The kinetics and mechanism of H2O2 decomposition at the U3O8 surface in bicarbonate solution

In the event of nuclear waste canister failure in a deep geological repository, groundwater interaction with spent fuel will lead to dissolution of uranium (U) into the environment. The rate of U dissolution is affected by bicarbonate (HCO3−) concentrations in the groundwater, as well as H2O2 produced by water radiolysis. To understand the dissolution of U3O8 by H2O2 in bicarbonate solution (0.1–50 mM), dissolved U concentrations were measured upon H2O2 addition (300 μM) to U3O8/bicarbonate mixtures. As the H2O2 decomposition mechanism is integral to the dissolution of U3O8, the kinetics and mechanism of H2O2 decomposition at the U3O8 surface was investigated. The dissolution of U3O8 increased with bicarbonate concentration which was attributed to a change in the H2O2 decomposition mechanism from catalytic at low bicarbonate (≤5 mM HCO3−) to oxidative at high bicarbonate (≥10 mM HCO3−). Catalytic decomposition of H2O2 at low bicarbonate was attributed to the formation of an oxidised surface layer. Second-order rate constants for the catalytic and oxidative decomposition of H2O2 at the U3O8 surface were 4.24 × 10−8 m s−1 and 7.66 × 10−9 m s−1 respectively. A pathway to explain both the observed U3O8 dissolution behaviour and H2O2 decomposition as a function of bicarbonate concentration was proposed.


Introduction
The current strategy for the disposal of spent nuclear fuel is in a deep geological repository according to the majority of the international community. The repositories provide a long-term storage solution, yet the release of radioactive species from spent nuclear fuel into the environment from the repository is projected to occur in the future upon failure of the repository barriers. Therefore, it is necessary to develop safety models for the repositories to predict their performance when failure occurs and nuclear material is exposed to the local environment. The main pathway for radionuclide release is predicted to be caused by the ingress of groundwater into the repository and interaction of the groundwater with the surface of the spent fuel. Understanding the reaction mechanisms between groundwater and spent fuel is integral to the development of safety models. Such interactions between the groundwater and spent fuel will lead to dissolution of the UO 2 matrix which constitutes the majority of the spent fuel. 1 The solubility of U in groundwater is governed by the form of U (U (IV) , U (V) and U (VI) ), with the hexavalent U (VI) form being more soluble than U (IV) and U (V) . [2][3][4] Therefore, the presence of U (VI) facilitates U dissolution into the groundwater upon canister failure.
Under the reducing, anoxic conditions typically found in groundwater at repository depths, the solubility of U (IV) is very low, [5][6][7] and so signicant dissolution of the UO 2 spent fuel may not be expected. However, radiation from the spent fuel will cause radiolysis of fuel adjacent water leading to the formation of a complex water chemistry involving radical, ionic and molecular species in the form of both reductants (e aq À, H $ , H 2 ) and oxidants (OH $ , H 2 O 2 ). 8 This will signicantly affect the local redox chemistry of the water and the oxidation state of U.
Of the oxidants generated by radiolysis, it has been shown that H 2 O 2 is the dominant species in regards to U dissolution under deep geological repository conditions. 9,10 The interaction of H 2 O 2 with the UO 2 surface has been thoroughly studied due to its importance for U dissolution, and it has been proposed that there are two competing pathways, both of which involve the decomposition of H 2 O 2 at the UO 2 surface. 11 In this case, the H 2 O 2 decomposition does not directly cause U dissolution. The second is an oxidative decomposition reaction where H 2 O 2 oxidises U (IV) to U (V) (eqn (4)) and U (V) to U (VI) (eqn (5)) while itself being reduced to OH À (eqn (6)) in a redox couple.
Typically, groundwater also contains bicarbonate (HCO 3 À ) which has been shown to enhance the dissolution of U due to favourable complexation with U (VI) and stabilisation of the dissolution products: [13][14][15][16] U IV + HCO 3 À / U V (HCO 3 ) ads + e À (7) Therefore, the concentration of bicarbonate is believed to have a signicant effect on U dissolution and the rate of H 2 O 2 decomposition at the UO 2 surface.
Due to the importance of developing models to predict U dissolution into groundwater, various studies have been undertaken with UO 2 in simulated groundwater. A recent study by Kumagai et al. 17 has shown that increasing the oxygen content from UO 2 to UO 2.3 increased U dissolution and reduced the rate of H 2 O 2 decomposition at the oxide surface. As the dissolution of U is governed by the redox behaviour of the U atoms, it follows that the ratio of U (IV) , U (V) and U (VI) will have a signicant impact on both U dissolution as well as the H 2 O 2 decomposition pathway. Therefore, the form of uranium oxide that exists on the spent fuel oxide will have a large effect on the dissolution of U into the environment. Due to the radiolysis of spent fuel surface adjacent groundwater and the elevated temperatures from spent fuel decay, the formation of highly oxidised forms of U is expected i.e., where x > 0.3 for UO 2+x . However, there is still a lack of knowledge regarding the impact of higher oxidised forms of U on the mechanism of U dissolution and H 2 O 2 decomposition.
To investigate this, we adopted U 3 O 8 as an extreme case, which corresponds to UO 2.66 containing two U (V) atoms and one U (VI) atom. 18-20 U 3 O 8 has been observed on used nuclear fuel both in wet 21 and air 22,23 environments, and can be used as a highly oxidised form of uranium oxide for an examination of the effects of U valence on U dissolution. As the complexation of bicarbonate with U (VI) is thought to drive U dissolution by favourable complexation, the effect of U oxidation state on the dissolution of U in bicarbonate solution can be investigated by using U 3 O 8 . The H 2 O 2 decomposition mechanism is dependent on U oxidation and so can also be investigated using U 3 O 8 for comparison with UO 2 . The concentration of bicarbonate in groundwater is dependent on the location of the deep geological repository, and can range from $10 À4 M (Tono, Japan), 24 to $10 À3 M (Daejeon, South Korea), 25,26 to $10 À2 M (Forsmark, Sweden) 27 and so it is necessary to understand U dissolution and H 2 O 2 decomposition at uranium oxide surfaces over a range of bicarbonate concentrations.
Therefore, in this work, U dissolution from U 3 O 8 suspensions with H 2 O 2 as a function of sodium bicarbonate (NaHCO 3 ) has been investigated, and the mechanism of H 2 O 2 decomposition at the U 3 O 8 surface has been elucidated.

Materials
Two samples of U 3 O 8 powder were used in this study to investigate the reproducibility of the U dissolution tests. The rst U 3 O 8 powder (sample 1) was prepared by heating UO 2 powder to 750 C for 3 hours under a continuous ow of air. The second (sample 2) was prepared by dissolving U metal in 13 M HNO 3 (Fujilm Wako Pure Chemical, 60%) to form UO 2 (NO 3 ) 2 $(H 2 O) n , which was then heated under identical conditions to give a 96% yield of U 3 O 8 . The formation of U 3 O 8 was conrmed by XRD and the data was rened using the Rietveld method. 28 The average crystallite size was measured using the Scherrer equation: 29 where d is the mean crystallite size, l is the X-ray wavelength (1.5406Å), b is the full width at half maximum value, and q is the diffraction peak position. The crystallite sizes were calculated as 47 and 46 nm for sample 1 and 2 respectively. The orthorhombic lattice constants were also calculated from the diffractograms using Bragg's law 30 for orthorhombic structures (1/d hkl 2 ¼ h 2 /a 2 + k 2 /b 2 + l 2 /c 2 ) giving values of a ¼ 6.72, b ¼ 11.96 and c ¼ 4.15Å for sample 1 and a ¼ 6.71, b ¼ 11.95 and c ¼ 4.14 A for sample 2. The lattice constants were consistent with those for U 3 (Fig. 2). A decrease in the concentration of dissolved U can be seen in 5, 10 and 20 mM NaHCO 3 solution, suggesting deposition from solution of U onto the U 3 O 8 surface aer the initial addition of H 2 O 2 as highlighted by the second y-axis. As the extent of deposition increased with bicarbonate, it can be predicted that the deposits are uranium carbonates. Under the experimental conditions, the stable form of U in solution is UO 2 (CO 3 ) 3 4À and so the deposits may be UO 2 (CO 3 ) 3 . Another possibility is the formation of uranyl peroxide (UO 2 (O 2 )), where the increase in deposition with bicarbonate is due to an increase in dissolved U with bicarbonate and, therefore, an increase in uranyl peroxide.

Kinetics of H 2 O 2 decomposition
The dissolution of U from U 3  The calculated values of k from the linear region (from t ¼ 2 hours to the experiment end) are plotted against bicarbonate concentration in Fig. 5 and the value of k was found to be in the range between 0.4 to 1.6 Â 10 À5 s À1 . The decrease in the pseudo-rst order rate constant coincided with U deposition from solution indicating that the secondary phases that deposit on the surface of the U 3 O 8 may block the approach of H 2 O 2 to the surface. As the plots in Fig. 4 show linear behaviour, this suggests that these deposits are stable over the experimental timescale.

U dissolution with H 2 O 2 decomposition
The mechanism of U dissolution via H 2 O 2 decomposition can be investigated by analysing the extent of U dissolution as      dissolution of U from the surface, it follows that the two-step oxidation of U (IV) / U (V) / U (VI) for UO 2 would result in less dissolution of U than the one-step oxidation for U (V) / U ( relative to UO 2 is more pronounced than the measured dissolved U concentrations suggest. Using the ratios taken from the gradients, the contributions of catalytic (k cat ) and oxidative (k ox ) decomposition to the measured pseudo-rst order rate constant can be found and are plotted in Fig. 7 as a function of bicarbonate.
At low bicarbonate concentrations, the main pathway for H 2 O 2 decomposition is the catalytic decomposition mechanism as there is little U dissolution associated with H 2 O 2 decomposition. As k ox is low, this indicates that the U 3 O 8 is protected from H 2 O 2 by a surface layer. It is postulated that upon addition of H 2 O 2 to the bicarbonate solution, oxidative dissolution proceeds on the bare U 3 O 8 surface (Fig. 8). As this involves the oxidation of U (V) to U (VI) , it is likely that U (VI) forms a surface layer on the U 3 O 8 which protects against further oxidative dissolution as U (VI) is already fully oxidised, and due to the low concentration of bicarbonate the surface layer is stable.
The composition of the surface layer is thought to be in the hydroxide form (UO 2 (OH) 2 $xH 2 O) due to the formation of hydroxide from the oxidative decomposition of H 2 O 2 . Raman analysis of the U 3 O 8 surface aer removal from solution and vacuum drying showed spectra representative of U 3 O 8 only (Fig. 9). Peaks relating to U 3 O 8 were observed including the U-O A1 g stretching modes at 335 and 410 cm À1 , and the U-O E g stretching mode at 475 cm À1 . 37 As U 3 O 8 was the only phase observed, any surface layer that formed had been removed prior to Raman analysis. If the surface layer is in the hydroxide form, it is expected to decompose upon drying which would explain the observed results. Further studies are required to elucidate the composition of the surface.
As the bicarbonate concentration increases from 0.1 to 5 mM, the rate of catalytic H 2 O 2 decomposition decreases. This is caused by an increase in deposition from solution as seen in Fig. 2. As the pseudo-rst order rate constant decreases up to 5 mM, it can be said that the deposits do not catalyse H 2 O 2 decomposition to the extent that U 3 O 8 does.
Increasing the bicarbonate concentration > 5 mM changes the main H 2 O 2 decomposition mechanism pathway from catalytic to oxidative. At NaHCO 3 concentrations of 10 mM and above, at least 90% of the H 2 O 2 decomposed via oxidation of U (V) to U (VI) . This is due to increased dissolution of the U (VI) surface layer and exposure of the U 3 O 8 surface beneath leading an increase in k ox . As the value of k increases with bicarbonate, this suggests that the dissolution step is rate determining rather than the redox reaction. Therefore, dissolution experiments in solutions of higher bicarbonate concentrations are required to elucidate the true value of k for oxidative dissolution in this system. Interestingly, a study by Nilsson et al. on UO 2 dissolution in 10 mM NaHCO 3 with H 2 O 2 addition using pellets Fig. 7 The catalytic (k cat ) and oxidative (k ox ) pseudo-first order rate constants for H 2 O 2 decomposition on U 3 O 8 as a function of bicarbonate concentration.  showed that $14% of H 2 O 2 decomposition events occurred via oxidative dissolution while the value was even lower ($2%) on SIMFUEL. 38 This suggests that k cat is high in the case of the pellets indicating that the surface oxide that forms on the pellets is more protective than on the powders. The large discrepancy between the H 2 O 2 decomposition behaviour between UO 2 pellets and U 3 O 8 powder (and UO 2 powder) is a point that requires investigation.
To clarify the dependence of the catalytic and oxidative mechanisms on U 3 O 8 , the second order rate constants for H 2 O 2 decomposition were obtained for 0.1 mM and 50 mM solutions with U 3 O 8 (sample 2). The second order rate equation, can be used to obtain the second order rate constant by plotting the pseudo-rst order rate constant against the U 3 O 8 surface area to total solution volume ratio (Fig. 10). The second order rate constant in 0.1 mM bicarbonate was 4.24 Â 10 À8 m s À1 . At this concentration, the decomposition was shown to be almost completely catalytic, and so this can be attributed to the catalytic decomposition reaction pathway shown in eqn (1)-(3). At 50 mM, the value of the measured second order rate constant was 8.44 Â 10 À9 m s À1 , and as the ratio of oxidative decomposition was $90%, we can estimate the oxidative decomposition rate constant to be 7.60 Â 10 À9 m s À1 for the pathway shown in eqn (5) and (6). These values are within the range described in the literature for catalytic decomposition (3.6 Â 10 À8 to 5 Â 10 À11 m s À1 ) and oxidative decomposition (1.4 Â 10 À7 to 2.0 Â 10 À10 m s À1 ) of H 2 O 2 at the UO 2 surface. 39 The pseudo-rst order rate constant measurement for 0.1 mM and 50 mM bicarbonate solutions using U 3 O 8 sample 1 (shown in Fig. 5) are included in Fig. 10 showing the reproducibility of the data using different U 3 O 8 powders.

Proposed pathway for U 3 O 8 dissolution by H 2 O 2 in NaHCO 3 solution
From the experimental results, a proposed pathway to explain the observed behaviour of U 3 O 8 in bicarbonate solution with H 2 O 2 is summarized, and a schematic is provided in Fig. 11. At low bicarbonate concentrations upon H 2 O 2 addition, oxidative decomposition of H 2 O 2 occurs at the exposed U 3 O 8 surface forming a surface layer comprised of U (VI) that provides protection against further oxidative dissolution. The decomposition of H 2 O 2 proceeds via catalytic decomposition, and so the rate of U dissolution is low. The surface layer protects the U 3 O 8 in bicarbonate concentrations up to 5 mM, and the H 2 O 2 decomposition mechanism remains catalytic and U dissolution remains low. At 10 mM bicarbonate, the concentration of bicarbonate is sufficient to induce dissolution of the surface layer, and the surface layer does not fully protect the U 3 O 8 which is exposed leading to oxidative decomposition of H 2 O 2 and an increase in U dissolution. At higher bicarbonate concentrations, the surface layer is further dissolved, and oxidative decomposition of H 2 O 2 and dissolution of U proceeds at higher rates.