Fermi surface reconstruction in FeSe under high pressure

We report Shubnikov–de Haas (SdH) oscillation measurements on FeSe under high pressure up to $P=16.1$ kbar. We find a sudden change in SdH oscillations at the onset of the pressure-induced antiferromagnetism at $P\sim 8$ kbar. We argue that this change can be attributed to a reconstruction of the Fermi surface by the antiferromagnetic order. The negative $d{T}_c/dP$ observed in a range between $P\sim 8$ and 12 kbar may be explained by the reduction in the density of states due to the reconstruction. The ratio of the transition temperature to the effective Fermi energy remains high under high pressure: $k_B{T}_c/E_F\sim 0.1$ even at $P=16.1$ kbar.

Figure 1

High-pressure phase diagram of FeSe. Structural transition at $T_s$, AFM transition at $T_N$ $(T_u$ in [20]), and SC transition at $T_c$. $T_c$ is defined as the temperature where the resistance becomes zero. Different symbol shapes correspond to different samples. Data for samples other than A, B, and K are from [20].

Figure 2

(Left) Magnetic-field dependence of the resistance in sample K for different pressures. (Right) The same data in a log-log plot.

Figure 3

(a) SdH oscillations in FeSe at different pressures as a function of inverse field 1/$B$, and (b) FT amplitudes of the oscillations (dotted lines) and MEM PSDs (solid lines). $[P$ (kbar), $T$ (K), sample] = (0, 0.046, B), (4.4, 0.38, K), (8.9, 0.36, K), (10.2, 0.43, B), (11.3, 0.64, B), (14.1, 1.1, K), and (16.1, 1.65, A) from bottom to top. High-$T$ data are shown for high pressures because of lower $B_{c2}$, i.e., wider normal-state field ranges.

Figure 4

(a) SdH frequency, (b) effective mass, (c) effective Fermi energy, and (d) $k_B{T}_c/E_F$ in FeSe as a function of pressure. Solid and hollow symbols are based on FT and MEM analyses, respectively. They are consistent with each other and reveal the meaningful pressure evolution of those quantities. Error bars for $F$ and $m^{*}$ are based on standard deviations of measurements and fitting errors, respectively. Errors for the last two quantities are derived from those. Different symbol shapes correspond to different samples as in Fig. 1.

Figure 5

(Upper) Example of simulated signal. (Lower) FT amplitude (dotted curve) and MEM PSD (solid curve).

Figure 6

(Upper) Example of simulated signal. (Lower) FT amplitude (dotted curve) and MEM PSD (solid curve).

Figure 7

Estimated amplitude vs $A_1$ and $A_2$. Circles and squares are square roots of integrated powers, while + and x marks are square roots of peak heights.

Figure 8

SdH oscillations in FeSe at $P=4.4$ kbar.

Figure 9

(a) FT amplitudes and (b) and (c) MEM PSDs.

Figure 10

Temperature dependence of oscillation amplitudes estimated by FT and MEM. Solid curves are fits to the Lifshitz-Kosevich formula to estimate effective masses [48].