Evidence for Half-Metallicity in $n$-type ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$

High quality ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$ single crystals have been investigated by magnetization, electron transport, and Andreev reflection spectroscopy. In the ferromagnetic ground state, the saturation magnetic moment of each unit cell corresponds to an integer number of electron spins ($3{\mu }_B/{\mathrm{Cr}}^{3+}$), and the Hall effect measurements suggest $n$-type charge carriers. Spin polarizations as high as 97% were obtained from fits of the differential conductance spectra of ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$/Pb junctions with the modified Blonder-Tinkham-Klapwijk theory. The temperature and bias-voltage dependencies of the subgap conductance are consistent with recent theoretical calculations based on spin active scatterings at a superconductor–half-metal interface. Our results suggest that $n\text{−}{\mathrm{HgCr}}_2{\mathrm{Se}}_4$ is a half-metal, in agreement with theoretical calculations that also predict undoped ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$ is a magnetic Weyl semimetal.

Figure 1

(a) Schematic band diagrams for ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$ in the ferromagnetic (left) and paramagnetic (right) states based on the first principles calculations reported in Refs. [4, 16]. (b) Magnetization curves at several temperatures from $T=2$ to 300 K. The inset shows an optical micrograph for a typical ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$ crystal. (c) Hall resistivity ${\rho }_{xy}$ curve taken at $T=1.6\mathrm{K}$. The small curvature near the zero field is due to the anomalous Hall effect (AHE). (d) Temperature dependence of longitudinal resistivity ${\rho }_{xx}$ at magnetic fields ${\mu }_{0}H=0$, 4, 8 T.

Figure 2

(a) Sketch of the ${\mathrm{HgCr}}_2{\mathrm{Se}}_4/\mathrm{Pb}$ planar junctions used for Andreev reflection spectroscopy. (b) Differential conductance $G(V)=dI/dV$ of the sample $J1$ plotted as a function of dc bias. (c) Normalized differential conductance $G(V)/G_N(V)$ spectra for sample $J2$. Here, the bias $V$ is defined as the dc voltage of the Pb electrode relative to the $\mathrm{Cr}/\mathrm{Au}$ electrode at the bottom of the ${\mathrm{HgCr}}_2{\mathrm{Se}}_4$ crystal, and $G_N(V)$ is normal state differential conductance recorded with a small field (${\mu }_{0}H=0.2\mathrm{T}$) applied.

Figure 3

Normalized conductance spectra (symbols) before symmetrization [(a),(b), for samples $J1$ and $J2$] and after symmetrization [(c)–(f), for samples $J1–J4$]. Their fits to the modified BTK theory are shown by lines. The measurement temperature was 1.6 K for all panels.

Figure 4

(a) Normalized conductances at small bias voltages of ${\mathrm{HgCr}}_2{\mathrm{Se}}_4/\mathrm{Pb}$ junctions at low temperatures ($T=0.3\mathrm{K}$ for sample $J1$ and 1.6 K for samples $J2–J4$; symbols). Each of them can be fitted to a quadratic function, i.e., $G(V)\sim G_0+b{V}^2$ (solid lines). The fit of the conductance spectra of sample $J2$ to the modified BTK theory (dotted line) is also shown for comparison. (b) Temperature dependence of the zero-bias conductance $G_0$ (symbols), which roughly follows a cubic dependence on $T$ (dotted line). It is also close to the calculations of the spin active scattering model with spin singlet-triplet mixing angle $\vartheta =\pi /4$ (solid line) [52]. The values derived from the modified BTK fit (solid line at the bottom) deviate strongly from the experimental curve. The data were taken from sample $J2$ in zero magnetic field.