Pseudogap-Driven Sign Reversal of the Hall Effect

We present a calculation of the Hall coefficient in $2H\mathrm{\text{−}}{\mathrm{TaSe}}_2$ and $2H\mathrm{\text{−}}{\mathrm{Cu}}_{0.2}{\mathrm{NbS}}_2$ based on their electronic structure extracted from angle-resolved photoemission spectra. The well-known semiclassical approach, based on the solution of the Boltzmann equation, yields the correct value for the normal-state Hall coefficient. Entering the charge density wave state results in the opening of the pseudogap and redistribution of the spectral weight. Accounting for this allows us to reproduce the temperature dependence of the Hall coefficient, including the prominent sign change, with no adjustable parameters.

Figure 1

Evolution of the Fermi surface of ${\mathrm{TaSe}}_2$ and ${\mathrm{Cu}}_{0.2}{\mathrm{NbS}}_2$ with temperature. The Fermi surface of ${\mathrm{TaSe}}_2$ changes topology with cooling (a, b). Absence of changes and uniform spectral weight distribution in the spectra of ${\mathrm{Cu}}_{0.2}{\mathrm{NbS}}_2$ (c),(d). Solid black lines in (a),(c),(d) are the tight-binding fit to the data. For the CDW-reconstructed Fermi surface (b) different types of guidelines represent the spectral weight distribution: the stronger line corresponds to the higher spectral weight. The relative contribution to ${\sigma }_{xy}$ from different parts of the Fermi surface at different temperatures is shown in the irreducible part of the Brillouin zone for ${\mathrm{TaSe}}_2$ (e) and ${\mathrm{Cu}}_{0.2}{\mathrm{NbS}}_2$ (f).

Figure 2

The temperature dependence of the Hall coefficient. The Hall coefficient in ${\mathrm{NbS}}_2$ has a weak temperature dependence, while in ${\mathrm{TaSe}}_2$ it changes sign (a). The Hall coefficient of ${\mathrm{Cu}}_{0.2}{\mathrm{NbS}}_2$, calculated in approximation of the equally “populated” bands, agrees well with the directly measured one [23], while in the case of ${\mathrm{TaSe}}_2$ one should take into account the spectral weight redistribution and the opening of the pseudogap (b). The discrepancy between two experimental curves for ${\mathrm{TaSe}}_2$ is due to the charge density wave suppression by impurities.