Diamond Annual Review 2019/20

44 45 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 MagneticMaterials Group Beamline I16 fundamental reasons for this large difference in the corrosion rates of the two materials. At Bristol University epitaxial films of both UN and U 2 N 3 with a thickness of ~ 200 nm were grown 2 for the experiments on I16 at Diamond Light Source. Whereas the crystal structure of UN is the simple rocksalt structure, that of U 2 N 3 is considerably more complicated, being the so-called 'bixbyite structure', and a perspective on this is shown in Fig. 1. There are two independent sites for the uranium atom (U 1 and U 2 ), whereas only one exists in UN. A number of methods, including photoemission experiments 2 , have been used to estimate the valency of these materials, and for UN this is ~ 3+, i.e. U(III), but for U 2 N 3 the valency is higher. Such methods are not site selective, so leave open the question of the valency at each individual site. This is important as the U(VI) valent state is highly soluble in water, so if that is present in at least one of the sites of U 2 N 3 , this could explain its high corrosion rates. The magnetic properties give one clue to the valency; for example, U(VI) has no 5 f electrons so cannot be magnetic. Resonant X-ray Diffraction with the beam energy tuned to the uranium M 4 edge at 3.726 keV showed that U 2 N 3 is antiferromagnetic and the magnetic wave-vector was determined for the first time (no single crystals of U 2 N 3 have been prepared previously), but the precise magnetic configuration remains ambiguous. Neutrons from the ISIS spallation source were used to try and answer this question – the use of neutrons and resonant X-rays being a powerful combination 3 . In an effort to extract further information on the valency and bonding of the two separate uranium sites 'Diffraction absorption experiments' were performed at the U M 4 edge on a number of Bragg reflections of the U 2 N 3 film. The reflections have different contributions from the two independent uranium atoms, as the atomic sites have different symmetries. For strong Bragg reflections, in which the scattering fromboth the U 1 and U 2 atoms are in- phase, or one set is absent, the expected result is a dispersive curve that reflects the combined effect of both the real (f o + f') and imaginary (f'') parts of the uranium scattering factor. Such a curve is shown in the green curve in Fig. 2 for the (112) reflection, in which only the U 2 atoms participate. (Scattering from nitrogen is neglected, as it is far weaker than that from uranium; in addition, there is no edge sensitive to nitrogen in the energy range covered). However, for reflections in which the strong Thomson scattering (from the 86 core electrons of uranium) is reduced by the cancellation between the two separate uranium atoms, a very interesting profile is shown in Fig. 2 that is precisely the energy profile at the M 4 edge of the imaginary part (f'') from the uranium atoms. This profile reflects the fact that around one of the uranium sites is an aspherical charge density , which involves the uranium 5 f electrons. For example, for the (013) reflection, which is forbidden and has no contribution from the Thomson (spherical) charge density, this aspherical part is the only contribution to the scattering intensity. Similarly, for the (002) and (022), in which the strong spherical charge density contributions almost cancel, the aspherical part is also observed. From the pattern of the intensities in Fig. 2, it becomes clear that any aspherical contribution from the U 1 sites must be small, suggesting that these sites may possibly have the U(VI) valency, in which there are no occupied 5 f states. This effect has been observed before, mainly at the K edge of the transition metals 4 . However, at the K edge with the d transition metals there is the possibility of both dipole and quadrupolar transitions, making the identification of the underlying physics complicated. For the U M 4 edge this ambiguity is removed; the transition is definitely of dipole symmetry illuminating an aspherical shape known as a charge quadrupole . The non- centrosymmetric coordination of this distribution around the uranium nucleus then couples to the imaginary scattering factor (f'') giving rise to scattered intensity, with a distinctive energy profile, at the Bragg position. Such charge quadrupoles have been observed previously in the actinides, but they are associated with effects related to the magnetism 5 . In this case, the charge quadrupoles are temperature independent, and so not related to the magnetic order at ~ 75 K. They are certainly induced by covalency, probably mixing between the uranium 5 f states and the nitrogen 2 p states. In conclusion, these experiments strongly suggest that the U 1 site may have a significantly higher valency, quite possibly U(VI), and this is responsible for the rapid corrosion rates of U 2 N 3 . In addition, these experiments have opened the way for more quantitative modelling in such systems, based on the observation of an aspherical 5 f charge distribution around the U 2 atom References: 1. Lawrence Bright E. et al. Comparing the corrosion of uranium nitride and uranium dioxide surfaces with H 2 O 2 . J. Nucl. Mater. 518 , 202-207 (2019). DOI: 10.1016/j.jnucmat.2019.03.006 2. Lawrence Bright E. et al. Epitaxial UN and α-U 2 N 3 thin films. Thin Solid Films 661 , 71-77 (2018). DOI: 10.1016/j.tsf.2018.07.018 3. Wadley P. et al. Antiferromagnetic structure in tetragonal CuMnAs thin films. Sci. Rep. 5 , 17079 (2015). DOI: 10.1038/srep17079 4. Kokubun J. et al. Anisotropic resonant X-ray scattering: Beauty of forbidden reflections. The European Physical Journal Special Topics 208 , 39 (2012). DOI: 10.1140/epjst/e2012-01605-4 5. Santini P. et al. Multipolar interactions in f-electron systems: The paradigm of actinide dioxides. Rev. Mod. Phys. 81 , 807 (2009). DOI: 10.1103/RevModPhys.81.807 Funding acknowledgement: EPSRC Grant No. 1652612 Corresponding author: Gerard Lander, Institute Laue Langevin, Grenoble, lander@ill.fr Figure 1: A projection of the bixbyite structure of U 2 N 3 which has a cubic unit cell with a o = 10.67 Å and space group #206 with 32 uranium atoms in the unit cell. The nitrogen atoms are shown in red. 8 of these called U 1 (in blue) have C 3i symmetry and 24 uranium atoms, called U 2 (in grey) have C 2 symmetry. Both U atoms have a non-centrosymmetric local environment, but this is more exaggerated (as can be seen from the figure) in the case of the U 2 (grey) atoms. Figure 2: Energy profiles of various Bragg reflections from the U 2 N 3 film. The profiles are independent of temperature. A normal “energy dispersive” curve is shown in green from the (112) reflection. The other profiles represent reflections in which the strong (spherical) Thomson scattering from the two uranium sites cancels, or almost cancel. They represent so-called anisotropic resonant X-ray scattering 4 and show that there is an aspherical charge density associated with the U 2 sites. Such a charge density is almost certainly due to covalency between the uranium 5f electrons and the 2p states of nitrogen. Understanding the structure of uraniumnitride Related publication: Lawrence Bright E., Springell R., Porter D. G., Collins S. P. & Lander G. H. Synchrotron X-ray scattering of magnetic and electronic structure of UN and U 2 N 3 epitaxial films. Phys. Rev. B 100 , 134426 (2019). DOI: 10.1103/PhysRevB.100.134426 Publication keywords: Actinides;Thin films; Single crystals; Resonant elastic X-ray diffraction U ranium mononitride (UN) has many possible advantages over uranium dioxide as a nuclear fuel, including being safer, stronger and more thermally conductive and having a higher temperature tolerance. Further fundamental research on UN and uranium sesquinitride (U 2 N 3 ) is required to understand their potential as advanced nuclear fuels. Formed on the surface of UN, U 2 N 3 dissolves in water. This has serious implications for the safe disposal of spent fuel, but what causes the dissolution? A possible reason is the presence of the highly soluble U(VI) valent state in U 2 N 3 . Researchers used a technique called Resonant X-ray Scattering to investigate U 2 N 3 . The Materials and Magnetism beamline (I16) combines X-ray diffraction to study atomic structure with spectroscopy to observe electronic structure. Its combination of high X-ray flux and low background detectors allows for the detection of weak signals in the energy range of interest for this experiment. Their results indicate that one of the uranium sites in U 2 N 3 might be U(VI) valent, which is known to be highly soluble in water. In addition, they found strong evidence for covalency involving the 5 f states in uranium. The specific signal of covalency has not been observed in actinides previously, somay open new avenues for the better characterisation of thesematerials. Developing theoretical models to explain these experimental results should lead to a deeper understanding of the properties of the materials. Uranium nitride is a possible 'accident tolerant fuel' that has many characteristics, such as a higher uranium density and a high thermal conductivity, making it a better choice than the conventional fuel UO 2 . However, the oxidation of UN progresses through the formation of a surface layer of U 2 N 3 , and that material has been found to corrode some 25 times more rapidly than UN itself 1 . Our experiments have endeavoured to find some