Spin-orbit driven Peierls transition and possible exotic superconductivity in ${\mathrm{CsW}}_2{\mathrm{O}}_6$

We study ab initio a pyrochlore compound, ${\mathrm{CsW}}_2{\mathrm{O}}_6$, which exhibits a yet unexplained metal-insulator transition. We find that (1) the reported low-$T$ structure is likely inaccurate and the correct structure has a twice larger cell; (2) the insulating phase is not of a Mott or dimer-singlet nature, but a rare example of a three-dimensional Peierls transition, with a simultaneous condensation of three density waves; (3) the spin-orbit interaction plays a crucial role, forming well-nested bands. The high-$T$ (HT) phase, if stabilized, could harbor a unique $e_g+i{e}_g$ superconducting state that breaks time reversal symmetry, but is not chiral. This state was predicted in 1999, but not observed. We speculate about possible ways to stabilize the HT phase while keeping the conditions for superconductivity.

Figure 1

The band structure obtained for the HT phase in the GGA and GGA+SOC calculations. Positions of the high-symmetry points in the Brillouin zone are shown in Fig. 2.

Figure 2

The Fermi surface as obtained in the GGA+SOC calculation for the high-temperature phase.

Figure 3

(a) Real and (b) imaginary parts of the noninteracting susceptibility ${\chi }_0\left(\mathbf{q}\right)$ calculated from the high-$T$ band structure. Isosurfaces with ${\chi }_0\left(\mathbf{q}\right)=92\%{\chi }_0^{\text{max}}$ are shown.