Formation of Solid Thorium Monoxide at Near-Ambient Conditions as Observed by Neutron Reflectometry and Interpreted by Screened Hybrid Functional Calculations

Oxidation of a ~1000Å sputter-deposited thorium thin film at 150°C in 100ppm of flowing oxygen in argon produces the long-sought solid form of thorium monoxide. Changes in the scattering length density (SLD) distribution in the film over the 700 min experiment measured by in-situ, dynamic neutron reflectometry (NR) shows the densities, compositions and thickness of the various thorium oxides layers formed. Screened, hybrid density-functional theory calculations of potential thorium oxides aid interpretation, providing lattice dimensions and energetics of oxygen migration. NR provided evidence of the formation of substoichiometric thorium oxide, ThOy (y<1) at the interface between the unreacted thorium metal and its dioxide overcoat which grows inward, consuming the thorium at a rate of 1.5 Å/min while y increases until reaching 1:1 oxygen-to-thorium. Its presence indicates that kinetically-favored solid-phase ThO can be preferentially generated as a majority phase under the thermodynamicallyfavored ThO2 top layer at conditions close to ambient. Introduction Thorium is one of only two actinides with commercial applications independent of its radioactive nature. Thorium is used to impart high strength and creep resistance in magnesium alloys. [1] [2] It also has utility as an oxide as a catalyst [3] and in high-quality lenses for cameras and scientific instruments. [4]. The phase diagram for thorium and oxygen shows that only one oxide phase is thermodynamically present at ambient pressure: thorium dioxide. [5] ThO2 has one of the highest melting point (3300 °C) of all oxides [6] and thus used in heat-resistant ceramics [3] and in mantles of portable gas lights. The availability of 5f orbitals diffuse enough to be involved in molecular bonding enables unusual chemical compounds. Thorium atoms can also bond to more atoms at one time than any other element. For instance, in the compound thorium aminodiboranate, thorium has a coordination number of fifteen. [7] The most important future application of thorium may be as an advanced nuclear fuel. [8] Current research in nuclear power generation is aimed at reduction of Pu and minor actinides in the spent light water reactor’s fuel stockpiles. Considerable research efforts are underway to evaluate the suitability of Th as a nuclear fuel. However, both of the simplest forms consideredthe metal or the dioxide—have significant disadvantages. [9] Therefore, production and characterization of new thorium-based materials for nuclear fuels is of great importance. Could a metastable, solid ThO be such a material? The diatomic molecule, ThO, is well known. It can be formed during the vaporization of ThO2. [10] Recently, ThO has been produced through laser ablation of Th metal in the presence of oxygen, [11] [12] [13] and has been characterized in both the gas phase and in a cryogenic matrix. [11] [14] [15] Nevertheless, the clear demonstration of solid ThO as a dominat phase at ambient conditions has not been observed though it has been long-sought. [16] [17] Evidence for the production under mild heating of relatively stable, solid-phase ThO is here demonstrated. It is further shown that ThO can constitute the majority of the thorium oxide phases present. The detection method is neutron reflectometry (NR). NR is known as a noncontact, nondestructive, Å-resolution analytical technique for characterizing chemical speciation of thin films, including nuclear materials. [18] This paper also describes how the NR method was extended as a time-resolved (i.e., dynamic), in situ tool to identify the presence, stoichiometry and growth rate of subsurface layers under controlled oxidation conditions. Specifically, NR was employed to understand the oxidation of thorium in ~100ppm of oxygen at 150°C. Both the growth rate of the ThO2 surface layer and of a substoichiometric ThOw (w≤1) layer formed between the ThO2 top layer and the unreacted thorium metal beneath it were measured. To aid in interpretation of the experimental observations screened hybrid-functional calculations were performed on various hypothetical thorium-oxygen structures. This work, also reported herein, provides evidence that a stable ThO layer can formed for kinetic, rather than, thermodynamic reasons. Experimental and Computational Methodologies Thorium Thin-Film Deposition A nominal 1000 Å thick thorium film was deposited on a ~7.62-cm diameter, 1-cm thick crystalline quartz substrate for the neutron reflectometry experiments by DC-magnetron sputtering. Special care was taken to achieve low impurity content and small thickness variation in the sputtered film. The cryopumped, high-vacuum deposition chamber achieved a base pressure <2 x10-4 Pa as determined by an ion gauge prior to the deposition. The sputter working gas was high purity argon which pressure could be adjusted by its flow rate. The system pressure during sputtering was a compromise to achieve stable plasma (this requires higher pressure) and lowering the pressure to increase the mean free path of the sputtered atoms. Both factors promote good film adhesion to the substrate and high film density. The pressure was adjusted to be ~0.35 Pa as determined using a “0.1 torr” Baratron (capacitance manometer). In the DCmagnetron sputtering system, a Meivac MAK sputter gun with a 10 cm diameter thorium target (raw material from Nuclear Fuel Services) was used in the “sputter-up” configuration. The target-to-substrate distance was adjustable from about -2 cm (that is, the target could be adjusted to be above the level of the substrate) to 15 cm. Prior to film deposition the thorium target was sputtered in the -2 cm position to prevent sputtered atoms from reacting the surface of the substrate. This presputtering for ~5-7 min removes the native oxide layer and ensures that the thorium deposited on the substrate is as oxygen-free as possible. Prior studies had shown that no atoms are deposited during this cleaning operation. Since the native thorium oxide is an insulator its removal from the surface of the target can be therefore detected as the gradual decrease and leveling off of the magnitude of the target voltage. Thickness uniformity and surface smoothness are essential for the NR measurements. Three factors contributed to thickness uniformity. First, a planetary system was employed. That is, the substrate was spun around its central axis as the sample tray was slowly rotated over the target at 15°/s for 20 revolutions. In addition, a large substrate-target distance (10 cm) was chosen. Lastly, a large target was employed. For the same target-substrate distance, sputtering from a large target is inherently capable of producing better uniformity. This is because the atoms emerge from a distributed source. Due to the distribution of the magnetic field behind the target in the MAK sputter gun most of the sputtered atoms emerge from a circular race track about 6 cm in diameter on the surface of the thorium target. The magnets confine electrons in the plasma to a region near the surface of the target creating high density plasma at low pressures. Previous studies had shown that this sputtering configuration produces thickness variations of less than 5% over a diameter of 5 cm on the substrate. Oxidation Experiments and In-Situ Neutron Reflectometry Measurements Time-resolved in-situ NR measurements were performed on the thorium thin film as it was slowly oxidized in a sealed, ceramiccoated, custom-built chamber with controlled oxygen content. The chamber was equipped with low neutron absorbing quartz windows, a two-channel oxygen sensor (Servomex, Inc.), and a gas handling system. Controlled oxidizing environments were created by mixing high-purity Ar and O2 gases at the desired concentrations and heating the substrate (up to 150±3oC). The concentration of O2 was measured both upstream and downstream from the reaction chamber. Once the sample was inside the reaction chamber it was purged using 100% Ar gas. After the O2 concentration achieved a stable value below 5 ppm, indicating that the reaction chamber was leak-tight, the sample was heated to 150oC. At this point a NR spectrum (Run “0”) was collected. Afterwards, O2 was added to the system until a concentration of ~100 ppm O2 in Ar was reached and the subsequent NR spectra were collected. The time-of-flight (TOF) Surface Profile Analysis Reflectometer (SPEAR) instrument at the Los Alamos National Laboratory Lujan Neutron Scattering Center was used to obtain NR data. [19] The neutron beam is produced from a spallation source and, after moderation by liquid H2, is directed onto the sample at a very low angle while the specular reflection is recorded by a TOF position-sensitive detector. During a NR experiment, a collimated beam of neutrons is directed on a planar substrate at a small angle, θ (usually 0.5 – 4 deg.), and the ratio of the numbers of the elastically scattered to the incident neutrons is measured in the specular geometry. This ratio is defined as the reflectivity or reflectance, R(Q), and is reported as a function of the magnitude of neutron momentum transfer vector, Q,      sin 4   Q Q  (1) where λ is the neutron wavelength. The momentum transfer vector Q  is defined as the difference between outgoing and incoming neutron’s wavevectors. For the specular geometry, it is perpendicular to the sample surface. The wavelength of a neutron is determined from its velocity v by measuring the time it takes to travel the length of the instrument and by using de Broglie expression,   h mnv , where mn is the neutron mass. In the present TOF NR measurements, the neutron wavelength range varied from 4.5 to 16 Å. For the data presented in this manuscript NR for the entire Q-range was covered by measurements performed at 4 different angles of incidence (i.e., ~0.5, 1.0, 2.0 and 4.0 degrees), and the reflectivity curves were combined together. In the NR experiments only specular scattering was analyzed, although the “off-specular” signal was also recorded. The “off-specular” data provides the neutron intensity distribution as a function of the component of the neutron momentum transfer vector parallel to the sample’s surface. This information can provide additional insight to extend the interpretation of the specular reflectivity measurements regarding in-plane correlations of the samples studied. [20] It is important to note that, since NR data are normalized to the incident neutron intensity, the measured SLD values are absolute and do not need to be arbitrarily scaled. Differences in scattering length densities (SLDs) and the thickness of the various layers in the sample result in neutrons being reflected from the interfaces. The SLD value, the product of the number density of atoms and their nuclear coherent scattering lengths, is unique for a particular chemical composition and specific structure of the measured film. Thus it can be considered a fingerprint of the material. SLD can be calculated according the Equation 2: