Effect of low dose irradiation of heavy ion on electrochemical corrosion and IASCC behavior of austenitic steel

The impacts of low dose irradiation on the behavior of electrochemical corrosion and irradiation assisted stress corrosion cracking for 321 stainless steel were studied using Fe2+ ion irradiation to simulate neutron radiation damage in primary circuit environment of pressurized water reactor. Low dose irradiation can improve the pitting resistance and reduce the cracking tendency of the alloy in B-Li solution to a certain extent, which was related to the δ phase content on the near-surface of the sample: The higher δ phase content on the near-surface of the 2 dpa irradiated sample was observed by grazing incident X-ray diffraction. In addition, the pits was significantly increased near micro-cracks for the unirradiated sample, indicating that the existence of pits induced the initiation of cracks. The research results provided an important reference for the failure mechanism of irradiation assisted stress corrosion cracking of core components.


Introduction
321 stainless steel (SS) is a derivative of 304SS, which inhibits crack growth and CrC precipitation by adding Ti, thereby obtaining better grain boundary stress corrosion resistance [1] .Because of these advantages, 321SS is generally applied in the secondary circuits [2] and the internals of pressure vessel [3] in WWER (the PWR in Russian), such as WWER-440 from Novovoronezh NPP [4] .However, despite the presence of Ti, stress corrosion cracking (SCC) or even irradiation assisted stress corrosion cracking (IASCC) still occur in 321 SS with the increased service time [5] .In addition, Nuclear Research Institute in Czech Republic [6] have shown that irradiated 321SS is prone to SCC.SCC of austenitic steel caused by high dose irradiation combined with tensile stress and hydrochemical environment has become a key problem of the operation safety for nuclear power.
At present, since neutron irradiation samples are scarce and have high residual radioactivity while ion irradiation samples has low residual radioactivity and short test period, a great deal researches is devoted to understand the influence mechanism of IASCC on austenitic stainless steel such as local deformation [7,8] , irradiation hardening [9][10][11] , irradiation swelling [12,13] , irradiation induced segregation (RIS) [14][15][16][17][18][19][20][21][22][23][24] and irradiation induced precipitation [25][26][27][28] , using ion irradiation to simulate neutron irradiation damage.In addition to irradiation and tensile stress conditions, corrosion/oxidation environment is also the necessary condition to induce SCC/IASCC.For instance, Z. Jiao [7] et al. suggested that local deformation and the pits at grain boundaries jointly initiate IASCC, while local deformation alone would not cause cracking.In paper [29][30][31][32][33][34] , the model of slip-film rupture-oxidation is established to illustrate the SCC crack propagation mechanism of LWR components.Quej-Ake L M [35] et al. further pointed out the mechanism of pitting corrosion in SCC crack initiation by reviewing the influence of electrochemical parameters of carbon steel pipes on internal and external SCC.SUN Baozhuang [36] et al. also believed that stress concentration occurred near pitting pits, which was conducive to the initiation of microcracks.Luo [37] et al. studied the pitting corrosion resistance of 316 NG, 304 NG and 321 SS by electrokinetic anode polarization test and salt spray test.The results showed that 321 stainless steel had the weakest pitting corrosion resistance, which may be the direct cause of SCC/IASCC of 321 stainless steel.In summary, the electrochemical theory validly interpreted the mechanisms of SCC/IASCC processes of metallic materials.
The process of electrochemical corrosion and SCC of 321SS is different from that of other stainless steels because Ti can induce the generation of delta phase [38] .However, the impacts of δ phase on the behavior of electrochemical corrosion and SCC is not completely unified.Liang Liang [39] et al. reported a little δ-ferrite in martensitic-steels was conducive to the formation of corrosion film and the reduction of corrosion rate.P.E.Manning [40] et al. suggested the pits were more prone to initiated in δ/γ phase boundary in Cl -solution.N.D. Tomashov [41] et al. similarly agreed with this view and suggested that the phenomenon was attribute to the Cr precipitation during the non-equilibrium process of the δ phase formation.Other studies [42,43] have attributed this phenomenon to the transition of the δ to the brittle σ phase during heat treatment and the Cr-rich δ phase leading to the surrounding Cr-poor region.In addition, M. Wang [44] et al. indicated that the nickel-based alloy 718 was easily formed a protective inner Cr oxide film at the δ/matrix interface, which may be the cause of the delta phase to relieve stress corrosion cracking.In contrast, Wen-Feng Lu [45] et al. suggested that δ phase were distributed in a consecutive network if the samples contain high δ-ferrite content, providing a consecutive SCC-growth path, which contributed to the sensitivity of cracks.Previous studies on the corrosion resistance of δ phase mainly focused on alloys with high δ-phase content, while there is less data on alloys with low δ-phase content, such as 321 SS.It is necessary to understand the effect of δ phase on electrochemical and SCC behavior under radiation conditions because it is common for nuclear grade austenitic SSs to contain a small amount of δ-ferrite.
The irradiation dose of 1~3 dpa is the initial dose of IASCC for SSs in PWRs environment.However, there are few reports on the electrochemical corrosion and IASCC of low dose heavy ion irradiated austenitic SS in simulated primary circuit solution environment of PWRs.Although Jyoti Gupta et al. [46] reported that the crack number density of 304L irradiated to 3 dpa using Fe 5+ was greater than that of the sample irradiated to 7 dpa, they did not conduct further analysis.Thus, whether low dose heavy ion irradiation affects the electrochemical corrosion and IASCC behavior of alloy is still unclear.In this paper, 321 SS was irradiated to 2 dpa with Fe 2+ ion at 450 ℃ to simulate neutron radiation damage at 320 ℃, comparing with the unirradiated sample.The electrochemical experiment was conducted at 25 ℃ while slow strain rate tensile (SSRT) test was carried out at 320 ℃ in a mixed electrolyte solution of 2.2 ppm Li and 1200 ppm B (the solution in the primary circuit of the PWR).The potentiodynamic polarization (PP) curves, AC electrochemical impedance spectroscopy (EIS) datas and IASCC sensitivity datas for the samples were obtained.The pitting, IASCC behavior and their relationship of 321 SS were studied.The result of this paper may provide reference for the electrochemical behavior and the IASCC mechanism of austenitic SS irradiated to low dose under a simulated primary-coolant environment.

Material
The material in this study was commercial grade 06Cr18Ni11Ti steel (the substitute of 321 SS in Russia).The chemical composition was shown in Table 1.The 321 SS was in the pipe form with a inside diameter of 57 mm and a outer diameter of 64 mm.The samples with the dimensions of 5×5×2 mm 3 were cut from the as received material for electrochemical tests.At the same time, the flake tensile samples was cut for SSRT test.The specific size (millimeter scale) is shown in Figure 1.The material was treated with solution annealing.The metallographic structure of the material shows typical austenite structures, as shown in Figure 2. The average grain size was about 22 μm and a small amount of ferrite phases were distributed in the austenitic matrix.

Irradiations test
Prior to irradiation, all the specimen surfaces were polished by silicon carbide sandpaper up to 5000#.Then the samples were electrochemically polished with a mixture of 95% C2H5OH and 5% HClO4 at -30 ℃ to eliminate the scratches and stress layer produced during mechanical polish.In additon, the tensile sample needs to be polished on both sides, because we want to count the crack information of the irradiated and unirradiated surfaces, but considering the preparation process of the electrochemical sample, it only needs to be polished on one side, and the two samples are used to distinguish the irradiation dose.After irradiation until the beginning of the next experiment, the specimens were stored with a high elastic capsule and a vacuum cabinet to prevent surface oxidation.Irradiation experiments were carried out using 2.5 MeV Fe 2+ ions via 1.7 MV tandem accelerator in the Accelerator Laboratory of Wuhan University [47] .The irradiation temperature was monitored with a K-type thermocouple which was contacting a corner of the specimen.Since the dose rate of ion irradiation was much higher than that of neutron irradiation (~10 -4 dpa/s for iron ions in comparison to ~10 -8 dpa/s for neutrons), Fe 2+ radiation temperature of 450 ℃ was used to compensate for this difference, that was, the microstructure information was similar to that obtained by neutron irradiation at 320 ℃ [46] .The half-peak fluence of the irradiated alloy was approximately 3.08×10 15 ions/cm 2 , corresponding to the half-peak damage dose of 2 dpa.The damage depth of 2.5 MeV Fe 2+ ions in 321SS was about 1.3 μm. Figure 3 showed the damage distribution calculated by using the SRIM-2011 software [48] .

SSRT test
To evaluate the cracking sensitivity of the samples before and after irradiation, SSRT tests were carried out with a strain rate of 5E-7 s -1 in simulated primary water environment of PWR using the CFS-30 tensile testing device with autoclave and water circuit, as shown in Figure 4. Tests were terminated when the sample attained 4% plastic strain and a total time of the test was approximately 2 days.The parameters of simulated primary water used for the test were 30 cc/kg H2 STP, 1200 ppm B and 2.2 ppm Li.In addition, the temperature and the pressure was maintained at 340 ℃ and 15 MPa respectively during the test in an autoclave (3.6 L capacity).The environmental conditions should be allowed to stabilize for 12 hours at least prior to conduct the target straining and the load was pre-stretched to 0.9 with a strain rate of 5E-6 s -1 in order to more clearly statistical cracks.The displacement sensor LVDT (Linear Variable Displacement Transducer) locating on the traction line of the autoclave was used to measure displacements.The computerized data acquisition system collected and recorded load and displacement datas in the interval of 10 s.

Electrochemical tests
The corrosion resistances of specimens were tested by CS350H electrochemical measurement system, as shown in Figure 5.The epoxy resin was used to embedded the samples after the insulated copper wires conncected the specimens to keep only the tested surface areas exposed.A classical three-electrode cell was use to carry out the electrochemical measurements, in which the platinum foil as the counter electrode, the testing samples as the working electrodes and a saturated calomel electrode (SCE, used in salt solution) was chosen as the reference electrode.The potentials were calibrated to standard hydrogen electrode (SHE).Due to the existance oftrace amounts of surface contaminants during sample preparation, the immersion time before the open circuit potential (OCP) test was prolonged from 1h to 4 h to obtain a stable OCP, and then the OCP was tested for 0.5 hour.Choose OCP as the initial potential for the AC impedance test in order to reach equilibrium at the interface of electrode/electrolyte, and the 5 mV of amplitude potential was set.The electrochemical impedance spectroscopy (EIS) were obtained from the frequency scan range of 10 mHz to 100 kHz.The PP tests were performed from a scanning voltage (with respect to the OCP) range of -0.5 V to 1.5 V with a scanning rate of 0.5 mV/s.All the electrochemical tests were conducted in a mixture of 1200 ppm B 3+ and 2.2 ppm Li + at 25 ℃.All of the electrochemical tests on the unirradiated specimen had been carried out in same condition for at least 3 times while the EIS of the irradiated sample was tested 3 times at least aimed to obtain reliable datas.

Microstructural analysis
The composition of phase for tested samples was analyzed by grazing incident X-ray diffraction (GIXRD) using X'pert Powder X-ray diffractometer.The electrons emitted from the cathode of the light pipe were gathered into an electron beam by the focusing cover.The electron beam obtained high speed under the high voltage field and bombarded the anode copper target.The energy of the electron beam was finally converted into heat and X-rays, which was monochromated to a wavelength of 0.154 nm.The X-ray scanning range was from 20 to 90 degree and the resolution was 0.05 degree.In order to correspond to the microstructure at irradiation damage region, the incident angle of X-ray was chosen as 4 degree by the penetration depth equation: 0 = 1/ (where is the absorption factor and ρ is the material density).
A Phenom XL modeled scanning electron microscope (SEM) was utilized to observe the surface morphologies after electrochemical corrosion and count crack information after SSRT tests for the irradiated and unirradiated samples.A range of 0.01 mm 2 was randomly selected in the tensile area of the samples to conduct the statistics on the number of cracks and number density to characterize the IASCC sensitivity.A total of 5 areas of each sample were selected for crack information statistics and then the average value was taken so as to improve the reliability of the statistical results.

Statistical results of crack information
Figure 6 and 7. showed the information of cracks in irradiated and unirradiated samples.As a first result, the number and density of cracks for the irradiated sample was lower than those in unirradiated sample.For the unirradiated sample, the crack number and average number density were 93 and 108 cracks/mm 2 , respectively, while for the irradiated sample, the values were 72 and 83 cracks/mm 2 , respectively.In addition, although the number of cracks in the unirradiated sample was higher, the crack length was lower than that of the irradiated sample on the whole, and the average crack length was 5.82 μm, while 8.33 μm in the unirradiated sample.

Electrochemical test results
Figure 8. showed the OCP curves of the unirradiated and irradiated specimens.For the unirradiated specimen, a steady degradation trend with immerse time for OCP was observed, and the selected point of OCP was -51.2 mV.While for the irradiated specimens, the slight OCP increased with little variation in the curve can be found and the selected points of OCP were 22.6 mV.In addition, the OCP of irradiated sample was always higher than that of unirradiated sample.Figure 10 showed the equivalent electrical circuits (EEC) model of the EIS for the 321 SS samples according to the previous studies in electrolytes [4] .When considering the low-frequency region as a semicircle, the EEC adopted the film resistance model, as shown in Figure 10a, while a straight line corresponded to the diffusion impedance model, as shown in Figure 10b.In these figures, the Rs is the B-Li solution resistance; the Rct is the resistance for charge transfer; the Rf is the passivating film resistance on the alloy surface and the Zw is the Warburg resistance.The CPE is the constant phase angle element replacing the ideal capacitance due to impurities, grain boundaries, dislocation and other factors.At high frequencies, the Rct and charging/discharging of the first CPE1 assesses the adsorption/desorption trend of anions on the metal surface in response to weak AC potential perturbations.At low frequencies, the Rf and the second CPE2 reflects the electrochemical behavior of the passivating film for the film resistance model, while the Zw represents the spread of metal ions at the alloy/solution interface for the diffusion impedance model.Table 2 listed the values of each EEC element fitted by the EIS for the irradiated and unirradiated samples.Compared to other electrical elements, the value of Rs was very small in Table 2, indicating that the influence of solution resistance on the experimental process can be ignored.The similar Rs values of two samples indicated that the samples were tested in essentially the same testing environment.Compared with Rf, the Rct values were much smaller.For passivated films, Rp can be calculated by Rp= Rct + Rf, in this case, Rp = Rf.
Therefore, in the film resistance model, the Rp changed with dose as follows: 2 dpa > 0 dpa, which was consistent with the variation trend of the Zw in the diffusion impedance model.

Table 2
Quantitative datas fitted with equivalent circuits from EIS. .Figure 11.showed the PP curves of 321 SS samples.The active/passive behavior was observed in the unirradiated sample.The first "active nose" was formed at 0.139 μA/cm 2 which was its passive current density (Ipass) and then the curve moved towards the passive region.Afterwards, The curve shifted towards the active region and then the second "active nose" was formed at 2.097 μA/cm 2 which was re-passivated current density.After that, the curve again shifted towards the passive region and then the passivation film was ruptured at the break down potential (Eb) of 1056.6 mV.After the Eb, the current density increaseed rapidly.It was important to note that the PP curve of the sample irradiated to 2 dpa did not show any passivation or pitting behavior, and there was also no sudden increase in the current density.Table 3 listed the calculated corrosion kinetics parameters such as Tafel slopes of the anode and cathode (βa and βc) and corrosion current density (Icorr) from the PP curves according to tafel extrapolation.Along with the dose increased from 0 dpa to 2 dpa, the anodic slope (βa) decreased from 690.88 mV to 238.25 mV, while the cathode slope (βc) increased after irradiation.The value of Icorr varied with the change of irradiation dose.The irradiated sample had the lowest value of 0.003 μA/cm 2 and the unirradiated sample had a higher value of 0.067 μA/cm 2 .In addition, the polarization resistance (Rp) was obtained by means of the Stern-Geary equation, in: The highest value of the Rp = 15093 kohm/cm 2 was achieved for the irradiated sample.The second was the unirradiated sample with a value of 827 kohm/cm 2 .The variation trend of Icorr with dose was just opposite to the Rp.Table 3 Kinetic parameters from PP curves.

Microscopic observation results
Figure 12. showed the GIXRD diagrams of the samples irradiated to 0 dpa and 2 dpa.Three face-centered-cubic (fcc) austenitic diffraction peaks of γ(111), γ(200) and γ(220) was observed at two-theta angles of 43.58°, 50.79°and 74.70°, respectively.In addition, the body-centered-cubic (bcc) ferrite diffraction peak of δ(110) was also observed at 44.48°.When the irradiation dose reached 2 dpa, the diffraction peak intensity of the δ phase was the highest.Figure 13.showed the SEM images of the surface morphology of the samples after SSRT test.The first result was that a large number cracks appeared on the irradiated and unnirradiated surface.Then the information of cracks in two samples had carried on the detailed statistics, as shown in figure 6 and 7.The second result was that there were a few pits on the irradiated surface, in which micro-cracks initiated and growed (Figure 13a), while no pits were found in unirradiated sample except cracks (Figure 13b).Figure 14 showed the   Figure 15.showed the SEM micrographs of unirradiated and irradiated 321 SS specimens after electrochemical testing.On the one hand, the unirradiated sample showed large size pits on the SS surface.On the other hand, the material was only slightly corroded and the surface was depressed downward without forming pits after irradiation.

IASCC sensitivity
For neutron irradiated samples, the percentage of intergranular fracture (IG %) and section elongation are usually used to characterize IASCC sensitivity.For ion irradiated samples, the information of cracks after SSRT test is usually used to characterize IASCC sensitivity because the ion damage layer is very shallow.In this paper, the number and the number density of cracks were used to characterize the IASCC sensitivity, as shown in Figure 6.The number and density of cracks in 2 dpa sample was lower than that in 0 dpa sample, indicating that the sample irradiated to 2 dpa was less IASCC sensitive, that is, more IASCC resistant.At the same time, the higher crack number also resulted in a lower average crack length of the unirradiated sample, as shown in Figure 7.

The electrochemical corrosion behavior
OCP reflected the corrosion tendency of the material surface, which was affected by the microstructure (element composition and defects, etc).The δ phase content was the highest at a low dose (2 dpa), as shown in Figure 12, which reduced the corrosion tendency of the sample via providing a stable source of Cr, leading to a positive shift in OCP with time and the higher OCP (Figure 8).The unirradiated sample had lower δ phase content (Figure 12), which led to difficulties in the formation of passivation film at the initial stage of the test.Thus, the OCP shifted towards negative potential (Figure 8).In addition, the OCP reported in this paper are 50-100 mV higher than those reported by Luo [49] et al. and Mohsin Rafique [50] et al., the possible reasons were as follows: the corrosion resistance of 321 SS was superior compared with 304 SS, which was attributed to the increased δ phase content due to the addition of Ti elements; And the prolonged immersion prior to the electrochemical test resulted in a slow oxidation of the sample in the electrolyte, which increased the OCP; Moreover, the electrolyte used in this experiment was similar to the water environment in the primary circuit of PWRs, which did not contain aggressive ions such as Cl -etc compared with the above literature [49,50] .
The Nyquist graph were plotted using the real and imaginary impedance that revealed all kinds of electrochemical mechanisms arising on the surface of SS in response to the weak AC.If the image of low frequency region was viewed as a semicircle in the Nyquist graph (Figure 9), the semicircular radius corresponding to the real part of the impedance indicated different resistance values (including passivation film and charge transfer) and the larger the radius, the larger the Rf.The incomplete semicircles of both 0 dpa and 2 dpa samples were attributed to the non-ideal capacitance, which demonstrated that the surface of the samples did not form a completely ideal double-layer structure.If the image of the low frequency region was viewed as a straight line, the slope of lines represented the impedance caused by the diffusion of metal cations at the electrode/solution interface, that was, the warburg resistance (Zw), and the higher the slope, the greater the Zw, indicating that the substance transfer was more difficult.According to the above information, the corrosion resistance of materials reflected in Nyquist diagram was 2 dpa > 0 dpa.
All samples' Nyquist graph were divided into two parts (Figure 9), which showed that there were two different mechanisms occured during test.The EEC model can assess the pitting resistance of the samples (Figure 10), where the Rs represents the ability of the solution into the alloy substrate.The Rct represents the ability of the charge transfer through the B-Li solution and the metal-solution interface of the electrode during the electrode process under the electrochemical corrosion potential (Ecorr) and it can be affect by the surface nature of the matrix and the type of ions present in the solution.The Rp calculated in EEC of the film resistance model was also used to characterize the corrosion resistance, that is, the higher the value, the better the pitting resistance, and the conclusion about the corrosion resistance was also as follows: 2 dpa > 0 dpa (Table 2), which had been confirmed by Nyquist diagram.In addition, the Zw was also used to evaluate the pitting resistance of 321 SS in the EEC of diffusion impedance model and obtained similar conclusions with the film impedance model on the corrosion resistance (Figure 10).
The ranking of the pitting resistance of the different samples corresponding to their irradiation doses was: 2 dpa > 0 dpa, according to the Icorr and the Rp values obtained by PP curve, as shown in Table 3.The βa values decreased and the βc values increased after the irradiations (Table 3), suggesting that the anodic reactions increased and the cathodic reactions decreased on the irradiated sample surfaces, that is, the formation rate of passivated films was increased.In other words, the corrosion resistance was improved, which was consist with the conclusion obtained from the EIS test.In addition, the surface morphology of irradiated and unirradiated samples after electrochemical corrosion also confirmed the above conclusions about pitting resistance.

Reason analysis of difference in corrosion resistance
The GIXRD pattern showed that the radiation dose mainly affects the δ phase content in the 321 SS, and low dose irradiation processes with heavy ion likely accelerate the formation of the δ phase.The variation trend of δ phase in Figure 12 was consistent with the pitting corrosion resistance level reflected in Figure 15 and the IASCC resistance level reflected in Figure 6, that is, the 2 dpa sample containing the highest δ content had the best pitting and IASCC resistance.One possible reason was that the Cr-rich δ phase promoted the formation of Cr2O3 [33,39,44,[51][52][53] .The other reason was that the Ti element in the δ phase preferentially combined with C to form TiC to avoid the precipitation of CrC [1] , which further improved the pitting resistance.Therefore, in order to further analyze the influence mechanism of δ effect to IASCC and pitting corrosion, the atomic resolution TEM studies will be beneficial.
Figure 14.showed that numerous pits were distributed near the crack.The formation of the pits introduced stress concentration at its bottom, and micro-cracks will occur under the action of external tensile stress.When the tensile stress was large enough so as to reach a threshold stress, micro cracks merged, so as to induce crack initiation.Thus, the experimental phenomenon of higher IASCC resistance of the irradiated sample may be attributed to the improved pitting resistance of 321 SS by low dose irradiation using heavy ion.

Conclusions
In the current experiment, electrochemical tests, SSRT tests and microscopic analysis of 321 SS have been performed before and after irradiation, the main conclusions were as follows: (1) The pitting resistance and IASCC resistance of sample irradiated to 2 dpa were better than that of the unirradiated sample, indicating that low dose irradiation using heavy ion may improve the corrosion properties of 321SS.In addition, it is necessary to carry out heavy ion irradiation experiment at high dose to further understand the dose effect on corrosion properties.
(2) The δ phase content was consistent with the pitting resistance of the samples, indicating that the δ phase may promote the pitting resistance of 321 SS by accelerating the formation of Cr oxide film on the surface and reducing the precipitation of CrC.
(3) Numerous pits were found near some cracks, indicating that the pits may be one of the main factors of IASCC initiation, which further indicated that low dose heavy ion irradiation can improve the IASCC resistance of 321 SS by improving its pitting resistance.

Figure 1 .
Figure 1.(a) small sample for electrochemical test, (b) tensile sample for the SSRT test.

Figure 4 .
Figure 4.The Cor-Force tensile testing device from Shanghai for SSRT testing.

Figure 6 .
Figure 6.Number and density of cracks in irradiated and unirradiated samples.

Figure 7 .
Figure 7.The crack length distribution in irradiated and unirradiated samples.

Figure 8 .
Figure 8.The OCP curves of unirradiated and irradiated 321 SS.

Figure 9 .
Figure 9. showed the Nyquist graphs for the unirradiated and irradiated 321 SS specimens.The Nyquist plots of all samples consisted of two parts, the image in the high frequency region corresponding to the left of the figure showed a semicircle while the image in the low frequency region corresponding to the right of the figure showed a semicircle trend or a straight line (may be attributed to the large radius of the semicircle).On the one hand, if the low frequency region curve as a semicircle, the sample irradiated to 2 dpa had the largest semicircle.On the other hand, if the low-frequency curve was viewed as a straight line, the sample irradiated to 2 dpa had the highest slope (tangent line).

Figure 10 .
Figure 10. the EEC models obtained from the EIS spectra (a) the film impedance model, (b)the diffusion impedance model.
pits near cracks on the surface of unirradiated sample.A large number of pits created more opportunities for the initiation of cracks.

Figure 14 .
Figure 14.The pits near the cracks for the unirradiated sample.