Insulator-metal transition of highly compressed carbon disulfide

Abstract

We present integrated spectral, structural, resistance, and theoretical evidences for simple molecular CS 2 transformations to an insulating black polymer with three-fold carbon atoms at 9 GPa, then to a semiconducting polymer above 30 GPa, and finally to a metallic solid above 50 GPa. The metallic phase is highly disordered 3D network structure with four-fold carbon atoms at the carbon-sulfur distance of ~1.70Å. Based on first principle calculations, we present two plausible structures for the metallic phase: α-chalcopyrite and tridymite, both of which exhibit metallic ground states and disordered diffraction features similar to that measured. We also present the phase/chemical transformation diagram for carbon disulfide, showing a large stability field of the metallic phase to 100 GPa and 800 K.

High-pressure layered structure of carbon disulfide

“…Above that pressure the HP1 layered structure clearly prevails, its en- thalpy remaining lower than that of β-cristobalite by a substantial amount, exceeding the estimated computational errors of about 10 meV/molecular unit, up until 160 GPa-a regime where, however, carbon coordinations larger than four will take over. 7 The same HP1 structure which we found here appears in SiS 2 at much lower pressures than in CS 2 as is natural given the shorter bond lengths and smaller compressibility of the carbon compound.…”

“…The tridymite structure shows slightly better agreement with experiment but represents a thermodynamically less likely candidate than β-cristobalite since its calculated DFT enthalpy is 0.3-0.4 eV/molecule higher. 7 The electronic structure of either fourfold coordinated crystal structure agrees with the observed metalization in the region from 40 GPa upwards. 7,9 Besides that, in more recent experiments performed on the high pressure metallic phases of CS 2 , resistivity showed evidence of superconductivity at 4-6 K over a broad pressure range from 50 to 172 GPa.…”

“…5, at P=50 GPa all mode frequencies are real and positive, confirming the mechanical stability of the structure. Comparison with calculated phonons of the β-cristobalite structure 7 shows that modes of the HP1 layered structure are slightly stiffer, although there is a fair amount of overall similarity. Phonon calculations for HP2, prohibitively expensive because of the large 36-atom unit cell, were not attempted, also given the uncertain stability of this phase.…”

“…In spite of that intrinsic thermodynamical metastability, solid CS 2 phases do exist, clearly for kinetic reasons, and are reported at ambient pressure not to decompose in measurable times, at least below ≈ 560 K, 7,8 a temperature rising even further at high pressures, possibly up to 1000 K at 70 GPa. 7 At high pressures but low temperatures, X-ray data have shown evidence of structural transitions of CS 2 from the molecular phase to polymeric phases with C-S coordinations rising from two (Cmca) to three (CS 3 ) to four (CS 4 ), 7 and further. With the exception of the CS 3 phase near 20 GPa, the high pressure structural behavior has been postulated so far to be similar to that of CO 2 , implying (not unreasonably) that CS 4 could be β-cristobalite.…”

“…In CS 2 a detailed interpretation of high pressure experimental data is further complicated by a large amount of structural disorder, particularly in the CS 4 phase where only broad rather than sharp Bragg peaks are present in the diffraction pattern. 7 With that rationalization of the fourfold coordinated state of CS 2 near 50 GPa through simple analogy with CO 2 , density functional theory (DFT) studies have considered β-cristobalite (I42d, also referred to as Chalcopyrite) and tridymite (P 2 1 2 1 2 1 ), for comparison with experimental data. Lacking sharp diffraction peaks or other distinctive features, that comparison nevertheless appears somewhat elusive.…”