1 1} planes and were as large as 300 nm diameter. 3.4.1. Threshold H concentrations for visible cavity formation There have been relatively few previous investigations of cavity formation in ceramics associated with H implantation. In general, the implanted hydrogen concentrations required to produce visible cavities were comparable to helium ion irradiation. Further work is needed to investigate the potential role of chemical bonding and the diffusion kinetics for implanted hydrogen. The threshold con-

centration for cavity formation due to H ion irradiation at room temperature was not accurately determined in the present study due to a limited number of investigated fluences. However, hydrogen irradiation at 650oC generally appeared to induce cavity formation at significantly lower H concentrations compared to room temperature irradiation. The threshold fluence to produce visible cavities at 650 °C in the present study ranged from 1  1021 m 2 (2 at.% H peak implanted concentration) for MgAl2O4 to 1  1022 m 2 for Si3N4 (20 at.% H at peak) with the other ceramics lying at intermediate fluences. The threshold fluence to produce visible cavities at room temperature was 5  1021 m 2 (15 at.% H peak implanted concentration) for MgAl2O4. From a comparison of the calculated implanted H profile (Fig. 1) and the observed widths of the cavity bands for the various irradiation conditions (e.g., Figs. 13 and 15–18), the threshold H concentration for cavity formation was found to be 1 at.% in all three oxide ceramics irradiated at 650 °C and >10 at.% for MgAl2O4 irradiated at room temperature. Table 3 compares the threshold H concentration for visible cavity formation from the present study and prior literature results [44,53,77]. Hojou et al. [90] found that formation of cavities in a-SiC at room temperature required a 15-keV H ion fluence of 5  1022 m 2, which was about fifty times higher than the corresponding fluence for low energy He ion irradiation. Muto et al. [75] reported relatively high fluences (1–3  1022 m 2; 100 at.% implanted gas) for inducing visible cavity formation in SiC irradiated at room temperature with either low-energy (12–19 keV) H or He ions, with threshold H concentrations about double the corresponding threshold He concentrations. Bubble formation was preceded by irradiation-induced amorphization in these room temperature irradiations. Two previous reports on 15 keV H2 ion irradiated Al2O3 systematically investigated the effect of irradiation temperature on bubble formation between room temperature and 1000 °C [53,77]. The authors observed spatially uniform production of cavities following irradiation at room temperature and 400 °C, with threshold fluences for cavity formation of 1.2  1021 H/m2 (10 at.% H) and 0.6  1021 H/m2 (5 at.% H), respectively [53,77]. At an irradiation temperature of 650 °C, the threshold fluence for cavity formation decreased further to 1.8  1020 H/m2 (1.5 at.% H) and the cavities were found to preferentially form in the interior of (0 0 0 1) dislocation loops. Some cavities were also observed to be preferentially  0 0} prism plane loops following irradiation associated with {1 1 at 650 °C [77]. As the irradiation temperature was increased to 800 and 1000 °C, heterogeneous cavity formation was detected for fluences as low as 1.2  1020 H/m2 (1 at.% H) [53,77], and the cavities were located at the interior of dislocation loops located  0 0} and {1 1 2  0} prism planes [77]. on (0 0 0 1) basal and {1 1 3.4.2. Evidence for long-range H migration beyond the irradiation zone Fig. 19 shows an example of the grain boundary cavities near and beyond the peak H implantation region in MgAl2O4 irradiated with 1 MeV H ions to high fluence. The grain boundary cavity size ranged from 10 to 150 nm, with even larger cavities observed at grain boundary triple points. Although the cavity formation was most pronounced within the irradiated region, some of the grain boundary cavities extended significantly beyond the ion irradiated region. This suggests that there might be rapid diffusion of hydrogen along grain boundaries in MgAl2O4. Further work on hot pressed or other forms of polycrystalline MgAl2O4 are needed to examine the possible effects of chemical reactions on grain boundary migration and cavity formation. Fig. 20 shows the low-magnification microstructure for MgAl2O4 irradiated with simultaneous beams of 0.4 MeV H and 0.36 MeV He ions at 650 °C. The irradiation produced calculated

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

17

beyond the implanted ion region (2.7 micron depth for the H ions in Fig. 20) suggests that the hydrogen was highly mobile within grain boundaries and may have chemically reacted with the nonspinel grain boundary phase present in the sintered polycrystalline MgAl2O4 specimen. Using a typical bulk diffusion coefficient for hydrogen in single crystal ceramics of DH  1  10 18 m2/s at 650 °C [91], matrix hydrogen diffusion distances of only 0.2 lm would be anticipated for the 8 h ion irradiation. Conversely, the apparent high grain boundary diffusivity suggested by Fig. 20 is consistent with effective diffusion coefficients of DH  1  10 14 m2/s at 650 °C measured for polycrystalline ceramics [91]. 3.5. Blistering mechanisms

Fig. 19. Large grain boundary cavities near or beyond the peak implantation region in MgAl2O4 irradiated with 1 MeV H ions at 650 °C to a fluence of 1.7  1022 m 2 (3.1 dpa, 30 at.% H at peak). A grain boundary triple point is located near the top of the micrograph.

Fig. 20. Cross-section microstructure of MgAl2O4 irradiated at 650 °C with simultaneous dual beams of 0.4 MeV H and 0.36 MeV He at a H/He flux ratio of 10 and a cumulative H fluence of 1  1022 m 2. Pronounced cavity formation is visible in the grain boundary within and beyond the H and He implanted regions.

peak He levels of 5 at.% and 3 dpa at 1 lm depth and peak H levels of 20 at.% and 1.8 dpa at 2.7 lm depth. In addition to the extensive matrix cavity formation associated with the implanted He ions at a depth near 1 lm and the implanted H ions at a depth of 2.7 lm, cavities were observed along grain boundaries in the unirradiated region to depths in excess of 20 lm from the implanted region. The higher magnification photo in upper part of Fig. 20 shows an example of grain boundary cavities observed at a depth of 15 lm. The observation of grain boundary cavities far

The TEM observations indicate that blistering and surface exfoliation in H and He irradiated ceramics at room temperature and 650 °C is due to progressive cracking at the ligatures between highly pressurized gas bubbles that form in the implanted ion region. The blistering is not associated with coarsening of gas bubbles, as proposed in an early gas pressure model [28]. The present observations support some aspects of the lateral compressive stress model [92], which predicts that blistering occurs when internal stresses at the implantation-produced cavities exceed the yield strength. In particular, the strong geometric relationship between cavity platelet habit plane orientation and the specimen surface is an indication of an important role of lateral stress. However, prior experimental studies have found that the power law relationship between the blister radius (r) and the implantation depth (R) is only slightly superlinear [10] (or even sublinear in the case of H-implanted Si [37]), whereas the lateral stress model predicts r  R1.5. Several aspects of the current observations are also generally consistent with the pressurized gas platelet model proposed by Chen et al. [45], where highly pressurized platelets are nucleated by atomic clustering of interstitial He or H between adjoining low-index planes. A similar model based on analysis of the exfoliation kinetics of hydrogen implanted and annealed Si and SiC has measured an activation energy of 3.5 eV for SiC blistering over a wide dose range, and the rate-limiting step was attributed to rupture of lattice bonds at the periphery of the pressurized disk-shaped cavities [15]. The similarity in the blistering threshold gas concentrations for H and He implantations suggests both gaseous species have limited matrix solubility and diffusivity in these ceramic materials for the temperatures investigated in this study (unlike the case for many metals, where the implanted H dose for blistering can be ten times higher than the He dose due to H permeation [26]). In the present study, cavities were typically observed to form on low-index (low surface energy) planes that were oriented with habit planes nearly parallel to the irradiated surface. This geometric alignment of cavities near the peak implanted gas region serves as the key first step for blistering that occurs during extended irradiation. The lowest surface energy planes are {0 0 1} for MgO, {1 1 1} for MgAl2O4, (0 0 0 1) for Al2O3, and (0 0 0 1) for a-SiC [93]. As shown in Fig. 5, He irradiation of polycrystalline Al2O3 at 650 °C produced cavities on (0 0 0 1) habit planes even for crystallographic orientations that were 25 degrees away from the preferred configuration. If suitably oriented low index planes are not present in a particular grain of a polycrystalline irradiated sample, then planar cavity formation may be induced on other higher-index (higher surface energy) planes located approximately parallel to the irradiated surface. Localized stresses associated with the ion implantation geometry play a significant role in the observed cavity habit planes, as previously discussed in Section 3.3.1. The biaxial lateral compressive stress (due to constraint of the swollen irradiated region by

18

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19

the unirradiated substrate) and accompanying tensile out-of-plane stress favors nucleation of vacancy clusters parallel to the irradiated surface due to preferential migration of radiation-induced vacancies to the regions of highest out-of-plane strain near the peak implantation zone [58]. In the present study, it was confirmed there was a strong tendency for cavities in the H and He peak implanted region to be aligned nearly parallel to the irradiated surface (Figs. 4, 6, 8, 15, 17 and 18) although cavity alignment with the irradiated surface was not observed in some cases such as H implanted Al2O3 and MgO at 650 °C (Figs. 13 and 16). In two room temperature irradiation cases (He irradiated Al2O3 and H irradiated MgAl2O4; Figs. 6 and 17), the cavity geometric configuration was found to be approximately parallel to the irradiated surface independent of the specific crystallographic orientation of the polycrystalline grains whereas in other cases cavity formation tended to be preferentially associated with the low index (low surface energy) plane that was most nearly parallel to the irradiated surface. Studies of hydrogen-implanted Si have reported that cavities occur predominantly on (0 0 1) habit planes for [0 0 1] oriented samples, on (1 1 1) habits for [1 1 1] orientations, and on a mixture of {1 1 1} and {1 0 0} habit planes for [1 1 0] orientations [37]. Without consideration of implantation stresses, cavity formation would be expected to occur preferentially on the close-packed {1 1 1} planes that have the lowest surface energy in Si; it is worth noting that [1 1 0] oriented Si has the highest threshold dose for hydrogen blistering [37], due to the distribution of the implanted H into diskshaped cavities on multiple {1 1 1} and {0 0 1} planes.

4. Conclusions In all of the H and He implanted specimens, blistering and surface exfoliation occurred at 650 °C above a critical fluence of 3 and 10  1021 m 2, respectively (10–20 and 20–50 at.% peak implanted gas concentration, respectively). The threshold fluences for blistering and exfoliation exhibited a rather weak dependence on type of ceramic, and decreased only slightly as the temperature was raised from room temperature to 650 °C. The threshold fluences for blistering and surface exfoliation were generally lower for helium irradiation compared to hydrogen irradiation, whereas the threshold implanted H and He atomic concentrations were comparable for inducing either blistering or exfoliation. The blister and exfoliation threshold fluences of the ceramic specimens were similar to or slightly lower than that of metals that have limited H and He solubility. The experimental observations indicate that both lateral stress and gas pressure effects are important for blistering and exfoliation. The threshold concentration of implanted H or He to produce visible cavities was typically 1–5 at.% at room temperature and 650 °C. It is remarkable that cavity formation was in general easily induced in the room temperature irradiations, considering the lack of long-range vacancy mobility in the ceramic materials at this temperature. In most cases, somewhat higher gas concentrations were needed to induce visible cavities at room temperature compared to 650 °C. In the case of SiC, cavity formation occurred after the completion of the crystal-to-amorphous phase transition and required slightly less gas concentrations at room temperature compared to 650 °C. Cavity formation was generally of higher density and more spatially uniform for He compared to H irradiation. In all cases except He-irradiated MgAl2O4 at 650 °C, the matrix cavity formation was limited to the ion-implanted region. For He-irradiated MgAl2O4 at 650 °C, the visible cavities extended from the He-implanted region to the irradiated surface. Long-range diffusion of hydrogen along grain boundaries was observed in MgAl2O4 irradiated at 650 °C. Introduction of additional ionizing radiation during the

He implantation had a relatively minor effect of bubble evolution in irradiated SiC, Al2O3 and MgAl2O4. Three general types of cavity morphology were observed, depending on the material and irradiation conditions: (A) Spatially isolated three-dimensional cavities (either randomly dispersed or heterogeneously nucleated), (B) clusters of small cavities located on low index habit planes, and (C) two-dimensional cavity platelets. The cavities were often associated with dislocation loops, particularly for H ion irradiation. Preferential cavity formation occurred on {001} and {1 1 0} habit planes in MgO, {1 1 1} and  0 0} in Al2O3 and (0 0 0 1) habit {1 1 0} in MgAl2O4, (0 0 0 1) and {1 1 planes in a–SiC. In general, cavity morphologies (B) and (C) were preferentially located on low index (low surface energy) planes that were nearly parallel to the irradiated surface. Room temperature irradiation of polycrystalline MgAl2O4 and Al2O3 with 1 MeV H or He ions created cavities that were preferentially oriented parallel to the irradiated surface (rather than being associated with matrix crystallographic orientations), which may be an indication of substantial ion implantation stresses associated with room temperature irradiation affecting the nucleation of cavities. Acknowledgments The microstructural observations were supported by the Office of Fusion Energy Sciences, U.S. Department of Energy. The ion beam irradiations and analysis of TEM results were performed using funding provided by the Materials Sciences and Engineering Division of the Office of Basic Energy Sciences, U.S. Department of Energy. The author thanks J.W. Jones and A.M. Williams for sample preparation, S.W. Cook and J.D. Hunn for performing the ion irradiations, and Y. Zhang for manuscript review and updated SRIM calculations. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32]

S.J. Zinkle, E.R. Hodgson, J. Nucl. Mater. 191–194 (1992) 58. S. Yamamoto et al., J. Nucl. Mater. 283–287 (2000) 60. S.J. Zinkle, Fusion Eng. Des. 74 (2005) 31. G.P. Pells, J. Nucl. Mater. 155–157 (1988) 67. N. Casal et al., J. Nucl. Mater. 417 (2011) 1316. C. Kinoshita, K. Fukumoto, K. Fukuda, F.A. Garner, G.W. Hollenberg, J. Nucl. Mater. 219 (1995) 143. C. Kinoshita, S.J. Zinkle, J. Nucl. Mater. 233–237 (1996) 100. S.J. Zinkle, S. Kojima, J. Nucl. Mater. 179–181 (1991) 395. K. Hojou, S. Furuno, T. Soga, K. Izui, J. Nucl. Mater. 179–181 (1991) 411. S.K. Das, M. Kaminsky, G. Fenske, J. Appl. Phys. 50 (1979) 3304. K.L. Wilson, Nucl. Fusion special issue: Data compendium for plasma-surface interactions (1984) 85. M. Bruel, Nucl. Instr. Meth. B 108 (1996) 313. R.B. Gregory, T.A. Wetteroth, S.R. Wilson, O.W. Holland, D.K. Thomas, Appl. Phys. Lett. 75 (1999) 2623. T. Hochbauer, A. Misra, M. Nastasi, J.W. Mayer, J. Appl. Phys. 92 (2002) 2335. S.W. Bedell, W.A. Lanford, J. Appl. Phys. 90 (2001) 1138. S.J. Zinkle, C. Kinoshita, J. Nucl. Mater. 251 (1997) 200. S.T. Murphy et al., Solid State Ionics 180 (2009) 1. J. Lefevre, J.M. Costantini, S. Esnouf, G. Petite, J. Appl. Phys. 106 (2009) 083509. M.B. Lewis, W.R. Allen, R.A. Buhl, N.H. Packan, S.W. Cook, L.K. Mansur, Nucl. Instr. Meth. B 43 (1989) 243. F. Gao, W.J. Weber, M.B. Posselt, V. Belko, Phys. Rev. B 69 (2004) 245205. J.F. Ziegler, J.P. Biersak, U. Littmark, The Stopping and Range of Ions in Solids, Pergamon Press, New York, 1985. S.J. Zinkle, C.P. Haltom, L.C. Jenkins, C.K.H. DuBose, J. Electron Microsc. Tech. 19 (1991) 452. S.J. Zinkle, Nucl. Instr. Meth. B 91 (1994) 234. S.J. Zinkle, J. Nucl. Mater. 219 (1995) 113. S.J. Zinkle, L.L. Snead, Nucl. Instr. Meth. B 116 (1996) 92. G.M. McCracken, Rep. Prog. Phys. 38 (1975) 241. W. Primak, J. Appl. Phys. 34 (1963) 3630. W. Primak, J. Luthra, J. Appl. Phys. 37 (1966) 2287. W. Primak, J. Appl. Phys. 48 (1977) 1556. W. Primak, J. Nucl. Mater. 74 (1978) 84. D.M. Gruen, B. Siskund, R.B. Wright, J. Chem. Phys. 65 (1976) 363. C.J. McHargue, G.C. Farlow, M.B. Lewis, J.M. Williams, Nucl. Instr. Meth. B 19 (20) (1987) 809.

S.J. Zinkle / Nuclear Instruments and Methods in Physics Research B 286 (2012) 4–19 [33] L. Romana, G. Fuchs, G. Massouras, P. Thevenard, Radiat. Eff. Defects Solids 115 (1990) 139. [34] E.L. Fleischer, M.G. Norton, Heterogen. Chem. Rev. 3 (1996) 171. [35] S. Furuno et al., Nucl. Instr. Meth. B 127 (128) (1997) 181. [36] R. Fromknecht, J.-P. Hiernaut, H. Matzke, T. Wiss, Nucl. Instr. Meth. B 166–167 (2000) 263. [37] B. Terreault, Phys. Status Solidi A 204 (2007) 2129. [38] H. Schroeder, H. Ullmaier, J. Nucl. Mater. 179–181 (1991) 118. [39] B.N. Singh, H. Trinkaus, J. Nucl. Mater. 186 (1992) 153. [40] H. Trinkaus, B.N. Singh, J. Nucl. Mater. 323 (2003) 229. [41] W.E. Lee, M.L. Jenkins, G.P. Pells, Philos. Mag. A 51 (1985) 639. [42] W.R. Allen, S.J. Zinkle, J. Nucl. Mater. 191–194 (1992) 625. [43] R. Yamada, S.J. Zinkle, G.P. Pells, J. Nucl. Mater. 209 (1994) 191. [44] K. Hojou, S. Furuno, K.N. Kushita, N. Sasajima, K. Izui, Nucl. Instr. Meth. B 141 (1998) 148. [45] J. Chen, P. Jung, H. Trinkaus, Phys. Rev. Lett. 82 (1999) 2709. [46] S. Muto, T. Tanabe, S. Igarashi, Phys. Scr. T108 (2004) 19. [47] M.T. Robinson, J. Nucl. Mater. 216 (1994) 1. [48] A. Mangiarotti et al., Nucl. Instr. Meth. B 580 (2007) 114. [49] S.J. Zinkle, G.P. Pells, J. Nucl. Mater. 253 (1998) 120. [50] L.L. Snead, Y. Katoh, S. Connery, J. Nucl. Mater. 367–370 (2007) 677. [51] J.H. Harding, K.J.W. Atkinson, R.W. Grimes, J. Am. Ceram. Soc. 86 (2003) 554. [52] Y. Katano, H. Ohno, H. Katsuta, J. Nucl. Mater. 155–157 (1988) 366. [53] N. Sasajima, T. Matsui, K. Hojou, H. Otsu, Nucl. Instr. Meth. B 148 (1999) 745. [54] M.A. van Huis et al., Nucl. Instr. Meth. B 216 (2004) 149. [55] J. Chen, Z.Y. He, P. Jung, Acta Mater. 54 (2006) 1607. [56] W.G. Wolfer, F.A. Garner, J. Nucl. Mater. 85&86 (1979) 583. [57] S.J. Zinkle, J. Nucl. Mater. 191–194 (1992) 645. [58] M. Nastasi et al., Appl. Phys. Lett. 86 (2005) 154102. [59] F.W. Clinard Jr., G.F. Hurley, L.W. Hobbs, J. Nucl. Mater. 108&109 (1982) 655. [60] S.J. Zinkle, J. Am. Ceram. Soc. 72 (1989) 1343. [61] S.J. Zinkle, Dislocation loop formation in ion-irradiated polycrystalline spinel and alumina, in: R.E. Stoller, A.S. Kumar, D.S. Gelles, (Eds.), 15th Int. Symp. on Effects of Radiation on Materials, ASTM STP 1125, American Society for Testing and Materials, Philadelphia, 1992, p. 749. [62] K. Fukumoto, C. Kinoshita, F.A. Garner, J. Nucl. Sci. Technol. 32 (1995) 773. [63] L.K. Mansur, W.A. Coghlan, J. Nucl. Mater. 119 (1983) 1. [64] H. Ullmaier, J. Nucl. Mater. 133–134 (1985) 100. [65] S.J. Zinkle, Irradiation spectrum and ionization-induced diffusion effects in ceramics. in: I.M. Robertson, et al., (Eds.), Microstructure Evolution During Irradiation, Materials Research Society, Pittsburgh, 1997, p. 667. [66] K. Yasuda, C. Kinoshita, R. Morisaki, H. Abe, Philos. Mag. A 78 (1998) 583.

19

[67] E.A.C. Neeft, R.P.C. Schram, A. van Veen, F. Labohm, A.V. Fedorov, Nucl. Instr. Meth. B 166–167 (2000) 238. [68] B.J. Kooi, A. van Veen, J.T.M. De Hosson, A.V. Fedorov, F. Labohm, Appl. Phys. Lett. 76 (2000) 1110. [69] K. Nakata, S. Kasahara, S. Shimanuki, Y. Katano, H. Ohno, J. Kuniya, J. Nucl. Mater. 179–181 (1991) 403. [70] J. Chen, P. Jung, H. Trinkaus, Phys. Rev. B 61 (2000) 12923. [71] P. Jung, H. Klein, J. Chen, J. Nucl. Mater. 283–287 (2000) 806. [72] C.H. Zhang, S.E. Donnelly, V.M. Vishnyakov, J.H. Evans, J. Appl. Phys. 94 (2003) 6017. [73] C.H. Zhang, S.E. Donnelly, V.M. Vishnyakov, J.H. Evans, T. Shibayama, Y.M. Sun, Nucl. Instr. Meth. B 218 (2004) 53. [74] S. Leclerc, M.F. Beaufort, A. Declemy, J.F. Barbot, Appl. Phys. Lett. 93 (2008) 122101. [75] S. Muto, T. Tanabe, T. Muruyama, Mater. Trans. 44 (2003) 2599. [76] W.R. Allen, J. Nucl. Mater. 199 (1993) 121. [77] S. Furuno, N. Sasajima, K. Hojou, K. Izui, H. Otsu, T. Matsui, J. Nucl. Mater. 258– 263 (1998) 1817. [78] M. Sasase, Y. Ito, K. Yasuda, R. Ishigami, S. Hatori, A. Yamamoto, Nucl. Instr. Meth. B 209 (2003) 179. [79] S. Leclerc, M.F. Beaufort, A. Declemy, J.F. Barbot, Nucl. Instr. Meth. B 242 (2006) 399. [80] C. Tromas, V. Audurier, S. Leclerc, M.F. Beaufort, A. Declemy, J.F. Barbot, J. Nucl. Mater. 373 (2008) 142. [81] D. Stewart, Y.N. Osetsky, R.E. Stoller, J. Nucl. Mater. 417 (2010) 1110. [82] H. Ullmaier, Nucl. Fusion 24 (1984) 1039. [83] W.J. Weber, N. Yu, L.M. Wang, N.J. Hess, J. Nucl. Mater. 244 (1997) 258. [84] L.L. Snead, S.J. Zinkle, J.C. Hay, M.C. Osborne, Nucl. Instr. Meth. B 141 (1998) 123. [85] M.D. Rechtin, Radiat. Effects 42 (1979) 129. [86] K. Hojou, K. Izui, J. Nucl. Mater. 160 (1988) 147. [87] S.J. Zinkle, V.A. Skuratov, D.T. Hoelzer, Nucl. Instr. Meth. B 191 (2002) 758. [88] Y. Chen, M.M. Abraham, H.T. Tohver, Phys. Rev. Lett. 37 (1976) 1757. [89] R. Gonzalez, Y. Chen, J. Phys.: Condens. Matter 14 (2002) R1143. [90] K. Hojou, S. Furuno, K.N. Kushita, H. Otsu, K. Izui, J. Nucl. Mater. 191–194 (1992) 583. [91] J.D. Fowler, R.A. Causey, D. Chandra, T.S. Elleman, K. Verghese, Tritium diffusion in ceramic CTR materials, in: J.S. Watson, F.W. Wiffen (Eds.), Radiation Effects and Tritium Technology for Fusion Reactors, CONF-750989, USERDA, Gatlinburg, TN, 1976, p. 423. [92] E.P. EerNisse, S.T. Picraux, J. Appl. Phys. 48 (1977) 9. [93] D.M. Saylor et al., J. Am. Ceram. Soc. 87 (2004) 724.