Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2024 Feb 27;9(10):12135–12145. doi: 10.1021/acsomega.3c10481

Oxygen Isotope and Fluorine Impurity Signatures during the Conversion of Uranium Ore Concentrates to Nuclear Fuel

Aaron M Chalifoux †,‡, Cody Nizinski †,‡, Miguel Cisneros §, Travis Tenner ∥, Benjamin Naes ∥, Kimberly N Wurth ∥, Erik Oerter §, Michael Singleton §, Gabriel J Bowen ⊥, Luther W McDonald IV †,*
PMCID: PMC10938325  PMID: 38496959

Abstract

graphic file with name ao3c10481_0007.jpg

Within the front end of the nuclear fuel cycle, many processes impart forensic signatures. Oxygen-stable isotopes (δ18O values) of uranium-bearing materials have been theorized to provide the processing and geolocational signatures of interdicted materials. However, this signature has been minimally utilized due to a limited understanding of how oxygen isotopes are influenced during uranium processing. This study explores oxygen isotope exchange and fractionation between magnesium diuranate (MDU), ammonium diuranate (ADU), and uranyl fluoride (UO2F2) with steam (water vapor) during their reduction to UOx. The MDU was precipitated from two water sources, one enriched and one depleted in 18O. The UO2F2 was precipitated from a single water source and either directly reduced or converted to ADU prior to reduction. All MDU, ADU, and UO2F2 were reduced to UOx in a 10% hydrogen/90% nitrogen atmosphere that was dry or included steam. Powder X-ray diffraction (p-XRD) was used to verify the composition of materials after reduction as mixtures of primarily U3O8, U4O9, and UO2 with trace magnesium and fluorine phases in UOx from MDU and UO2F2, respectively. The bulk oxygen isotope composition of UOx from MDU was analyzed using fluorination to remove the lattice-bound oxygen, and then O2 was subsequently analyzed with isotope ratio mass spectrometry (IRMS). The oxygen isotope compositions of the ADU, UO2F2, and the resulting UOx were analyzed by large geometry secondary ion mass spectrometry (LG-SIMS). When reduced with steam, the MDU, ADU, and UO2F2 experienced significant oxygen isotope exchange, and the resulting δ18O values of UOx approached the values of the steam. When reduced without steam, the δ18O values of converted ADU, U3O8, and UOx products remained similar to those of the UO2F2 starting material. LG-SIMS isotope mapping of F impurity abundances and distributions showed that direct steam-assisted reduction from UO2F2 significantly removed F impurities while dry reduction from UO2F2 led to the formation of UOx that was enhanced in F impurities. In addition, when UO2F2 was processed via precipitation to ADU and calcination to U3O8, F impurities were largely removed, and reductions to UOx with and without steam each had low F impurities. Overall, these findings show promise for combining multiple signatures to predict the process history during the conversion of uranium ore concentrates to nuclear fuel.

Introduction

The front end of the nuclear fuel cycle involves many steps to convert natural uranium to usable fuel for nuclear reactors. The typical processes involved are mining of the uranium ore and then milling of the ore into a more workable powder. The uranium is then dissolved in nitric acid and extracted to purify and concentrate the uranium in solution.1,2 Next, the uranium is precipitated as a uranium ore concentrate (UOC) in the form of uranyl peroxide (UO4), ammonium uranyl carbonate (AUC, (NH4)4UO2(CO3)3), ammonium diuranate, (ADU, (NH4)2U2O7), magnesium diuranate (MDU, MgU2O7), or sodium diuranate (SDU, Na2U2O7). These UOCs are generally calcined to UO3 or U3O8 for easier transport to conversion facilities in which the calcined material is then converted to UF6 via reactions with HF and F2.1,2 The resultant UF6 can be directly reduced to UO2 or converted into AUC or ADU before reduction. Often, steam is added during the reduction of the UOCs or UF6 products to improve the purity of the final UO2.3−5

Each processing step within this cycle may produce or affect nuclear forensic signatures by altering the morphology, the abundance and/or distribution of impurities,6−8 and/or the stable isotope ratios of the materials.2,9−13 Recently, analysis of 18O/16O stable isotope ratios (or δ18O values, equivalent to [Rmeasured/RVSMOW – 1] where R = 18O/16O, and VSMOW stands for Vienna Standard Mean Ocean Water, where δ18O VSMOW = 0‰) of U-oxide materials has garnered attention due to the possibility of these ratios providing information about their process history and provenance. A 2001 study by Pajo et al. found that δ18O values of collected UO2 pellets matched closely that of the local meteoric waters (rain and snow) where the pellets originated.14 More recently, Klosterman et al. examined the fractionation between uranyl peroxide and ADU with the water, from which these compounds were precipitated, along with the changes in δ18O values after calcination in dry air and an inert nitrogen atmosphere.12,13 Klosterman et al. subsequently expanded these studies by calcining U3O8 using two experimental pathways: (1) U3O8 was calcined with N2 + H2O (vapor), where H2O vapor δ18O = −8.8‰, and (2) U3O8 was calcined with N2 + O2 + H2O (vapor), where O2 δ18O = +23.2‰ and H2O δ18O = −26.2‰. Experiment 1 studied oxygen exchange between U3O8 and water vapor; experiment 2 studied U3O8 oxygen exchange with water vapor and O2.15 During calcination with N2 + H2O, Klosterman et al. found oxygen isotope equilibration between the U3O8 and the steam occurred rapidly above 600 °C.15 When the U3O8 was calcined with N2 + O2 + H2O, Klosterman et al. found that U3O8 preferentially exchanged oxygen with atmospheric O2 at 800 °C, U3O8 exchanged oxygen with atmospheric O2 and water vapor at 600 °C, and U3O8 preferentially exchanged oxygen with water vapor at 400 °C.15

More recently, Assulin et al. studied oxygen isotope exchange between U3O8 and atmospheric O2 at a range of temperatures and found that oxygen exchange between U3O8 and O2 was minimal below 100 °C, partial exchange occurred between 100 and 200 °C, and isotopic equilibrium was reached above 400 °C.16 They also found that preferential incorporation of 16O by U3O8 led to U3O8 attaining δ18O values lower than O2. Expanding this work, they calcined three U3O8 samples of differing δ18O values to UO2 between 500 and 700 °C for 2, 4, and 6 h.17 From these experiments, they found that the final δ18O values of the UO2 were dictated by the isotopic equilibrium between the water generated during the reduction and the resultant UO2.17 They also found that the UO2 products were slightly more depleted in 18O content than the starting U3O8 materials, while the water vapor formed was around 11‰ more enriched.17

Collectively, these experiments provide valuable information about the oxygen isotope fractionation of UOCs during calcination in inert or atmospheric conditions at elevated temperatures. However, knowledge is still lacking with respect to oxygen isotope fractionation and impurity effects for many other UOC conversion pathways. Therefore, the purpose of this study is to examine the oxygen isotope fractionation from the conversion of MDU, ADU, and UO2F2 to UOx for dry and steam-assisted reductions and also to generate complementary impurity information, such that when the data are combined, they can provide meaningful predictions related to the process history of nuclear fuel.

Experimental Section

Preparation of Starting Uranium Compounds

The preparation of the MDU, ADU, and UO2F2 materials was described in previous publications.18,19 Briefly, all materials were prepared from UO3 (Alfa Inorganics Ventron, lot 80105, 102770). For MDU, UO3 was dissolved in a minimal amount of 8 M nitric acid and heated to 108 °C until the excess water evaporated. This was done twice using waters either depleted in 18O (δ18O = −15.3 ± 0.1‰; hereafter, the resulting material is referred to as dMDU) or enriched in 18O (δ18O = +21.7 ± 0.1‰; hereafter the resulting material is referred to as eMDU).18 The dMDU and eMDU materials were precipitated by adding a MgO slurry.

The UO2F2 was also prepared from the UO3 stock by adding HF. Half of the uranyl fluoride solution was used for the ADU precipitation, in which NH4OH was added until a pH of 9 was achieved. The other half of the UO2F2 was directly dried and again split in half for reductions with or without steam.19

Reduction of MDU, ADU, and UO2F2 to UOx

The reduction of MDU, ADU, and UO2F2 to UOx was reported previously.19,100 The MDU and UO2F2 were directly reduced to UOx with and without steam in a 10% H2 and 90% N2 atmosphere; the ADU was first calcined at 800 °C in a nitrogen atmosphere under 300 mL/min of N2 to produce U3O8, and the U3O8 was subsequently reduced to UOx with or without steam in a H2/N2 atmosphere.

Regarding the steam, the approximate water vapor concentration and the δ18O value were both measured using a Picarro L2130-i isotope analyzer that was connected to a tube furnace outlet. The average water vapor concentration in a 50 mL min–1 wet N2 stream coupled with a 440 mL min–1 dry UHP N2 stream was 5240 ± 40 ppm, analyzed by cavity ring down spectroscopy (CRDS) measurements taken over 30 min.19 For a flow of 60 mL·min–1, the water vapor concentration was 6290 ± 50 ppm;19 the δ18O value of the water vapor was −25.11 ± 0.36‰.19 We did not measure the δ18O value of the liquid water from which the vapor was generated; however, 13 tap water samples collected between 2013 and 2023 from the University of Utah campus near the lab discussed herein have an average δ18O value of −15.86‰ (±0.41‰) (waterisotopes.org, projects 00058, 00059, 00064).20 Based on H2O liquid–vapor oxygen isotope fractionation, the theoretical δ18O value of the vapor in equilibrium with this water (liquid) at 25 °C is −25.16‰;21 the calculated water vapor δ18O value closely approximates the measured water vapor δ18O value. The water vapor concentration of the dry atmosphere was not measured. However, all gases were passed through VICI Metronics gas purifiers to remove any water vapor that may have been present in the high-purity gas tanks.

Following synthesis, the composition of the materials was verified by using p-XRD. Most of the materials did not reach complete reduction to UO2. This was expected for MDU but was anomalous for ADU and UO2F2. The MDU was reduced to UO2, U4O9, and MgU2O6. Qualitatively, it appeared that there might be less MgU2O6 in the materials reduced with steam, but quantification using Rietveld refinement was not possible due to the extensive peak overlap. The ADU materials reduced with and without steam were nearly identical, comprising roughly 93% UO2 and 7% U4O9.

In contrast, steam significantly affected the final composition of reduced UO2F2. Without steam, the final product consisted of 71% UO2F2, 27.2% U3O8, and 1.7% UO2. UO2F2 reduced with steam fully converted to UO2 in some replicates but was a mixture of UO2 with U3O8 and U4O9 in other replicates. Full details of the p-XRD data are available in Chalifoux et al. and Nizinski et al.22,100

Bulk Fluorination of UOx from MDU for Oxygen Isotope Measurements

The O2 analyzed by fluorination-IRMS is predominantly oxygen within the UOx crystal lattice and other oxides that are present in the powder; however, this technique may also measure oxygen of UOx oxidation/alteration phases (e.g., schoepite) and/or adsorbed water. For oxygen isotope measurements, ∼16–20 mg of UOx was loaded into nickel capsules and the capsules were placed into nickel reaction vessels attached to a vacuum line. Two measurements of each replicate were performed, and the averages of the δ18O values, O wt %, and yield were taken. Fluorination was completed in two steps: (1) samples were initially fluorinated with 18 Torr of chlorine trifluoride (ClF3) at room temperature (∼25 °C) for 5 min to remove adsorbed water on UOx; (2) prefluorination ClF3 and reactant products were removed from the vessels, and samples were subsequently fluorinated with 400 Torr of ClF3 and reacted at 550 °C for 4 h. The final fluorination step liberates the lattice-bound oxygen. O2 was passed through four liquid nitrogen traps to remove condensable gases and a heated NaCl trap to remove F2 and then collected on a molecular sieve at liquid nitrogen temperature. The molecular sieve was then warmed to room temperature to release the O2, and the pressure of the O2 was measured on a calibrated manometer to determine the oxygen concentration of fluorinated materials. The O2 was transferred to an Isoprime PrecisION IRMS (Elementar) operating in a dual-inlet mode for oxygen isotope measurements.

Oxygen isotope ratios were calibrated relative to VSMOW. Two reference materials were analyzed: (1) a SiO2 standard (NBS-28: δ18O = 9.6‰) and (2) a UO2 standard (CRM-125A: δ18O = −9.63‰).23 Reference materials were analyzed in duplicate on separate analysis days, and a two-point calibration based on the reference material δ18O values was used to correct measured δ18O values from unknown U–O-bearing materials. The precision for δ18O values was determined based on the average absolute differences of known and measured δ18O values of reference materials (±0.71 and 0.62‰, 2σ). The precision of oxygen concentrations was calculated based on the average absolute differences of the measured and stoichiometric oxygen abundance of reference materials (±1.49 and 1.64 wt %, 2σ).

LG-SIMS for Oxygen Isotope Measurements and F Impurity Mapping

To prepare samples for particle analysis by large geometry secondary ion mass spectrometry (LG-SIMS), materials from ADU and UO2F2 reduction reactions were received as powders at Los Alamos National Laboratory (LANL). Subaliquots of the powders were added to 20 mL glass scintillation vials partially filled with electronics-grade n-hexane (97%) and then ultrasonicated to suspend the particles. For each material, 50–200 μL amounts of suspension were pipetted onto a 1” diameter N-type P-doped single-side-polished Si wafer (⟨100⟩ orientation; 1–100 ohm-cm resistance). After applying, the n-hexane evaporated, evenly dispersing and electrostatically adhering the micron-sized particles to the wafer surface.

The particle-laden wafer samples were analyzed by using the Cameca IMS 1280 LG-SIMS at LANL. A primary beam of Cs+ ions was focused in “Gaussian” mode. Common analytical parameters were as follows: impact energy: 10 keV; sample voltage: – 10 keV; field aperture: 4000 μm; contrast aperture: 400 μm; entrance slit: 243 μm; and energy slit bandpass: 50 eV. An electron gun was not employed, as the doped Si wafer provided charge compensation.

For the LG-SIMS oxygen isotope analysis of individual particles, the following settings were used. A field magnification of 133 × 133 μm (M. Area: 60) was set by the transfer optics. A fixed magnet axial mass of ∼16.9 was employed, and secondary ion signals from 16O and 18O were detected simultaneously (e.g., multicollection) using a Faraday cup (FC) detector and an electron multiplier (EM) detector, respectively. A multicollection exit slit width of 500 μm was used, and with all aforementioned instrument parameters, the mass resolving power was ∼2000. Prior to each analysis, a 400 × 400 μm rastered area around the targeted particle was sputter-cleaned by the primary ion beam for 30 s, using a 10 nA beam current. For the analysis, a primary ion beam raster of 7 × 7 μm and a current of 100 pA was used to generate signals of 5E4 to 2E5 counts per second (cps) on the EM detector for 18O, and 2.5E7 to 1E8 cps on the FC detector for 16O, per particle. The analysis duration was 120 s, split into ten 12 s cycles. A total of 20 to 32 single particle data were collected per sample. Uncertainties of individual data are calculated as the expanded standard error (2SE) of the 10 cycles per analysis. Uncertainties of averaged δ18O values are calculated as the expanded standard deviation (2SD) from sample data sets. Raw data from samples were corrected to accurate values through calibration measurements of CRM 125-A reference UO2 particles.23 For each day of sample analysis, brackets of 6–10 CRM 125-A particle analyses were collected at the beginning, middle, and end of the session. The instrument bias was calculated as the difference between the raw averaged δ18O per bracket and the reference δ18O value of CRM 125-A; once determined, raw data from samples were then instrument bias-corrected for accuracy. A two-point linear regression was employed for calculating the instrument bias for data collected between brackets using timestamps of analyses from samples to calculate the bias. It is recognized that there could be additional uncertainties to the bias for the materials analyzed with different species than CRM 125-A (UO2), but at the time of this reporting, no other oxygen isotope reference materials for other U-oxide species exist.

Impurity distributions of fluorine24 were also investigated via LG-SIMS. For these measurements, the transfer optics were set to a field magnification of 50 × 50 μm (M. Area: 160). The exit slit width was 405 μm, and the overall mass resolving power was ∼2000. For each analysis, the primary ion beam was rastered over a 150 × 150 μm area, using a 10 pA primary ion beam current. Secondary ion signals from 18O (50–500 cps), 19F (500–400,000 cps), and 238U16O (50–1000 cps) were measured by hopping the magnet and detecting with one axial EM. Note that the measured 18O signal was only used to confirm the spatial locations of 238U16O signal from particles. The total analysis duration was 285 s, which consisted of counting times of 125, 5, and 125 s for 18O, 19F, and 238U16O, respectively, and included 10 s magnet settling times between each magnet hop. Prior to each measurement, the area analyzed was sputter-cleaned for 60 s with a primary ion beam current of 10 nA. For each sample, a total of nine measurements were collected in a 3 by 3 grid. For each analysis, a map of secondary ion signals coming from the particles distributed within the area was generated, using Cameca’s Automated Particle Measurement (APM) software. For each particle identified, the APM software reported the corresponding 19F/238U16O ratio. The uncertainties of individual data are reported as counting statistics errors. Uncertainties of averages from APM maps are not calculated because (1) data scatter is intimately tied to counting statistics and not to the repeatability of measurements; and (2) most samples show significant heterogeneity of 19F/238U16O values, as reported later. Note that because (1) 19F and 238U16O have significantly different count rates, due to their different signal generation efficiencies, and (2) no reference material was available to calibrate for 19F and 238U16O signal, the absolute 19F/238U16O values of data are almost certainly inaccurate. However, the measured 19F/238U16O values are still highly useful for comparing the relative F impurity abundance differences between samples.

The isotope homogeneity of particle data sets was evaluated using a counting statistics-based model reported in Tenner and co-workers.25 Specifically, it has been observed that for certified reference materials that are assumed to be isotopically homogeneous, the scatter of data is directly related to the analytical precision or counts per particle. The scatter of isotope ratio data is modeled as a function of counts per particle by eq 1:

graphic file with name ao3c10481_m001.jpg 1

where the avg. isotope ratio is that of a data set or any desired isotope ratio to be modeled. N(avg)counts and D(avg)counts correspond to all possible count combinations of the numerator and denominator of the avg. isotope ratio or any desired isotope ratio to be modeled. The term n approximates a Gaussian distribution, probability = erf(n/√2). Here, an n value of 3.5 was applied to the model, which corresponds to 99.9% of data falling within the model bounds if a material is isotopically homogeneous. For all N(avg)counts and D(avg)counts combinations of the model, the D(avg)counts values are plotted on the x-axis, and the isotope ratio values calculated from eq 1 are plotted on the y-axis. Note that modeled 18O/16O ratios are easily converted to delta-notation.

While samples were made in triplicate in the initial study, the extensive sample preparation, instrument calibration, and sample measurement time limited the number of replicates that were analyzed by LG-SIMS. As such, only single replicates of the starting UO2F2, the ADU, the ADU calcined to U3O8, the ADU-U3O8 reduced to UOx without steam, the ADU-U3O8 reduced to UOx with steam, and the UO2F2 reduced to UOx without steam were investigated by LG-SIMS. Two replicates of the UO2F2 reduced to UOx with steam were analyzed, as the previously reported p-XRD data showed wide variations in the composition of the final product. Replicate 1 was roughly 38% U3O8, 38% U4O9, and 24% UO2, while replicate 2 was nearly 100% UO2.

Results and Discussion

Fluorination of UOx from MDU

Following fluorination, oxygen yields between 122.59 and 130.91% were measured, which reveals that each UOx compound is not stoichiometric UO2, but rather a mixture of phases with oxygen reservoirs. P-XRD analyses of the UOx samples support this conclusion by providing clear evidence that full reductions to UO2 were not completed after a 6 h reduction time.

UOx samples reduced from dMDU and eMDU without steam had δ18O values of +9.91 ± 0.67 and +19.25 ± 0.67‰, respectively (Table 1 and Figure 1). Unfortunately, the bound oxygen on the MDU could not be directly measured, but the composition of the hydrates on the dMDU and eMDU were previously measured using TGA-IRIS. Chalifoux et al. found that dMDU and eMDU water groups had δ18O values of −5.22 ± 0.40 and +5.39 ± 1.38‰, respectively.14 The higher δ18O values of the UOx are attributed to 16O loss during calcination/reduction of the MDU and 18O enrichment in the resulting UOx. However, even though MDU and UOx products have different absolute δ18O values, the difference in the δ18O values of eMDU and dMDU, and UOx’s made from eMDU and dMDU (without steam), remains constant.

Table 1. Bulk Fluorination δ18O Values, Oxygen Content, and Fluorination Yield of UOx from MDU Reduced with and without Steam Presenta.

sample day and average δ18O (‰) δ18O error (2σ) O (wt %) O error (2σ) yield (%)
UOx-dMDU w/H2O day 1 –15.14 0.71 15.71 1.49 132.54
day 2 –14.70 0.62 15.32 1.64 129.27
average –14.92 0.67 15.52 1.57 130.91
UOx-dMDU w/o H2O day 1 9.91 0.71 15.41 1.49 130.08
day 2 9.91 0.62 15.05 1.64 126.99
average 9.91 0.67 15.23 1.57 128.54
UOx-eMDU w/H2O day 1 –12.06 0.71 14.74 1.49 124.40
day 2 –11.40 0.62 14.31 1.64 120.78
average –11.73 0.67 14.53 1.57 122.59
UOx-eMDU w/o H2O day 1 19.50 0.71 14.78 1.49 124.71
day 2 19.01 0.62 14.81 1.64 124.97
average 19.26 0.67 14.80 1.57 124.84
Open in a new tab
a

Errors are reported as 2σ.

Figure 1.

Figure 1

Open in a new tab

Comparison of bulk δ18O values of UOx samples synthesized from dMDU and eMDU. UOx reduced with steam is in black, while UOx reduced without steam is in red. The δ18O value of the steam is represented by a blue dashed line. During reduction, if steam is present, the δ18O values of UOx samples trend toward those of the steam. If the reduction is completed without steam, then the δ18O values increase due to the volatilization of the lighter 16O.

These findings are similar to those from previous dry atmosphere studies executed by Klosterman et al., where they observed that calcination and reduction of ADU also enriched the final products in 18O, thus increasing the reported δ18O values.13 The enrichment of the calcined and reduced products in 18O is interpreted as the result of preferential loss of 16O, matching similar isotope effects observed with silicate and magnesian minerals.13,26−29

In contrast to the dry atmosphere reductions, UOx materials reduced with steam have lower δ18O values, suggesting significant oxygen isotope exchange between the U-oxide and the water vapor during reduction. The UOx samples made from dMDU and eMDU in the presence of steam have δ18O values of −14.92 ± 0.67‰ and −11.73 ± 0.67‰, respectively (Table 1 and Figure 1). Interestingly, these samples did not maintain the 10‰ difference observed between the starting MDU hydrates and the UOx reduced in a dry atmosphere. The similar δ18O values between UOx samples reduced with steam suggest that significant UOx-steam oxygen exchange is occurring, especially considering the δ18O: −25.11‰ value of the steam.

Another interesting finding is that UOx reduced in an atmosphere containing steam had higher δ18O values compared to that of the steam. These differences can be explained through two potential reactions that occurred during the reduction of the MDU to UOx. First, the starting material, MDU, has to lose oxygen and water during the reduction reaction. In this regard, Klosterman et al. previously showed that the reduction of ammonium diuranate (ADU) to UOx in an inert atmosphere raised the δ18O relative to the starting material,13 which could explain the higher δ18O values compared to that of the steam. In the second reaction, the UOx products exchange oxygen with the water vapor. Klosterman et al. and Assulin et al. previously studied these exchange reactions for U3O8.16,30 Both observed a decrease in δ18O relative to the steam, and both authors noted that this could be related to uranium forming more stable bonds with the lighter 16O, resulting in lower δ18O values of the calcined products at elevated temperatures (relative to the oxygen isotope composition of H2O).

In the present studies, both the reduction and exchange reactions occur, likely simultaneously. Hence, the δ18O is being raised as oxygen and water are removed from the starting MDU and the δ18O is being lowered during the exchange reaction with the water vapor. While the initial δ18O values of the MDU are unknown, ADU reductions in inert atmospheres caused the δ18O to rise by about 17‰.13 In contrast, the prior exchange studies with U3O8 showed that the δ18O was −5 to −10‰ lower than the water vapor. If both reactions are occurring, then the expected δ18O would be about +7 to +12‰ above the water vapor, which is in line with the measurements in this study. Ultimately, a combination of many isotope effects could have occurred, but the overall isotopic composition of UOx during reduction appears to be primarily governed by the exchange with steam, while all other effects are of smaller significance.

LG-SIMS Oxygen Isotope Results of Products from UO2F2 and ADU Reductions

LG-SIMS δ18O values from individual particles are plotted as a function of their 16O counts in parts S1 and S2. All materials have particle data that plot near, but outside of counting statistics-based models of homogeneity for the averaged value of data sets. The enhanced scatter of data relative to the models could reflect limits of instrument performance with respect to operating parameters and the sample matrix and/or could indicate subtle but real oxygen isotope variability at the particle level. The observation that variations in oxygen isotope ratios among uranium oxide microparticles by SIMS are larger than counting statistics-based uncertainty metrics was also noted by Tamborini et al.31 Typical 2SD uncertainties of particle data sets reported here are ±5‰, which is similar to those reported in previous SIMS studies of uranium oxide particles31,32 when the reported isotope ratios and uncertainties are converted to delta-notation.

The LG-SIMS oxygen isotope data reported here from the UO2F2 starting material particles have an averaged δ18O value of 15.2 ± 5.8‰ (Figure 2a and Table 2). Reduction of this material in a dry atmosphere had a minimal impact on the oxygen isotope composition, with the resulting UOx having a δ18O value of 10.4 ± 5.0‰. However, when the UO2F2 was reduced with 18O-depleted steam (δ18O: −25.11‰) it significantly exchanged oxygen isotopes with the steam, resulting in a δ18O value of −16.2 ± 4.2 and −19.9 ± 5.9‰ for replicates 1 and 2, respectively (Figure 2b). Note that replicate one was comprised of a mixture of UO2 with U3O8 and U4O9, while replicate two was nearly pure UO2, perhaps reflecting a relationship between the completeness of the reaction and the extent of oxygen isotope exchange with the steam.

Figure 2.

Figure 2

Open in a new tab

Comparison of LG-SIMS δ18O values of (A) UO2F2, (B) UO2F2 reduced to UOx with and without steam, and (C) ADU synthesized from UO2F2 and its subsequent calcination and reduction products. The blue line represents the δ18O value of steam for the reductions that included steam.

Table 2. LG-SIMS δ18O Values of UO2F2, ADU, and Subsequent UOx Were Reduced with and without Steam Presenta.

sample δ18O (‰) δ18O error (2σ)
UO2F2 15.2 5.8
UO2F2-UOx w/H2O (rep 1) –16.2 4.2
UO2F2-UOx w/H2O (rep 2) –19.9 5.9
UO2F2-UOx w/o H2O 10.4 5.0
UO2F2-ADU 5.5 4.5
ADU-U3O8 1.9 6.7
ADU-U3O8-UOx w/H2O –17.5 5.8
ADU-U3O8-UOx w/o H2O 3.1 7.4
Open in a new tab
a

Errors are reported as the expanded standard deviation (2σ).

Precipitation of ADU from the uranyl fluoride solution resulted in a slight depletion of 18O relative to the starting UO2F2, as the particle data have an averaged δ18O value of 5.5 ± 4.5‰. The reason for this depletion is unknown, but other studies have shown that the fractionation of oxygen isotopes during the precipitation of uranium ore concentrates is complex and time-dependent.18,33 Calcination of the ADU to U3O8 in an inert atmosphere resulted in a minimal change to the oxygen isotopic composition, as the particle data have an averaged δ18O value of 1.9 ± 6.7‰. Reduction of this material in a dry atmosphere also resulted in minimal change, with an averaged δ18O value of 3.1 ± 7.4‰. In contrast, the reduction of the U3O8 to UOx in a steam atmosphere resulted in extensive exchange of the oxygen isotopes with steam resulting in an averaged particle data set δ18O value of −17.5 ± 5.8‰.

Collectively, the LG-SIMS oxygen isotope analysis of UO2F2 and ADU reductions with steam revealed the same pattern as that observed from the bulk fluorination of MDU reduced with steam. Specifically, when steam is present, a significant exchange of the oxygen isotopes with the steam occurs, resulting in a δ18O signature near that of the steam. The uncertainties of the LG-SIMS data are much larger than that of the bulk fluorination coupled to IRMS, but this is expected because the mass of the material analyzed (a few picograms per micron-sized particle) and the corresponding signal for LG-SIMS oxygen isotope analyses are each significantly lower than the abundance of material analyzed and the corresponding signal generated during IRMS. Also of note is that the LG-SIMS and bulk fluorination measurements may be limited in accuracy due to the lack of matrix-matched reference materials to calibrate instrumentation. To date, only a single U-oxide has been qualified as a working reference material through a round robin analysis, CRM 125-A, a UO2.23 With the development of new oxygen isotope standards for UO3, U3O8, and U4O9, improved accuracy could be achieved for these particular matrices, and it may be possible to more accurately bias correct raw data from mixed oxide speciation systems that are often encountered by industrially produced materials. Based on these initial studies, if enough samples are available, then bulk fluorination provides the most precise results. Nonetheless, LG-SIMS can make measurements on limited sample sizes and on a particle-by-particle basis. LG-SIMS particle data can be compared to a counting statistics-based model of expected scatter for a homogeneous material because it allows for identifying significant outlier data that may represent a different isotope source.

Signatures of Process History through Impurities and Potential Links to Oxygen Isotopic Ratios

In addition to oxygen isotope ratios, other signatures of the process history are present within UOCs and final UOx products. For example, Chalifoux previously showed that the morphologies of UOx reduced from MDU with or without steam were qualitatively similar using a lexicon of descriptors to describe the morphology but quantifiably different using convolutional neural networks for image classification.100 Combining such information with oxygen isotope data from the same material may provide more robust assessments about the process history of materials. For example, if it can be independently determined that a material was produced by steam-assisted reduction, via morphology or other characteristics, and if steam-assisted reduction to the final UOx product also imparts the oxygen isotope signature of the steam itself (which may not be exact, due to incomplete exchange and other fractionation effects described above), then it could confidently be interpreted that such a UOx material has δ18O values related to that of the local meteoric water. If true, then δ18O values of such UOx materials could be used as a geolocational signature, which could narrow down and/or rule out where an interdicted material originated. In contrast, if it could be determined via morphology and other characteristics that UOx materials were synthesized in dry reactions with/without intermediate calcination steps, it would allow for determining that oxygen isotope values represent those of starting materials and/or fractionation effects during calcination, offering valuable nuclear forensic information.

Regarding trace element signatures of process history, LG-SIMS APM maps of 19F and 238U16O signals from particles studied here provide visual distinctions of fluorine impurity abundances and distributions among materials from UO2F2 and ADU reductions with and without steam. In addition, the measured 19F/238U16O ratios of particle populations show differences in averaged values and distributions of values between samples (Figures 3–6). To provide a frame of reference, 19F and 238U16O signals from CRM 125-A particles were mapped by LG-SIMS, since it is a UO2 reference material and presumably has low F impurities. Indeed, Figure 3a shows relatively little 19F signal from particles, and the average of measured 19F/238U16O values from the CRM 125-A particle population is 0.11. The vast majority of data fall within the counting statistics model of expected scatter for this value (red curves in the right panel of Figure 3a). However, there are some data points with higher 19F/238U16O values outside of the model bounds, which is corroborated by a small number of particles with elevated 19F values in the LG-SIMS maps (e.g., Figure 3a, leftmost panel). In contrast to CRM 125-A, as expected, the UO2F2 starting material has a high abundance of 19F measured in particles (Figure 3b). The average 19F/238U16O value of the UO2F2 particle population, 71, is over 600 times greater than that for CRM 125-A, with many data having values above the model bounds for predicted data scatter about this value (Figure 3b).

Figure 3.

Figure 3

Open in a new tab

LG-SIMS example APM maps of 19F and 238U16O signals and plots of particle 19F/238U16O ratios versus 238U16O counts for (a) CRM 125-A and (b) UO2F2 starting material. Each map is 256 × 256 pixels, and for each particle identified, all signal information from its constituent pixels was used to calculate its 19F/238U16O value and count statistics uncertainty. Note that for all measurements, 19F was counted for 5 s, and 238U16O was counted for 125 s. As such, the absolute ratios are inaccurate, but the relative comparisons between the samples are still relevant. Shown in the plots as red curves is the counting statistics-based model of expected scatter about the averaged data set value (e.g., eq 1) if a material is isotopically homogeneous.

Figure 6.

Figure 6

Open in a new tab

Box and whisker plots of all LG-SIMS APM particle 19F/238U16O data from (A) UO2F2 starting material, (B) UOx materials directly reduced from the UO2F2 starting material, (C) materials from the ADU reduction reactions, and (D) CRM 125-A. Middle bars are the average and bottoms and tops of boxes represent the 25th and 75th percentiles of data, respectively. Note that for all measurements 19F was counted for 5 s, and 238U16O was counted for 125 s. As such, the absolute ratios are inaccurate, but the relative comparisons between samples are still relevant.

LG-SIMS F impurity abundances and distributions for UOx materials that were directly reduced from the UO2F2 starting material are shown in Figure 4. Reduction without steam led to the formation of UOx particles with fluorine concentrations even higher than those of the UO2F2 starting material. The averaged 19F/238U16O of the dry-reduced UOx particle data set is 197, and the particles are significantly heterogeneous in F abundance, with many data above and below the counting statistics-based model (Figure 4a). These results match previous findings by Nizinski et al. where they measured the same samples by scanning electron microscopy–energy dispersive spectroscopy (SEM–EDS),19 and it was found that elemental F/U ratios of the dry-reduced UOx are higher than those of the UO2F2 starting material. In stark contrast, when the UO2F2 was reduced with steam, the abundance of 19F is vastly diminished in map images, and the resulting 19F/238U16O values of UOx materials are similar to, if not lower than, those of CRM 125-A (Figure 4b,c). Here, the differences in averaged 19F/238U16O values and the range of values between replicates 1 and 2 likely reflect the degree of completeness of reduction. Specifically, replicate 2 was the highest purity sample from XRD, being nearly 100% UO2, while replicate 1 converted less fully to UO2, with observed residual U3O8 and U4O9.

Figure 4.

Figure 4

Open in a new tab

LG-SIMS example APM maps of 19F and 238U16O signals and plots of particle 19F/238U16O ratios versus 238U16O counts for materials that were directly reduced from the UO2F2 starting material. See the caption of Figure 3 for other details.

LG-SIMS F impurity abundances and distributions for materials precipitated from the UO2F2 starting material as ADU, calcined as U3O8 from the ADU, and reduced to UOx from the U3O8 with and without steam, are shown in Figure 5. The initial precipitation of the UO2F2 as ADU substantially improved the purity of the product, as the averaged 19F/238U16O is 4.7 (Figure 5a). Calcination of the ADU to form U3O8 lowered the average 19F/238U16O even further to 1.1, although the range of values is similar to those of the ADU (Figure 5b). Reduction of the U3O8 to UOx did not lead to a significant change in averaged 19F/238U16O values (1.1 and 0.7 respectively, for dry and steam-assisted reductions, respectively), but the proportion of outlier data with high 19F/238U16O values was appreciably diminished. The slightly elevated averaged 19F/238U16O values of the UOx products from this processing route, relative to those of the UOx materials formed by direct UO2F2 reduction with steam, may also reflect incomplete conversion, as their compositions are each 93% UO2 and 7% U4O9, based on the aforementioned p-XRD data. However, direct reduction from UO2F2 with steam may simply lead to higher purity UOx, regardless of the conversion rate. This is evidenced by the replicate 1 data from Figure 4b, which are roughly 38% U3O8, 38% U4O9, and 24% UO2 based on p-XRD data. Although this material is significantly less converted than all other reduced UOx materials investigated, it still has a very low averaged 19F/238U16O value of 0.3.

Figure 5.

Figure 5

Open in a new tab

LG-SIMS example APM maps of 19F and 238U16O signals and plots of particle 19F/238U16O ratios versus 238U16O counts for materials from the ADU reduction reactions. See Figure 3 caption for other details.

A box and whisker plot summary of all 19F/238U16O particle data collected by LG-SIMS mapping is shown in Figure 6. When compared to the LG-SIMS averaged oxygen isotope data from the same materials shown in Figure 2, there is a positive correlation with respect to the UO2F2 starting material and the UOx materials produced via direct reduction, meaning that (1) the starting UO2F2 material and UOx reduced dry have relatively high δ18O and 19F/238U16O values, and that (2) the UOx materials reduced with steam have relatively low δ18O and 19F/238U16O values (e.g., Figures 2a,b and 6a,b). For the UOx materials reduced via the ADU reaction sequence, this correlation is not as obvious. Of these materials, the ADU has the highest averaged δ18O and 19F/238U16O values, and the reduced UOx with steam has the lowest averaged δ18O and 19F/238U16O values. However, the intermediate calcination step to U3O8 and subsequent dry reduction to UOx appear to have F impurities largely removed (e.g., relative low averaged 19F/238U16O values), even though their averaged δ18O values are still relatively high (Figures 2c and 6c).

Overall, the relationships between oxygen isotope ratios and F impurities appear to provide meaningful insights into the process history of U-O-bearing materials (e.g., Figures 2 and 6), especially when combined with p-XRD information, as it relates to the completeness of material conversion. However, there is much more to investigate if the goal is to establish an expanded and well-defined parameter space that is used for predicting the process history of unknown UOCs and nuclear fuels as they are received (e.g., interdictions or otherwise). Areas for further exploration include other types of reactions, different oxygen isotope values/ranges of materials, measuring other types of impurities, and how they are influenced by processing, for systems that are analogous to real-world UOC production.

Conclusions

Several UOx materials were synthesized from the reduction of MDU, UO2F2, and ADU with and without steam. When no steam was present, the oxygen isotopic compositions of the reacted materials resembled that of the starting material when comparisons were available. When steam was present during reduction reactions, the uranium oxides extensively exchanged oxygen with the steam, yielding oxygen isotope compositions similar to those of the steam. All UOx materials reduced with steam have slightly higher δ18O values than that of the steam. This is likely due to multiple reactions contributing to the final δ18O value, including, but not limited to, the fractionation of oxygen isotopes, as water and oxygen are removed from the MDU during reduction, and the completeness of exchange of oxygen between the water vapor and the UOx product. Oxygen isotopic measurements by bulk fluorination yielded the most precise results, while LG-SIMS was able to analyze oxygen isotopic content on a particle-by-particle basis, albeit with lower precision due to far less material analyzed (and, therefore, a lower measured signal).

Isotope mapping of F impurities by LG-SIMS provides compelling evidence to identify when steam was or was not used during processing and, subsequently, when the oxygen isotopic signature could be related to steam-assisted versus dry reduction to UOx. This demonstrates that combining multiple signatures (oxygen isotopes, speciation, impurities, and morphology) within UOC-UOx materials has the potential to reveal important nuclear forensic signatures, such as geolocation or feedstock source of material.

Acknowledgments

The MDU and UO2F2 synthesis was supported by the Department of Homeland Security (DHS) under project 2016-DN-077-ARI102. The reduction of the MDU to UOx using dry and wet atmospheres was funded by the Department of Energy’s National Nuclear Security Administration, Office of Defense Nuclear Nonproliferation Research and Development. The reduction of the UO2F2 and all subsequent LG-SIMS analysis was supported by the National Technical Nuclear Forensics Center (NTNFC) within Countering Weapons of Mass Destruction (CWMD), formerly the Domestic Nuclear Detection Office (DNDO), of the Department of Homeland Security. The bulk fluorination including the participation of Miguel Cisneros, Erik Oerter, and Michael Singleton at LLNL was conducted under the auspices of DOE Contract DE-AC52-07NA27344, release number LLNL-JRNL-855739. Through Los Alamos National Laboratory, this document is approved for unlimited release under LA-UR-23-31529. Through Pacific Northwest National Laboratory, this document is approved for unlimited release under PNNL-SA-192129.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.3c10481.

  • LG-SIMS individual particle δ18O values versus 16O counts from the UO2F2 starting material and from UOx materials reduced directly from the UO2F2 starting material and LG-SIMS individual particle δ18O values versus 16O counts from materials representing the ADU reduction reactions (PDF)

The authors declare no competing financial interest.

Supplementary Material

ao3c10481_si_001.pdf (475.7KB, pdf)

References

  1. Oliver A. J.Conversion of Natural Uranium. In Uranium for Nuclear Power: Resources, Mining and Transformation to Fuel, 1st ed.; Hore-Lacy I.. Ed.; Elsevier, 2016.
  2. Pastoor K. J.; Kemp R. S.; Jensen M. P.; Shafer J. C. Progress in Uranium Chemistry: Driving Advances in Front-End Nuclear Fuel Cycle Forensics. Inorg. Chem. 2021, 60 (12), 8347–8367. 10.1021/acs.inorgchem.0c03390. [DOI] [PubMed] [Google Scholar]
  3. Knudsen I. E.; Hootman H. E.; Levitz N. M.. A Fluid-Bed Process for the Direct Conversion of Uranium Hexafluoride to Uranium Dioxide. Nucl. Sci. Eng. 1963, 20, 259–265, 10.13182/NSE64-A19567. [DOI]
  4. Brandberg S. G. The Conversion of Uranium Hexafluoride to Uranium Dioxide. Nuclear Technology 1973, 18 (2), 177–184. 10.13182/NT73-A31286. [DOI] [Google Scholar]
  5. Lindman N. The kinetics of the elimination of fluorine from uranylfluoride/uraniumdioxide pellets. J. Nucl. Mater. 1977, 66 (1–2), 23–36. 10.1016/0022-3115(77)90130-1. [DOI] [Google Scholar]
  6. Kayzar-Boggs T. M.; Kinman W. S.; Bostick D. A.; Cardon A.; Foley R. R.; Hexel C. R.; King-Lopez R.; Lindvall R. E.; Marks N.; Sharp M. A. Exploring the use of thorium isotope compositions and concentrations as nuclear forensic signatures for uranium ore concentrates. J. Radioanal. Nucl. Chem. 2021, 327, 877–889. 10.1007/s10967-020-07534-6. [DOI] [Google Scholar]
  7. Marks N.; Borg L.; Eppich G.; Gaffney A.; Genneti V.; Hutcheon I.; Kristo M.; Lindvall R.; Ramon C.; Robel M.. Technical report on the behavior of trace elements, stable isotopes, and radiogenic isotopes during the processing of uranium ore to uranium ore concentrate; Lawrence Livermore National Lab. (LLNL): Livermore, CA (United States), 2015. [Google Scholar]
  8. Rolison J. M.; Druce M.; Shollenberger Q. R.; Kayzar-Boggs T. M.; Lindvall R. E.; Wimpenny J. Molybdenum isotope compositions of uranium ore concentrates by double spike MC-ICP-MS. Applied geochemistry 2019, 103, 97–105. 10.1016/j.apgeochem.2019.03.001. [DOI] [Google Scholar]
  9. Abbott E. C.; Brenkmann A.; Galbraith C.; Ong J.; Schwerdt I. J.; Albrecht B. D.; Tasdizen T.; McDonald L. W. Dependence of UO2 surface morphology on processing history within a single synthetic route. Radiochim. Acta 2019, 107 (12), 1121–1131. 10.1515/ract-2018-3065. [DOI] [Google Scholar]
  10. Schwerdt I. J.; Hawkins C. G.; Taylor B.; Brenkmann A.; Martinson S.; McDonald L. W. Uranium oxide synthetic pathway discernment through thermal decomposition and morphological analysis. Radiochim. Acta 2019, 107 (3), 193–205. 10.1515/ract-2018-3033. [DOI] [Google Scholar]
  11. Nizinski C. A.; Hanson A. B.; Fullmer B. C.; Mecham N. J.; Tasdizen T.; McDonald L. W. Effects of process history on the surface morphology of uranium ore concentrates extracted from ore. Miner. Eng. 2020, 156, 106457 10.1016/j.mineng.2020.106457. [DOI] [Google Scholar]
  12. Klosterman M. R.; Oerter E. J.; Chakraborty S.; Singleton M. J.; McDonald L. W. Fractionation of Oxygen Isotopes in Uranium Oxides during Peroxide Precipitation and Dry Air Calcination. ACS Earth and Space Chemistry 2021, 5 (6), 1622–1630. 10.1021/acsearthspacechem.1c00112. [DOI] [Google Scholar]
  13. Klosterman M. R.; Oerter E. J.; Deinhart A.; Chakraborty S.; Singleton M.; McDonald L. W. Oxygen Kinetic Isotope Effects in the Thermal Decomposition and Reduction of Ammonium Diuranate. ACS Omega 2021, 6 (45), 30856–30864. 10.1021/acsomega.1c05388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Pajo L.; Mayer K.; Koch L. Investigation of the oxygen isotopic composition in oxidic uranium compounds as a new property in nuclear forensic science. Fresenius J. Anal Chem. 2001, 371 (3), 348–352. 10.1007/s002160100983. [DOI] [PubMed] [Google Scholar]
  15. Klosterman M. R.; Oerter E. J.; Singleton M. J.; McDonald L. W. IV Oxygen Isotope Fractionation in U3O8 during Thermal Processing in Humid Atmospheres. ACS Omega 2022, 7, 3462–3469. 10.1021/acsomega.1c05838. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Assulin M.; Yam R.; Kossoy A.; Elish E.; Shemesh A. Equilibrium fractionation of oxygen isotopes in the U3O8–atmospheric oxygen system. ACS omega 2022, 7 (50), 47040–47047. 10.1021/acsomega.2c06148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Assulin M.; Yam R.; Grego-Shnaiderman A.; Kodosha Y. E.; Lulu-Bitton N.; Elish E.; Shemesh A. Oxygen Isotope Alterations during the Reduction of U3O8 to UO2 for Nuclear Forensics Applications. ACS Omega 2023, 8 (37), 33651–33657. 10.1021/acsomega.3c03903. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Chalifoux A. M.; Singleton M.; Cisneros M.; Oerter E. J.; Hanson A. B.; McDonald L. W. IV Challenges in correlating oxygen stable isotope ratios of hydrates on uranium ore concentrates to process waters. Radiochim. Acta 2023, 111, 613–620. 10.1515/ract-2023-0154. [DOI] [Google Scholar]
  19. Nizinski C. A.; Olson J.; Chalifoux A. M.; Kurtyka N.; Athon M.; Tenner T.; McDonald L. W. IV Identification and elemental impurity analysis of heterogeneous morphologies in uranium oxides synthesized from uranyl fluoride precursors. ACS Omega 2023, 8, 16896–16906. 10.1021/acsomega.3c00699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Chalifoux A. M.; Gibb L.; Wurth K. N.; Tenner T.; Tasdizen T.; McDonald L. W. Morphology of uranium oxides reduced from magnesium and sodium diuranate. Radiochimica Acta 2024, 112 (2), 73–84. 10.1515/ract-2023-0221. [DOI] [Google Scholar]
  21. Waterisotopes Database. 2017. http://waterisotopesDB.org (accessed 2023).
  22. Horita J.; Wesolowski D. J. Liquid-vapor fractionation of oxygen and hydrogen isotopes of water from the freezing to the critical temperature. Geochim Cosmochim Ac 1994, 58 (16), 3425–3437. 10.1016/0016-7037(94)90096-5. [DOI] [Google Scholar]
  23. Nizinski C. A.; Olson J.; Chalifoux A. M.; Kurtyka N.; Athon M. T.; Tenner T.; McDonald L. W. Identification and Elemental Impurity Analysis of Heterogeneous Morphologies in Uranium Oxides Synthesized from Uranyl Fluoride Precursors. Acs Omega 2023, 8 (19), 16896–16906. 10.1021/acsomega.3c00699. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Oerter E. J.; Singleton M.; Pili E.; Klosterman M. R.; Shemesh A.; Agrinier P.; Deinhart A.; Yam R.; Assulin M.; Elish E.; et al. The oxygen stable isotope composition of CRM 125-A UO2 standard reference material. Appl. Geochem. 2022, 146, 105470 10.1016/j.apgeochem.2022.105470. [DOI] [Google Scholar]
  25. Fauré A.-L.; Rodriguez C.; Marie O.; Aupiais J.; Pointurier F. Detection of traces of fluorine in micrometer sized uranium bearing particles using SIMS. Journal of Analytical Atomic Spectrometry 2014, 29 (1), 145–151. 10.1039/C3JA50245G. [DOI] [Google Scholar]
  26. Scott S.; Naes B.; Christian J.; Foley B.; Tenner T.; Kuhne W.; Wurth K.; Shehee T.; Lawson S.; Ajo H.. New Particle Working Standards for NWAL Particle Laboratory Calibration and Quality Control-Operational Engineering for an Aerosol-Based Production Platform for the Synthesis of Plutonium-Containing Reference Particulate Materials; Savannah River Site (SRS); Savannah River National Laboratory: Aiken, SC (United States), 2023. [Google Scholar]
  27. Sharp Z.Principles of Stable Isotope Geochemistry, 2nd ed.; 2017, 10.25844/h9q1-0p82. [DOI] [Google Scholar]
  28. Davis A. M.; Hashimoto A.; Clayton R. N.; Mayeda T. K. Isotope mass fractionation during evaporation of Mg2Si04. Nature 1990, 347 (6294), 655–658. 10.1038/347655a0. [DOI] [Google Scholar]
  29. Girard J.-P.; Savin S. M. Intracrystalline fractionation of oxygen isotopes between hydroxyl and non-hydroxyl sites in kaolinite measured by thermal dehydroxylation and partial fluorination. Geochim. Cosmochim. Acta 1996, 60 (3), 469–487. 10.1016/0016-7037(95)00421-1. [DOI] [Google Scholar]
  30. Clayton R. N.; Mayeda T. K. Kinetic isotope effects in oxygen in the laboratory dehydration of magnesian minerals. J. Phys. Chem. A 2009, 113 (10), 2212–2217. 10.1021/jp808621n. [DOI] [PubMed] [Google Scholar]
  31. Klosterman M. R.; Oerter E. J.; Singleton M. J.; McDonald L. W. IV Oxygen Isotope Fractionation in U3O8 during Thermal Processing in Humid Atmospheres. ACS omega 2022, 7 (4), 3462–3469. 10.1021/acsomega.1c05838. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Tamborini G.; Phinney D.; Bildstein O.; Betti M. Oxygen isotopic measurements by secondary ion mass spectrometry in uranium oxide microparticles: A nuclear forensic diagnostic. Anal. Chem. 2002, 74 (23), 6098–6101. 10.1021/ac0259515. [DOI] [PubMed] [Google Scholar]
  33. Pajo L.; Tamborini G.; Rasmussen G.; Mayer K.; Koch L. A novel isotope analysis of oxygen in uranium oxides: comparison of secondary ion mass spectrometry, glow discharge mass spectrometry and thermal ionization mass spectrometry. Spectrochimica Acta Part B: Atomic Spectroscopy 2001, 56 (5), 541–549. 10.1016/S0584-8547(01)00219-1. [DOI] [Google Scholar]
  34. Klosterman M. R.; Oerter E. J.; Chakraborty S.; Singleton M. J.; McDonald L. W. IV Fractionation of oxygen isotopes in uranium oxides during peroxide precipitation and dry air calcination. ACS Earth and Space Chemistry 2021, 5 (6), 1622–1630. 10.1021/acsearthspacechem.1c00112. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao3c10481_si_001.pdf (475.7KB, pdf)

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES