Multiscale Speciation of U and Pu at Chernobyl, Hanford, Los

Article pubs.acs.org/est

Multiscale Speciation of U and Pu at Chernobyl, Hanford, Los Alamos, McGuire AFB, Mayak, and Rocky Flats Olga N. Batuk,‡ Steven D. Conradson,*,‡,† Olga N. Aleksandrova,§ Hakim Boukhalfa,‡ Boris E. Burakov,∥ David L. Clark,‡ Ken R. Czerwinski,⊥ Andrew R. Felmy,# Juan S. Lezama-Pacheco,∇ Stepan N. Kalmykov,○,◆ Dean A. Moore,# Boris F. Myasoedov,◆ Donald T. Reed,‡ Dallas D. Reilly,# Robert C. Roback,‡ Irina E. Vlasova,○ Samuel M. Webb,¶ and Marianne P. Wilkerson‡ †

Synchrotron-SOLEIL, L’Orme des Merisiers, Saint-Aubin - BP48, 91192, France Los Alamos National Laboratory, Los Alamos, New Mexico 87545, United States § Ural Federal University, Mira Street 19, Ekaterinburg 620002 Russia ∥ V.G. Khlopin Radium Institute, 28, 2-nd Murinskiy Ave., St. Petersburg 194021, Russia ⊥ University of Nevada, MSM 245, 4505 S. Maryland Pkwy, Las Vegas, Nevada 89154, United States # Pacific Northwest National Laboratory, PO Box 999 MSIN: K8-96, Richland, Washington 99352, United States ∇ Environmental Earth System Sciences Department, 473 Via Ortega, Stanford University, Stanford California 94305-4216, United States ○ Radiochemistry Division, Chemistry Department, Lomonosov Moscow State University, Leninskie Gory, Moscow 119991, Russia ◆ Frumkin Institute of Physical Chemistry and Electrochemistry of RAS, Leninsky av. 31, Moscow 119071, Russia ¶ SLAC National Accelerator Laboratory, 2575 Sand Hill Road, Menlo Park, California 94025, United States ‡

S Supporting Information *

ABSTRACT: The speciation of U and Pu in soil and concrete from Rocky Flats and in particles from soils from Chernobyl, Hanford, Los Alamos, and McGuire Air Force Base and bottom sediments from Mayak was determined by a combination of X-ray absorption fine structure (XAFS) spectroscopy and X-ray fluorescence (XRF) element maps. These experiments identify four types of speciation that sometimes may and other times do not exhibit an association with the source terms and histories of these samples: relatively well ordered PuO2+x and UO2+x that had equilibrated with O2 and H2O under both ambient conditions and in fires or explosions; instances of small, isolated particles of U as UO2+x, U3O8, and U(VI) species coexisting in close proximity after decades in the environment; alteration phases of uranyl with other elements including ones that would not have come from soils; and mononuclear Pu−O species and novel PuO2+x-type compounds incorporating additional elements that may have occurred because the Pu was exposed to extreme chemical conditions such as acidic solutions released directly into soil or concrete. Our results therefore directly demonstrate instances of novel complexity in the Å and μm-scale chemical speciation and reactivity of U and Pu in their initial formation and after environmental exposure as well as occasions of unexpected behavior in the reaction pathways over short geological but significant sociological times. They also show that incorporating the actual disposal and site conditions and resultant novel materials such as those reported here may be necessary to develop the most accurate predictive models for Pu and U in the environment.

■

INTRODUCTION

production and testing as well as weapons and reactor accidents, a hazard that is amplified by the public perception of radioactive elements in general and Pu in particular. The

The chemical speciation of potentially hazardous subsurface contaminants is a crucial parameter in environmental risk assessment and remediation because it determines their transport, toxicology, and ultimate disposition. This tenet especially applies to uranium and plutonium contamination that arguably constitute the most intractable environmental restoration problems at legacy sites from nuclear weapons © XXXX American Chemical Society

Received: December 18, 2014 Revised: March 25, 2015 Accepted: March 27, 2015

A

DOI: 10.1021/es506145b Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

example, Manhattan Project era waste dumps, is not available and also because many of the samples were obtained via remediation and not research projects. The variety of original sources that subsequently became soil or concrete contaminants includes: (1) metal turnings in lathe coolant and leaks from a HNO3-based purification line at Rocky Flats; (2) a nuclear armed missile explosion and fire at McGuire AFB; (3) samples around 1 km away from the reactor at Chernobyl; (4) a sample from Reservoir 17 at Mayak; (5) a hot spot that may have been waste from R&D operations during and shortly after the Manhattan project from the Los Alamos TA-21 disposal area; and (6) samples from the acidic Z-9 and neutral Z-12 disposal crib areas at Hanford. A more detailed description of the sample provenance can be found in the Supporting Information.

paragenesis of at least U after its release is, however, an exceedingly complicated problem, both thermodynamically and kinetically,1−6 with less being known about Pu because there are no natural systems to study. In laboratory systems, under oxic conditions Pu can transform to PuO2+x (actually PuO2+x−y(OH)2y·z(H2O) compounds7−9) and, often at elevated temperature, uranium to U(VI) oxyhydroxides10 at rates dependent on the particle morphology and other reaction conditions. In the environment, weathering induced oxidation of U-containing minerals and U and Pu contaminants to, respectively, VI and IV on the decades time scale has been observed5,11,12 on at least some occasions. This “rapid” conversion of the material in a variety of original forms to the thermodynamically favored species identified in the laboratory was the basis for the successful restoration of the Rocky Flats site.12,13 Demonstrating that Pu was present in the expected PuO2+x form and that the transport had occurred as colloids validated models that calculated safe residual levels, thereby reducing the total amount of contaminated material deemed hazardous enough to require removal and disposal to manageable quantities. Analogously, in identifying their history, origin, and intent for nuclear forensics, assigning a useful time to their release also presumes that these weathering reactions will begin almost immediately and proceed to completion in a regular way over weeks to at most a few years. This predictability of Pu/U-O2-H2O in the laboratory contrasts with the natural systems for U where the presence of other elements that can react to form alteration products that inhibit oxidation and other surface reactions by impeding diffusion, the radiation field, and the simple mechanics of the two-phase solid−liquid system combine to give almost overwhelming complexity to the species whose releases have caused them to become environmental contaminants.14 This plethora of possible species must then also be incorporated in evaluations of actual and potential nuclear accidents.15 However, even while including the possibility of substitutions and other disorder, models are still constrained to describe these species in the context of known, crystalline compounds. After examining U and Pu in soil samples from six different locations in the U.S. and Russia, some in bulk but most as single particles using synchrotron microprobes, we herein present results indicating that this type of description is incomplete. These field-derived samples give us insight as to the extent and rate of interaction in the environment, atomic scale correlations with other atoms, and allow us to examine the assumptions currently being used to predict their short and long-term fate and transport. Over these geologically inconsequential but societally important time periods of decades we therefore observe both minimal and substantial momentum in the formation of the equilibrium species as well as finding novel species that would have resulted from unpredicted interactions of Pu and U species with other components of their waste streams and surroundings. The objectives of this study were to exploit our access to samples from so many sites and sources to explore the range of species and correlations displayed by U and Pu contamination including the identification of novel types; determine the extent to which these data signified their origins and history; and to elucidate whether and under what conditions these species might have been transformed by chemical reactions after their release. We will therefore not speculate on detailed mechanisms of their formation, which would be difficult in any event because for many sites the necessary information on, for

■

MATERIALS AND METHODS Brief Site Background and Sampling History. The Rocky Flats Environmental Technology Site (RFETS) fabricated components of fissile material for U.S. nuclear weapons.12,13 The data shown here came from soils underneath or adjacent to an asphalt pad designated 903 that was used for storing drums of Pu-contaminated lathe coolant, collected around 1998, and from the concrete floor of a building where separations were performed involving solutions with high concentration HNO3, obtained around 2002, with measurements performed shortly after the samples were obtained. This site was a US Department of Energy (DOE) remediation project. These were bulk measurements with a beam of mm dimension that could nevertheless be placed on different portions of the samples. The releases of material from McGuire AFB16−18 in 1960 and Chernobyl19−21 in 1986 involved high-temperatures, melting, fires, and explosions that consumed, respectively, a missile with a nuclear warhead and a nuclear reactor core. Surface soil samples from Chernobyl containing the particles shown here were collected in summer, 1986, 0.5−1.5 km west to northwest from the reactor in the Red Forest and in 1990 in the Western Plume area, following which they were stored in the laboratory. All the particles were separated from their host soils in 1990. We analyzed a total of 19 U, Zr, and mixed U/Zr particles. The fire and explosion fragmented these bulk solids and liquid melt into particles of varying sizes that in some cases combined the various constituents before condensing and solidifying. Particles from McGuire were collected from soil cores in 2007. Here we show results from particles obtained from 2 of these. During the time plutonium was produced at Mayak a large number of water reservoirs at this south Ural site22−25 were contaminated with irradiated fuel processing wastes. The Ucontaining particles we analyzed were found in bottom sediment samples collected in 2010 from one of these ponds, Reservoir 17, which at least for several years prior had been oxic (450 mV). We report XRF mapping results for 14 of these and EXAFS for 3 with a number of others giving XAFS spectra of varying quality. The Los Alamos TA-21 waste site26,27 was created in the years following the Manhattan project. Record keeping was incomplete, an unmapped intact truck as well as other artifacts were found during its excavation. We found 2 disk-like uranium-containing particles and 13 plutonium particles with a variety of sizes and compositions in soil samples obtained from the excavation in a single small volume of soil that had B


Article


Figure 1. χ(R) EXAFS from particles and bulk samples for U, Pu, and Fe from the indicated locations: Fourier transform moduli (left) and real components (right). The data are solid lines and the curve-fits are dashed. (a, b) Representative U L3 edge χ(R) spectra from two of the Chernobyl particles. (c, d, e) Representative U L3 edge χ(R) spectra from three of the Mayak particles showing the different U oxide species. (f) A representative Pu L3 edge χ(R) spectrum from the composite McGuire particle. (g) A representative PuO2+x-like Pu L3 edge χ(R) spectrum from one of the Hanford Z-9 crib particles. (h, i) Bulk Pu L2 edge χ(R) spectra from samples from the Z-9 (h, PuO2+x-like) and Z-12 (i, non-PuO2+x) Hanford cribs. (j) Bulk PuO2+x-like Pu L2 edge χ(R) spectrum from a Rocky Flats soil sample. (k) A representative non-PuO2+x-type Pu L3 edge χ(R) spectrum from a Pu particle in a Los Alamos soil sample. (l) A representative Pu L3 edge χ(R) spectrum from the mixed Los Alamos Pu/Fe particle, PuO2+x-like structure plus Fe shell at R = 3.35 Å. (m) A representative Fe K edge χ(R) spectrum from the same Los Alamos mixed Pu/Fe particle. The transform is taken over k = 3−9 Å−1 for all spectra to allow direct comparisons of both position and amplitude in this figure although the actual range of the curve-fit analyses of these spectra shown in the Supporting Information varies depending on the quality of the data.

samples were described as acidic, high salt, and acid rich with significant amounts of CCl4, tributyl phosphate and derivatives, lard oil, and NO3− as cocontaminants, with a possible correlation of the Pu with P.32,33 In addition to XAFS on a bulk sample, μ-XRF images from Z-9 soils show large numbers of highly regular squares with sides ∼20 μm containing small amounts of Pu but no other elements within the range of the detector system (Z > K). Some of these squares had much more substantial PuO2 domains on their surfaces of varying thickness and extent. We found 12 such squares with Pu

shown high activity during routine surveys and was rumored to be contaminated with laboratory R&D waste. Hanford crib soil samples contained wastes from Pu separations and recovery processes that were intentionally disposed in the cribs into the beginning of the 1960s. The Z-9 crib was excavated in 1974−1976 because the amount of Pu was so large that criticality events had become a concern. These samples came from the surface of the Z-9 crib, collected and archived prior to its mining operation, and from the surface of the Z-12 crib that was studied in 2004 and 2006.28−32 Z-9 C


Article


to the shortest range spectrum from these samples, longer ranges were used when possible to give the results in the Supporting Information. UO2+x, U3O8, and U(VI) oxyhydroxides give easily distinguished EXAFS, whereas the distinction between UO2 and UO2+x is more subtle34 in that, as with PuO2+x,8 it involves disorder in the U−U pair and a reduction in the EXAFS amplitude, the introduction and growth of a new peak that is the uranyl-type oxo, or U-oxo contribution that can overlap with the principal U−O shell around 2.36 Å, and finally the formation of other peaks from O neighbors with distances between 2.7 and 3.4 Å. The absence of a U−U or Pu−Pu contribution at R = 3.5−4.5 Å demonstrates that the U or Pu is monomeric instead of in oxide or oxyhydroxide form. Although these trends can be difficult to observe and compare in spectra measured over this limited energy range, an estimate of the relative value of “x” is obtained from the numbers of oxo and U near neighbor atoms obtained from curve-fits (Supporting Information Table S1). This table and 15 full spectra from these and additional samples are presented in the Supporting Information. X-ray diffraction measurements were attempted with a number of samples, with the only successes (Figure 2) being the >200 μm particles from McGuire AFB that had been annealed in the fire. This finding of widespread crystalline disorder in particles up to 50−100 μm in size is perturbing since it suggests that diffraction patterns from bulk samples may be sampling only a fraction of the material. This disorder is also evident in that the low index reflections are complete circles whereas the particles could have been expected to be single grains or at least highly textured over the limited volume sampled by the 2 μm beam. Pu and U Metal or Alloy Origin. Real space Fourier transforms of their EXAFS spectra indicate (Figure 1j, Supporting Information Figure S11, and12) that the Pu speciation in Rocky Flats soil contaminated with machining fluid used for cutting Pu and concrete contaminated from a Pu, fire is PuO2+x. Pu(U)O2+x and UO2+x are also found by both EXAFS (Figure 1f and S7) and XRD (Figure 2) that shows the Fm3m diffraction pattern with lattice constants within the error range for these compounds in particles from McGuire AFB that were released when a fire consumed a nuclear armed missile. Two μm resolution XRF element maps show (Figure 3) that one particle is best described as a conglomerate of a Pu-rich and a U-rich particle that melded without significant mixing plus some small Fe and Ga spots (Supporting Information Figure S16), whereas in a second particle that was presumably subjected to higher temperatures for a longer time the Pu and U are intermingled to give a more homogeneous mixture (Figure 3). U−Zr Oxides at Chernobyl. Like these samples from McGuire and Rocky Flats, materials from Chernobyl were also exposed to high temperatures that promoted reactions and mixing between the UO2 and the Zr, air, water, graphite, and other materials. The U particles collected from Chernobyl exhibit homogeneous (U, Zr)Ox solid solutions (16−18 wt % Zr) on the μm length scales of these analyses according to the XRF mapping and quantitative electron probe microanalysis. This would result from the high temperature promoting the combining of the fuel and cladding materials to give (U, Zr)Ox with varying U:Zr (up to 52 wt % Zr) ratios with some stainless steel inclusions. On the Å length scale of the EXAFS a different result is obtained. The Zr and U EXAFS can be fit with, respectively only Zr neighbors at 3.67 Å and only U neighbors at 3.87 Å, implying significant inhomogeneity. This clustering

sufficient to record EXAFS. The Z-12 crib material was described as low salt and near neutral, with NO3− and F− listed as the only cocontaminants.33 The XAFS of Z-12 was recorded on a bulk sample. XRF, XAFS and XRD Data Collection and Data Analysis. All the XRF, XAFS and XRD measurements were performed on beamlines 2−3 (microprobe) and 11−2 (bulk) at the Stanford Synchrotron Radiation Lightsource. The X-ray optics at beamline 2−3 focuses the beam to a 2 μm fwhm spot. Details of the data acquisition and analysis and the results of the curve-fitting as χ-functions and Fourier transforms are presented in the Supporting Information accompanying this report in Figures S1−S15 and the metrical parameters are summarized in Table S1. Calculation of the Correlations. The spatial correlations for the μ-XRF element maps were calculated using the Pearson function that measures the strength of association between a pair of variables. For the correlations between experimental data and the fitting results we used the root-mean-square deviation. This function is a frequently used measure of the differences between values predicted by a model or an estimator and the values actually observed from the data being modeled or estimated. Quantitative Electron Probe Microanalysis. Quantitative electron probe microanalysis of U and Zr in Chernobyl samples was carried out at the V.G. Khlopin Radium Institute using an SEM-microprobe “CamScan-4DV”, at 20 kW, current 10 nA. Synthetic ZrSiO4 and UO2 standards were used for the evaluation of the recorded data. The standard software package ZAF-4/FLS was used for the analysis.

■

RESULTS AND DISCUSSION Based on the objectives of this report, this discussion begins with the two samples of Pu from Rocky Flats and two from McGuire AFB that were known to have originally been in metallic form and the U that was combined with the Pu in the fire that caused the release at McGuire. This is followed by a description of the U contamination that was also the product of a high temperature event−the explosion and fire at Chernobyl− that also showed only UO2+x as the highest U valence. Next are the numerous U oxide particles from a single sample from a pond at Mayak that exhibit a wider range of valences. Finally, U and Pu from the Z-9 cribs at Hanford, the LANL TA-21 site, and a purification line at Rocky Flats, are presented, all known or suspected to have been released in acidic solutions that would have reacted with the components of the soil or concrete. The Rocky Flats results12 and some earlier bulk Pu data from Hanford32 are included here because of their value in comparisons with information from other sites. For some of the sites the results from some or even most of the particles evaluated are not shown, including ones where the element correlations or even certain aspects of the speciation are different. Reiterating, our purpose is to show the range of behavior and not an exhaustive list of all of the variations. Figure 1 showing the EXAFS data displays a different order. Insofar as many of these species exhibit the Pu/UO 2+x disordered fluorite structure, scanning down the figure through the spectra facilitates the comparison of the R = 1.8 and 3.7 peaks indicative of AnO2, first for the U and then for the Pu among the different samples, with the final Fe EXAFS from a particle containing both Pu and Fe in somewhat separate domains. This comparison dictated that the Fourier transformation range of the spectra included in the figure conform D


Article


whereas particles where U is the only element with Z ≥ Ca show more oxo and are assigned to UO2+x (Figure 1a and b and Supporting Information Table S1).34 It is important to note that none of the spectral signatures of U(VI) are observed, which for EXAFS typically constrains the amount that could be present to less than 10%. Insofar as the initial step in U paragenesis is oxidation so that subsequent alteration phases are uranyl compounds, the absence of U(VI) may perhaps have occurred originally because of the high temperature and subsequently because further oxidation from weathering may have been inhibited by the Zr or a surface modification involving, for example, carbon.37,38 U Oxides at Mayak. The 14 U-containing particles analyzed from the single soil sample obtained from Mayak ranged in size from ∼5 up to 50 μm (Figure S4 Supporting Information), with high U counts typically found for those ≤8 μm. The fluorite-structured UO2+x with varying x displayed at Chernobyl is also exhibited by many of the particles from the single sample of PA “Mayak” Reservoir 17 bottom sediment. Some Mayak particles, however, contained U in higher valences. Representative EXAFS spectra (Figure 1c−e) and their curve-fits (Supporting Information Table S1) unambiguously show the three UO2+x, U3O8, and U(VI)-oxyhdroxide species, with no indications of mixed structures that would indicate that the former are becoming the latter even after decades in the pond, as has been observed at some sites.10,11,39 This demonstration of inhibited or retarded oxidation of the U, perhaps similar to that at Chernobyl, indicates that it is likely that these U species in these materials are the same as when they were originally deposited from various waste streams at Mayak, in contrast to, for example, the U at the Fernald processing site that showed increasing amounts of U(VI) with longer exposure times.11 This variation in U speciation extends to the correlations with other elements. Although these three particles had correlations with Ca and Fe ≤ 0.13, others were as high as >0.7, with varying degrees of correlation between these other two elements (Supporting Information Figures S17− S19). The incomplete documentation for the site and Reservoir 17 in particular preclude developing a more detailed correlation between the particle composition and the original disposal and subsequent site history. There were, however, occasions when treatments rendered the pond both highly acidic and alkaline, which might have promoted surface modification of the UO2 particles that could affect subsequent oxidation. However, whether the inhomogeneity of the speciation in this small volume of the sample is because these U-oxides were deposited from a number of different sources that were mixed, as we postulate, or from reactions after deposition that somehow discriminated between these particles despite their identical conditions is uncertain. It is nevertheless of interest to compare these results with some of the extensive literature on UO2 paragenesis.2−6,40 The combination of relatively pure U oxides of varying valence and degrees of correlation with other elements signifying alteration phases is consistent with these prior reports that described the decomposition of UO2 by oxidation of the surface that promoted the formation of the uranyl alteration compounds via crystallization or precipitation with other cations and inward along grain boundaries that fractured the bulk UO2 to release small particles. What differs here then is that there is no source of bulk UO2. It is highly unlikely that the radiation field was sufficient to effect reduction.40 The 137Cs γ rays emitted by this sample confirm

Figure 2. μ-X-ray diffraction patterns taken from the McGuire particles shown in Figure 3. (a) Diffraction from Pu-enriched area of the inhomogeneously mixed particle. (b) Diffraction from U-enriched particle area for the same particle. (c) Diffraction from the homogeneously mixed McGuire particle.

of the cations on the nanometer scale is typical of mixed U−Zr oxide.35,36 There are also particles of almost pure Zr-oxide (Figure 4). The separate components of the missile and reactor as well as variations in the temperature attained by individual particles are therefore manifested in the distributions of the elements composing the particles. The U in the mixed U/Zr particles shows few or no oxo neighbors, consistent with UO2, E


Article


Figure 3. μ-XRF maps of the Pu−U composite (upper) and homogenized (lower) particles from McGuire and their Pu−U spatial correlations.

and minerals, oxidation is retarded after an initial burst of U release because, on the scale up to 10 μm or greater, diffusion is inhibited by the accumulation of uranyl alteration compounds and the UO2 itself is masked as the higher oxides adhere to and cover the surface instead of dissolving.2−6 Given the probability that the original U was largely homogeneous and in the form of dispersed particles ≤10−50 μm in size like these and that the environmental conditions were identical for all of these particles because of their proximity, there is again the issue of what relatively subtle effect, perhaps limited to the surface of the particles, caused this substantial difference in behavior. Pu and U from Acidic Solution. In contrast to these other materials, disposal of Pu and U in waste streams sufficiently acidic to react with concrete from a purification process line at Rocky Flats or soil minerals in the Hanford cribs or a disposal site at Los Alamos appears to have promoted both unusual speciation and morphology. The Pu particles found in the XRF maps of a smear of soil on tape from a few grams collected from one hot spot at the Los Alamos TA-21 site that most likely originated as R&D waste exhibit association with Fe in all three possible ways; typical surface complexes or precipitates on Febased mineral phases, isolated Pu particles with proximity but no contact, and single particles composed of Fe−Pu mixtures (Figure 5). We interpret these results to suggest that the variation in particle composition is due to Pu deposition onto the chemically diverse soil materials when the putatively acidic solution was neutralized by reacting with the soil components, releasing or activating some that could then possibly combine with the Pu that was also originally dissolved. The U particles from the Los Alamos site also exhibit substantial alteration. Two ∼100 μm diameter disks are not simple U-oxyhydroxides but instead are homogeneous 0.3:0.7 U:Zn mixtures with a trace of Cu (Figure 6a), similar to particles reported at Hanford.41 Uranyl (Figure 6b) alteration products can therefore involve additional elements from the laboratory− perhaps even galvanization from pipes−as well as from soils,

Figure 4. (a) Tricolor μ-XRF maps for three particles collected at Chernobyl, and (b) corresponding Zr−U correlations showing μmscale homogeneous mixing in a particle with comparable amounts of U and Zr, a Zr-rich particle that also contains a U- and an Fe-rich domain, and a particle that is almost pure Zr.

the origin of these materials from spent fuel processing, indicating that the U source term would have been relatively homogeneous high valence U oxides created during the probable oxidative dissolution of the original UO2. This poses the question of the subsequent formation of the UO2 in these particles, especially insofar as the pond has been oxidizing for many years. A second question vis-à-vis the studies of minerals and laboratory systems is the homogeneity of the U speciation in the individual particles. The range of U−Fe−Ca correlations as well as U species (Supporting Information Figures S17− S19) implies that chemical processes were proceeding at a significant rate on the decades time scale. In laboratory systems F


Article


Figure 5. Two-color μ-XRF maps (Pu = red, Fe = green) of Los Alamos TA-21 Pu particles. (a), a mixed Pu−Fe particle, one of two found in the measured samples. (b) Pu particle on an Fe-based mineral grain, one of six similar particles. (c) separated, noncorrelated Pu and Fe particles, one of three similar particles.

Figure 6. (a) Cu−U and Zn−U spatial correlations for a U-containing Los Alamos particle (excitation energy is 18 100 eV), the Pearson function was applied to calculate R for these correlations. (b) XANES of the same particle compared with that of schoepite as a prototype uranyl oxyhydroxide compound.

Figure 7. (a) Pu map of a Hanford Z-9 crib sample (sensitivity to only Pu results from subtracting the counts measured at 18000 eV that is below the Pu L3 absorption edge from those at 18100 eV that is above it). The scale shows the differential Pu count rate. (b) the Pu−P and Pu−O curve-fits to the Fourier filtered, non-PuO2 wave in the data, from which the much higher quality of the Pu−P fit−it is under the red line of the Pu−P fit at lower k−can be discerned.

implying that they can form quickly and are subsequently stable. A sample showing Pu-containing particles from the Hanford Z-9 crib (Figure 7a) indicates an even more complicated and specific process. Although the spectrum of a bulk Z-9 sample (Figure 1h) gives PuO2+x, a particle from a separate sample gives a different result (Figure 1g). It contains many regular, ∼20 μm square particles composed of elements lighter than Ca, most likely a silicate or a phosphate that could have formed from the breakdown of the tributyl phosphate from the PUREX process. These particles show barely detectable amounts of Pu that are uniformly distributed either within or, less likely, on the surface of the particle. A few of them, however, show large Pu deposits. This finding of two populations of Pu that are distinct in both the correlation of their distribution with respect to the particles and their quantity imply that the low concentration Pu was incorporated during the formation of the particles whereas

the high concentration surface precipitates formed after the square substrates and in a way that favored growth over nucleation, that is by a slow process involving Pu diffusion or transport. Since the maps from all of these samples never show any signal in the areas between the larger particles, colloidal or otherwise diffuse Pu must be negligible. That the preponderance of the Pu occurs in μm-scale particles that were most likely not its original form implies that Pu was mobilized and subsequently reacted, nucleating on the crystal surfaces as a surface complex followed by the accumulation of additional Pu on this surface complex to give the bulk species observed here. G


Article

Environmental Science & Technology Molecular scale speciation in addition to this μm-scale morphology further supports the idea that a highly reactive source or subsequent term can promote chemistry that can often be distinct and unexpected. In addition to the Zn:U(Cu) particles, an entirely novel Pu species occurs in what is apparently a Pu-enriched particle obtained from concrete exposed to HNO3-Pu solutions from a purification line at Rocky Flats because the XANES measured at a particular location for the large beam gave a much higher count rate and a spectrum unique in both its shape and high energy compared to other locations (Figure 8). The Pu in the Hanford Z-12 crib

Figure 9. Pu L3 EXAFS on mixed Fe−Pu TA-21 particle: analysis of the anomalous peak in the spectrum by comparing the curve-fits with O and Fe and the corresponding residuals.

compounds (Figure 1g).9 However, closer examination shows a non-PuO2 feature around R = 3 Å that gives a 99% correlation with the data when fit with a P (cannot be distinguished from Si, another possibility) shell but only 90% with O (Figure 7b), indicating its origin in a P(Si) second shell neighbor. The 3.45−3.54 Å distances show that the P is O-bridged to the Pu within the PuO2 lattice, giving substitutional disorder analogous to other oxides but unprecedented if perhaps not unexpected for Pu. The P would be incorporated in concert with the mobilization of PuO2+x that subsequently recrystallized into the large domains on the particle faces, implying that the substrate controls this process via, for example, the high concentration of phosphate at the surface of the crystal, and is more than merely a passive template for the sorption of PuO2+x. The process would be similar to that already reported, except in the field rather than the lab, and therefore coupled to local conditions such as an acidic waste stream that could also have reacted with soil minerals.44,45 Neutralization would terminate the reaction, but with the P retained in the PuO2 as a signature of not only the substrate but also the history of the material. Environmental Implications. These results clearly identify both expected and novel aspects of the environmental chemistry of U and Pu beyond those previously reported from laboratory and mineral studies. One important finding is that the samples from none of these sites showed fluorescence from U and Pu above the detection limit between the particles; these elements were always concentrated in particles without any sign of a more diffuse population. Of course, free U and Pu are required for the chemical reactions in the soil that we have postulated based on these results, but their concentrations would be low, corresponding with the decades time scales of these reactions. Alternatively, the transformations could occur rapidly during rare occasions of extreme conditions. On the Åscale of speciation, in some samples, particles that were within mm of each other for decades that most likely began as homogeneous species or originated in the same process exhibit different speciation with no indication of convergence to their presumed thermodynamic minima, implying that the observed species are the source terms that were, for example, inhibited from oxidation. In other samples they are present in unusual, non-PuO2+x forms, or as these expected oxyhydroxides but with additional elements whose incorporation would have resulted from transitory local conditions and chemical activation resulting from the waste stream or perhaps in combination with chemistry enhanced by surface phenomena. Most

Figure 8. Bulk Pu L2 XANES measured on a Rocky Flats concrete sample contaminated via leaks from an acid purification line compared with those of Pu(IV) and Pu(VI) standards. The two experimental spectra giving different results were measured at two different locations in the same Rocky Flat bulk concrete sample, indicating that at least the Pu from the first run had occurred as a Pu-enriched particle instead of being diffused through the material.

sample, in contrast to the Z-9 crib, is well fit with only a single shell, an O/F at a short 2.22 Å Pu−O/F distance (Supporting Information Table S1) with no indication of Pu neighbor as in PuO2+x.8,32 Similarly, the EXAFS of the Los Alamos non-Feassociated Pu particles (Figure 5c) show only a single O shell at 2.31 Å (Figure 1k and Supporting Information Table S1). The Pu species in these samples are therefore mononuclear and not PuO2+x-like compounds, analogous to the non-UO2 U(IV) species that have been reported and therefore posing the question of whether this analogy extends to the biogenic origin of the latter.42 The Pu in the Fe−Pu particle from TA-21 is PuO2-like with Pu−O/Pu distances of 2.31/3.75 Å (Figure 1 and Supporting Information Table S1). The large extra peak in χ(R) at R = 3.1 Å that does not occur in PuO2+x is fit over 40% better with an Fe shell at 3.35 Å (72% of its spectral weight) than with O (only 51%) (Figure 9). This distance is within the range expected for an Fe within the fluorite lattice. This result, that Fe is on occasion incorporated into PuO2+x, is corroborated by a reanalysis of the EXAFS of PuO2 precipitated by reduction with Fe that with this new structure motif now also shows a similar Fe neighbor shell.43 The Fe in this particle is goethite-like, a typical soil mineral, without any evidence of a Pu neighbor incorporated into its structure. These results imply a crystallographically heterogeneous mixed Fe−Pu oxide with Fe > Pu− consistent with the relative count rates−and with composition fluctuations giving Pu-enriched domains containing sufficient Pu to adopt the fluorite structure. The Pu EXAFS from Hanford Z-9 crib particles display the overall shape and Pu−O/Pu distances of PuO2+x(F)-type H


Article

Environmental Science & Technology Table 1. Significant Properties and Results location

samples

source

principal species

distinguishing characteristics

Chernobyl

U: with Zr pure U

reactor explosion

UO2 UO2+x

no oxidation past UO2 less oxidation with Zr

Hanford

Pu: Z-9 bulk and particle Z12 crib

acidic high salt

PuO2+x

incorporated P into PuO2+x

near neutral low salt

Pu(IV) mononuclear

mononuclear species

Pu: isolated

r&d waste

Pu(IV) mononuclear

multiple species at same location

Los Alamos

Fe surface Fe particle U:

PuO2+x Uranyl with Zn

mononuclear Pu incorporated Fe into PuO2+x in mixed Pu:Fe particle: U:Zn alteration phase

McGuire

Pu and U

missile fire

PuO2+x UO2+x

mixed PuUO2+x diffraction quality

Mayak

U

separations

UO2+x U3O8 U(VI) oxyhydroxide

multiple valences and alteration products at same location in homogeneous particles

Rocky Flats

Pu: 903B pad purification

disposal purification

PuO2+x Unique (XANES)

importantly, they not only provide the molecular scale chemical speciation information that was incomplete in prior reports of complexity in their environmental chemistry but also suggest possible mechanisms for it that appear to recapitulate the source term and subsequent history after release.46−49 Although these time periods are virtually instantaneous on the geological scale, they are significant on the regulatory one. These results demonstrate that duplicating and predicting the behavior of U and Pu at a specific site are likely to require experiments that duplicate both the site conditions and the chemical form of the contaminant when first released to promote the formation of unexpected or even unknown species that may be present. In addition to F− that we have already reported,32 a number of these samples either directly show or point to the presence of additional cationic elements as second neighbors in the PuO2 and UO2 lattices, a substitutional pattern previously unknown for Pu. This is not necessarily detrimental to environmental regulation and control; the observed or implied stability and diminished solubility of such ternary oxides44,45 could actually be advantageous in many remediation, restoration, and licensing efforts (see Table 1).

■

novel Pu(VI) species

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +33 1 69 35 91 80; e-mail: steven.conradson@ synchrotron-soleil.fr. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

We acknowledge Pavel M. Stukalov (1959-2010) the former head of the Environmental Control Laboratory of PA “Mayak” for his assistance with Reservoir 17 sediments and Kaiser-Hill LLC for contributing plutonium-contaminated soils and concretes. Los Alamos National Laboratory is operated by Los Alamos National Security, LLC, for the National Nuclear Security Administration of U.S. Department of Energy under Contract DEAC52-06NA25396. Financial support was provided by the Los Alamos LDRD program, the U.S. Department of Energy’s Office of Biological and Environmental Research (BER), as part of BER’s Subsurface Biogeochemistry Research Program (SBR) which originates from the SBR Scientific Focus Area (SFA) at the Pacific Northwest National Laboratory (PNNL). Advanced interpretation of the data was supported by the Heavy Element Chemistry Program, Chemical Sciences, Biosciences, and Geosciences Division, Office of Basic Energy Sciences. Other financial support includes the Russian Basic Research Foundation (project 10-03-01029-a) and Ministry of Education and Science of Russian Federation (projects 02.740.11.0853 and 11.519.11.5011). Portions of this research were carried out at the Stanford Synchrotron Radiation Lightsource, a Directorate of SLAC National Accelerator Laboratory and an Office of Science User Facility operated for the U.S. Department of Energy Office of Science by Stanford University.

ASSOCIATED CONTENT

S Supporting Information *

Complete figures of the EXAFS data and fit results for 15 samples including Zr for Chernobyl and Fe for the Los Alamos TA-21 Pu−Fe particle, XRF maps of additional elements for the McGuire AFB particles, element correlations and XRF maps of some Mayak particles, and a table of the curve-fitting results for the 15 EXAFS spectra as described in the text. This material is available free of charge via the Internet at http:// pubs.acs.org. I


Article


■

(19) Burakov, B. E.; Anderson, E. B.; Shabalev, S. I.; E, S. E.; Ushakov, S. V.; M, T.; J-Y, B.; Winter, P.J.; Duco, J. The Behaviour of Nuclear Fuel in First Days of the Chernobyl Accident, Scientific Basis for Nuclear Waste Management XX, 1997; Gray, W. J.; Triay, I. R., Eds.; Material Research Society, 1997; pp 1297−1308. (20) Salbu, B. Speciation of radionuclidesAnalytical challenges within environmental impact and risk assessments. J. Environ. Radioact. 2007, 96 (1−3), 47−53. (21) Salbu, B.; Krekling, T.; Lind, O. C.; Uoghton, D. H.; Drakopoulos, M.; Simionovici, A.; Snigireva, I.; Snigirev, A.; Weitkamp, T.; Adams, F.; Janssens, K. V.; A, K. High energy X-ray microscopy for characterisation of fuel particles. Nucl. Instrum. Methods 2001, A 467−468, 1249−1252. (22) Myasoedov, B. F.; Drozhko, E. G. Up-to-date radioecological situation around the M ̀ ayak’ nuclear facility. J. Alloys Compd. 1998, 271−273 (0), 216−220. (23) Stukalov, P. M. Radioactive Pollution of Industrial Pond at PA ″Mayak″ Old Swamp. Review on results of research 1949−2006. Chapter 1, 2007. (24) Christensen, G. C.; Romanov, G. N.; Strand, P.; Salbu, B.; Malyshev, S. V.; Bergan, T. D.; Oughton, D.; Drozhko, E. G.; Glagolenko, Y. V.; Amundsen, I.; Rudjord, A. L.; Bjerk, T. O.; B, L. Radioactive contamination in the environment of the nuclear enterprise ’Mayak’ PA. Results from the joint Russian-Norwegian fiekd work in 1994. Sci. Total Environ. 1997, 202, 237−248. (25) Børretzen, P.; Stansdring, W. J. F.; Oughton, D. H.; Dowdall, M.; Fifield, L. K. Pu and U atom ratios and cencentration factors in reservoir 11 and Asanov swamp, Mayak PA: An application of accelerator mass spectrometry. Environ. Sci. Technol. 2005, 39, 92−97. (26) McGehee, E. D. Decommissioning the Manhattan project: Historic preservation at Los Alamos. Trans. Am. Nucl. Soc. 2009, 100. (27) Mulkin, R. Characterization of Transuranic Solid Waste from a Plutonium Processing Facility; LA-5993-MS; Los Alamos National Laboratory: Los Alamos, 1975; pp 1−24. (28) Mashal, K.; Harsh, J. B.; Flury, M.; Felmy, A. R.; Zhao, H. T. Colloid formation in Hanford sediments reacted with simulated tank waste. Environ. Sci. Technol. 2004, 38 (21), 5750−5756. (29) Cantrell, K. Transuranic Contamination in Sediment and Groundwater at the U.S. DOE Hanford Site, PNNL 18640; Pacific Northwest National Laboratory: Richland, WA, 2009; pp 1.1−6.5. (30) Dai, M.; Buesseler, K. O.; Pike, S. M. Plutonium in groundwater at the 100K-Area of the U.S. DOE Hanford Site. J. Contam. Hydrol. 2005, 76 (3−4), 167−89. (31) Kaya, J. H.; Stebin, R. S.; Orr, R. D. Rapid, quantitave analysis of americium, curium and plutonium isotopes in Hanford samples using extraction chromatography and precipitation plating. J. Radioanal. Nucl. Chem. 1995, 194 (1), 191−196. (32) Felmy, A. R.; Cantrell, K. J.; Conradson, S. D. Plutonium contamination issues in Hanford soils and sediments: Discharges from the Z-Plant (PFP) complex. Phys. Chem. Earth 2010, 35 (6−8), 292− 297. (33) Cantrell, K. J. Riley, R. G. A Review of Subsurface Behavior of Plutonium and Americium at the 200-PW-1/3/6 Operable Units, PNNLSA-58953; Pacific Northwest National Laboratory: Richland, WA, 2008; pp 1−12; A.1-C.12. (34) Conradson, S. D.; Manara, D.; Wastin, F.; Clark, D. L.; Lander, G. H.; Morales, L. A.; Rebizant, J.; Rondinella, V. V. Local structure and charge distribution in the UO2-U4O9 system. Inorg. Chem. 2004, 43 (22), 6922−6935. (35) Villella, P.; Conradson, S. D.; Espinosa-Faller, F. J.; Foltyn, S. R.; Sickafus, K. E.; Valdez, J. A.Degueldre, C. A. Local atomic structure in cubic stabilized zirconia. Phys. Rev. B 2001, 64 (10). (36) Walter, M.; Somers, J.; Bouexiere, D.; Rothe, J. Local structure in solid solutions of stabilised zirconia with actinide dioxides (UO2, NpO2). J. Solid State Chem. 2011, 184 (4), 911−914. (37) Nelson, A. J.; Felter, T. E.; Wu, K. J.; Evans, C.; Ferreira, J. L.; Sickhaus, W. J.; McLean, W. Uranium passivation by C+ implantation: A photoemission and secondary ion mass spectrometry study. Surf. Sci. 2006, 600 (6), 1319−1325.

REFERENCES

(1) Wronkiewicz, D. J.; Bates, J. K.; Gerding, T. J.; Veleckis, E.; Tani, B. S. Uranium release and secondary phase formation during unsaturated testing of UO2 at 90-degrees-C. J. Nucl. Mater. 1992, 190, 107−127. (2) Wronkiewicz, D. J.; Bates, J. K.; Wolf, S. F.; Buck, E. C. Ten-year results from unsaturated drip tests with UO2 at 90 degrees C: Implications for the corrosion of spent nuclear fuel. J. Nucl. Mater. 1996, 238 (1), 78−95. (3) Buck, E. C.; Wronkiewicz, D. J.; Finn, P.; A.Bates, J. K. A new uranyl oxide hydrate phase derived from spent fuel alteration. J. Nucl. Mater. 1997, 249 (1), 70−76. (4) Deditius, A. P.; Utsunomiya, S.; Ewing, R. C. Alteration of UO2+x under oxidizing conditions, Marshall Pass, Colorado, USA. J. Alloys Compd. 2007, 444, 584−589. (5) Baker, R. J. Uranium minerals and their relevance to long term storage of nuclear fuels. Coord. Chem. Rev. 2014, 266, 123−136. (6) Plasil, J. Oxidation-hydration weathering of uraninite: The current state-of-knowledge. J. Geosci. 2014, 59 (2), 99−114. (7) Conradson, S. D.; Begg, B. D.; Clark, D. L.; Den Auwer, C.; Espinosa-Faller, F. J.; Gordon, P. L.; Hess, N. J.; Hess, R.; Keogh, D. W.; Morales, L. A.; Neu, M. P.; Runde, W.; Tait, C. D.; Veirs, D. K.; Villella, P. M. Speciation and unusual reactivity in PuO2+x. Inorg. Chem. 2003, 42 (12), 3715−3717. (8) Conradson, S. D.; Begg, B. D.; Clark, D. L.; den Auwer, C.; Ding, M.; Dorhout, P. K.; Espinosa-Faller, F. J.; Gordon, P. L.; Haire, R. G.; Hess, N. J.; Hess, R. F.; Keogh, D. W.; Morales, L. A.; Neu, M. P.; Paviet-Hartmann, P.; Runde, W.; Tait, C. D.; Veirs, D. K.; Villella, P. M. Local and nanoscale structure and speciation in the PuO2+x−y(OH)2y·zH2O system. J. Am. Chem. Soc. 2004, 126 (41), 13443−13458. (9) Conradson, S. D.; Begg, B. D.; Clark, D. L.; den Auwer, C.; Ding, M.; Dorhout, P. K.; Espinosa-Faller, F. J.; Gordon, P. L.; Haire, R. G.; Hess, N. J.; Hess, R. F.; Keogh, D. W.; Lander, G. H.; Manara, D.; Morales, L. A.; Neu, M. P.; Paviet-Hartmann, P.; Rebizant, J.; Rondinella, V. V.; Runde, W.; Tait, C. D.; Veirs, D. K.; Villella, P. M.; Wastin, F. Charge distribution and local structure and speciation in the UO2+x and PuO2+x binary oxides for x ≤ 0.25. J. Solid State Chem. 2005, 178 (2), 521−535. (10) Burns P. C., F R. Uranium: Mineralogy, geochemistry and the environment. Rev. Mineral. 1999, 38. (11) Morris, D. E.; Allen, P. G.; Berg, J. M.; ChisholmBrause, C. J.; Conradson, S. D.; Donohoe, R. J.; Hess, N. J.; Musgrave, J.; A.Tait, C. D. Speciation of uranium in Fernald soils by molecular spectroscopic methods: Characterization of untreated soils. Environ. Sci. Technol. 1996, 30 (7), 2322−2331. (12) LoPresti, V.; Conradson, S. D.; Clark, D. L. XANES identification of plutonium speciation in RFETS samples. J. Alloys Compd. 2007, 444, 540−543. (13) Clark, D. L.; Choppin, G. R.; Dayton, C. S.; Janecky, D. R.; Lane, L. J.; Paton, I. Rocky Flats closure: The role of models in facilitating scientific communication with stakeholder groups. J. Alloys Compd. 2007, 444, 11−18. (14) Runde, W. The chemical interaction of actinides in the environment. Los Alamos Sci. 2000, 26, 1184−1188. (15) Burns, P. C.; Ewing, R. C.; Navrotsky, A. Nuclear fuel in a reactor accident. Science 2012, 335 (6073), 1184−8. (16) Rademacher, S. E. The Influence of Heterogeneity in Gamma Spectroscopy Analysis of Soils Contaminated with Weapons Grade Plutonium at the BOMARC Missile Accident Site, McGuire AFB NJ, IERA-SD-BR-SR-2001-0006; Air Force Institute for Environment, Safety, and Occupational Risk Analysis: Brooks Air Force Base, TX, 2001; pp 1−54. (17) Bowen, J.; Glover, S.; Spitz, H. Morphology of actinide-rich particles released from the BOMARC accident and collected from soil post remediation. J. Radioanal. Nucl. Chem. 2012, 296 (2), 853−857. (18) Chaparro, O. M. Heterogeneity effects in plutonium contaminated soils. M. S. Dissertation, Air force Institute of Technology, Wright-Patterson Air Force Base, OH, 2009. J


Article

Environmental Science & Technology (38) Lind, O. C.; S, B.; Skipperud, L.; Janssens, K.; Jaroszewicz, J.; De Nolf, W. Solid state speciation and potential bioavailability of depleted uranium particles from Kosovo and Kuwait. J. Environ. Radioact. 2009, 100 (4), 301−307. (39) Shoesmith, D. W. Fuel corrosion processes under waste disposal conditions. J. Nucl. Mater. 2000, 282 (1), 1−31. (40) Utsunomiya, S.; Ewing, R. C.; Wang, L. M. Radiation-induced decomposition of U(VI) phases to nanocrystals of UO2. Earth. Planet. Sci. Lett. 2005, 240 (2), 521−528. (41) Singer, D. M.; Zachara, J. M.; Brown, G. E. Uranium speciation as a function of depth in contaminated Hanford sedimentsA microXRF, micro-XRD, and micro- and bulk-XAFS study. Environ. Sci. Technol. 2009, 43 (3), 630−636. (42) Alessi, D. S.; Lezama-Pacheco, J. S.; Stubbs, J. E.; Janousch, M.; Bargar, J. R.; Persson, P.; Bernier-Latmani, R. The product of microbial uranium reduction includes multiple species with U(IV)-phosphate coordination. Geochim. Cosmochim. Acta 2014, 131, 115−127. (43) Ding, M.; Conca, J. L.; den Auwer, C.; Gabitov, R. I.; Hess, N. J.; Paviet-Hartmann, P.; Palmer, P. D.; LoPresti, V.; Conradson, S. D. Chemical speciation of heterogeneously reduced Pu in synthetic brines. Radiochim. Acta 2006, 94 (5), 249−259. (44) Towle, S. N.; Bargar, J. R.; Brown, G. E., Jr; Parks, G. A. Sorption of Co(II) on metal oxide surfaces: II. Identification of Co(II)(aq) adsorption sites on the (0001) and (1102) surfaces of αAl2O3 by grazing-incidence XAFS spectroscopy. J. Colloid Interface Sci. 1999, 217 (2), 312−321. (45) Bargar, J. R.; Brown, G. E., Jr; Parks, G. A. Surface complexation of Pb(II) at oxide-water interfaces: II. XAFS and bond-valence determination of mononuclear Pb(II) sorption products and surface functional groups on iron oxides. Geochim. Cosmochim. Acta 1997, 61 (13), 2639−2652. (46) Radioactive Particles in the Environment: Sources, Particle Characterization and Analytical Techniques, IAEA-TECDOC-1663; International Atomic Energy Agency, 2011; pp 1−77. (47) Salbu, B. Radionuclides released to the environment following nuclear events. Integr. Environ. Assess. Manage. 2011, 7 (3), 362−4. (48) Salbu, B. Hot particlesA challenge within radioecology. J. Environ. Radioact. 2001, 53, 267−268. (49) Kersting, A. B. Plutonium transport in the environment. Inorg. Chem. 2013, 52, 3533−3546.

K


Multiscale Speciation of U and Pu at Chernobyl, Hanford, Los

Recommend Documents