Real-Space Coexistence of the Melted Mott State and Superconductivity in Fe-Substituted $1T\mathrm{\text{−}}{\mathrm{TaS}}_2$

We have performed high-resolution angle-resolved photoemission spectroscopy of layered chalcogenide $1T\mathrm{\text{−}}{\mathrm{Fe}}_x{\mathrm{Ta}}_{1-x}{\mathrm{S}}_2$ which undergoes a superconducting transition in the nearly commensurate charge-density-wave phase (melted Mott phase). We found a single electron pocket at the Brillouin-zone center in the melted Mott phase, which is created by the backfolding of bands due to the superlattice potential of charge-density-wave. This electron pocket appears in the $x$ region where the samples show superconductivity, and is destroyed by the Mott- and Anderson-gap opening. Present results suggest that the melted Mott state and the superconductivity coexist in real space, providing a new insight into the interplay between electron correlation, charge order, and superconductivity.

Figure 1

(a) Schematic electronic phase diagram of $1T\mathrm{\text{−}}{\mathrm{Fe}}_x{\mathrm{Ta}}_{1-x}{\mathrm{S}}_2$ derived from transport properties [23] as a function of temperature and $x$, where CCDW, NCCDW, SC, and AL represents the commensurate CDW, nearly commensurate CDW, superconductivity, and Anderson localization, respectively. For $x=0.02$ and 0.03, the SC critical temperature $T_c$ is 2.8 and 2.6 K, respectively. (b) EDCs of pristine $1T\mathrm{\text{−}}{\mathrm{TaS}}_2$ measured at 30, 300, and 360 K along the $\Gamma M$ direction. (c) Corresponding second-derivative ARPES intensity $1T\mathrm{\text{−}}{\mathrm{TaS}}_2$ plotted as a function of wave vector and $E_{\mathrm{B}}$.

Figure 2

(a) 2D ARPES intensity plot of pristine $1T\mathrm{\text{−}}{\mathrm{TaS}}_2$ measured at $T=360\mathrm{K}$. Orange circles correspond to the normal-state FS. (b) EDCs at five representative $k_{\mathrm{F}}$ points (A–F) shown in (a), measured at $T=360$, 300, and 30 K. (c) Temperature dependence of the EDCs at point B. (d) Symmetrized EDCs near $E_{\mathrm{F}}$.

Figure 3

(a) ARPES intensity of $1T\mathrm{\text{−}}{\mathrm{Fe}}_x{\mathrm{Ta}}_{1-x}{\mathrm{S}}_2$ measured along the $\Gamma M$ direction for $x=0$, 0.02, and 0.05, plotted as a function of wave vector and $E_{\mathrm{B}}$ in the normal metallic phase ($T=350\mathrm{K}$). (b) Direct comparison of the band dispersions for $x=0$, 0.01, 0.02, and 0.05 at $T=350\mathrm{K}$. The energy positions of the bands are determined by tracing the peak position of EDCs. (c) Comparison of ARPES spectra at the $k_{\mathrm{F}}$ point along the $\Gamma M$ direction. (d) $x$ dependence of EDCs around the $\Gamma$ point in $1T\mathrm{\text{−}}{\mathrm{Fe}}_x{\mathrm{Ta}}_{1-x}{S}_2$ at $T=30\mathrm{K}$. (e) ARPES intensity plot as a function of a 2D wave vector for $x=0.02$. Intensity is obtained by assuming the hexagonal symmetry.

Figure 4

(a) Various gap energies (the Mott gap ${\Delta }_{\mathrm{Mott}}$, the NCCDW gap ${\Delta }_{\mathrm{NCCDW}}$, and the AL gap ${\Delta }_{\mathrm{AL}}$) and density of states (DOS) at $E_{\mathrm{F}}$ of an electron pocket plotted as a function of $x$. The ${\Delta }_{\mathrm{Mott}}$ value was estimated by the peak position of the EDC at the minimum-gap $k$ locus of the weakly dispersing lower Hubbard band (LHB), while the ${\Delta }_{\mathrm{NCCDW}}$ and ${\Delta }_{\mathrm{AL}}$ values were determined by the EDC peak position at the $k_{\mathrm{F}}$ point along the $\Gamma M$ direction in the normal phase (note that the gap value does not strongly depend on the particular selection of the $k_{\mathrm{F}}$ point because of its nearly isotropic nature). The DOS at $E_{\mathrm{F}}$ is estimated by the intensity at $E_{\mathrm{F}}$ of EDCs integrated along the $\Gamma M$ direction, normalized to the 0.4-eV peak. (b) Schematic band diagram derived from the present ARPES experiment as a function of temperature and $x$. Dashed curve represents the absence of the band.