Electronic Instability in a Zero-Gap Semiconductor: The Charge-Density Wave in $({\mathrm{TaSe}}_4{)}_2\mathbf{I}$

We report a comprehensive study of the paradigmatic quasi-1D compound $({\mathrm{TaSe}}_4{)}_2\mathrm{I}$ performed by means of angle-resolved photoemission spectroscopy (ARPES) and first-principles electronic structure calculations. We find it to be a zero-gap semiconductor in the nondistorted structure, with non-negligible interchain coupling. Theory and experiment support a Peierls-like scenario for the charge-density wave formation below $T_{\mathrm{CDW}}=263\mathrm{K}$, where the incommensurability is a direct consequence of the finite interchain coupling. The formation of small polarons, strongly suggested by the ARPES data, explains the puzzling semiconductor-to-semiconductor transition observed in transport at $T_{\mathrm{CDW}}$.

Figure 1

(a) Crystal structure of nondistorted $({\mathrm{TaSe}}_4{)}_2\mathrm{I}$ showing the natural cleavage plane. The conventional tetragonal unit cell is outlined. (b) A projection of the structure onto the cleavage plane. $t_{\parallel }$ and $t_{\perp }$ are the intrachain and interchain nearest-neighbor hopping integrals, respectively, in the tight-binding model discussed in the text. ARPES intensity maps illustrate the band dispersion (c) parallel (along $XP$) and (d) perpendicular to the chains, for $k_{\parallel }=0.75\pi /c$. (e) 3D BZ. (f),(g) Constant energy maps extracted at the top of the Ta $5d_{z^2}$ band ($E_{\mathrm{B}}=0.3\mathrm{eV}$, $T=100\mathrm{K}$) in two perpendicular planes, as depicted in the 3D BZ. Dashed lines in (d) and (g) are guides to the eye. Photon energy in (f) is set to $h\nu =88\mathrm{eV}$, that corresponds to the dashed line in (g). The second derivative is used in (g).

Figure 2

Schematic constant energy surfaces from (a) a 2D and (b) a 3D TB model. (c) CE map measured at $E_{\mathrm{B}}=0.6\mathrm{eV}$ with photon energy $h\nu =88\mathrm{eV}$. On the momentum distribution curve [$k_{\perp }=\pi /(a/\sqrt{2})$], stars and arrows indicate the main band and replicas corresponding to the periodicity of the undistorted crystal, respectively. (d) Theoretical contours derived from a 3D TB model. Projected 3D (solid line) and 2D (dashed) BZs are also indicated.

Figure 3

(a) Electronic band structure of nondistorted $({\mathrm{TaSe}}_4{)}_2\mathrm{I}$ calculated from first principles. The inset shows the Fermi surface. The blue arrow depicts the CDW vector. (b) Distorted vs nondistorted band structures along $\Gamma Z$. The horizontal lines indicate the Fermi level in the intrinsic (dotted line) and $n$-doped (plain line) cases. (c) EDCs along the $\Gamma Z$ direction ($h\nu =110\mathrm{eV}$, $T<T_{\mathrm{CDW}}$), sampled around $\pi /c$ (dotted line) with momentum step $0.01{\AA }^{-1}$. Natural iodine vacancies yield an effective $n$ doping that allows us to probe CB. Small polaron formation suppresses the QP weight at $E_F$ and transfers the spectral weight at a higher $E_p$ binding energy.