Oxygen Isotope Alterations during the Reduction of U3O8 to UO2 for Nuclear Forensics Applications

The fabrication of UO2 from U3O8 is an essential reaction in the nuclear fuel cycle. The oxygen isotope fractionation associated with this reaction has significant implications in the general field of nuclear forensics. Hence, the oxygen isotope fractionation during the reduction of U3O8 to UO2 was determined in the temperature range of 500–700 °C and for a duration of 2 to 6 h under a high-purity H2 atmosphere. Three U3O8 samples, possessing a different oxygen isotopic composition, were used to investigate key parameters involved with the fractionation during the reduction process. All UO2 products did not maintain the original isotope composition of the starting U3O8 under all conditions. The results show that the system UO2–H2O attains isotope equilibrium at 600 °C, provided the reduction process lasts at least 4 h or more. At 600 °C, UO2 was isotopically depleted by 2.89 ± 0.82‰ compared to the U3O8 from which it was produced. We find that the H2O formed during the reduction plays a major role in determining the final δ18O of UO2 prepared from U3O8. The isotope equilibrium of the system UO2–H2O at 600 °C was calculated, indicating that δ18O of the H2O was enriched by about 11‰ relative to the UO2 due to the uranium mass effect. These findings could potentially have important implications for nuclear forensics, as they provide a new method for determining the history of UO2 samples and tracing back their production process.


INTRODUCTION
Uranium dioxide pellets (UO 2 ) are a major end-product in the nuclear fuel cycle, beginning with uranium ores.The illicit trafficking of uranium from one of the production stages has occurred in the past, attracting significant scientific attention in nuclear forensics.Therefore, an ongoing effort has been dedicated to developing new analytical techniques, focusing on finding novel signatures to allow improved and credible attribution of uranium compounds found outside of regulatory control.
UO 2 in the reduction process of U 3 O 8 to UO 2 is usually applied at high temperatures (above 500 °C) under a reducing environment (e.g., pure H 2 or a mixture of H 2 /Ar), 1 and the chemical reaction is given in eq 1.
The mechanism and kinetic model for this reaction at 510 °C identified the rate-limiting step as the desorption of water vapors from the external surface of the U 3 O 8 particles while conversion proceeds. 2The study of the diffusion of oxygen in the natural uraninite−H 2 O at a temperature range of 50−700 °C suggested an initial extremely fast-path diffusion mechanism that overprints the oxygen isotopic composition of the entire crystal, independent of temperature, and a slower volume-diffusive mechanism dominated by defect clusters. 3lume diffusion was also considered the primary process controlling oxygen diffusion in polycrystalline and singlecrystalline UO 2 in the temperature range of 605−750 °C. 4 Depth profiling beneath the UO 2 pellet surface at room temperature indicates oxygen diffusion into the UO 2 lattice, while water species diffusion occurs along grain boundaries, behaving as high-diffusivity paths. 5he use of the oxygen isotopic composition ( 18 O/ 16 O� expressed in δ notation as δ 18 O vs VSMOW) as a signature for geolocation 6−8,11−13 and the process history 9,10,15−18 of uranium oxides for nuclear forensics applications 14 have gained much interest in recent years.The δ 18 O variation of naturally occurring uranium ore minerals is in the range of −32 to +11‰. 8−6 On the other hand, it has been documented that uranium oxides of different geographic origins have significantly different 18 O/ 16 O ratios, which correlate with local rainwater isotopic composition. 6he comparison of several oxygen extraction methods from UO 2 , U 3 O 8 , and UO 3 compounds has revealed that different amounts of water molecules present in the uranium oxides must be removed, as they interfere with the isotope measurement 11 and exchange reactions with humid atmospheres, affecting the oxygen fractionation. 15Klosterman et al. 15 synthesized U 3 O 8 from uranium peroxide at 300−1000 °C, reporting fractionation of about −22‰ up to about −5‰, respectively, between the oxide and atmospheric oxygen with a retrograde isotope effect. 16The exposure of U 3 O 8 and UO 2 to humidity in an oxidizing atmosphere showed that hydrated uranium oxide grows as a secondary mineral (metaschoepite) on aged U 3 O 8 and UO 2. It was suggested that δ 18 O values of the metaschoepite hydration water are likely to reflect those of the water vapor to which the sample was exposed. 12,13Thus, the oxygen isotopes of metaschoepite mineral hydration water may retain information on the mineral formation location.
Hattori and Halas 19 theoretically investigated the fractionation factors α (defined as the isotopic ratio between two compounds) in the systems UO 2 −H 2 O and UO 3 −H 2 O over a temperature range of 0−1000 °C, finding that uraninite is consistently depleted with 18 O compared to its associated H 2 O across this temperature range.Yong-fei 20 developed a theoretical model to calculate the oxygen isotopic fractionation factors of metal oxides, including uraninite, over the same temperature range.However, the two models 19,20 produced different results, with a difference of approximately 10‰ in 1000 ln α (UO 2 −water) at temperatures up to 300 °C and a difference of ∼2‰ at higher temperatures between 300 and 600 °C.Fayek and Kyser 9 found fractionation factors similar to those reported by Hattori and Halas 19 at high temperatures but closer to those predicted by Yong-fei 20 at low temperatures.
This study focuses on the oxygen isotope change resulting from the manufacturing processes of UO 2 from U 3 O 8 .UO 2 powders were synthesized in the temperature range of 500− 700 °C for 2−6 h under high purity H 2 to follow the process of isotope partition among U 3 O 8 as the source material, UO 2 , and H 2 O as products of the reduction reaction.

MATERIALS AND METHODS
Three U 3 O 8 samples with different oxygen isotopic compositions were used as starting materials for this study.Two were synthesized from uranyl nitrate hydrate (UNH).U 3 O 8 -I has a δ 18 O value of 10.68 ± 0.27‰, and U 3 O 8 -II has a δ 18 O value of 8.52 ±0.22‰.These two materials were synthesized by changing the cooling rate to achieve the difference in their δ 18 O values. 17The third sample, U 3 O 8 -III, has a δ 18 O value of 4.75 ± 0.47‰ and is a natural uranium commercial U 3 O 8 material purchased from CETAMA, France (commercially known as "Chanterelle"), usually used as a calibration standard for elemental impurities in uranium.
The samples (∼100 mg each) were weighed in an alumina crucible and placed at room temperature in a stainless-steel furnace before applying a vacuum (∼10 −5 Torr) for 12 h.The samples were calcined for 2, 4, and 6 h at a temperature range of 500−700 °C under high purity (99.999%)H 2 1 atm, followed by cooling the reactor to room temperature via shutting down the furnace.The three U 3 O 8 samples were placed together in the reactor (the UO 2 products are marked as samples 1UO 2 −27UO 2 in Table 1).One set of experiments was conducted where each U 3 O 8 sample was placed alone at 600 °C for 4 h (samples 31UO 2 −33UO 2 were those placed separately; Table 1).Table 1 presents all experiments' temperatures, samples, and experimental conditions calcined time.
XRD analysis (Rigaku, Ultima III) was performed on samples weighing several milligrams under an atmospheric environment by continuous scanning at 40 kV/40 mA in the range of 10−80°at a rate of 2°/min.The analyzed samples were the starting materials (U 3 O 8 -I, U 3 O 8 -II, and U 3 O 8 -III) and those prepared at 500 °C; 2 h, 500 °C; 4 h, and 600 °C; 2 h.
Oxygen isotopic analyses of the U 3 O 8 and UO 2 samples were conducted using an isotope ratio gas-chromatographymass spectrometer (irmGCMS, Thermo Scientific Delta Plus Advantage) and an IR CO 2 laser (10.6 μm, New Wave Research 25 W).The method is described in detail elsewhere. 14,17,18,21

XRD.
The XRD diffractograms of the samples prepared at 500 °C for 2 h present a mix of UO 2 and U 3 O 8 (Figure 1) phases, while samples prepared at 500 °C for 4 h and 600 °C for 2 h present a single UO 2 phase (Figure 2).The results (Table 1) will be evaluated according to the two groups of experiments.The first is the reduction of a single U 3 O 8 sample in the reduction furnace at 600 °C for 4 h, and the second is where the three original U 3 O 8 samples were reacted together at different times and different temperatures.δ 18 O of UO 2 obtained from U 3 O 8 of the first group is lower by 2.1, 3.7, and 3‰ than the original values for samples 31UO 2 , 32UO 2 , and 33UO 2 , respectively.The results from the second group show a general trend of conversion to a common isotope value, which is unique to each temperature.The difference among samples is larger at a reaction time of 2 h and decreases with increasing reaction time.The samples prepared from U 3 O 8 -I and U 3 O 8 -II did not retain the original δ 18 O values and showed a similar trend toward depleted values, while the UO 2 prepared from U 3 O 8 -III presented an opposite trend toward a small enrichment.At a calcination temperature of 500 °C for 4 h (Figure 4), none of the samples retained the original δ 18 O values, showing similar trends toward depletion.At 600 and 700 °C, the δ 18 O values obtained for all three samples showed enriched and similar end values.The δ 18 O values of the samples that were reduced at 500, 600, and 700 °C for 6 h (Figure 5) exhibited identical values within the standard deviations, with U 3 O 8 -I and U 3 O 8 -II being 6 and 4‰ depleted, respectively, from the original U 3 O 8 values.The δ 18 O value of sample U 3 O 8 -III converged to the same value as that of U 3 O 8 -I and U 3 O 8 -II, which was very close to its original value.

DISCUSSIONS
The conversion of U 3 O 8 to UO 2 under a H 2 atmosphere involves releasing water from the U 3 O 8 (eq 1).The conversion process was studied by Pijolat et al. 2 and Alfaro et al. 1 in an open system at high temperatures under variable flow and variable mixtures of gases, H 2 /N 2 /He, and water vapor.Full conversion to UO 2 within several minutes up to 1 h was achieved under flow conditions.In contrast to the open system flow experimental configurations, our reduction of U 3 O 8 was carried out within a closed system, maintaining a constant pressure of high purity H 2 and a significant molar excess of H 2 (2.14 mmol relative to the U 3 O 8 0.11 mmol) at temperatures of 500, 600, and 700 °C.We attribute the partial conversion of U 3 O 8 to UO 2 at 500 °C for 2 h to the difference between the closed-versus open-flow experimental setups.UO 2 does not retain the original δ 18 O value of the U 3 O 8 starting materials at all temperatures and reaction times.Nevertheless, at high temperatures (≥600 °C) and longer reaction times (≥4 h), δ 18 O converges to a common value when the three starting materials are reacted together and fully converted, indicating the establishment of an isotope equilibrium between the reaction products, UO 2 and H 2 O. Thus, the similar δ 18 O values obtained from the three different original materials represent an isotope equilibrium with the generated mixture of the three waters, and therefore, a UO 2 − H 2 O fractionation factor cannot be deduced from combined material experiments.
A different set of reduction experiments, where samples 31UO 2 −33UO 2 were placed separately at 600 °C for 4 h, was conducted to resolve this problem.It allows calculating the oxygen isotopic composition of the formed water through mass balance considerations and the fractionation factor between U 3 O 8 and UO 2 .The isotopic composition of the water will be calculated in the next section.

Oxygen Isotopic Composition of H 2 O.
The oxygen isotope mass balance can be used to calculate the oxygen isotopic composition of the formed water molecule, considering a closed system case where a single sample is converted   from U 3 O 8 to UO 2 under an H 2 atmosphere (samples 31UO 2 , 32UO 2 , and 33UO 2 ).The isotope mass balance equation is given by where m represents the mass of the material indicated in the parenthesis, and The calculated δ 18 O (Hd 2 O) values for samples 31UO 2 , 32UO 2 , and 33UO 2 are 16.81, 19.57, and 13.56‰, respectively.
We further apply the mass balance calculation to a mixture of 3 samples to test our assumptions and calculations.Samples 13UO 2 , 14UO 2 , and 15UO 2 were placed together in the reactor and prepared under the same temperature, time, and pressure conditions as the single sample reaction above (Table 2).The final δ 18 O (Hd 2 O) , which results from mixing three different waters, is calculated using eq 4, which originates from solving a set of 3 separate mass balance equations.The result, 16.64‰, agrees well if we consider this water to be a mixture of the previously calculated waters of samples 31UO 2 , 32UO 2 , and 33UO 2 .Further, based on eq 4, in a closed system containing water molecules with an average δ 18 O value of 16.64 ± 3.01‰, the theoretical expected δ 18 O of the UO 2 samples is 5.14‰.The measured δ 18 O values of samples 13UO 2 , 14UO 2 , and 15UO 2 were similar: 5.34 ± 0.34, 5.54 ± 0.27, and 5.55 ± 0.13‰, respectively.We conclude that the final δ 18 O of UO 2 prepared from U 3 O 8 is determined by isotope equilibrium between the products of the reduction reaction, UO 2 , and H 2 O.

Isotope Fractionation.
The XRD measurements show that the initial U 3 O 8 was fully converted to UO 2 and H 2 O in our analytical setup, except for the 500 °C and 2 h reaction.Therefore, the fractionation factor between U 3 O 8 and UO 2 cannot be calculated because these two phases were not sampled along the reaction path and do not co-exist in isotope equilibrium at the final stage.It is interesting to note that although the two phases, U 3 O 8 and UO 2 co-exist, the final δ 18 O converges to a similar value.This may indicate that the isotope exchange between U 3 O 8 , UO 2 , and H 2 O is faster than the reduction reaction.
However, the difference in δ 18 O (Δ 18 O) between UO 2 and H 2 O can be calculated (eq F035). (F035) Equation F035 was applied to the average δ 18 O values of samples 13UO 2 , 14UO 2 , and 15UO 2 (5.48 ± 0.12‰) and the average δ 18 O value of the calculated formed water (16.64 ± 3.01‰).The calculated Δ 18 O at 600 °C in this work is −11‰.This result, where the water is isotopically enriched relative to the solid, can be explained by the uranium "mass effect" (i.e., the tendency of uranium to hold the lighter atoms because of energy considerations) at the first steps of the reaction, where oxygen is being released from U 3 O 8 , to form the UO 2 and the enriched H 2 O.Our experimental results show depletion values relative to the starting materials, which are further supported by the theoretically calculated Δ 18 O at 600 The average Δ 18 O value is 2.89 ± 0.82‰ (n = 3), suggesting that the UO 2 products are depleted by ∼ 2.89‰ relative to the initial U 3 O 8 at 600 °C.

CONCLUSIONS
Three U 3 O 8 samples with different starting δ 18 O values were reduced to UO 2 under a high-purity H 2 atmosphere at a temperature range of 500−700 °C for 2, 4, and 6 h.We find that the final δ 18 O of UO 2 prepared from U 3 O 8 is determined by the reaction of the formed H 2 O and the synthesized UO 2 .The system UO 2 −H 2 O reaches isotope equilibrium at 600 and 700 °C in the longer reduction times (≥4 h).UO 2 products are isotopically depleted relative to the U 3 O 8 from which they were formed by 2.89 ± 0.82‰ at 600 °C.The H 2 O formed during the reaction is enriched by about 11‰ relative to the UO 2 due to the uranium mass effect.In nuclear forensic investigations, a crucial aspect revolves around understanding the relationship between materials seized outside of regulatory control.Hence, our findings can serve as an additional tool to shed light on the fabrication process and facilitate material linkage when U 3 O 8 and UO 2 are present on the scene.However, additional research is required to comprehensively comprehend the mechanisms underlying the observed isotope fractionation and to fully exploit this technique in the field of nuclear forensics.

■ AUTHOR INFORMATION
heated overnight at 80 °C under a high vacuum.Prefluorination was performed thrice for the entire cell with 80 Torr of BrF 5 .Samples were reacted by laser heating in a 90 Torr of BrF 5 atmosphere.The liberated oxygen was purified by liquid nitrogen traps, concentrated on a 5 Å molecular sieve, cooled in liquid nitrogen, and transferred to the mass spectrometer through a gas chromatograph column for isotope measurement in a continuous flow mode.The international SiO 2 standard NBS-28 (δ 18 O = 9.58‰)22 was used for consistency and calibration in each batch.The measured values are expressed in δ-notation in peril, relative to Vienna Standard Mean Ocean Water (VSMOW).The long-term standard deviation (SD) for NBS-28 was 0.42‰.All the samples were run at least in triplicate, and the SD is reported for each sample.

Figure 1 .
Figure 1.XRD diffractograms of the samples prepared at 500 °C for 2 h showing a mixed phase of UO 2 and U 3 O 8 .

Figure 2 .
Figure 2. XRD diffractograms of the samples prepared at ≥500 °C for 4 h showing a complete reduction to a single UO 2 phase.

3 . 2 . δ 18 O of U 3 O 8 .
The average δ18 O values of the starting materials (U 3 O 8 ) and the UO 2 samples (3−5 measurements for each sample), prepared at the temperatures range of 500− 700 °C for 2−6 h, and of NBS-28 samples, are presented in Table1.The standard deviations (SD) are within the size of the symbols in the figure.The δ 18 O values of UO 2 samples as a function of calcination time are plotted for each calcination temperature in Figures3−5.The δ 18 O precision of the UO 2 is identical to the routinely measured NBS-28 standards.

Figure 3 .
Figure 3. δ 18 O (in ‰ relative to VSMOW) values for the starting materials at room temperature (U 3 O 8 at 25 °C) and UO 2 samples at 500 °C for 2 to 6 h.The SD is within the symbol size.

Figure 4 .
Figure 4. δ 18 O (in ‰ relative to VSMOW) values for the starting materials (red, blue and black triangles) at room temperature (U 3 O 8 at 25 °C) and UO 2 samples at 600 °C for 2−6 h from the single sample experiments (red, blue, and black empty circles) and from the group reduction experiments (red, blue, and black lines).The SD is within the symbol size.

Figure 5 .
Figure 5. δ 18 O (in ‰ relative to VSMOW) values for the starting materials at room temperature (U 3 O 8 at 25 °C) and UO 2 samples at 700 °C for 2−6 h.

Table 1 .
188O (in ‰ Relative to VSMOW) Values of U 3 O 8 and UO 2 Samples Prepared at Different Temperatures 22iefly, U 3 O 8 samples (1026−1651 μg), UO 2 samples (1240−1700 μg), and SiO 2 samples (NBS-28, standard material for quality check and calibration; 260−450 μg) were placed in nickel cups in a stainless-steel chamber and a The highlighted row corresponds to the δ 18 O values of the starting U 3 O 8 samples.bNBS-28 has an assigned isotope value of 9.58 ± 0.09‰ as an international standard.22 18O is its oxygen isotopic composition.The values of the parameters δ 18 O (Ud 3 Od 8 ) , m (Ud 3 Od 8 ) , and δ 18 O (UOd 2 ) in eq 2 are measured, while the values of the parameters m (UOd 2 ) and m (Hd 2 O) are calculated to determine δ 18 O (Hd 2 O) .Table2summarizes the first step of calculations, which yields the mass of the produced UO 2 and formed H 2 O.

Table 3
18mmarizes the calculated δ18O of the total formed H 2 O. : δ 18 O f �δ18O of the final mixture of the formed water molecules (‰).m f �the mass of the total formed water molecules (mg).δ18Oi /δ 18 O iii /δ 18 O iii �δ 18 O of U 3 O 8 -I, U 3 O 8 -II, and U 3 O 8 -III, respectively (‰).m i /m ii /m iii �mass of U 3 O 8 -I, U 3 O 8 -II, and U 3 O 8 -III, respectively (mg). where

Table 2 .
Calculation of the UO 2 and Formed H 2 O Mass

Table 3 .
20lculated δ18O of the Total Formed H 2 O °C, published by Hattori and Halas19and Yong-fei20(−6.4 and −8‰, respectively).The data in Table2also allow estimating the isotopic difference between the initial U 3 O 8 and the final UO 2 (eq 6).