Lithium oxide: a quantum-corrected and classical Monte Carlo study.

Extensive Monte Carlo simulations of lithium oxide, Li2O, an important material for fusion applications over a wide range of temperatures have been performed. In the low temperature range 1-500 K, quantum path-integral corrections to the enthalpy and unit cell size were determined. We show that classical Monte Carlo underestimates both these quantities and the difference between unit cell parameters with and without quantum corrections is large enough that such corrections should be included in any comparison between theory and experiment. Over the range 300-1000 K, the formation energies of Schottky and Frenkel defects are calculated and compared with those from direct free energy minimisation in the quasiharmonic approximation, which also includes quantum corrections; the Monte Carlo results highlight the onset of failure of the quasiharmonic approximation even at modest temperatures and suggest only a small variation of the defect enthalpies with temperature. Several possible diffusion mechanisms are identified. While an interstitialcy mechanism activates at around 900-1000 K, lithium vacancy migration dominates from 500 K. The estimated migration energy of the Li-vacancy jump (0.28 eV) agrees very well with the most recent NMR study. At temperatures above 1000 K, the superionic phase transition and subsequent melting are simulated and there is good agreement with available experimental data. Our simulations predict a rapid rise in the heat capacity and the thermal expansion coefficient which continues up to the melting point which leaves two interesting questions for future experimental studies: (i) whether above the superionic transition the heat capacity and the thermal expansion coefficient in antifluorite Li2O rise up to the melting point, as in our simulations, or fall, as observed in several fluorites, and (ii) the subsequent change in the heat capacity during melting.

Magnetic and thermodynamic properties of face-centered cubic Fe-Ni alloys.

A model lattice ab initio parameterized Heisenberg-Landau magnetic cluster expansion Hamiltonian spanning a broad range of alloy compositions and a large variety of chemical and magnetic configurations has been developed for face-centered cubic Fe-Ni alloys. The thermodynamic and magnetic properties of the alloys are explored using configuration and magnetic Monte Carlo simulations over a temperature range extending well over 1000 K. The predicted face-centered cubic-body-centered cubic coexistence curve, the phase stability of ordered Fe3Ni, FeNi, and FeNi3 intermetallic compounds, and the predicted temperatures of magnetic transitions simulated as functions of alloy composition agree well with experimental observations. Simulations show that magnetic interactions stabilize the face-centered cubic phase of Fe-Ni alloys. Both the model Hamiltonian simulations and ab initio data exhibit a particularly large number of magnetic configurations in a relatively narrow range of alloy compositions corresponding to the occurrence of the Invar effect.

Magnetic cluster expansion simulation and experimental study of high temperature magnetic properties of Fe-Cr alloys.

We present a combined experimental and computational study of high temperature magnetic properties of Fe-Cr alloys with chromium content up to about 20 at.%. The magnetic cluster expansion method is applied to model the magnetic properties of random Fe-Cr alloys, and in particular the Curie transition temperature, as a function of alloy composition. We find that at low (3-6 at.%) Cr content the Curie temperature increases with the increase of Cr concentration. It is maximum at approximately 6 at.% Cr and then decreases for higher Cr content. The same feature is found in thermo-magnetic measurements performed on model Fe-Cr alloys, where a 5 at.% Cr alloy has a higher Curie temperature than pure Fe. The Curie temperatures of 10 and 15 at.% Cr alloys are found to be lower than the Curie temperature of pure Fe.

Surface diffusion and surface growth in nanofilms of mixed rocksalt oxides.

We have modelled the surface diffusion and growth of BaO and SrO both in the homoepitaxial and heteroepitaxial (BaO on SrO and SrO on BaO) cases. The diffusion proceeds most favourably by an exchange mechanism involving the surface layer. When impurities are adsorbed on the surface this can lead to intermixing between the layers. This strongly suggests that ionic materials may not be grown on a substrate with a similar structure without significant intermixing. Island growth begins with the formation of individual clusters which grow and merge together.

Dopant incorporation into garnet solid solutions--a breakdown of Goldschmidt's first rule.

Atomistic simulations suggest trace elements are more soluble in a 50:50 pyrope (Mg3Al2Si3O12)-grossular (Ca3Al2Si3O12) garnet mixture than in either end-member; consistent with partitioning experiments, and, contrary to Goldschmidt's first rule, large trace element cations may substitute for Mg2+, small trace elements for Ca2+.