Pressure-induced Insulator-Metal Transitions in Solid Xenon and Hydrogen: A First-Principles Quasiparticle Study.

H CHACHAM,XJ ZHU,SG LOUIE
DOI: https://doi.org/10.1103/physrevb.46.6688
1992-01-01
Abstract:We report a quasiparticle study of the pressure-induced isostructural insulator-metal transition in solids by the mechanism of band-gap closure. Two examples are investigated: solid Xe and molecular solid hydrogen. The band gaps are calculated with a first-principles quasiparticle approach, in which the electron self-energy operator is expanded to first order in the screened Coulomb interaction, viz., the GW approximation. For the case of solid xenon, the crystal structure has been experimentally established to be hexagonal-close-packed (hcp) in the vicinity of the observed metallization pressure of 132(+/-5) GPa. Our calculation for solid xenon yields a metallization pressure of 128 GPa, in good agreement with experiment. The theoretical results further quantitatively explain all the salient features observed in the experimental optical spectra at metallization. For molecular solid hydrogen the structure has yet to be determined definitively. Our calculations are carried out for structures in which hydrogen molecules assume an hcp arrangement. The quasiparticle results are compared with those from Hartree-Fock (HF) and local-density approximation (LDA) calculations. In addition to the well-known HF overestimate and the LDA underestimate of the band gap, we find that both HF and LDA predict a linear behavior in the band gap versus density, whereas the quasiparticle results do not show such a linearity. This difference results from a significant increase in the dielectric screening with density, which gives rise to a strong and nontrivial dependence of the self-energy correction to the LDA band gap on density. We have also studied the effects of orientational disorder of the H-2 Molecules within a virtual-crystal model. We find that at a given density, the minimum band gap increases monotonically and nonlinearly with orientational disorder. A simple tight-binding picture provides a convenient way to understand the variation of the hydrogen band gap both with pressure and with disorder. Our calculations predict a metallization pressure of 151 GPa for the hcp phase if the molecules are perfectly aligned along the c axis, and 300 GPa if there is no orientational order. At present, a definitive conclusion is difficult to draw as regards the metallization of molecular solid hydrogen. More details of the crystal structure, especially those concerning the orientational ordering of molecular H-2 at megabar pressures, are needed.
What problem does this paper attempt to address?