contents

Chemistry at the interface

Exploring thermal- and radiation-induced interactions of water on uranium dioxide surfaces

Most plans for the disposition of surplus and waste nuclear materials involve storage in sealed containers where the evolution of gases from reactions of adsorbed species, most notably water, can present both pressure and flammability hazards. Despite efforts such as calcining the material and handling in "dry" gloveboxes to minimize the water content before packaging, both residual moisture and readsorbed water are likely to be present in the final containers. Given the anticipated temperature excursions during transportation and storage, this water may thermally desorb, increasing the pressure, or thermally dissociate to produce hydrogen (H2) gas, increasing flammability hazards. In addition, radiation from the nuclear material may induce radiolysis of the water both in the gas phase and in the adsorbed state with likely products being H2, oxygen (O2), and hydrogen peroxide (H2O2).

For high-surface-area samples such as those typical of the materials disposition programs described in the article by Kirk Veirs on Page 7, the overwhelming majority of the water would be in the adsorbed state. Predictive modeling over the storage lifetimes of such containers requires understanding the thermal- and radiation-induced chemistry that occurs in multiple phases (gas, solid, and interface). There has been considerable study of the radiolysis of the homogeneous phases of water. For the heterogeneous adsorbed state, the sparser existing literature indicates that significant changes in radiolysis yields and product distributions can occur. To better understand the relative importance of the thermal- and radiation-induced chemistry, we have studied the interactions of water on single crystals of uranium dioxide (UO2) where the experiments have been designed to specifically address the issues of chemistry at the interface.

Normal view of the UO2 (111) surface. Red spheres are oxygen; black spheres are uranium. The blue hexagon is the plane through the uranium atoms and indicates the three-dimensional nature of the surface.

All of the studies presented here were conducted using single-crystal samples of uranium dioxide in an ultrahigh vacuum environment. Currently only thorium dioxide (ThO2) and uranium dioxide are available as macroscopic single crystals of the actinide oxides. Such an approach offers a number of advantages for understanding the fundamental mechanisms of the interactions of water with actinide oxide surfaces. Single crystals, prepared and characterized using standard surface analytical techniques, have well-defined structures and compositions with minimal defects, making them ideal for molecular-level studies.

Uranium dioxide, as with many of the actinide dioxides, adopts a fluorite lattice, where uranium is formally U4+ and eight-fold coordinated and oxygen is formally O2- and four-fold coordinated. The (111) fluorite surface, used for most of the studies discussed here, is composed of one seven-fold uranium and two three-fold oxygens. This surface is stoichiometric, nonpolar, and formed by cleaving the fewest bonds. The (111) is the lowest-energy termination and does not reconstruct-meaning the surface atoms are essentially in locations consistent with an ideal termination of the bulk lattice-whereas for higher-energy surfaces such as the (100) and (110) the surface atoms often rearrange to lower the overall free energy. As such, studies of the (111), with its known geometric structure, are ideal for comparisons between experiment and theory.

In addition to the obvious benefits of preparing, characterizing, and maintaining clean samples, working in ultrahigh vacuum also aids in isolating the elementary chemical steps in the thermal- and radiation-induced chemistry of water on surfaces. The absence of a gas phase, except during precisely controlled exposures, minimizes the effects of readsorption, which can greatly complicate kinetic analyses. Vacuum also eliminates any contributions from gas-phase radiolysis.

The thermal desorption of adsorbed deuterium oxide (D2O) from uranium oxide (UO2) (III).

Temperature programmed desorption (TPD) was the principal experimental technique used to investigate the thermal chemistry of water on uranium dioxide surfaces. In TPD, the surface is exposed to controlled quantities of water at low temperatures, the surface is then heated, and gaseous product evolution is monitored with a mass spectrometer. Evolved gas-phase products are directly determined and binding energies and reaction kinetics can be obtained from an analysis of the desorption temperature profiles. Essentially, the more strongly bound a species, the higher the temperature required to desorb it.

The TPD results for clean, minimally defected uranium dioxide surfaces, exposed to deuterium oxide (D2O), are relatively straightforward. The only desorbing species is water; we didnÕt detect any deuterium gas (D2), O2, or other species that would indicate a dissociative adsorption of water. Two distinct states for adsorbed water were observed: low-temperature nonsaturating, and high-temperature saturating. The low-temperature nonsaturating state can be attributed to the formation of condensed multilayers of water on the surface with a binding energy of 11 kilocalories per mole (kcal/mole) consistent with the sublimation of water ice. The high-temperature saturating state can be attributed to the direct interaction between water and the uranium dioxide surface. A simple kinetic analysis of this monolayer state yields a binding energy of approximately 13Ð15 kcal/mole. This value compares favorably with both ab initio electronic structure calculations and experimentally determined binding energy of water with isostructural plutonium dioxide.

The TPD results indicate that the majority of water interacts molecularly on clean uranium dioxide (111) surfaces with a binding energy only slightly greater than water with itself. However, this small change can have important implications in "real-world" environments. A steady-state kinetic model can be used to determine the vapor pressure of water needed to form the direct, monolayer film. Based on the TPD-derived binding energy of 13Ð15 kcal/mole, a water pressure of only a few millitorr is enough to saturate the monolayer film at room temperature. This pressure is below that used in most gloveboxes, so most processed material will, except in the driest boxes, readsorb one layer of water. Most of the standards developed for nuclear materials storage require accounting for at least this level of water loading when designing the containers.

In addition to the thermal chemistry of the actinide oxides, there exists the potential for radiation-induced chemistryÑor radiolysis. All of the isotopes of the actinides are radioactive and thus can emit a variety of particles and ionizing radiation including alpha (a), beta (b), and gamma (g) particles; neutrons; recoil and fission atoms; and secondarily produced electrons, ions, and x-rays. These decay particles, for example, the approximately five-million-electronvolt (MeV) a typical of the more common actinides including uranium-238 and plutonium-239, transfer their considerable energy to the surrounding media, inducing reaction and displacing atoms. The short-lived primary radiolytic products such as ions, electrons, atoms, and free radicals initiate further ion-molecule and free-radical reactions ultimately leading to more stable radiolysis products. These final radiation-induced products are often quite distinct from the thermal chemical products.

While the objective is to understand the radiation chemistry that occurs at the surfaces of radioactive materials, there are practical and scientific advantages to using external radiation sources on either low-activity or surrogate substrates. Time and safety issues are the most obvious. For a modestly active material such as plutonium-239, approximately 10,000,000 alpha particles traverse a square centimeter of surface per second. Assuming ten radiolytic events at the surface per alpha particle, it would take 10,000,000 seconds (about four months) to convert a monolayer (approximately one quadrillion sites per square centimeter). While relatively short compared to storage timescales, this time is inconveniently long for most laboratory studies.

The evolution of deuterium gas (D2) and oxygen (O2) following electron irradiation of deuterium oxide (D2O)-covered uranium dioxide (UO2) (III).

By necessity, high-surface-area materials are typically used to increase signals when relying on the intrinsic radioactivity. Electron guns and ion beams have currents ranging from nanoamps to microamps with typical beam sizes on the order of several millimeters allowing for monolayer conversions in seconds to hours. This accelerated rate allows for the use of planar, single-crystal substrates. Alpha particles lose more than 99 percent of their energy to the surrounding media by ionization events and the production of secondary electrons. The mean energy of secondary electrons is on the order of 100 electronvolts (eV) with approximately 10,000 secondary electrons produced per alpha decay event. The alpha produces only about 200 lattice-atom displacements; the majority of the lattice damage, more than 2,000 displacements, results from the recoil atomÑa 70 kiloelectronvolt (keV) thorium-234 for uranium-238. Both ionization and displacement effects can be simulated by using convenient low-energy laboratory sources.

Low-energy electrons, 100Ð500 eV, were used to simulate the ionization-induced chemistry of water adsorbed on uranium dioxide surfaces. The primary radiation-induced products are D2 gas, O2 gas, and small amounts of water. The absolute and relative amounts of D2 and O2 produced are functions of the temperature and coverage of water. The highest yields occur for multilayer water. D2 is produced in hyperstoichiometric amountsÑfor example, the D2:O2 ratio is at least 5; ideally it would be 2 for D2O (deuterium oxide or heavy water). The cause of the oxygen deficiency in the neutral products is poorly understood.

Hyperstoichiometric hydrogen production (with respect to O2) is observed in the radiolysis of pure gas- and liquid-phase water. In these cases, the oxygen balance is maintained through the formation of H2O2. No peroxide gas evolved during irradiation although small amounts were retained in the multilayer ice as observed during TPD following irradiation. For monolayer and lower coverage, the overall yields decreased, but the D2:O2 ratio actually increased (exceeding 10). No peroxide was observed for monolayer coverages. Although we are awaiting more definitive experiments, it is likely that some of the "missing" oxygen becomes incorporated into the uranium dioxide substrate.

In addition to neutral species, positive ion desorption during irradiation was monitored as well. For a nominally clean uranium dioxide surface, electron irradiation yielded oxygen (O), hydrogen (H), and fluorine (F) cations. The H+ came from the small adsorption of water from the background gases in the system, and the F+ was attributed to impurities in the fluxes used to produce the uranium dioxide crystal. This result highlights the potential sensitivity of ion desorption to minority low-concentration species. A D+ ion was the only species observed from multilayers of water, which is consistent with previous work (heavy water was used to distinguish between traces of background water in the vacuum chamber).

The thermal desorption of deuterium gas (D2) following deuterium oxide (D2O) exposure to sputter-damaged uranium dioxide (UO2) (III).

For submonolayer amounts of water, D+ and OD+ were the primary products. The D+ and OD+ signals persisted to surface temperatures up to 600 kelvin (K). From the thermal studies discussed above, all molecular water was observed to desorb by 300 K. To determine the origin of the OD species, isotopically labeled 18O was incorporated into the surface by ion sputtering. Upon exposure to D216O, 18OD ions were observed, indicating that at least some of the oxygen comes from the substrate as the result of dissociation of water at minority defect sites not observed in the TPD studies.

The observation of OD+ suggests that defects on the uranium dioxide surface could have a dramatically different chemistry than the majority sites. To further explore this, we intentionally introduced defects into the surface by low-energy argon (Ar) ion sputtering. Ion bombardment of surfaces produces numerous defects at the surface and the near-surface region. In addition to the sputtered surface species, which yield surface vacancies, the collisional cascade creates lattice defects such as vacancy or interstitial (Frenkel) pairs.

The collisional cascades created by 5 keV Ar+ ions are very similar, differing principally in extent, to those produced by a recoil atom. In compounds, preferential sputtering of one element is often observed. In general, lower-mass species are more likely to be sputtered. For metal oxides, the preferential sputtering of oxygen becomes more pronounced as the metal becomes heavier. For uranium dioxide, oxygen depletion of the near-surface region has been observed using x-ray photoelectron spectroscopy (XPS) after ion bombardment.

TPD experiments were performed for both annealed (minimal defects) and sputtered (highly defected) uranium dioxide surfaces following exposure to heavy water. Significant deuterium gas desorption was observed above 400 K for the defective surface but was not observed for the annealed surface. The hydrogen evolution from the defective uranium dioxide surface likely resulted from the dissociative adsorption of heavy water at defect sites to form adsorbed hydroxyls. There were no other high-temperature desorption peaks observed at m/z (mass-to-charge ratio) 20 or 32 to suggest that there is a high-temperature reaction that leads to the gas-phase evolution of a product containing the oxygen from the water. This oxygen could, therefore, remain at the surface and heal the defects that lead to D2 gas generation.

To provide experimental evidence of the reoxidation of the sample by water, TPD experiments were performed for successive heavy-water exposures to the sputtered sample surface. To minimize thermally healing the defects, the TPD heating ramp for this series of experiments was terminated at 500 K, at which point the majority of D2 gas had desorbed but was still about 200 K lower than where the defects were observed to be thermally annealed. After the first heavy-water dose there was a significant decrease in the amount of hydrogen desorption for the second and subsequent exposures. After five exposures, the TPD for the sputtered and annealed surfaces were indistinguishable. Water therefore can heal defects, presumably oxygen vacancies, with the attendant evolution of hydrogen gas.

The irradiation results indicate that over prolonged periods, significant radiolytic production of H2 and O2 can occur. In all cases, H2 was produced in excess of O2, which is consistent with previous work on gas generation from plutonium dioxide powders. As the absolute yields decrease and relative H2:O2 yields increase with decreasing water loading, these results suggest that minimizing moisture content would be beneficial not only in reducing the pressure hazards, but also in ensuring that the gas composition is outside the flammability limits for hydrogen and oxygen mixtures. We have also identified an additional channel for hydrogen production: surface defects. Because the damage resulting from ion sputtering is similar to that caused by the recoil atom formed during radioactive decay, the thermal chemistry at defects may be an indirect but significant factor in the radiation chemistry of adsorbed species on alpha-emitting materials.

This article was contributed by Jeffrey Stultz (currently at U.S. Borax) and Mark Paffett and Stephen Joyce, Chemistry Division. J.S. acknowledges postdoctoral support from the Seaborg Institute.


NMT | LANL | DOE
Phone Book | Search | Help/Info

L O S  A L A M O S  N A T I O N A L   L A B O R A T O R Y
Operated by the University of California for the US Department of Energy

Questions? - Copyright © UC 1998-2000
- For conditions of use, see Disclaimer