Effect of carrier doping on the uniform and FFLO excitonic phases in the three-chain Hubbard model for Ta2NiSe5

We investigate the effect of carrier doping on the excitonic phases in the three-chain Hubbard model for Ta2NiSe5 which is known as a candidate material for the excitonic insulator (EI) [1,2]. The model consists of a Ni chain responsible for a valence (f) band and adjacent two Ta chains responsible for two-fold degenerate conduction (c) bands and has been developed for describing the phase transition from the semiconductor to the excitonic insulator accompanied by the orthorhombic to monoclinic structural transition of Ta2NiSe5 at ambient pressure [3]. One of the authors and coworkers have investigated the same model in the semimetallic case corresponding to Ta2NiSe5 under pressure [4] and have found that the Fulde-Ferrell-Larkin-Ovchinnikov (FFLO) excitonic state characterized by the condensation of excitons with finite center-of-mass momentum is realized in the case with a large band-overlapping (high pressure) where two c bands and one f band cross the Fermi level in contrast to the case with a small band-overlapping (low pressure) where one c band and one f band cross the Fermi level and the uniform excitonic state corresponding to the EI is realized as the same as the semiconducting case [5,6,7]. Near the phase transition to the FFLO excitonic state, the enhanced excitonic fluctuations are found to mediate the c-f interband Cooper pairs with finite center-of-mass momentum resulting in the FFLO superconductivity [8], which is expected to be realized in the semimetallic Ta2NiSe5 under high pressure [9]. In the present paper, we examine the uniform and FFLO excitonic phases in the presence of carrier doping as a straight forward extension to the previous works where the non-doped case is exclusively investigated [3,5-8]. We obtain the excitonic phase diagrams as functions of the number of electrons per unit cell n and the temperature T, where we set the intersite c-f Coulomb interaction V=0.4eV and the energy gap between the conduction and valence band D=0 which corresponds to the boundary between the semiconducting and the semimetallic cases. We see an apparent asymmetry between the hole doping (n<2) and the electron doping (n>2) due to the difference in degeneracy between the valence and conduction bands.

[1] Y. Wakisaka et al., Phys. Rev. Lett. 103, 026402 (2009).
[2] T. Mizokawa et al., J. Supercond. Nov. Magn. 22, 67 (2009).
[3] T. Kaneko, T. Toriyama, T. Konishi and Y. Ohta, Phys. Rev. B 87, 035121 (2013).
[4] A. Nakano et al., IUCrJ 5, (2018).
[5] T. Yamada, K. Domon and Y. Ōno, J. Phys. Soc. Jpn 85, 053703 (2016).
[6] K. Domon, T. Yamada and Y. Ōno, J. Phys. Soc. Jpn 85, 065005 (2016).
[7] K. Domon, T. Yamada and Y. Ōno, J. Phys. Soc. Jpn 87, 054701 (2018).
[8] T. Yamada, K. Domon and Y. Ōno, J. Phys. Soc. Jpn 88, 064701 (2019).
[9] K. Matsubayashi et al., J. Phys. Soc. Jpn. 90, 074706 (2021).