No abstract

In situ x-ray diffraction and Raman spectroscopy have been carried out to pressures of 93.6 and 63.2 GPa, respectively, to explore the pressure-induced phase transformation of CoFe 2 O 4 spinel. CoFe 2 O 4 adopts a distorted tetragonal spinel structure at one atmosphere. At a pressure of ϳ32.5 GPa, both x-ray diffraction and Raman spectroscopy indicate that CoFe 2 O 4 transforms to the orthorhombic CaFe 2 O 4 structure, which remains stable to at least 93.6 GPa. The bulk modulus (K 0 ) of the tetragonal and the high-pressure polymorphs were calculated to be 94͑12͒ and 145͑16͒ GPa, respectively, with KЈϵ4. Upon release of pressure the orthorhombic phase persists and appears to be structurally metastable. At zero pressure, laser induced heating leads to a significant transformation back to the tetragonal phase. The high-pressure orthorhombic phase at one atmosphere is 14.7% denser than the tetragonal phase.

In this work, we present a simple approach for devising order parameters (OPs) for atomic systems based on pattern recognition techniques. It exploits the fact that all crystalline substances are characterized by a unique “signature” cell (SC) which is constructed using a central atom and its nearest NSC neighbors in a given crystal. The algorithm measures the local degree of similarity between a SC and the system to be analyzed. The best fit of a SC to NSC atoms surrounding a given atom in the system is determined by maximizing a fictitious energy of binding among those atoms and the SC atoms. The fictitious potential energy is designed to give maximum attractive energy for maximum overlap. The maximum binding energy of interaction attained in this process is used as a measure of similarity between the crystal structure and the system (i.e., as an OP). The proposed method provides a unified and intuitive approach for constructing relevant OPs for a given system. We used these OPs to characterize the order of different phases in the Lennard-Jones system and in a model silicon system. It is shown that these novel OPs give a more complete description and a better understanding of the structural order in amorphous silicon than conventional OPs.

Virtual Gibbs ensembles (VGEs), a class of pseudo-ensemble Monte Carlo methods, are specialized in this work to simulate phase equilibrium for systems for which conventional direct methods (such as two-box Gibbs ensembles and one-box interfacial ensembles) are inadequate. It is shown that by removing the mass/volume balance constraints of conventional Gibbs ensembles, the resulting VGEs can be used to effectively simulate systems wherein (a) the number of molecules or the composition of one of the phases is to be kept constant and (b) the initial choices for the volume and number of molecules would preclude direct methods to convergence to a stable two-phase state. Applications of VGEs are presented for the simulation of vapor–liquid, vapor–solid, and liquid–solid equilibrium in single component systems and multicomponent sorption equilibria of gases in polymer melts. VGE simulations of solid–fluid coexistence entail the combination of elements of both interfacial ensembles and Gibbs ensembles.