High-pressure band structure and superconductivity of bcc and fcc lithium

Superconductivity in lithium under high pressure investigated with density functional and Eliashberg theory

“…33 but deviates significantly from 0.19 reported in Ref. 32. A possible reason for slightly larger in the present study is that the LDA exchange correlation tends to underestimate the equilibrium volume and hence overestimate the overall phonon frequencies.…”

“…The calculated T c increases monotonically for pressures from 25 to 32 GPa, but its slope as a function of pressure becomes much smaller from 32 to 33 GPa. In view of the substantial phonon softening in fcc Li at pressure close to 33 GPa, it is not unreasonable to attribute the difference between the predicted and observed T c for pressure [30][31][32][33] GPa to the neglect of anharmonic effects. It has been shown theoretically 55 and experimentally 56 that in the "hightemperature" superconductor MgB 2 , anharmonicity plays a significant role in reducing T c .…”

“…This feature is in contrast to fcc Li, in which the electronic bands are flat and almost parallel to the Fermi level near the L point. 32,35,36 Thus the spherical FS in fcc Li is distorted anisotropically and has parallel necks at the boundary of the first BZ. The Fermisurface nesting between these necks has been shown to be the origin of strong EPC in fcc Li.…”

“…15 There have been extensive theoretical studies of superconductivity in the fcc phase of Li under pressure. 26,[28][29][30][31][32][33][34][35][36][37] In particular, it has been suggested that superconductivity in fcc Li is driven by Fermi-surface nesting 30,31 and Kohn anomalies, 30 resulting in an unusually high T c . On the other hand, no theoretical investigation has been made on superconductivity of the high-pressure cI16 phase based on firstprinciples calculation of the phonon spectra and the electronphonon spectral function.…”

“…Theoretical works on pressure-induced superconductivity so far, however, have focused on the fcc phase only. [26][27][28][29][30][31][32][33][34][35][36][37] The only exception is the study on fcc and cI16 by Christensen and Novikov 26 based on a rigid muffintin approximation-which proved to be inadequate for Li ͑Ref. 34͒-and without calculation of the phonon spectra for cI16.…”