Dimensionality-driven orthorhombic $\mathrm{MoT}{\mathrm{e}}_2$ at room temperature

We use a combination of Raman spectroscopy and transport measurements to study thin flakes of the type-II Weyl semimetal candidate $\mathrm{MoT}{\mathrm{e}}_2$ protected from oxidation. In contrast to bulk crystals, which undergo a phase transition from monoclinic to the inversion symmetry breaking, orthorhombic phase below $\sim 250\mathrm{K}$, we find that in moderately thin samples below $\sim 12\mathrm{nm}$, a single orthorhombic phase exists up to and beyond room temperature. This could be due to the effect of $c$-axis confinement, which lowers the energy of an out-of-plane hole band and stabilizes the orthorhombic structure. Our results suggest that Weyl nodes, predicated upon inversion symmetry breaking, may be observed in thin $\mathrm{MoT}{\mathrm{e}}_2$ at room temperature.

Figure 1

(a) Structure of orthorhombic ($\gamma$ or $T_d$) and monoclinic ($\beta$ or $1T^{\prime }$) phases of $\mathrm{MoT}{\mathrm{e}}_2$. (b) Right inset: Optical image of thin flake device capped with hBN to protect from sample oxidation. $\mathrm{MoT}{\mathrm{e}}_2$ is outlined with a dashed line. Scale bar is $10\mu \mathrm{m}$. Main panel: Normalized temperature-dependent resistivity of $\mathrm{MoT}{\mathrm{e}}_2$ bulk crystal and thin flakes. Upper traces are offset for clarity. Kink and hysteresis between cooling and warming at 250 K corresponds to a first-order $\beta \text{−}\gamma$ phase transition. Hysteresis becomes less visible in thinner flakes. Left inset: Percentage resistivity difference between cooling and warming in the middle of the hysteresis region (marked by purple arrows) as a function of flake thickness.

Figure 2

(a) Main panel, top two traces: Raman spectra of 50- and 20-nm $\mathrm{MoT}{\mathrm{e}}_2$ at room temperature resembling the $\beta$ phase. Lower traces: Spectra of 50-nm $\mathrm{MoT}{\mathrm{e}}_2$ at 140 K ($\gamma$ phase) and thinner flakes at 294 K. $\gamma \text{−}\mathrm{MoT}{\mathrm{e}}_2$ shows additional peaks at $\sim 12$ and $\sim 130\mathrm{c}{\mathrm{m}}^{-1}$, indicating thin samples are in the $\gamma$ phase at room temperature. (b) Raman mode positions vs thickness for $\gamma \text{−}\mathrm{MoT}{\mathrm{e}}_2$ flakes at 294 K. Corresponding mode positions for the 50-nm sample at 140 K are marked by dashed lines.

Figure 3

(a) Temperature evolution of Raman spectra for the 4.5-nm sample upon cooling and heating from room temperature. (b) Mode positions and areal intensities vs temperature after Lorentzian fits. Traces have been offset for clarity. No abrupt changes at $\sim 250\mathrm{K}$ corresponding to a first-order $\beta \text{−}\gamma$ phase transition are observed.

Figure 4

(a) Temperature-thickness phase diagram. Below a critical thickness, a single $\gamma$ phase is stabilized at all temperatures up to 400 K. The high-temperature $\beta$ phase undergoes a crossover with reducing thickness, likely due to a slowly changing structure. (b) Main panel: Proposed mechanism for crossover. An out-of-plane hole band is shifted to lower energy upon cooling in thick samples and confinement in thin samples. Red and blue traces are reproduced from Kim et al. [25] and correspond to bulk $\mathrm{MoT}{\mathrm{e}}_2$ in the $\beta$ and $\gamma$ phase, respectively. Gray traces correspond to the $\beta$ band confined to thickness 20, 10, 7, and 5 nm. Inset: $\beta$ band energy vs thickness at the $\mathrm{\Gamma }$ point, showing crossing below the $\gamma$-phase energy (dashed blue line) at $\sim 5\mathrm{nm}$.