Anomalies in non-stoichiometric uranium dioxide induced by pseudo-phase transition of point defects

A uniform distribution of point defects in an otherwise perfect crystallographic structure usually describes a unique pseudo phase of that state of a non-stoichiometric material. With off-stoichiometric uranium dioxide as a prototype, we show that analogous to a conventional phase transition, these pseudo phases also will transform from one state into another via changing the predominant defect species when external conditions of pressure, temperature, or chemical composition are varied. This exotic transition is numerically observed along shock Hugoniots and isothermal compression curves in UO2 with first-principles calculations. At low temperatures, it leads to anomalies (or quasi-discontinuities) in thermodynamic properties and electronic structures. In particular, the anomaly is pronounced in both shock temperature and the specific heat at constant pressure. With increasing of the temperature, however, it transforms gradually to a smooth cross-over, and becomes less discernible. The underlying physical mechanism and characteristics of this type of transition are encoded in the Gibbs free energy, and are elucidated clearly by analyzing the correlation with the variation of defect populations as a function of pressure and temperature. The opportunities and challenges for a possible experimental observation of this phase change are also discussed.

💡 Research Summary

The paper investigates a novel type of transition in non‑stoichiometric uranium dioxide (UO₂₊ₓ) that the authors term a “pseudo‑phase transition.” In a perfect crystal a uniform distribution of point defects defines a single pseudo‑phase; however, when external variables such as pressure, temperature, or composition are varied, the dominant defect species can switch abruptly, producing effects that closely resemble conventional phase transitions. Using first‑principles density‑functional theory (DFT) with LSDA + U to treat the strongly correlated 5f electrons of uranium, the authors calculate formation energies of the most relevant point defects—oxygen vacancies (O_v) and uranium interstitials (U_i)—over a wide pressure range while keeping the fluorite lattice unchanged. A quasi‑annealing scheme is employed to avoid metastable electronic states.

From the defect formation energies they construct Boltzmann‑weighted defect concentrations n_i = exp(−ΔG_i/k_BT) and embed these into the Gibbs free energy of the non‑stoichiometric system: G(T,P,x) = G₀(T,P) + Σ ΔG_i n_i, where G₀ is the free energy of the perfect lattice and the second term accounts for defect contributions. By varying T, P, and the deviation x, they track how the relative magnitude of ΔG_i changes and consequently which defect dominates.

The central finding is that at low temperatures and for hypostoichiometric compositions (x < 0), a sharp switch occurs around 39 GPa: below this pressure oxygen vacancies dominate, while above it uranium interstitials become the prevailing defect. This switch manifests as a quasi‑discontinuous volume collapse on the pressure–volume Hugoniot, a pronounced kink in the shock temperature, and a marked anomaly in the specific heat at constant pressure (C_P). The bulk sound velocity and thermal expansivity also show subtle kinks. Importantly, the magnitude of these anomalies scales with the stoichiometry deviation; larger |x| produces a more pronounced effect.

Temperature plays a crucial role in smoothing the transition. At higher temperatures the Boltzmann factor reduces the sensitivity of defect populations to ΔG_i differences, broadening the transition zone and turning the quasi‑discontinuities into smooth cross‑overs. Consequently, the anomalies in thermodynamic quantities diminish and eventually disappear, consistent with the analytic nature of the Gibbs free energy (all derivatives remain finite). The authors illustrate this behavior with isothermal compression curves from 300 K to 3000 K, showing the progressive widening of the defect‑population crossover.

A “pseudo‑phase diagram” on the T–P plane is constructed, delineating regions where O_v or U_i dominate for two representative compositions (x = −0.02 and x = −0.1). The boundary shifts only weakly with temperature, confirming that compression is the primary driving force. The diagram also predicts that at absolute zero a true first‑order‑like jump could occur, but even then only the first derivative of G would be discontinuous; higher‑order derivatives remain finite because defect concentrations are fixed at the transition point.

The paper discusses experimental prospects. Metastable hypostoichiometric UO₂₊ₓ with x = −0.02 has been synthesized at room temperature, and there are indications that x = −0.05 could be produced at elevated temperatures. Detecting the pseudo‑phase transition would require high‑precision measurements of C_P (where the kink is most pronounced) or shock‑wave experiments capable of resolving the temperature kink. High‑pressure diamond‑anvil cell studies could also capture the volume collapse.

Overall, the work introduces the concept of defect‑driven pseudo‑phase transitions, showing that abrupt changes in dominant point‑defect species can generate thermodynamic anomalies analogous to conventional phase transitions without any change in the underlying crystal symmetry. This bridges defect physics and high‑pressure thermodynamics, offering a new route to tailor material properties—particularly in nuclear fuel oxides—by engineering defect populations.