Study on fluctuations of interface-enhanced superconductivity in ultrathin FeSe/SrTiO3 by the Nernst effect

TL;DR

This study addresses why interface-enhanced superconductivity arises in ultrathin FeSe films on SrTiO$_3$ by probing superconducting fluctuations with the Nernst effect. Measurements on ~2.5 and ~5 nm films show normal-state Nernst signals similar to bulk FeSe, suggesting STO-derived carriers are confined to a few interfacial layers. Below $T^* \approx 1.2\,T_{ ext{c}}^{\text{onset}}$, the Nernst signal from superconducting fluctuations emerges, and its magnetic-field dependence of $\alpha_{xy}^{2D}$ favors amplitude fluctuations over phase fluctuations, consistent with Gaussian and vortex-fluid descriptions and a two-layer parallel model. Together with a Meissner response below $T_{ ext{c}}^{\text{zero}}$ and a lack of a robust pre-formed-pair pseudogap tied to superconductivity, these results suggest the pseudogap originates from non-superconducting electronic states rather than superconductivity.

Abstract

Ultrathin FeSe films on SrTiO3 substrate show interface-enhanced superconductivity. However, how the superconductivity is established including superconducting fluctuations remains unclear. This study investigates the Nernst effect, which is sensitive to superconducting fluctuations, in ultrathin FeSe films on SrTiO3. Temperature dependence of Nernst signals in the normal state is similar to bulk FeSe, suggesting that the electrons of SrTiO3 are transferred only to a few layers near the FeSe/SrTiO3 interface. The Nernst effect caused by SC fluctuations was observed only below T ~ 1.2 Tconset within our measurement resolution, which is similar to other Fe chalcogenide systems. Our results suggest that the pseudogap in monolayer FeSe/STO possibly originates in other electronic states rather than superconductivity.

Figures (6)

• Figure 1: (a) Schematic picture of setup used for Nernst effect measurement in this study. (b) $V_{xy}$ as a function of $\Delta T$ at 38 K and 9 T in a 2.5 nm film.
• Figure 2: (a) Temperature dependences of normalized sheet resistance $R_{sq}$ at 0 T in the grown films. (b) Nernst signals $N$ as a function of temperature up to 90 K at 9 T in the 2.5 nm film.
• Figure 3: Temperature dependence of (a) $R_{sq}$ and (c) $N$ at 0--9 T in the 2.5 nm film. (b) and (d) show the corresponding measurements results obtained for the 5 nm film. (e) and (f) plot the $N$ and $S\ \mathrm{tan}\theta$ at 9 T for the 2.5 nm and 5 nm films, respectively, for comparison.
• Figure 4: (a), (b) Temperature dependence of transverse thermoelectric coefficients divided by magnetic fields $\alpha_{xy}^{2D} /B$ in 0.4, 1, 3, 5, and 9 T of 2.5 nm and 5 nm films, respectively. The insets show the expanded plots near $T^*$. Solid lines are just the connection between the neighbouring points.
• Figure 5: Analysis of SC fluctuations for the 5 nm film. Red points in (a) and (b) denote $\alpha_{xy(\mathrm{SC})}^{2D}$ as a function of temperature at 0.4 T and magnetic field at 22.8 K. The black curves in (a) and (b) were calculated from Eq. (\ref{['eq:fluidmodel']}) using parameters in Table (\ref{['table:fitpara']}). The blue curve in (a) exhibits the fitting result for Eq.(\ref{['eq:gaussian']}), while that in (b) is adjusted curve of Ref.Michaeli_2009 with $B_{\mathrm{peak}}=4.7$ T.