3.1. Irradiation-Induced Growth and Swelling
Irradiated polycrystalline α-U was initially characterized macroscopically in the late 1940s and early 1950s, and samples were observed to grow in length and/or increase in volume (swell) at a rapid rate when irradiated. Understanding the dimensional instability arising from radiation is important for reactor component design. The early results considered the sample deformation to be a function of material fabrication, working temperature and thermal history. Some studies [
32,
33] revealed that the swelling rate of α-U could be reduced with small additions of low solubility alloying elements, especially silicon, aluminum or carbon. It was also observed that thermal cycling produces damage in unirradiated uranium roughly like the radiation effects in appearance. Pugh [
34] theorized that the mechanism behind irradiation-induced elongation was the same as thermal cycling. Though it is now known that there is apparently no relationship, at that stage it was hypothesized that thermal cycling could be a stand-in for irradiation. In polycrystalline α-U, enormous irradiation-induced elongation was observed especially at high burnup. Note that the macroscopic growth in polycrystalline samples is not the same as the volume-conservative growth in single crystals, though the phenomena are related. We are careful to specify polycrystalline or single crystalline sample states in this discussion. The first detailed analysis of polycrystalline sample growth was presented by Kittel and Paine [
35] in 1958, on the irradiation growth of uranium that had been processed with different conditions. The irradiation was conducted at the Material Test Reactor (MTR) with irradiation temperatures up to 493 K and burnups up to 1.82 at% (no details were provided on sample purity). This study indicated that irradiation growth is a strong function of burnup. Around the same time, Paine and Kittel [
36] measured the volume-conservative irradiation growth rate with respect to burnup for single crystal α-U specimens, though the irradiation temperature is not well known. This work determined that single crystal specimens grow in the [010] direction, shrink in the [100] direction, and remain the same in the [001] direction, and that the overall volume of the crystal is conserved during irradiation growth. The discovery of irradiation growth in single crystal uranium immediately leads to a hypothesis for the multiple observed irradiation-induced deformation behaviors in polycrystalline material and the variation in behavior depending on the sample processing. The collective deformation behavior of multiple individual grains behaving anisotropically should be influenced by grain size and grain texture.
Initial studies motivated by macroscopic engineering concerns were succeeded by more scientific endeavors to understand the physics of the irradiation growth behavior. During 1952–1962, three different mechanisms were postulated. The first mechanism, proposed by Pugh [
34], combined the insights from Cahn [
37] on the observed twin and slip systems and the idea of local thermal expansion induced by fission spikes to propose that irradiation growth resulted from irreversible plastic deformation in the heated region surrounding a fission track. Interestingly, Pugh’s theoretical predictions on anisotropic dimensional change were experimentally observed in most of the single crystal swelling studies [
38]. The second mechanism, proposed by Seigle and Opinsky [
39], hypothesized that the lattice imperfections generated in α-U during irradiation could anisostropically diffuse to and be eliminated at grain boundaries and free surfaces. In this case, the interstitial atoms would add to the lattice, producing an extension, and vacancies would subtract from the lattice, producing a contraction of the grain at its boundaries. If sufficient diffusion occurs and the flux remains unbalanced, macroscopic growth could be observed. Using simplified considerations of the interstitial positions in the α-U lattice and the possible diffusion routes, they theorized that the maximum interstitial migration would be along the [010] direction and the maximum vacancy migration would be along the [100] direction, implying lattice growth along [010] and lattice shrinkage along [100]. This is qualitatively the same as the observed single crystal deformation, but the actual anisotropic diffusion behavior of interstitials and vacancies separately and their interacting behavior in a crystal remains to be determined to confirm or reject this mechanism. The third mechanism, proposed by Buckley [
40], hypothesized that the irradiation-induced elongation and contraction of uranium crystals are the result of the anisotropic condensation of vacancy and interstitial defects in planar clusters within the volume associated with a U-235 fission event. Based on this hypothesis, planar clusters of interstitial atoms increase the dimension of a crystal in the [010] direction by creating extra (010) planes, and planar clusters of vacancies erode existing (100) planes, causing contraction in the [100] direction. Around the same time, Hudson et al. [
41] performed the first thin film transmission electron microscopic measurements and observed the formation of dislocation loops in α-U after neutron irradiation.
To date, the details of irradiation growth in α-U remain unclear, including the exact mechanism or mechanisms of irradiation growth in single crystals, the single-crystal irradiation growth behavior at elevated temperatures, and the impact of interphase boundaries, grain boundaries and solid solution impurities. Loomis and Gerber [
42] carried out a detailed investigation and studied the irradiation-induced dimensional change in single crystals and lineage crystals (crystals with low-angle grain boundaries) over a wide range of temperature (shown in
Figure 6). Note that in
Figure 6, the growth coefficient is defined as the ratio of change in length to the initial length per fraction of U fission atoms. They also reported the markedly enhanced dimensional changes in the [010] and [100] directions below the irradiation temperature of 50 K. The enhanced dimensional change was hypothesized to be the result of increased point-defect mobility due to the CDW-state observed below 43 K. However, it remains unclear whether irradiation-induced crystal growth completely stops at high temperature. Hudson observed that the presumed dislocation loops were randomly dispersed at low temperature but showed a very marked coplanar distribution at high temperatures (>623 K). The advanced characterization of α-zirconium, which also exhibits anisotropic growth, has suggested that defect clustering and interstitial and vacancy loop formation is most likely in that material [
43]. They proposed this is to be a sequential result of the mechanism of movement into rows by glide, as the ease of deformation by slip increases with temperature. Thus, it appears likely that the hypothesis of irradiation growth resulting from vacancy and interstitial clustering on preferential planes is correct.
By 1958, it was not yet clear whether the irradiation-induced dimensional change in polycrystalline material was a result of irradiation growth due to induced lattice defects, or whether it is swelling induced due to fission gases. In fact, in the 1950s, irradiation growth and swelling were not understood as separate mechanisms, and the observed changes in α-U were simply thought of as irradiation-induced dimensional changes. Later it was found that the swelling was responsible for majority of the dimensional change. At the time, there was a hypothesis that dimensional changes resulting from lattice defects would anneal out, as well. Kittel and Paine [
35] suggested that the irradiation-induced dimensional changes under 573 K were not due to swelling induced by fission gases, and further macroscale studies were carried out at different irradiation temperatures and burnups to elucidate the answer. A first meta-analysis on irradiation-induced swelling above 673 K was reported by Pugh [
44] in 1961. Pugh plotted the irradiation swelling versus burnup and observed significant spread in the data. It was presumed that this spread was due to differences in the sample geometry, irradiation conditions, and potentially other factors. Certain samples exhibited grain boundary cracking or catastrophic “breakaway” swelling. A breakthrough was made on the effect of irradiation temperature and α-U swelling when the data from Pugh’s analysis were re-analyzed by Granata and Saraceno [
45]. This analysis included the effect of irradiation temperature, the data exhibited a peak in swelling at intermediate temperatures. The focus then shifted towards capturing the dependence of swelling on irradiation temperature, which was previously only studied with respect to the burnup. This new understanding of the effect of temperature on swelling lead to additional careful microstructural investigations of irradiation effects on α-U.
The most comprehensive and systematic study on the effect of irradiation and temperature on α-U microstructure evolution was performed by Leggett et al. [
46]. In this work, irradiations of high-purity polycrystalline α-U (<150 ppm of impurity by weight) were performed over the range of 573–923 K and over a burnup range of 0.03–0.4 at% at a fast neutron flux up to 2 × 10
13 n/cm
2. Qualitatively different microstructures were observed depending on the irradiation temperature, potentially indicating that several different mechanisms contribute to irradiation damage in the material. The observed microscopic images at different irradiation temperatures are shown in
Figure 7 compared against the unirradiated microstructure (
Figure 7a). Below 623 K, a severely worked microstructure indicating plastic flow in the material was observed (
Figure 7b), though no major volume change was measured. Recent modeling indicates that stresses high enough to cause plastic deformation or void formation can result in polycrystalline α-U with extremely low levels of burnup [
47]. In the range of 673–773 K (
Figure 7c), a similar worked microstructure was again observed, but jagged voids on the order of 1 micron also appeared, apparently the result of grain boundary tearing. Recent modeling demonstrates that this is likely due to the weaking of grain boundary at high temperatures [
13]. In the range of 773–873 K (
Figure 7d,e), they observed a few deformation twins and crystallographically aligned pores on the order of few microns in size. However, the grain boundaries were essentially free of pores. Above 873 K, tiny spherical pores (again in the order of 1 micron) were observed throughout the material and more dominantly at the grain boundaries (
Figure 7f).
Figure 8 summarizes the complex irradiation swelling behaviors as a function of temperature.
Around the same time, Angerman and Caskey [
48] reported microstructural changes in α-U over the range of 403–823 K and at burnups of 0.05 and 0.77 at%. The reactor conditions and neutron flux were not provided. Overall, their observations were very similar to that of Leggett et al. though this study focused on a somewhat lower temperature range and the samples in this study were of lower purity (~1000 ppm impurities by weight). Like Leggett et al., Angerman and Casky also observed several twin systems in most grains and intersecting twins were observed in some areas. Slip within the grains also took place, as evidenced by the bending of the twins. At higher burnups, more distorted grain structures were observed, likely due to deformation by slip. Small fission gas bubbles were uniformly distributed in the samples irradiated below 673 K, similar to other observations [
44]. The authors observed the formation of irregular cavities in several samples irradiated above 723 K, which contributed significantly to the overall swelling. The metallographic analysis shown in
Figure 9 reveals large cavities with a range of length (10–200 microns), especially at high irradiation temperature. The cavities were mostly found in grain boundaries or junctions with angular incisions. Angerman and Caskey [
48] reported pronounced variations in microstructure and cavitation (formation of mechanical cavities/voids) with burnup, as shown in
Figure 9. At burnups below 0.1 at%, the grain structure of the uranium was easily recognizable regardless of the irradiation temperature, though some grain boundary cracking was observed at high irradiation temperatures. This grain boundary cracking is likely a result of thermal expansion stress at high temperature, hypothesized in [
49]. Thermal cycling experiments without irradiation found that polycrystalline samples deformed significantly and developed interior cracks with about 100 K temperature difference, and recent modeling has shown that thermal stresses arising from the negative thermal expansion coefficient is sufficient to cause cracking after a temperature change of about 100 K in irradiated α-U. At burnups of 0.15 to 0.3 at%, the grain structure became twisted and distorted as the result of deformation by slip. Uniform and fine “marbleized” structures were observed at burnups of 0.6 to 0.8 at%, though data are only available only at temperatures below 623 K.
The formation mechanisms of the multiple morphologies of irradiation-induced swelling within α-U (jagged cavities along grain boundaries, small aligned or non-aligned pores, massive cavities) is still an area of active research. Historically, the research community was unsure whether small pores (round spaces observed in the matrix) (observed in [
46]) were voids or fission gas bubbles; they were later confirmed to be voids [
48]. This uncertainty originated from the fact that no advanced electron microscopy was used in these investigations, generally because they were still new tools or not even invented yet. Fission bubble growth was argued to be negligible at the burnups for which these pores were observed [
44], though a few arguments were put forward to support the assertion that the observed pores were gas bubbles which then grew enormously, leading to high swelling above 723 K. Pugh [
44] suggested that the large “breakaway swelling” volumes were formed by the diffusion of gaseous fission products to alpha grain boundary cracks, ultimately driving plastic deformation of the matrix. Speight and Greenwood [
50] later proposed that dislocation sweeping could promote the breakaway condition by increasing the average bubble size through bubble coalescence. It was also demonstrated that enhanced swelling may arise in α-U for >823 K if the intergranular stresses due to irradiation growth are high enough to move dislocations. Conversely, later studies [
51] show clear evidence that certain porosity is due to mechanical cavitation and void formation. Recent modeling work has indicated that mechanical tearing along grain boundaries can arise from thermal stresses and irradiation growth stresses [
13,
47,
49], but the nucleation mechanism of intergranular voids remains unclear. Twin boundaries were implicated in the formation of crystallographically aligned voids [
46], though later work hypothesized they are void superlattices. Fission gases and vacancy agglomeration are both considered as mechanisms for void formation in α-U. To validate any of these hypotheses, the atomistic behavior of collision cascades and point defects in α-U and their collective behavior at diffusional length and time scales at low burnups must be studied with advanced characterization and simulation techniques. In addition, although several experimental investigations indicate that impurities affect the irradiation growth and swelling rate [
33,
51], there remains no comprehensive analysis of the role of these impurities in controlling irradiation growth and swelling behavior.
3.2. Irradiation Effect on Mechanical Properties
Although several attempts have been made to understand the effect of irradiation on mechanical properties of α-U, such as ductility, hardness and creep, the results of these studies have been contradictory. A common observation, however, is the increased hardness of α-U, which was hypothesized to be a result of fission product impurity, irradiation-induced voids or dislocation loop formation. In 1958, Kittel and Paine [
35] tested polycrystalline natural uranium in the Materials Testing Reactor and found that the hardness increased with burnup, and that annealing did not restore the hardness. Similar observations were also made by Loomis et al. [
42] on single crystalline α-U in 1964. However, a saturation in hardness was observed at a fluence of approximately 1 × 10
17 neutrons/cm
2. Both studies irradiated material at relatively low temperatures of up to 493 K and neither study provided information about the purity of the samples. Although radiation hardening is well known to occur due to the impediment of dislocation motion by defect loops, the reason for the observed yield strength saturation is unclear.
The effect of irradiation on the creep and ductility of single crystal α-U remains unclear. High-purity single crystals of α-U were irradiated and then mechanically tested, and the single crystals retained considerable ductility (
Figure 10a). Conversely, significantly lowered elongations were observed in polycrystalline α-U by Vorob’ev [
52] and Konobeevskii et al. [
53] (
Figure 10b). Konobeevskii et al. [
53] attributed such reduction to the formation of numerous internal cracks in polycrystalline metal as a result of irradiation. The initial study on creep testing in irradiated U was also reported by Konobeevsky et al. [
53], who irradiated the material and then creep-tested it out of pile. The creep rate was increased by a factor of 1.5 to 2 versus the un-irradiated material. Following this, Roberts and Cottrell [
54] conducted in situ creep testing in an irradiation environment, and observed a similar enhancement in creep during irradiation.
Although the mechanical behavior of irradiated α-U varies significantly within the literature, it appears that while single crystal material retains its ductility, the observed loss of strength and ductility in polycrystalline material is likely the result of the presence of grain boundaries. In fact, the possibility of superplasticity in irradiated polycrystalline α-U has been discussed in [
55,
56]. Though internal stress-driven plasticity (a phenomenon also found to be responsible for generating environmental superplasticity in other materials [
57]) is proposed to contribute to the observations in [
46], there is a need for more comprehensive mesoscale modeling of irradiation considering the effects of plasticity and grain-boundary interactions. Internal stress-driven plasticity under irradiation could arise from the anisotropic radiation growth of differently oriented grains, causing stresses high enough to initiate plastic flow. Further irradiation would continue to drive irradiation growth and stresses. Microcracks at grain boundaries could arise due to the presence of hard inclusions, insufficient slip and twinning processes to accommodate deformation across grain boundaries, or low grain boundary adhesion.
However, detailed comparisons of mechanical behavior in-pile and out-of-pile from the existing literature may not be meaningful, given the very different samples, irradiation and testing conditions. Many historical tests did not characterize the details in the microstructure or composition of the materials. Many of the tests on irradiated material were performed before the development of irradiation damage models (Kinchin–Pease, Norgett—Robinson–Torrens) that calculate displacements per atom (dpa), and thus neutron fluxes, fluences, and spectra need to be reported to obtain an accurate understanding of the amount of irradiation damage. Furthermore, the irradiation temperatures were historically not well controlled or measured. Finally, given the difference in the purity level, grain size and fabrication methods, all of which could affect the irradiation behavior and mechanical response, more detailed testing is needed.
3.3. Irradiation Effects on Thermal and Electrical Properties
Neutron irradiation increases the electrical resistivity and thermal conductivity of α-U, but the specific relationship is not well understood. A direct correlation was observed between the irradiation-increased resistivity and single crystal irradiation growth in all the available studies, albeit with limitations on the burnup range investigated. The effect of thermal neutron irradiation on the electrical resistivity of lineage single crystals (<600 ppm impurities) was studied by [
42] in the CP-5 reactor up to 550 K. For a constant irradiation temperature, electrical resistivity increased with irradiation dose and was also proportional to the increase in irradiation growth strain. However, the increase in resistivity with irradiation dose is reported to saturate [
42], contrary to the dimensional changes induced by irradiation growth.
Though the very few results about the effects of irradiation on the thermal properties of α-U are non-conclusive given the uncontrolled testing performed in 1960s, the existing results and estimates do reveal that the thermal conductivity of irradiated α-U is impacted. The overall dependence is governed by various phenomena including lattice resistance imposed by defects, microstructural damage on the grain boundaries, porosity development and the presence of fission gases. Brailsford and Major [
58] theoretically estimated the reduction in thermal conductivity as well as electrical resistivity due to irradiation in the form of an equation (Equation (1)), considering various key factors,
where
R is the % reduction in the thermal conductivity,
D0 and
D1 are the initial and final density (g/cm
3) (evaluated through any applicable swelling model) and b is burnup (at %). The burnup dependence implicitly accounts for changes due to lattice strains, lattice defects and fission gas bubbles and voids. Though the estimate considers several factors, several simplifying assumptions were made. Although the analysis considers the increased resistivity from irradiation effects, the effect of dislocation loops were totally neglected [
59]. However, the effect of dislocation loop dynamics is expected to play a vital role in the irradiation growth rate. Given the strong, suggestive results of the dependence of electrical resistivity on the irradiation growth rate and similar mechanisms responsible for thermal resistivity, more comprehensive models need to be developed.
To date, there is no reliable experimental analysis to validate Equation (1). However, recent DFT studies by Peng et al. [
60] estimate the effect of U vacancies, U interstitials and Zr substitution on the thermal conductivity. Although Equation (1) does not include temperature dependence, there is reasonable order-of-magnitude agreement between its predictions and the DFT studies. Among the three types of defects studied by DFT, U vacancies have the most substantial impact on thermal conductivity. Similar to the case of irradiation effects on dimensional changes and mechanical properties, further experimental investigations are required to understand the effect of irradiation on α-U’s thermal conductivity and electrical resistivity under carefully controlled conditions, including specification of the initial purity, the metallurgical condition of the specimen and the details of the irradiation (temperature, flux, fluence and neutron spectrum). Simultaneous measurements of the electrical resistivity and thermal conductivity during irradiation would be valuable to test the validity of the Wiedemann–Franz law for irradiated specimens, which forms the basis of Brailsford and Major’s theoretical model.
The effect of irradiation on the superconductivity of α-U is a fundamental area of interest that could receive more attention. Existing results suggest that irradiation could affect the superconductivity of materials in two different ways: First, extended defects, such as column-shaped amorphous regions, act as strong, correlated pinning centers and usually increase the critical current density substantially. Second, point defects influence the microscopic parameters responsible for superconductivity, thus affecting the critical temperature and the critical fields. Analysis of the irradiation effects could provide a strong fundamental basis to characterize the intrinsic superconductivity of uranium as well as quantify the impact of irradiation-induced defects.