A. Formation of <c>-component loops and threshold incubation dose (TID)

The TID for the formation of <c>-component loops is plotted as a function of irradiation temperature for both WQ and OA samples in Fig. 26, which shows that the incubation dose is strongly temperature and initial microstructure dependent. Furthermore, TID decreases with an increase of temperature.

FIG. 26.

TID for the formation of <c>-component loops plotted as a function of temperature for both WQ and OA samples. A difference in TID can be seen for the two different starting microstructures.

It is well established now that the <c>-component loops are faulted vacancy loops and lie in the basal plane with a Burgers vector of $b = {1 \over 6} \lt 20\overline 2 3 >$ and form after an incubation dose. ^{Reference Griffiths8,Reference Griffiths and Gilbert15} Formation of these loops and the corresponding TID depend on several factors such as the presence of alloying elements, the generation of point defects and their diffusion due to both irradiation and thermal contributions. ^{Reference Griffiths8,Reference De Carlan, Régnard, Griffiths, Gilbon, Bradley and Sabol21,Reference Griffiths, Loretto and Smallman42,Reference Gilbon, Simonot, Garde and Bradley43}

The increase in TID with decrease of temperature can be attributed to the diffusion of alloying elements, which is decreased or even arrested at lower temperature. Interstitial solutes such as Fe play an important role by locally decreasing the stacking fault energy of Zr, which would assist the nucleation of <c>-component loops. ^{Reference Griffiths8} The incubation period before the appearance of <c>-component loops has been linked to a critical interstitial solute concentration by Griffiths et al. ^{Reference Griffiths, Gilbert, Fidleris, Van Swam and Eucken44} The volume of interstitial impurity concentration increases by radiation induced dissolution of alloying elements, either from intermetallic particles or from the β phase. ^{Reference Griffiths8,Reference Griffiths, Gilbert, Fidleris, Van Swam and Eucken44} The WQ microstructure does not contain intermetallic precipitates before irradiation, but Fe is predominantly found in the β phase present along α grain boundaries. From the β phase, Fe is dispersed into the α phase during irradiation, which has been reported to occur in Excel after neutron irradiation. ^{Reference Griffiths8} Assuming that the lowering of stacking fault energy is the mechanism of <c>-component loop nucleation, an early dissolution and redistribution of alloying elements will result in an earlier nucleation of <c>-component loops and hence in a lower TID. Other than SPP, an effect of temperature has been observed on TID.

A variation in the TID for the formation of <c>-component loops has been due to the presence of SPPs in the initial microstructure. The differences seen between OA and WQ samples as shown in Fig. 27 can be explained in terms of the dissolution and redistribution of alloying elements out of the SPPs and into the Zr matrix. In the case of OA samples, alloying elements are concentrated within the α phase in the form of SPPs, whereas in the case of WQ samples, these elements are mostly concentrated in the grain boundary β phase. The presence of alloying elements in the form of solid solution causes a decrease in the stacking fault energy, hence assisting the formation and stability of <c>-component loops. ^{Reference Griffiths8} During irradiation, the pre-existing SPPs undergo radiation induced dissolution and the alloying elements (Nb, Mo, and Fe) are redistributed into the matrix. This redistribution of alloying elements into the matrix has been confirmed after 1 dpa at 400 °C, where irradiation induced reprecipitation of SPPs occurs, as shown in Fig. 13. As the alloying elements are redistributed into the matrix, the stacking fault energy decreases and hence <c>-component loops start to nucleate in the presence of an appropriate vacancy concentration. The dissolution of SPPs can be well explained by the model developed by Nelson et al., ^{Reference Nelson, Hudson and Mazey45} which is based on two facts, ^{Reference Nelson, Hudson and Mazey45} (i) irradiation can enhance solute diffusion within the host matrix and (ii) irradiation can destroy (dissolve) the precipitates.

FIG. 27.

Comparison of mean loop size plotted as a function of dose at all irradiation temperatures in both OA and WQ samples.

The TID for nucleation of <c>-component loop observed in this study is consistent with previous studies of neutron and ion irradiation, where it is attributed to the dissolution and redistribution of alloying elements into the matrix. ^{Reference Griffiths46} The point defects generated as a result of cascade collapse behave in different manners, depending on the initial microstructural environment. Dissolution of alloying elements is a thermally assisted phenomenon and occurs more readily at higher temperatures (400 °C) than at lower irradiation temperatures (100 °C). STEM mappings on OA sample at 400 and 300 °C show that Fe undergoes irradiation induced dissolution and redistribution. At 400 °C, it is reasonable to assume that Fe leaves the SPPs at a lower dose than at 300 °C, resulting in a lower TID. At 100 °C, there is no, or perhaps rather, limited dissolution and redistribution of Fe, which is reflected in the higher TID for the formation of <c>-component loops.

The results of EDS mapping are consistent with the loop nucleation observations in the OA samples irradiated at 400 and 300 °C. The EDS mapping on a SPP shows a reduction in Fe and Mo in the pre-existing SPP, compared to that found in the preirradiation SPP. Fe seems to play the most important role in the nucleation of <c>-component loops, as it dissolves into the matrix at both 400 and 300 °C. Since it is an interstitial diffuser, compared to the substitutional diffusers Nb and Mo, it is reasonable to conclude that Fe will diffuse into the matrix more rapidly at lower doses, compared to the slow diffusing elements Mo and Nb. ^{Reference Griffiths8,Reference Hood30}

The present results show the low temperature formation of <c>-component loops in Excel alloy, which are usually observed only at relatively high temperatures (>300 °C). ^{Reference Griffiths8} Low temperature (100 °C) formation of <c>-component loops has previously been observed in Zr-Excel alloy with a different microstructure (∼60% α and 40% β, solution treated at 855 °C followed by water quenching), which was attributed to the dissolution of SPPs in the α phase and the higher solute content of Excel which stabilized <c>-component loops at 100 °C. ^{Reference Idrees, Yao, Sattari, Kirk and Daymond23} However, the microstructure studied here does not contain such second phase particles. It is difficult to determine why <c>-component loops formed during the irradiation at such a low temperature in the present study, as no EDS microanalysis could be carried out on these samples. It is speculated, however, that Sn, which is present in the form of solid solution, may assist the formation of <c>-component loops, as has been suggested by Griffiths and Gilbert. ^{Reference Griffiths and Gilbert15}

Preferential nucleation of <c>-component loops at the pre-existing <c>-component dislocation structures is observed here and has been reported previously for Zr alloys. ^{Reference Holt, Causey, Christodoulou, Griffiths, Ho, Woo, Bradly and Sabol47,Reference Griffiths, Gilbon, Regnard and Lemaignan48} The <c>-component loops spread out in the basal plane and their multiplication occurs primarily by the helical climb of pre-existing <c>-type dislocations. The acceleration of the elongation rate of Zr pressure tube components with increasing fluence has been attributed to an increase in the density of <c>-component dislocations, mainly by a helical climb multiplication mechanism. Griffiths et al. ^{Reference Griffiths, Holt and Rogerson49} have also linked the accelerated breakaway irradiation growth to the increase in the <c>-component loop density by the helical climb multiplication phenomenon.

B. Mean loop size and defect density

Figures 8(a) and 8(b) show the variation of loop size and areal density as a function of both dose and temperature in WQ samples, respectively. The smallest observed loop size occurs at 400 °C, which is perhaps rather strange. The loop size at 400 °C can be attributed to several factors, however, it is hard to establish a clear conclusion as to the primary contribution. Smaller loop sizes at higher temperature (400 °C) in the WQ sample can be explained by a Sn-cluster to vacancy interaction mechanism. From Figs. 9 and 10, it is evident that Sn undergoes significant radiation induced redistribution, leading to cluster formation at both 400 and 300 °C in WQ samples. Sn forms tiny clusters in the α phase as it diffuses away from the grain boundary, leaving an extended denuded zone on the α–β grain boundary, where <c>-component loops both nucleate and grow. Sn cluster formation shows a significant effect on the growth and morphology of <c>-component loops. It has been explained by Hood ^{Reference Hood30} that the effects of individual Sn atoms present as solute and Sn clusters are opposite to each other. Sn clusters act as preferential trapping and recombination sites for vacancies, lowering the vacancy concentration of the matrix. As the Sn clusters were not observed in preirradiated microstructures, the probability of vacancy trapping or recombination is low at relatively lower doses where <c>-component loops nucleate and grow. In the current study, however, the exact dose at which the Sn forms these clusters is not known. The formation of Sn rich clusters occurs due to irradiation enhanced diffusion and generation of a high density of point defects. Since this is a thermally assisted mechanism, it is reasonable to assume that these clusters will form at a relatively lower dose at higher temperature of 400 °C, as compared to 300 °C. At higher doses where the Sn clustering occurs, vacancies may be trapped on these clusters reducing the vacancy concentration of the matrix, as observed by Buckley and Manthorpe. ^{Reference Buckley and Manthorpe50} Trapping and recombination of vacancies by Sn clusters hence result in decreased loop growth rates and therefore a smaller defect size. In the present case, irradiation induced clustering of Sn is more obvious at 400 °C (Figs. 9 and 10), than at 300 °C (Fig. 16). These tiny clusters of Sn act as strong trapping sites for irradiation induced vacancies, lowering the vacancy concentration. The limited supply of the vacancies to <c>-component loops results in a smaller loop size at 300 °C compared to that seen at 400 °C.

In addition, enhanced vacancy mobility at higher temperatures can play an important role in <c>-component loop morphology, where vacancies are mobile enough to move across larger distances. The annihilation of vacancies can then occur at alternative sinks such as grain boundaries, impeding the loop growth. Jostsons et al. ^{Reference Jostsons, Kelly, Blake, Farrel, Sprague and Kramer14} have discussed the stability of vacancy loops in Zr by the mechanism of thermal emission of vacancies. It has been pointed out that the vacancy loops shrink in size and even can be eliminated after nucleation in the cascade collapse due to vacancy thermal emission at higher irradiation temperatures. The vacancy thermal emission mechanism might have occurred at the early stages of nucleation of loops causing elimination of vacancy type loops after nucleation resulting in lower defect density at 400 °C.

Similarly, Dubinko and Klepikov ^{Reference Dubinko and Klepikov51} have proposed that vacancies can be emitted from defects (vacancy dislocation loops, voids, etc) by the occurrence of collision events in the vicinity of defects; subsequent diffusion (a thermally activated process) of the ejected vacancies under certain conditions (dose, dose rate, temperature, etc) may then result in annihilation of vacancy dislocation loops or voids. ^{Reference Dubinko and Klepikov51} This may explain the limited stability of vacancy type <c>-component dislocation loops at 400 °C.

The large mean loop size at 300 °C and the small loop size at 100 °C can be explained in terms of thermal and irradiation enhanced diffusion of irradiation induced point defects (both vacancy and self-interstitial atoms). Cascade collapse results in the bulk creation of Frenkel pairs of freely migrating point defects and point defect clusters. ^{Reference Dubinko and Klepikov51} The behavior of these point defects depends on their environment, such as temperature, impurities, and extended defects (i.e., point defect sinks such as dislocations and grain boundaries). ^{Reference Azevedo3} At 300 °C, the mobility of point defects is higher than at 100 °C, leading to a large mean free path of point defects. These point defects migrate across large distances in the lattice and consequently annihilate rapidly on sinks such as small point defect clusters, the sample surface, grain boundaries, and dislocations. At 100 °C, however, the mobility of point defects is reduced so that the elimination rate on sinks such as already formed small point defect clusters is decreased. This leads to a higher concentration of point defects in the material; and hence to a high nucleation rate of point defect clusters, which then leads to the higher density of loops at lower temperatures. In the present case, however, there are anomalies in the defect size and density behavior; the general trend in metals and alloys, that loop size increases with an increase of temperature and loop density decreases with an increase in temperature, has not been observed here.

Figure 12 shows the variation of loop size with both dose and temperature in OA samples; a trend of increase in loop size with increase of temperature and dose is seen. Loop size increases with dose increment at all irradiation temperatures. Cascade collapse results in the bulk creation of Frenkel pairs of freely migrating point defects (V/SIA) and point defect clusters. ^{Reference Dubinko and Klepikov51} The subsequent evolution of the microstructure is considered to be dictated by thermally activated processes such as point defect diffusion, ^{Reference Dubinko and Klepikov51} which as noted depends on the microstructure (temperature, impurities, and extended defects, i.e., point defect sinks such as dislocation and grain boundaries). At higher temperatures, the mobility of these point defects is high, leading to a larger mean free path. Hence point defects migrate across large distances and consequently annihilate at sinks, such as previously formed small point defect clusters, the sample surface, grain boundaries, and dislocations; this last sink then leads to a higher growth rate of dislocation loops at higher temperatures. At lower temperatures, however, the mobility of point defects is reduced so that the elimination rate to sinks is decreased. This leads to a slower growth rate of loops and to a higher concentration of point defects in the material. This high concentration of point defects in turn induces a high nucleation rate of point defect clusters, leading to a higher density of loops at lower temperatures.

The lower defect densities at 400 °C as compared to 300 and 100 °C in the OA sample can be attributed to the presence of Mo in the solid solution. The contribution of Mo to the behavior of Zr alloys is not well understood. ^{Reference Williams, Ells and Dixon52,Reference Carpenter and Walters53} Mo rich SPPs in Zr based alloys are thought to be beneficial in improving the corrosion resistance of Zr alloys, and coarsening of these SPPs may weaken corrosion resistance, however, Mo in solid solution, on the contrary, counter-balances the kinetics of microstructural coarsening by diminishing the diffusivity. Lee et al. ^{Reference Lee, Hwang, Yasuda and Kinoshita22} reported that the addition of 0.5 wt% Mo to Zr–1 wt% Nb–1% Sn–0.1% Fe decreased the initial defect nucleation rate and suppressed the growth rate of defects. These observations were attributed to retardation of diffusion by Mo. In our observations, at 400 °C, Mo dissolved completely from the SPPs into the matrix. The dissolved Mo seems to retard the loop nucleation rate, which is reflected in the density data at 400 °C. At lower temperatures, however, Mo does not dissolve completely from the SPPs into the matrix, hence defect densities are found to be higher than at 400 °C.

There is a significant difference in the defect size of <c>-component loops between the WQ and OA samples (Fig. 27), which shows the potential effects of lattice chemistry and distribution of alloying elements. The presence of alloying elements in the form of SPP results in smaller loop sizes. In the case of the OA sample, the radiation damage first dissolves and then redistributes the alloying elements from the SPPs. At 400 °C, there is thermally assisted precipitate dissolution, causing a higher solute concentration in the matrix. Mo reduces the defect nucleation and growth rate of irradiation induced defects in Zr. Also, very fine Sn rich clusters may act as preferential trapping and recombination sites for vacancies, as observed in the WQ samples. An increase in defect density and decrease in loop size by the addition of Nb in Zr have been reported by Buckley and Manthorpe. ^{Reference Buckley and Manthorpe50} It has been reported that due to the addition of substitutional solutes in Zr, the loop growth rate is decreased significantly, which was attributed to the change in the point defect mobility and enhanced vacancy–interstitial recombination due to the presence of Sn, as suggested by Hood. ^{Reference Hood30} In the case of the WQ sample, Sn undergoes irradiation induced precipitation, whereas Mo and Nb remain concentrated in the β phase. On the other hand, in the OA sample, Mo undergoes irradiation induced dissolution from the SPPs into the α phase matrix. As Sn is present in the α phase in both of the microstructures, it is Mo which must create the difference in the loop size between OA and WQ samples.

C. Irradiation induced clustering of alloying elements

STEM-EDS mapping on WQ samples shows the formation of Sn rich clusters at 400 and 300 °C, which was uniformly distributed in the form of solute in the α phase before irradiation.

Abriata et al. ^{Reference Abriata, Bolcich and Arias54} have reported that the Zr–Sn binary system with the Zr₄Sn + β_Zr → α_Zr peritectoid reaction at 962 °C has the potential for precipitation of Zr₄Sn intermetallic phase at room temperature for Sn concentrations above 3 wt%. In a recent reassessment of the Zr–Sn binary system, Toffolon-Masclet et al. ^{Reference Toffolon, Brachet, Servant, Legras, Charquet, Barberis and Mardon55} reported the solubility of Sn in α-Zr as 8.2 and 6.9% at 900 and 600 °C, respectively. Similarly, Carpenter and Walters ^{Reference Carpenter and Walters53} showed that precipitation does not occur in binary Zr–Sn alloys with less than 9% Sn. This implies that Sn precipitation would not occur before irradiation in Excel (3.5 wt% Sn). However, during irradiation, the substitutional Sn diffuses in Zr by a vacancy driven mechanism ^{Reference Hood and Schultz56} ; radiation induced diffusion may cause a supersaturation of Sn in localized areas, resulting in cluster formation.

Radiation induced precipitation of Sn has been reported in other Zr alloys after neutron irradiation. ^{Reference Griffiths8,Reference Woo and Carpenter57} For example, Griffiths ^{Reference Griffiths8} has reported the formation of ZrSn(Fe) precipitates in Excel alloys due to the supply of Fe from the β phase following neutron irradiation at about 416 °C. Similarly, the Fe content in the β phase might have increased in localized areas resulting in weak segregation to clusters. Extended Sn denuded zones at the grain boundary observed at 400 and 300 °C (Figs. 10 and 16) are presumably related to the extensive Sn precipitation in the interior of the α grains.

Nb precipitation was not observed either in α or β phases, rather it stays uniformly distributed in the form of solute in the β phase, despite being a substitutional element like Sn and Mo.

In the β phase, the Mo seems to form small clusters after irradiation at 300 °C. The difference in the clustering behavior of Nb and Mo can be related to their atomic radii, Table I. Although, Nb and Mo are both substitutional elements in Zr, due to the smaller atomic radius of Mo compared to Nb, Mo will have a higher diffusion rate in Zr, resulting here in the formation of clusters. The only possible Mo based intermetallic phase in β Zr is ZrMo₂, according to the phase diagram of Zinkevich and Mattern, ^{Reference Zinkevich and Mattern58} however, in the present study, the crystal structure of these clusters could not be determined.

TABLE I.

Summary of EDS data indicating status of element within the SPP. Atomic radii values adapted from Ref. Reference Hood30.

D. Irradiation induced destabilization of SPP

SPPs which are stable under purely thermal conditions can be destabilized under irradiation. In the current study, three types of irradiation effects have been observed on the SPPs as a function of temperature: (i) irradiation induced dissolution, (ii) irradiation induced precipitation, and (iii) irradiation induced amorphization. STEM-EDS results show that the large pre-existing precipitates undergo irradiation induced dissolution at 400 and 300 °C, whereas irradiation induced amorphization occurs at 100 °C. Irradiation induced precipitate dissolution can be explained by two mechanisms: (i) recoil dissolution and (ii) enhanced defect concentrations in SPPs or matrix under irradiation. Recoil dissolution is caused by the ejection of atoms in the core of the cascade, from a SPP into the matrix, resulting in the mixing of the precipitate and matrix atoms close to the SPP–matrix interface. This process strongly depends on the displacement rate and temperature. Thermodynamically, solute atoms tend to cluster themselves in the SPP, however, if the rate of ejection is higher than the rate of reclustering, the SPP undergoes irradiation induced dissolution. If the atomic flux crossing out of the precipitate is higher than the flux crossing into the precipitate, the precipitate dissolves, if the other way around, it grows. The dissolution rate is directly proportional to the displacement rate, which is higher in the case of ion irradiation (∼10⁻³ dpa/s) than neutron irradiation. ^{Reference Nelson, Hudson and Mazey45,Reference Russell59}

An EDS line scan in the sample irradiated at 400 °C shows that the SPP underwent irradiation induced dissolution. Although the exact change in the composition of the SPP with irradiation is not known, from the STEM mapping on the SPP it is evident that Fe and Mo leave the SPP preferentially, whereas there is still an elevated concentration of Nb present in the SPP. In contrast, at 300 °C, Fe is fully redistributed out from the SPP, whereas Mo and Nb are only partially dissolved from the SPP into the α matrix, as shown in Fig. 18. These observations suggest that there is a difference in elemental dissolution rates. The atomic flux ejected out from a precipitate during the recoil dissolution will depend not only on the displacement rate and temperature, but also on the nature of alloying element itself. The differences between the response of Fe, Mo, and Nb could perhaps arise due to differences in their atomic size. The STEM mapping results and the atomic radii ^{Reference Hood30} of major alloying elements are summarized in Table I. The absence of Fe/Mo and the presence of Nb in the irradiated precipitates as shown in Figs. 14 and 18 can also be attributed to the difference in diffusion rates between Fe, Mo, and Nb in a Zr matrix. ^{Reference Doriot, Gilbon, Bechade, Mathon, Legras and Mardon60} Fe has a small atomic radius and is a fast diffuser, present in the interstitial state in irradiated Zr alloys and hence may diffuse out of intermetallic precipitates during irradiation. ^{Reference Griffiths, Gilbert and Carpenter61}

At high temperature (400 °C), the SPPs dissolve and elements diffuse into the surrounding matrix by irradiation enhanced thermal diffusion of point defects. ^{Reference Doriot, Gilbon, Bechade, Mathon, Legras and Mardon60} Eventually, a supersaturation of solute occurs, resulting in irradiation induced precipitation of SPPs elsewhere. The dose required for solute supersaturation has been recorded as 1 dpa, based on when irradiation induced SPPs are first observed (Fig. 13). Supersaturation of solute occurs in the close vicinity of pre-existing SPPs that are undergoing irradiation induced dissolution. It should be noted that the dose rate seen here is high compared to that seen during typical neutron irradiation. There is thus a strong supply of solute atoms from the SPPs into the matrix and little time for diffusion of solutes further into the grain, which thus causes an early supersaturation of solute near the pre-existing SPPs, compared to other parts of grains. In a detailed study of impurity diffusion in α-Zr, Hood ^{Reference Hood30} demonstrated that smaller atoms diffuse faster as compared to larger atoms, hence the ready dissolution of Fe from the SPPs.

A wealth of literature is present on the behavior of Nb rich SPPs in Zr alloys. A coarsening of Nb rich SPPs in Zr–2.5Nb alloy has been reported by Nuttal and Faulkner ^{Reference Nuttall and Faulkner62} after electron irradiation, which was attributed to the thermally driven coarsening mechanism assisted by irradiation enhanced diffusion. Our observations, however, show dissolution rather than coarsening and can be explained with the theoretical treatment for irradiation stability of solute rich precipitates at equilibrium in a substitutional binary alloy proposed by Russel. ^{Reference Russell59} According to Russel's model, in the presence of excess vacancies produced by irradiation, precipitate growth is encouraged when the atomic volume of the precipitate exceeds that of the matrix, and conversely, dissolution is favored when the precipitate atomic volume is less than the matrix. ^{Reference Russell59,Reference Nuttall and Faulkner62} Kruger and Adamson ^{Reference Kruger and Adamson63} have investigated SPP behavior in several Zr based alloys including an alloy similar to Zr-Excel (named XLL) irradiated to fluences of (0.85–1.5) × 10²⁵ n/m² (E > 1 MeV) at ∼286 °C. Before irradiation, XLL contained spherical, intragranular precipitates that contained Mo, Nb, Fe, and Zr. These SPPs were identified as Mo₂Zr laves phase with a considerable concentration of Nb and Fe. Depletion and diffusion of Fe into the matrix were reported, however, the Nb and Mo concentration ratio remained unchanged. The appearance of precipitates was not appreciably changed by irradiation, making it difficult to determine the dissolution. ^{Reference Kruger and Adamson63} These observations are somewhat consistent with our results, as our TEM analysis did not show significant alteration in the appearance of the SPPs and that the Fe leaves the SPPs at 400 and 300 °C. In regards to Nb and Mo, our observations differ from those reported by Krugar and Adamson, ^{Reference Kruger and Adamson63} since STEM mapping and line scan analysis provide strong evidence of Nb and Mo dissolution. However, Mo seems to diffuse at a faster rate than Nb, which is feasible since it is unlikely that Mo and Nb diffuse at an identical rate ^{Reference Kruger and Adamson63} (experimental data are not available). The compositional gradient observed within the irradiated precipitates, and the reprecipitation of SPPs, makes the dissolution process obvious at 400 °C. In the current investigation, there was no direct evidence on Nb dissolution/redistribution at 300 °C, but minor dissolution of Nb has been inferred in the Zr–2.5Nb alloy after irradiation at 295 °C, ^{Reference Kruger and Adamson63} making it conceivable that dissolution of Nb may have occurred in Excel at 300 °C as well.

At 300 °C after 10 dpa, all of the alloying elements seem to still be present in the precipitates, with the exception of Fe. There may be limited irradiation induced dissolution of precipitates at 300 °C, but due to the relatively low temperature, significant changes in the composition of the precipitates are not seen. Irradiation induced dissolution, diffusion, and reprecipitation are thermally activated processes which are enhanced by irradiation, resulting in the differences between 400 and 300 °C. Hence it is reasonable to assume that at 300 °C, radiation induced dissolution occurs, but not to an extent to cause significant departure of the elements from the SPPs, as evident from the STEM mapping. The rate of dissolution may be slower than at 400 °C due to the more limited diffusion of point defects, which is also consistent with the appearance of <c>-component loops at a higher dose at 300 °C than at 400 °C.

E. Irradiation induced amorphization

In the present study, it is clear that SPPs remain crystalline at intermediate and higher temperature, due to a competition between the irradiation induced disordering and thermally driven reordering of the structure. At higher temperatures, thermal reordering is rapid enough to overcome radiation induced disordering; the SPP remains crystalline, however, the microchemistry changes due to dissolution effects. At lower temperatures, a ballistic interchange of alloying elements and Zr across the SPP–matrix interface may cause the stoichiometry changes in the SPP resulting in amorphization. ^{Reference Onimus, Béchade and Konings64}

Irradiation induced amorphization is dependent on several factors including but not limited to temperature, dose and dose rate, and irradiating ion species. ^{Reference Motta65} In ion irradiation, a high density of energy is deposited on collision cascades accumulating damage that is difficult to anneal with the available thermal assistance. Dose rates in ion irradiation are typically several orders higher than that during neutron irradiation, so annealing through thermal diffusion occurs with difficulty, causing easier amorphization under ion irradiation than under neutron irradiation. The ratio of collisional displacement to thermal jump is the most important factor in controlling the amorphization process. In the present case, ion irradiation at a high dose rate (∼10⁻³ dpa/s) and at a low irradiation temperature (100 °C) causes a significant increase in the ratio of collisional displacements to thermal jumps, making crystalline SPPs thermodynamically less favorable than the amorphous phase. Any deviation from the stoichiometry, the presence of extended defects such as dislocations, and stacking faults can all cause a large increase in free energy which facilitates the crystalline to amorphous transformation. ^{Reference Motta and Olander66}

Irradiation induced amorphization of different types of SPPs has been reported for neutron irradiation cases. ^{Reference Griffiths46,Reference Griffiths, Gilbert and Carpenter61,Reference Yang67,Reference Etoh and Shimada68} Intermetallic particles of Zr(Fe, Cr)₂ and Zr₂(Fe, Ni) in Zircaloy-2 and Zircaloy-4 undergo irradiation induced dissolution at higher temperatures (366–435 °C), and irradiation induced amorphization at low temperatures (<77 °C) during neutron irradiation. ^{Reference Griffiths, Gilbert and Carpenter61} In the present case, however, the SPPs have a different chemical composition compared to previously reported irradiation induced amorphization. Complete amorphization of Fe and Nb rich SPPs at ∼278 °C has been reported by Kruger and Adamson ^{Reference Kruger and Adamson63} in neutron irradiated Zr based alloys. These SPPs were reported to be smaller than the initially existing precipitates and to be partially depleted in Fe, indicating that dissolution and amorphization were occurring simultaneously.