Pressure-induced electronic transitions in samarium monochalcogenides

The pressure-induced isostructural insulator-to-metal transition for SmS is characterized by the presence of an intermediate valence state at higher pressure which cannot be captured by density functional theory. As a direct outcome of including the charge and spin fluctuations incorporated in dynamical mean-field theory, we see the emergence of insulating and metallic phases with increasing pressure as a function of changing valence. This is accompanied by significantly improved predictions of the equilibrium lattice constants and bulk moduli for all Sm monochalcogenides verifying experiments. Nudged elastic band analysis reveals the insulating states to have a finite quasiparticle weight, decreasing as the gap closes rendering the transition to be not Mott-like, and classifies these materials as correlated band insulators. The difference between the discontinuous and continuous natures of these transitions can be attributed to the closeness of the sharply resonant Sm-4$f$ peaks to the Fermi level in the predicted metallic states in SmS compared with SmSe and SmTe.

Figure 1

Density of states (DOS) for SmS obtained with (a) DFT, (b) DFT$+$DMFT for the Sm-4$f^6$ configuration corresponding to the ambient pressure insulating black phase [(c) is a magnified version of (b) showing the existence of a finite band gap], and (d) DFT$+$DMFT for the configuration corresponding to the onset of high-pressure metallicity with intermediate Sm-4$f$ valence of 5.24(2) in the golden phase.

Figure 2

Total energy $E_{\mathrm{tot}}$ curves obtained using density functional theory plotted as a function of the lattice constant $a$ for SmS as implemented in the plane-wave basis code vasp [39, 40] using the projector augmented wave (PAW) [41, 42] PBE [43] potentials with and without spin-orbit coupling (SOC) [44] optimized with the conjugate gradient algorithm for each value of $a$ with the energy cutoff of 520 eV, a $19\times 19\times 19$ Monkhorst $k$ grid [37], and a smearing of 0.05 eV with the tetrahedron method and Blöchl corrections.

Figure 3

Comparison between DFT and DFT$+$DMFT predictions of total energy $E_{\mathrm{tot}}$ curves plotted as a function of lattice constant $a$ and fitted with a Birch-Murnaghan [47, 48] equation of state for (a) SmS, (b) SmSe, and (c) SmTe. Arrows show the experimentally observed values for the respective lattice constants under ambient conditions.

Figure 4

Left column: density of states obtained with DFT$+$DMFT corresponding to the ambient pressure insulating states for (a) SmS, (c) SmSe, and (e) SmTe. Right column: the high-pressure intermediate valence metallic states with insets focusing on the behavior of Sm-4$f$ peaks around the Fermi level for (b) SmS, (d) SmSe, and (f) SmTe, respectively.

Figure 5

A comparison between the DOS for different values of Hund's coupling J for SmS with $J=0.3$ eV and (b) $J=0.8355$ eV, adapted from Ref. [33]. The choice of $J=0.3$ is based on correctly reproducing the hybridization between Sm-4$f$ and S-$p$ states and the band gap consistent with experimental observations.

Figure 6

(a) Total quasiparticle weight $Z$, (b) Sm-$4f$ electron occupancy $n_f$, and (c) total density of states at the Fermi level $\rho \left(E_F\right)$ along the NEB path going from the insulating state (Ins; $\lambda =0$) to the metallic state (Met; $\lambda =1$) for SmS, SmSe, and SmTe.

Figure 7

Left column: difference in the density of states obtained with DFT$+$DMFT at high temperature ($T=0.05$ eV; solid lines, unshaded) and room temperature ($T=0.026$ eV; dashed lines, shaded) corresponding to the ambient pressure insulating states for (a) SmS, (c) SmSe, and (e) SmTe. Right column: the high-pressure intermediate valence metallic states for (b) SmS, (d) SmSe, and (f) SmTe. Small shifts in the DOS are observed on lowering the temperature with no qualitative differences in trends compared with the higher temperature.