A correlated insulator at the surface of the polar metal Ca$_3$Ru$_2$O$_7$

Abstract

We investigate the electronic structure at the surface of the correlated oxide Ca$_3$Ru$_2$O$_7$, a low-symmetry ruthenate oxide which hosts an unconventional polar-metal phase. From a combination of angle-resolved photoemission spectroscopy and scanning tunneling spectroscopy measurements, we demonstrate that the surface hosts an insulating phase, a distinct departure from metallicity within the bulk. Utilizing quantitative low-energy electron diffraction in conjunction with electronic structure calculations, we show how this results from a combined surface structure relaxation and the impact of marked electronic correlations in this system. Our findings highlight the proximity of Ca$_3$Ru$_2$O$_7$ to an insulating metallic state, and illustrate how subtle structural distortions can control its emergent electronic phases.

Figures (8)

• Figure 1: (a) Unit cell of bulk Ca3Ru2O7 with lattice vectors labeled. (b) Magnetic and electronic phase diagram of Ca3(Ru_1-$x$Ti_$x$)2O7. P: Paramagnetic, -M: Metallic phase, -I: Insulating phase, AFM$_\mathrm{\emph{a}}$: Antiferromagnetic order with spins orientated along the a-axis, AFM$_\mathrm{\emph{b}}$: Antiferromagnetic order with spins orientated along the b-axis, M: Mixed magnetic phase and AFM$_\mathrm{\emph{G}}$: Antiferromagnetic type $G$ order, with spins canted along the c-axis. Data taken from Refs. Peng2013Peng2016. (c) and (d) ARPES measurements ($h\nu = 33$ eV, linear horizontal polarization, $T = 18$ K) taken along the $\mathrm{M}_y-\Gamma-\mathrm{M}_y$ direction. (e) Tunneling spectrum for the surface of Ca3Ru2O7 ($I_{\mathrm{set}} = 200$ pA, $V_{\mathrm{set}} = -200$ mV, $T = 4.2$ K) (pink line) and angle-integrated photoemission intensity extracted from panel d (purple line).
• Figure 1: (a) and (b) Top view of the Ca3Ru2O7 (001) surface and the two antiphase domains. Between antiphase domains the direction of the preserved glide line (dot-dashed line) is rotated by 90$^\circ$. (c) and (d) Simulated LEED patterns at an incident electron beam energy, $E_\mathrm{beam}$, of 175 eV for the two antiphase domains. As the direction of the preserved glide line is rotated by 90$^\circ$ between antiphase domains, the direction of the extinct LEED spots is also rotated by 90$^\circ$ between antiphase domains. (e) $R_p$ factor as a function of antiphase domain occupancy within the calculated $I(V,\mathbf{Q})$ spectra. The $R_p$ value has a minimum close to 100$\%$ occupancy, suggesting only a single antiphase domain is present within the region of the probing LEED spot. This is consistent with the observation of only a single extinct glide line within our LEED measurements.
• Figure 2: (a) Top-down view of the two-dimensional bulk unit cell of Ca3Ru2O7. (b) Low-temperature STM topography of the Ca3Ru2O7 surface ($I = 100$ pA, $V = -2$ V, $T = 4.2$ K). Inset shows the Fourier transform of the topography. The image was low-pass filtered to remove high‑frequency noise. (c) LEED pattern of Ca3Ru2O7 measured at $T = 10$ K and an incident electron beam energy, $E_{\mathrm{beam}} = 175$ eV. The locations of the present (1,0) and extinct (0,1) LEED spots have been highlighted by the white circles. The bulk unit cell is highlighted by the blue square, and the locations of the glide line is marked by the dot-dashed lines in panels (a) and (b).
• Figure 2: (a) and (b) Top view of bulk Ca3Ru2O7 and the two polar domains. Between polar domains the direction of the preserved glide line (dot-dashed line) is unchanged whilst the direction of the polar displacement of Ca atoms is reversed, as shown by the solid arrows. (c) and (d) Simulated LEED patterns at an incident electron beam energy, $E_\mathrm{beam}$, of 175 eV for the two polar domains. As the direction of the preserved glide line is unchanged between polar domains, the direction of the extinct LEED spots is also unchanged between antiphase domains. (e) $R_p$ factor as a function of polar domain occupancy within the calculated $I(V,\mathbf{Q})$ spectra. The $R_p$ value has a minimum at $\sim 50\%$ occupancy, suggesting an equal population of the two domains are present within the region of the probing LEED spot.
• Figure 3: (a) LEED image recorded at an incidence electron beam energy, $E_{\mathrm{beam}}$, of 250 eV. Locations of the (1,1), (2,2), (3,1) and (4,0) LEED spots are highlighted by colored circles. (b) Four selected experimental (colored points) $I(V,\mathbf{Q})$ spectra of Ca3Ru2O7. Calculated $I(V,\mathbf{Q})$ spectra correspond to the best-fit surface structure (purple lines) or the bulk structure taken from Ref. Yoshida2005 (gray lines). (c) The out-of-plane and in-plane displacements at the surface of Ca3Ru2O7. (d) $R_p$ values for all the $I(V,\mathbf{Q})$ spectra analyzed and for two different structural models: the best-fit surface structure (purple points) and the bulk structure from Ref. Yoshida2005 (gray points). The total $R_p$ value for each model is indicated by the dashed line and points are offset horizontally for clarity.