Sliding Charge Density Wave observed through Band Structure

Abstract

An incommensurate CDW may have the ability to slide, i.e., to generate an excess of current when the system is submitted to an external field. Sliding phenomenon is closely related to deformation of the periodic lattice distortion associated to the CDW. In principle, however, the sliding state can also be observed through the band structure. Here we show that broken symmetry in k-space is observed by Angle-Resolved Photoemission Spectroscopy (ARPES) in the sliding regime of TbTe$_3$, which could be consistent with theoretical predictions.

Figures (5)

• Figure 1: Scheme of displaced band structure of 1D CDW in the case of position and time-dependent phase. a) Gap opening at $\pm$ k$_F$ with constant phase with 2k$_F$=5/7. The gray curve correspond to the electronic dispersion of the metallic high temperature phase. b) Shift of the Fermi wave vector with $\delta k=\pm 1/2 \delta \phi(x)/\delta x$ in the case of contraction of the CDW wavelength ($\delta \phi(x)/\delta x>0$) and c) in the case of dilatation of the CDW wavelength ($\delta \phi(x)/\delta x<0$). d) Assymetric gap openning in the case of time-dependent phase with $\delta k=1/(2\hbar v_F)\delta \phi(t)/\delta t$.
• Figure 2: (a) Fermi surface of TbTe3 at RT and without current. The dashed lines are a guide to the eye of the Fermi surface sheets. The nesting vector $q_{CDW}$ is shown as well as red dashed line located at $k_x = -0.192$ Å$^{-1}$ where the band dispersion has been measured. (b) Band dispersion along the red dotted line in with the CDW gap. Zooms near EF are shown to better appreciate the gap. The CDW band gap shift versus applied current and a reference ungapped band have been analyzed in profiles with full and dashed lines respectively.
• Figure 3: Fermi surfaces of TbTe3 for different applied currents (integration window across the Fermi energy of 20 meV).
• Figure 4: (a) Detail at EF of the band dispersion at $k_x = -0.192$ Å$^{-1}$ for different applied currents ($-70$ mA (top row), $0$ mA (middle row), $+70$ mA(bottom row)). Since the intensity for positive and negative $k$ is different due to photoemission matrix elements, peak intensities have been normalized to the peak intensity at zero current. Red straight lines drawn at the band edges serve as guides to the eye, emphasising the horizontal (wavevector) shift of the bands. (b) Intensity contour plots for the left band (k$<$0, top row) and right band (k$>$0, bottom row). Positive current induces an upward shift (downward) on the left (right) band. Negative current induces an opposite behavior.
• Figure 5: Spectra evolution with applied current. (a, b) EDCs for 0 and $\pm 70$ mA applied currents. Spectra have been obtained from $k$ integration in the regions delimited by blue (left) and red (right) lines in Fig. \ref{['fig2']}(b). (c, d) Evolution as a function of the applied current of the closest peak to the Fermi level (around $\sim$-0.15 eV) for the left and right integrated line profiles, respectively. Error bars represent the standard deviation of the peak position. (e,f) Evolution as a function of the applied current of the semiconducting bands for the left and right integrated line dashed profiles in fig. 2. Error bars represent the standard deviation of the peak position.