Oxygen Isotope Fractionation in U3O8 during Thermal Processing in Humid Atmospheres.

The incorporation of oxygen isotopes from water into uranium oxides during industrial processing presents a pathway for determining a material's geographical origin. This study is founded on the hypothesis that oxygen isotopes from atmospheric water vapor will exchange with isotopes of oxygen in solid uranium oxides during thermal processing or calcination. Using a commonly encountered oxide, U3O8, the exchange kinetics and equilibrium fractionation with water vapor (in a concentration range of 50-55% relative humidity) were investigated using processing temperatures of 400, 600, and 800 °C. In an atmosphere containing only water vapor diluted in N2, oxygen isotope equilibration in U3O8 occurred within 12 h at 400 °C and within 2 h at 600 and 800 °C. Fractionation factors (1000lnα, U3O8-H2O) between the water and oxide were -12.1, -11.0, and -8.0 at 400, 600, and 800 °C, respectively. With both humidity and O2 present in the calcining atmosphere, isotopic equilibration is attained within 2 h at and above 400 °C. In this mixed atmosphere, which was designed to emulate Earth's troposphere, isotopes are incorporated preferentially from water vapor at 400 °C and from O2 at 600 and 800 °C. Rapid and temperature/species-dependent isotope exchange also elucidated the impact of retrograde exchange in humid air, showing a shift from O2-dependent to H2O-dependent fractionation as U3O8 cooled from 800 °C. These results confirm that uranium oxides inherit oxygen isotopes from humidity during thermal processing, illuminating an important mechanism in the formation of this forensic signature.

Introduction

The ubiquitous presence of oxygen bound to uranium during processing on the front end of the nuclear fuel and weapons cycles establishes a forensic significance for the stable isotopes of oxygen. Indeed, oxygen isotopes in uranium oxides have been identified in recent studies to provide information relevant to geolocation and processing history.[1,2] In the precipitation of uranyl peroxide from water of a known oxygen isotope composition, the precipitate was shown to incorporate oxygen isotopes from the aqueous solution during mineralization with a consistent fractionation.[1] When this compound was then calcined in an atmosphere of dry air, the original signature from process water was lost, and oxygen isotopes from atmospheric oxygen, O2, were rapidly incorporated. Another study showed that in the absence of an exchanging atmosphere (i.e., inert), the thermal decomposition of ammonium diuranate (ADU) between 400 and 800 °C was associated with a consistent loss of 16O.[2] While these studies highlight the potential for the use of oxygen isotopes as a forensic signature of industrially processed uranium oxides, one important atmospheric component was absent during the calcining and thermal decomposition stages: water vapor (i.e., humidity).

Water vapor can be present during many thermal processing steps in uranium production. On the front end of the fuel and weapons cycle, natural air containing a component of water vapor (at a natural concentration generally referenced by percent relative humidity, RH) may be circulated through furnaces during the calcination of yellowcakes to uranium oxides.[3] This procedure can be performed at a variety of temperatures, generally producing UO (3 ≤ x ≤ 3.5) below 600 °C and U3O8 above 600 °C.[4] The latter compound (U3O8) is preferred by many facilities because of its enhanced transportability and reduced reactivity.[5−7] As mentioned previously, laboratory-scale experiments following oxygen fractionation during calcination in dry air have been performed which showed a rapid incorporation of oxygen isotopes from atmospheric O2 between 300 and 1000 °C.[1] The oxygen isotope compositions found in these lab-synthesized oxides were generally more enriched in 18O than samples measured from industrial facilities.[8,9] Tropospheric oxygen has a known and globally stable isotope composition (δ18O = +23.8 ± 0.3‰ vs VSMOW: Vienna Standard Mean Ocean Water, defined below), which is significantly enriched in 18O when compared to meteoric water sources (generally measured with δ18O values between −30 and 0‰).[10,11] Therefore, we hypothesize that the discrepancy in isotope ratios between laboratory and industrial oxides could be a result of the presence of humidity in industrial calcining atmospheres, which would exchange oxygen isotopes with uranium oxides at elevated temperatures causing decreased 18O concentrations.

Water vapor in the form of steam is also widely utilized in fuel fabrication routes after uranium isotope enrichment of the gaseous hexafluoride (UF6). Reconversion of the gas back into a solid oxide is performed using either a wet route, in which UF6 is first dissolved into an aqueous solution prior to precipitation, or a dry route, in which UF6 is hydrolyzed by steam to solid UO2F2. The removal of fluoride impurities from the precipitated or hydrolyzed products from either route is greatly enhanced by the addition of steam during thermal decomposition and final reduction to UO2.[12] The presence of water vapor in these thermal processing steps again lends credence to the potential for exchange with the solid oxide, creating an oxygen isotope signature inherited from the process water. Because U3O8 is the intermediate product in the formation of UO2, understanding the exchange kinetics and fractionation of the former compound with oxygen from water vapor is important.[13]

The purpose of this study was to determine first whether oxygen from water vapor (as humidity) is incorporated and exchanged with an industrially relevant uranium oxide—U3O8—during thermal processing or calcination. Equilibrium fractionation and exchange kinetics were investigated by calcining U3O8 samples in an atmosphere containing only water vapor (diluted in N2) between 400 and 800 °C at various time intervals. Identical experiments were then conducted using a simulated industrial atmosphere, consisting of humidity with a known oxygen isotope composition diluted in dry air to determine whether a preferential or competitive fractionation occurs between water and O2. With these new insights into the rates and extents of oxygen incorporation into uranium oxides from more practical atmospheres, a better understanding of the oxygen isotope signature in real-world nuclear materials can be formulated.

Experimental Section

Synthesis

The precipitation and structural characterization of ADU have been previously reported.[2] Briefly, ammonia was introduced into a 0.1 mol·L–1 solution of uranyl nitrate (δ18Owater = −15.6‰) at 65 °C until a pH of 7.5 was attained. After a 30-min settling period, the precipitate was filtered and washed with deionized water. The ADU was dried at 80 °C for 16 h and then finely ground as a slurry with n-pentane using a zirconia mortar and pestle before storage in a vacuum desiccator.

Two separate U3O8 compounds were calcined from ADU which served as starting materials for aging experiments: U3O8_d was made by calcining ADU with vapor from H2O_d diluted in N2 (50–55% RH, discussed in further detail below) for 72 h at 800 °C; U3O8_e was made by calcining ADU with vapor from H2O_e diluted in N2 (50–55% RH) for 72 h at 800 °C. Phase purity of each U3O8 compound was confirmed via powder X-ray diffraction analysis using a Bruker D2 Phaser with a Cu Kα X-ray source (Figure S1). Scans were performed from 10 to 90° (2θ), at increments of 0.02 and 0.5 s per step.

Water Samples and Equipment

Two isotopically distinct water sources were used for vapor exchange experiments: H2O_d (liquid) with δ18O = −16.47 ± 0.04‰ (n = 3) and H2O_e (liquid) with δ18O = +1.0 ± 0.1‰ (n = 3). Humidity was introduced from either of these water sources using a setup similar to that employed by Oerter et al.[14] Depending on the type of exchange experiment, either ultrahigh purity (UHP) N2 or dry compressed air (20% O2 in N2, δ18OO2 = +23.2 ± 0.2‰, measured previously and reported in Ref (1)) was diverted through two separate gas flowmeters (Across International). One flowmeter directed the gas through a 20 L carboy at 22 °C within which a 36 cm length of water vapor permeable Accurrel PP V8/2HF polypropylene membrane tubing (8.65 mm outside diameter, 1.55 mm wall thickness, and 0.2 μm nominal porosity; Membrana GmbH, Wuppertal, Germany) was submerged. This wet gas line was then reconnected with a dry gas line emanating from the other flowmeter. The combined line was directed into a tube furnace (STF 1200, Across International). A hygrometer probe (± 5% RH accuracy; Linkstyle) was installed after the exit tubing of the furnace (setup shown in Figure ). Gas flowrates were kept identical for all experiments, with a 120 mL·min–1 flow through the wet line diluted with a dry 80 mL·min–1 flow. This flowrate generated a consistent RH at 22 °C between 50 and 55% as measured with the outlet probe.

Figure 1

Setup for calcination experiments. Water vapor diffuses through membrane tubing submerged in the reservoir and is subsequently diluted with a second stream of dry gas. The RH is controlled by flowmeters for each gas stream.

Calcination Experiments

Before each calcination experiment, the tube furnace was equilibrated in the designated atmosphere at the calcination temperature for at least 12 h with no samples present. The first set of experiments consisted of calcining U3O8_d in an atmosphere of H2O_e carried by N2 (50–55% RH). Each U3O8_d starting material was split into three fractions of approximately 30–50 mg and placed in platinum crucibles. The crucibles, supported by an alumina boat, were loaded into the tube furnace at room temperature. The atmosphere in the furnace was purged for at least 30 min with a wet flow set at 120 mL·min–1 and dry N2 flow at 500 mL·min–1. The dry flow was then set to 80 mL·min–1 for approximately 30 min to attain a RH between 50 and 55%. The furnace was then heated to 400, 600, or 800 °C at a ramp rate of 20 °C·min–1. After each designated heating time interval, the wet flow was ceased, and the dry flow increased to 500 mL·min–1 N2 while still at temperature. After a 10-min flushing, the furnace was shut off and rapidly cooled (cool-down rate illustrated in Figure S2) to room temperature. Each sample replicate was weighed (to an accuracy of ± 0.1 mg) before and after each experiment to confirm that the U3O8 had not been oxidized or hydrated (i.e., no weight change).

The second set of experiments used U3O8_e as the starting oxide, which was calcined in an atmosphere of H2O_d carried by air from the dry cylinder (50–55% RH). Triplicate samples were introduced into the furnace, and the atmosphere was equilibrated in the same manner as described above, with the use of air instead of N2. After each calcination between 1 and 72 h at 400, 600, or 800 °C, the furnace atmosphere was switched over to N2 at 500 mL·min–1 and flushed for 10 min prior to cooling. A set of retrograde exchange experiments were also performed by calcining U3O8_e in H2O_d carried by cylinder air (50–55% RH) at 800 °C for 1 h. After the one-hour time period, the furnace was shut off and cooled either rapidly (by lifting the insulated lid) or slowly (keeping the lid closed), while the samples were still in an atmosphere of humid air (cooling rates shown in Figure S2).

Fluorination and Oxygen Isotope Measurements

Fluorination at the University of Utah was performed using a technique previously described.[1] Uranium oxide samples of 2–4 mg were loaded into Inconel reaction tubes and heated for 1 h at 150 °C to remove adsorbed gases. The tubes were then prefluorinated at room temperature with 10–20 torr BrF5 for 3 min before loading with 80–130 torr of BrF5 for a 16-h fluorination at 550 °C. Oxygen produced during fluorination was cryogenically separated from other reaction products and subsequently converted into CO2 for offline isotope analysis. A silicate oxygen isotope standard (NBS-28, δ18O = 9.6 ± 0.1‰) was fluorinated in conjunction with each set of uranium oxides for calibration.

Oxygen isotope measurements of fluorinated samples were performed within the Stable Isotope Ratio Facility for Environmental Research (SIRFER) lab at the University of Utah. Oxygen isotope ratios of CO2 were measured using a Thermo Finnigan GasBench II device coupled to a Thermo Finnigan Delta Plus XL. The carbonate internal lab reference materials were weighed using a Sartorius microbalance, loaded into 5.9 mL flat-bottomed borosilicate vials (Labco), and capped with Labco butyl rubber septa. The vials were then flushed for 6 min on a PAL autosampler, with UHP helium at a flow rate of 50 mL·min–1. During flushing, the vials were kept in a heated (25 °C) aluminum block. The reference materials were reacted with 10 drops of 104% phosphoric acid (kept at 25 °C) to produce CO2 gas. The reference materials were allowed to equilibrate for 24 h before analysis. During analysis using the PAL autosampler, CO2 in 12 mL Labco gas vials (for both samples and reference materials) was collected using a sampling loop of 100 μL and then transported to the mass spectrometer. Five individual injections/analyses were made for each sample and reference material, and their average was taken as the final number. Three sets of internal reference materials were used to calibrate the system and for the normalization procedure to obtain isotope values for unknown samples. The reference materials were Carrara marble, LSVEC, and Marble-Std. Carrara and LSVEC were used as primary reference materials, and Marble-Std was used as a secondary reference material to cross-check the final number. Internal reference materials were calibrated against international standards NBS-18 and NBS-19.

Oxygen stable isotope values of liquid water were measured using a Picarro L2130-i laser water isotope analyzer at SIRFER, following the method demonstrated by Good et al.[15] Oxygen isotope values of water vapor produced by diffusion through membrane tubing were measured also using a Picarro L2130-i connected to the tube furnace outlet via a jacketed gas line held at 40 °C to prevent condensation. Measurements were performed with water vapor concentrations of approximately 20,000 ppmV in N2, with a wet flowrate identical to that employed during aging experiments (120 mL·min–1) at 23 °C.

Oxygen stable isotope values are reported in δ-notation relative to VSMOW: δ = (RSample/RVSMOW – 1) × 1000‰, where RSample and RVSMOW are the respective 18O/16O ratios of the sample and standard (RVSMOW = 0.00200520),[16] and values are reported in per mille (parts per thousand, ‰). Fractionation in water between vapor and liquid is reported as 1000lnα, which is a unitless factor calculated from the quotient of 18O/16O ratios (α = Rvap/Rliq). Fractionation between water vapor in the furnace atmosphere and U3O8 products is reported as either 1000lnα (α = RU3O8/RH2O(v)) or Δ18O, representing the difference between δ18O values (i.e., Δ18O = δ18OU3O8 – δ18OH2O(v)). Oxygen stable isotope values of CO2 samples were provided from SIRFER in δ-notation relative to Vienna Peedee Belemnite (VPDB) and subsequently converted relative to VSMOW using δ18OVSMOW = 1.03092 (δ18OVPDB) + 30.92.[17] Uncertainties for δ18O values of CO2 samples and standards measured at SIRFER were generally less than 0.1‰. The uncertainty reported for each sample (1σ) was propagated from each measurement uncertainty, the literature uncertainty of the fluorination standard (NBS-28 silica, δ18O = 9.6 ± 0.1‰), and the standard deviation of replicate analyses. Average oxygen yields for fluorination of NBS-28 silica and U3O8 samples were 99 ± 5 and 100 ± 3%, respectively.

Results and Discussion

Liquid–Vapor Isotope Fractionation in Humidity

The two distinct water vapors used for the exchange experiments were produced by diffusion of H2O_d and H2O_e through membrane tubing into N2 carrier gas. The vapors were measured at 23 °C after passing through the furnace to have δ18O values very closely related to equilibrium values (Table ). Fractionation (as 1000lnα, unitless) for H2O_d vapor and H2O_e vapors was −10.0 and −9.9, respectively, while calculated vapor–liquid fractionation at 23 °C is −9.49 based on the equation from Majoube et al.[18] The slightly larger fractionation observed among measured water vapors is likely attributable to a gas flow rate greater than the diffusion of water through membrane tubing, creating a slightly undersaturated wet gas stream with a higher concentration of 16O.

Table 1

Measured vs Calculated δ18O Values (‰) for Water Vapors Used in Exchange Experiments in This Studya

		calc	meas	calc–meas
water	δ18Oliq	δ18Ovap	δ18Ovap	δ18Ovap
H2O_d	–16.47 ± 0.04	–26.01	–26.2 ± 0.3	0.23
H2O_e	+1.0 ± 0.1	–8.50	–8.8 ± 0.2	0.27

Fractionation equations from the study by Majoube were used for equilibrium calculations.[18]

The setup employed for water vapor production in this study was modeled after that of Oerter et al. (2019), with one exception being the length of membrane tubing submerged in the water reservoir.[14] The researchers noted that isotopic equilibrium between vapor and liquid using the membrane tubing was not attained until a length of at least 200 cm was submerged. The minorly deviated fractionation in this case was expected, with a tube of only 36 cm. Nevertheless, both fractionation and water vapor concentration were consistent over experimental timeframes, with average standard deviations over 12 h of ± 0.27‰ and ± 41 ppmV, respectively. The removal of water vapor from each reservoir during these aging experiments would be associated with Rayleigh-type fractionation or a gradual depletion of 16O in the liquid over time because of faster diffusion of the lighter isotope through the membrane tubing and into the vapor phase. Oerter et al. measured the average enrichment in δ18O of water from a 19 L reservoir using a 500 mL·min–1 gas flow rate to be 0.0163‰·day–1. Given the lower flow rate used in this study (120 mL·min–1), overall experimental timeframes (approx. 15 days total use per reservoir), and similar reservoir size (20 L), the change in δ18O of the vapor in these experiments was assumed to be negligible.

ADU Calcination

ADU used to make U3O8 samples for aging experiments was previously determined to have a bulk δ18O value of −16 ± 1‰.[2] Calcination of this precipitate to make starting compounds for subsequent aging experiments was performed at 800 °C for 72 h in atmospheres containing either H2O_d or H2O_e diluted in N2. The resulting δ18O values were −43.3 ± 0.5‰ and −16.7 ± 0.5‰ for U3O8_d and U3O8_e, respectively. The seemingly unchanged isotope composition of U3O8_e (with respect to ADU) is a coincidental fractionation with water vapor used in the calcination (δ18OH2O(v) = −8.8‰). This is supported by previous work showing the calcination of the same ADU to result in a preferential loss of 16O under an inert atmosphere using the same heating rate and ultimate temperature—in which case the δ18O value of U3O8 was −0.7‰.[2]

Interestingly, fractionation factors between each of the resulting U3O8 compounds and the water vapor sources were different, with U3O8_d having Δ18O = −17.1‰ and U3O8_e having Δ18O = −7.9‰. A re-examination of the calcination procedures revealed a difference in methods; namely, termination of the calcination of U3O8_d was performed with water vapor still present using a slow cooling rate (see Figure S2) while the calcination of U3O8_e was ended by switching the atmosphere to dry N2 prior to a fast cool-down. The difference in fractionation here was thus likely a result of retrograde isotope exchange during cooling of U3O8_d, a phenomenon which will be discussed in more detail in the section “Retrograde Isotope Exchange”.

Fractionation with Water Vapor

Water vapor exchange experiments utilized U3O8_d as a starting material, which was subsequently aged between 400 and 800 °C in an atmosphere containing H2O_e in N2 (50–55% RH). Oxygen isotope compositions of U3O8 after 72 h calcinations were taken as equilibrium values (Figure ). Justification for equilibration within this timeframe came from a lack of statistically significant change in isotopic compositions when compared to products from shorter calcination times. At 400 °C, equilibration occurred between 2 and 12 h because δ18O values at 12 and 72 h were identical (See Table S1 for δ18O values). At and above 600 °C, isotope exchange with atmospheric humidity was fast enough that 2 h was a sufficient duration for equilibration. These results show a rapid exchange between oxygen isotopes in water vapor and U3O8 at the temperatures utilized here.

Figure 2

U3O8_d (δ18O = −43.3‰) was calcined in an atmosphere of vapor from H2O_e (δ18Ov = −8.8‰) diluted to 50–55% RH in N2. Samples heated at 600 and 800 °C appear to achieve equilibrium with water vapor within 2 h, while U3O8 calcined at 400 °C reaches equilibrium within 2–12 h.

As expected, higher temperature exposure led to smaller fractionation factors between U3O8 and water vapor. Theoretical fractionation factors for α-UO3–H2O and UO2–H2O from the study by Hattori and Halas and Zheng are smaller than those measured for U3O8-H2O in this work (Figure ).[19,20] Because U3O8 is a mixed-valence compound containing both UV and UVI, the bonding with oxygen, and thus fractionation, may be expected to fall between the more oxidized UO3 and more reduced UO2.[21,22] The experimentally determined fractionation for U3O8 in this case falls well out of range of the calculations for both compounds. The greater depletion of 18O in U3O8 here may suggest a more substantial “mass effect” (i.e., heavier elements tend to incorporate more of the light isotope and vice-versa) in uranium oxides than previously postulated.[20]

Figure 3

Oxygen isotope fractionation factors between U3O8 and water vapor compared to calculated fractionation for other U-oxides from equations provided by Hattori and Halas and Zheng.[19,20]

Fractionation with Humid Air

The addition of oxygen at a tropospheric concentration (i.e., 20%) to the humid calcination gas resulted in starkly different isotope fractionations among U3O8 products than was observed in an atmosphere containing only water vapor. When U3O8_e (δ18O = −16.7 ± 0.5‰) was calcined in an atmosphere containing 18O-depleted water vapor (δ18O = −26.2 ± 0.3‰) and 18O-enriched O2 (δ18O = +23.2 ± 0.2‰), fractionation was driven in opposite directions as a function of temperature (Figure ). At 400 °C, the δ18O value of U3O8 dropped below that of the water vapor and was at equilibrium within 2 h. In contrast, calcination at both 600 and 800 °C caused a significant increase in δ18O values. Equilibration at higher temperatures also occurred within 2 h, as the δ18O values from U3O8 products at this time interval were statistically indistinguishable from those aged for 72 h.

Figure 4

U3O8_e (δ18O = −16.7‰) was calcined in an atmosphere of vapor from H2O_d (δ18Ov = −26.2‰) diluted to 50–55% RH in dry air (δ18OO2 = +23.2‰). Samples calcined at 400, 600, and 800 °C all appear to equilibrate within 2 h.

The different direction of fractionation observed between 400 °C and the higher temperatures suggests that different mechanisms of oxygen exchange may be occurring. Previous work pertaining to corrosion and diffusion rates of oxygen in uranium and uranium oxides has illustrated temperature-dependent mechanisms. The corrosion of uranium has been shown in multiple studies to occur at different rates in the presence of water vapor, dry air, and humid air.[23−25] Haschke found that the presence of water vapor enhanced oxidation rates of uranium, except above 500 °C, where it was suggested that the reaction rates of U + H2O and U + O2 become indistinguishable.[24] Oxygen diffusion in uraninite has also been observed to be controlled by different mechanisms above and below 400 °C.[26]

Calcination of U3O8 in humid air also revealed distinct fractionations from those observed with water vapor alone or dry air (Figure ). Although heating at 400 °C in humid air did show a water-driven isotope exchange, the δ18O value was shifted more positive than the vapor-only fractionation because of the presence of O2. Likewise, calcination at 600 and 800 °C in humid air showed a negative shift in δ18O values because of the presence of water vapor. These shifts in equilibrium δ18O values in the mixed atmosphere support a mechanism in which both species (H2O and O2) are interacting with the uranium oxide. While studies involving oxygen diffusion and reactivity of U3O8 in humid air at the temperatures utilized here are limited in the literature, there have been a number of investigations into the catalytic ability of the compound that may provide useful insight. Heneghan et al. found U3O8 to be an extremely active catalyst for the oxidation of volatile organics between 478 and 604 °C, both in the presence and absence of O2, suggesting a redox mechanism aided by the mixed oxidation states of uranium.[27] A follow on to this work by Taylor et al. showed that isotopically labeled O2 from the atmosphere was incorporated into oxidation products at 596 °C, which revealed that O2 would diffuse into or onto U3O8 and exchange isotopes with other gaseous compounds.[28] Harris et al. then added water vapor to these catalysis experiments, which proved to be a very effective promoter for oxidation of organics by U3O8, making the compound active at temperatures as low as 250 °C in humid air as opposed to 350 °C in dry air.[29] Together, these studies support the interactivity between U3O8, H2O, and O2 at the temperatures used in this work, with an increasing contribution from O2 as temperatures increase.

Figure 5

Fractionation between U3O8 and water vapor [top] and fractionation between U-oxides and atmospheric O2 [bottom] during 72 hr calcining experiments. Data for fractionation between UOx and dry air (N2/O2, gray squares) are reproduced from Klosterman et al., with amorphous UO3 at 400 °C and U3O8 at 600 and 800 °C and calcination times of 20 h.[1]

Water vapor concentration may play a significant role in the final signature of a calcined uranium oxide. In this study, the vapor concentration during calcination experiments was kept the same (50–55% RH) throughout, as was the mass of exposed U3O8 (90–150 mg). With the 200 mL·min–1 overall flowrate employed, the exposure concentration over time for the U3O8 would have been between 0.21 and 0.35 mol H2O·mol U–1·min–1. Given industrial settings which would maximize the throughput of solids, this exposure is likely much greater than may be encountered in calcining operations. A re-evaluation of industrial U3O8 samples analyzed by Plaue and Dierick et al., which had δ18O values between −6.2 and +12.4‰, suggests a lack of equilibration with both atmospheric O2 and humidity.[8,9]

Retrograde Isotope Exchange

The combination of a temperature-dependent fractionating mechanism with rapid equilibration in the mixed atmosphere presented potential for an additional phenomenon known as retrograde exchange. If isotope exchange kinetics at temperatures lower than the aging temperature are sufficiently fast that they exceed cool-down rates, then ultimate δ18O values and fractionation will be reflective of a lower temperature. This effect was avoided in the products of calcination experiments discussed above by flushing the atmosphere with inert gas while still at the designated temperature prior to cooling. However, the potential impact of retrograde exchange is vital for a more comprehensive interpretation of the formation of oxygen isotope signatures, so the effect was investigated in this work.

As shown in Figure , U3O8 calcined at 800 °C for 1 h in an atmosphere of 18O-enriched O2 and 18O-depleted water vapor preferentially incorporated oxygen isotopes from O2. When the atmosphere was unchanged during cooling, the products were measured with much lower δ18O values, indicating incorporation of oxygen isotopes from water vapor. Both a fast cool-down rate (800–100 °C in approx. 20 min) and a slow cool-down rate (800–100 °C in approx. 200 min) showed significant drops in δ18O values in the final product (see Figure S2 for cool-down rates). The isotope composition of rapidly cooled U3O8 (δ18O = −26.4 ± 1.8‰) fell directly on the water vapor line, showing a significant change from equilibrium at 800 °C (δ18O = +14.1 ± 0.6‰). With a slower cooling rate, the δ18O value drops below the water vapor line to −36.3 ± 0.1‰, revealing a fractionation (Δ18O) of −12.2‰.

Figure 6

δ18O values for U3O8 (initial δ18O = −16.7‰) calcined for 1 h in a mixed atmosphere (18O-depleted H2O vapor + air with 18O-enriched O2). U3O8 was cooled from 800 °C in either an inert atmosphere or the mixed atmosphere at different rates.

As mentioned previously, U3O8_d was made by calcination in an atmosphere with H2O_d and slowly cooled from 800 °C with the vapor present. This led to a highly 18O-depleted product having a fractionation with the water vapor of −17.1‰. This fractionation is greater than that observed when U3O8 was cooled in humid air (water vapor and O2 present), once again showing a difference in isotope incorporation and interactivity in different atmospheres. Fractionation between U3O8 and water vapor at 400 °C without retrograde exchange was −5.6‰ in humid air and −11.9‰ (Δ18O) in vapor alone; therefore, the closure temperature (i.e., temperature at which retrograde isotope exchange ceases) during this slow cooling is almost certainly lower than 400 °C.

Conclusions

Oxygen isotope fractionation between U3O8 and water vapor and between U3O8 and humid air was observed to occur rapidly between 400 and 800 °C, and with different mechanisms for each atmosphere. Isotope exchange with water alone led to oxides more depleted in 18O than the vapor and also more depleted than theoretically calculated fractionation factors for UO3 and UO2. The addition of O2 to the humid atmosphere caused a change in the mechanism of exchange between 400 and 600 °C, with a preferential incorporation of oxygen isotopes from water at lower temperatures and O2 at higher temperatures. The speed of isotopic equilibration in humid air also revealed the dramatic effect of retrograde exchange during cooling. While U3O8 incorporated oxygen from atmospheric O2 at 800 °C, if water was still present in the atmosphere during cooling, the oxygen isotope composition of the compound was quickly altered as a result of exchange with water. These results provide novel insight into oxygen isotope fractionation in uranium oxides and can help inform the interpretion of signatures encountered in industrial materials.