Introduction

During reactor operation, the uranium-22.5 atomic% zirconium (U-22.5at.%Zr) system nominally operates between 500 and 700 °C with thermal gradients near 35 °C/mm, and the peak temperature at the center of the cylindrical fuel [1, 2]. The primary phenomena impacting the alloy under these reactor conditions include fuel-cladding chemical interaction, fuel swelling, fuel-cladding mechanical interactions, and constituent redistribution [3,4,5,6,7,8]. These phenomena are impacted by material properties, such as heat of transport and diffusivity, which are in turn dictated by crystal structure and phase morphology [9,10,11,12]. Due to the harsh reactor environment, in situ crystallographic and morphological studies are non-existent, therefore relying on pre- and post-irradiation experiments to comprehend the microstructural evolution of the fuel.

While the equilibrium U–Zr phase diagram is best described by Rough and Bauer [13, 14], there is room for furthering the fundamental understanding [15,16,17] to ensure that novel fabrication techniques result in a known microstructure. Typically fabricated (i.e., cast, extruded, and annealed) U-22.5at.%Zr fuel observed at room temperature is comprised of a binary structure of α-U (Cmcm), with < 1at.%Zr solubility [18, 19], and δ-UZr2 (P6/mmm), with a composition of U-66.3at.%Zr [23], with Zr inclusions dispersed throughout the matrix [13,14,15, 20,21,22]. The morphology of the α-U and δ-UZr2 phases is determined by composition, fabrication methods, and cooling rates [23, 24], whereas the Zr inclusions remain relatively unaffected by heat treatments [15]. As the initial microstructure may impact phase and crystallographic evolution in a reactor [3], characterization should be performed on each uniquely fabricated U–Zr alloy.

This study characterizes the un-irradiated microstructure of U-22.5at.%Zr fuel foils intended for separate effects irradiation testing, which enables the deconvolution of reactor power, temperature, and composition using unique irradiation vehicle designs and fuel geometries [25,26,27]. Two types of foils were investigated, including a heavily rolled fuel foil and an annealed (following rolling) fuel foil. Rolling and subsequent heat treating is a novel fabrication technique for U–Zr alloys, requiring characterization to quantify microstructural differences and recovery. Characterization via transmission electron microscopy (TEM) and stereological methods included crystal structure and phase identification, as well as morphological quantification.

Materials and methods

A U-22.5at.%Zr ingot was drop cast via arc-melting into a Cu mold in an inert atmosphere glovebox [15, 25]. Sample-A was harvested with a focused ion beam from a foil, fabricated by hot-rolling to a ~ 90% reduction, annealing for two hours at 800 °C, then cold-rolling to a final ~ 60% reduction of the ingot. Samples-B,C were harvested from the same foil following a two-hour anneal, after rolling, at 900 °C with a 1 °C/min cooling rate.

TEM bright field (BF) images, dark field (DF) images, and energy-dispersive X-ray spectroscopy (EDS) maps were collected (200 kV accelerating voltage) on all samples using a FEI Titan ChemiSTEM FEG-STEM. Selected area electron diffraction (SAED) was performed on various regions and analysis was completed using manual measurements in conjunction with CrysTBox [28]. Stereology was performed following methods described by Underwood [29] and Vander Voort [30]. Further details can be found in the Supplementary Materials.

Results

Sample-A has two major U–Zr phases, a U-rich phase and a Zr-rich phase as pictured in Fig. 1c, d, j, k, l, in a lamellar structure and one minor phase, Zr inclusions, as pictured in Fig. 1a, b. No minor elements were detectable, using EDS, in the bulk material of Samples-A, C. Figure 1j, k, l, shows a compositional gradient in the Zr-rich phase, with Zr content increasing from 15at.%Zr near the phase boundary to 63at.%Zr in the center. Figure 1j, k, l also shows that the U-rich portion of the lamellar structure has a consistent composition near U-15at.%Zr. The SAED patterns, Fig. 1f, h, identify the Zr-rich phase to be space group of P6/mmm (hexagonal δ-UZr2 [31]), and the U-rich phase to be space group of Cmcm (orthorhombic α-U [32]). The SAED patterns of δ-UZr2 and α-U show nanocrystalline grains in the bulk microstructure, indicated by the onset of rings in Fig. 1f, h. The δ-UZr2 and α-U lamellar structures are morphologically distorted near the minor phase, Fm-3 m (face-centered cubic, FCC) Zr inclusions, indicated by Fig. 1b, g, i. Line scans across the Zr inclusions, Fig. 1j, k, indicate a composition of U-85at.%Zr.

Fig. 1
figure 1

Heavily rolled U-22.5at.%Zr foil, Sample-A, showing: (a) Dark field TEM image with the locations of EDS line scans (red) corresponding to subfigures j–l. (b) Bright field TEM image with red SAED locations corresponding to diffraction patterns f–i. EDS maps of (c) U, (d) Zr, and (e) O. (f,g,h,i) SAED patterns with the zone axis and diffraction spots indicated. The locations of the patterns collected are indicated with the corresponding figure labels in (b). (j, k, l) EDS line scans of U and Zr corresponding to locations in (a)

Full size image

Figure 2a–g shows that Samples-B,C have the same phases, α-U [32] and δ-UZr2 [31] and dispersed inclusions, with a broadened and coarsened lamellar structure, as compared to Sample-A. The EDS line scans across the lamellar structure, Fig. 2h, i, j, shows that the Zr-rich phase has a composition near U-63at.%Zr and the U-rich phase has a composition near U-15at.%Zr, in agreement with Sample-A. The SAED patterns of the δ-UZr2 and α-U structure following the annealing, Fig. 2f, g, show no evidence of reorientation or nanocrystalline grains, which were present following rolling. The line scan in Fig. 2i indicates the inclusions in Sample-B remain near U-85at.%Zr. SAED of the Zr inclusions in Sample-C, shown in Fig. 2k, l, m, n, confirms that the Zr inclusion remains as Fm-3m following annealing.

Fig. 2
figure 2

Samples-B,C, U-22.5at.%Zr foils annealed at 900 °C for two hours followed by a 1 °C/min cooling rate. (a) Dark field TEM image of Sample-B with the locations of EDS line scans (red) corresponding to figures h–j. (b) Bright field TEM image with red SAED locations corresponding to diffraction patterns f and g. EDS maps of (c) U. (d) Zr. (e) O. (f–g) SAED patterns with the zone axis and diffraction points indicated on each figure. (h–j) EDS line scans corresponding to figure a, with Zr (blue) and U (yellow). (k) Dark field TEM image of Zr inclusion in Sample-C with red circle corresponding to the SAED in (n). EDS maps of (l) Zr. (m) U. (n) SAED pattern of Zr inclusion with the zone axis and diffraction points indicated

Full size image

Stereology results for Samples-A, B include volume fraction (VV, measured as PP), features per line (NL), and the mean intercept length, L3, and mean random spacing, σr, between lamellae are presented in Table 1. The phase fraction (PP) for δ-UZr2 in Sample-A, is greater than twice that of Sample-B. Inclusions in the rolled Sample-A are more prevalent; however, sample population is limited by sample size relative to bulk microstructure. The L3 and σr values verify a finer lamellar structure in Sample-A than observed in Sample-B, which has a broader and coarser lamellar structure.

Table 1 Stereology results for Samples-A,B, including: (a) the volume fraction for each phase, PP, using 300 randomly located points per image, (b) the number of features per line, NL, or the number of phase intercepts per μm of random test lines, (c) the mean intercept length for the δ-UZr2 phase, L3 = PP/NL [32], and (d) the mean random spacing between lamellae, σr = 1/NL [33]
Full size table

Discussion

All foils were comprised of two primary phases, δ-UZr2 [31] and α-U [32], in a lamellar structure. The lamellar structure in Sample-B lacks indications of reorientation or nano-grains observed in Sample-A, agreeing with expected results following a 900 °C anneal [15, 24]. Sample-B has a mean random spacing between lamellae over three times that of rolled Sample-A, appears coarsened, and increasingly globular. The differences in the lamellar formation and coarsening is expected to be due to different cooling rates between hot-rolling and annealing as well as deformation induced during rolling [24].

The EDS results of all samples show a semi-quantitative composition in α-U, near 15at.%Zr, measuring higher than the literature value, < 1at.%Zr solubility [18]. The δ-UZr2 phase in Sample-B was measured to be UZr2 at 63at.%Zr, indicating that the high Zr solubility in α-U is non-negligible. The EDS results of δ-UZr2 in Sample-B show the Zr content plateauing near the phase boundary interfaces. In contrast, the δ-UZr phase in Sample-A shows a compositional gradient, also peaking near 63at.%Zr in the center of the phase. While the peak composition of both samples is the same, the compositional plateauing was not apparent in the Sample-A. Although the electron interaction volume was consistent for all samples, at ~ 50–100 nm3, the finer lamellar structure in Sample-A could result in an oversampling of lamella co-located within the thickness of the sample (~ 50 nm), thus causing the compositional disparity, or supersaturation of U in the δ-UZr phase, to be difficult to quantify.

However, increasing U in the δ-UZr2 phase also results in differing phase fractions between Samples-A,B. Sample-B, comprised of 76.32% ± 0.02% α-U, 21.93% ± 0.03% δ-UZr2, and 1.75% ± 0.03%Zr inclusions, is in near agreement with the equilibrium system [13, 14, 33]. Sample-A, comprised of 41.99% ± 0.07% α-U, 53.47% ± 0.09% δ-UZr2, and 4.53% ± 0.05%Zr inclusions, differs significantly from the equilibrium system, indicating the supersaturation of U in deformed δ-UZr2. Further investigation into the relationship between defect concentration and phase fractions in U–Zr is of interest since the nuclear fuel is subjected to a defect-producing environment. This is particularly important since phases determine bulk behavior [9, 10].

Inclusions in U-22.5at.%Zr alloys have been both hypothesized and measured to be α-Zr (hexagonal close-packed, HCP) stabilized by a minor constituent, such as O or N [5, 34, 35], typically originating from the Zr-rind on the edge of a casting [36]. Alternate findings suggest that the Zr inclusions may form during solidification [37] and result in an FCC structure, hypothesized to be caused by unknown differences in casting techniques [38]. The Zr inclusions in this study were found to be FCC, rather than HCP α-Zr. The first observation of an HCP to FCC transformation in Zr was made in 2002 during ball-mill mechanical attrition of pure Zr [39] and again in 2017 during the cold-rolling of pure Zr [40]. Therefore, additional research is needed to determine whether the FCC structure was present due to the casting technique or if the inclusions underwent a stress-induced phase transformation during cold-rolling.

Conclusions

TEM-SAED indexing, TEM-EDS analysis, and stereology were used to quantify the microstructure of U-22.5at.%Zr samples following a novel fabrication process of rolling and annealing. A sample was characterized following hot- and cold-rolling. Two additional samples were characterized following a post-roll anneal. The rolled sample had a supersaturation of U in the δ-UZr2 phase, resulting in a phase fraction of δ-UZr2 more than twice that of the equilibrium system. This finding suggests that the defect concentration of the alloy may have a higher impact on the microstructural evolution and Zr migration within the fuel than is commonly thought. Inclusions in the sample were FCC Zr, rather than HCP α-Zr, reinforcing the need for further studies on the formation and evolution of Zr inclusions to identify the mechanisms that determine the crystal structure.

The annealed sample was found to be in close agreement with the equilibrium system, typically fabricated by casting, extrusion, and heat treatment. Following annealing, the phase fractions returned to expected values predicted by the phase diagram of 21.93% ± 0.03% δ-UZr2 and 76.32% ± 0.02% α-U with dispersed FCC Zr inclusions. These findings show that post-rolling heat treatments can recover U–Zr fuel microstructures found in traditionally cast fuel slugs [20] with lamellar spacing dependent upon heat treatments.

In both samples, the solubility of Zr in α-U was identified to be higher, 15at.%, than the values historically reported. While this is not an absolute measurement of Zr solubility, it does represent the need for further studies, such as atom probe tomography, where precise quantification can be performed.