Transport anomalies across the quantum limit in semimetallic ${\text{Bi}}_{0.96}{\text{Sb}}_{0.04}$

We report on a study of electronic transport in semimetallic ${\text{Bi}}_{0.96}{\text{Sb}}_{0.04}$. At zero field, the system is a very dilute Fermi liquid displaying a $T^2$ resistivity with an enhanced prefactor. Quantum oscillations in resistivity as well as in Hall, Nernst, and Seebeck responses of the system are detectable and their period quantifies the shrinking of the Fermi surface with antimony doping. For a field along the trigonal axis, the quantum limit was found to occur at a field as low as 3 T. An ultraquantum anomaly at twice this field was detected in both charge transport and Nernst response. Its origin appears to lie beyond the one-particle picture and linked to unidentified many-body effects.

Figure 1

(a) Zero-field resistivity ${\text{Bi}}_{0.96}{\text{Sb}}_{0.04}$ as a function of $T^2$ for a current along the binary axis. The inset shows the data over a wider temperature range. (b) Temperature-dependence of the Seebeck coefficient at low temperatures for a temperature gradient along the binary axis.

Figure 2

(Color online) (a) Hall and (b) longitudinal resistivity as a function of magnetic field at two different temperatures. QL marks the quantum limit anomaly. Beyond this limit (i.e., in the ultraquantum regime), another anomaly is visible and marked as UQ. (c) presents a semilogarithmic plot of resistivity as a function of $B^{-1}$.

Figure 3

Upper Panel: Nernst signal as a function of inverse of the magnetic field. Numbers identify peaks associated with various Landau levels. Note the gradual emergence of new peaks due to Zeeman splitting at lower temperatures (see text). The lower panels present the position of identified landau levels for two different field orientations.

Figure 4

(a) Nernst data obtained for ${\text{Bi}}_{0.96}{\text{Sb}}_{0.04}$ in a 28 T resistive magnet. (b) Comparison of Nernst response in ${\text{Bi}}_{0.96}{\text{Sb}}_{0.04}$ and in Bi. The 0.45 K data obtained here and the 0.85 K data for Bi (Ref. 16) are plotted as a function of $B^{-1}$ normalized to the period of oscillations in each case.