Transition metal dichalcogenides (TMDs), whether in bulk or in monolayer form, exhibit a rich variety of charge-density-wave (CDW) phases and stronger periodic lattice distortions. While the actual role of nesting has been under debate, it is well understood that the microscopic interaction responsible for the CDWs is the electron-phonon coupling. The case of $1T-{\mathrm{TiSe}}_2$ is, however, unique in this family in that the normal state above the critical temperature $T_{\mathrm{CDW}}$ is characterized by a small quasiparticle band gap as measured by angle-resolved photoemission spectroscopy, so that no nesting-derived enhancement of the susceptibility is present. It has therefore been argued that the mechanism responsible for this CDW should be different and that this material realizes the excitonic insulator phase proposed by W. Kohn. On the other hand, it has also been suggested that the whole phase diagram can be explained by a sufficiently strong electron-phonon coupling. In this paper, in order to estimate how close this material is to the pure excitonic insulator instability, we quantify the strength of electron-hole interactions by computing the exciton band structure at the level of hybrid density functional theory, focusing on the monolayer. We find that in a certain range of parameters the indirect gap at $q_{\mathrm{CDW}}$ is significantly reduced by excitonic effects. We also stress the important role of the spin-orbit coupling in significantly reducing the band gap. We discuss the consequences of those results regarding the debate on the physical mechanism responsible for this CDW. Based on the dependence of the calculated exciton binding energies as a function of the mixing parameter of hybrid density functional theory, we conjecture that a necessary condition for a pure excitonic insulator is that its noninteracting electronic structure is metallic.

• Received 30 May 2018
• Revised 30 October 2018

DOI:https://doi.org/10.1103/PhysRevB.98.235106