The results of experiments on acoustic-electric transformation (AET) in high-quality single crystalsFe0.963 ± 0.005Se indicating the existences of the piezomagnetic effect in them are presented.
Let us say a few words about the ideology of the AET experiment. A transverse elastic wave with the wave vector $q = (0, 0, q)$ and the displacement vector $u = (u, 0, 0)$ enters into the sample through a delay line. This wave produces an electromagnetic field (EMF) which forces free electrons to move to compensate the initial disturbance. The resultant current and the resultant electrical component of the EMF are aligned along the x-axis. In the magnetic field $H = (0, 0, H)$ the Lorenz force applied to free electrons results in the appearance of the y-components of the current and the electrical field (the Hall components). So in the absence of magnetic order and the external magnetic field we can observe only x-component of the electrical field.
In spite of this in the absence of an external magnetic field were recorded signals of AET, in FeSe single crystals, observed in the Hall-polarization. The signals appear below the temperature of solidification of the silicon oil ($T \sim 110K$), used as bonded material.
Two effects may provide the Hall current in the absence of the magnetic field. They are the anomalous Hall effect (AHE) and piezomagnetic (PZM) effect and both of them are realized in magnetically ordered media. Our estimates indicate that the appearance of the signals can not be explained by AHE. As for PZM we obtained a quite reasonable estimate that is in one order less than the maximum know value for the PZM modulus which was measured in CoF2 [1].
The explanation of our results in terms of the PZM effect immediately raises the question on the type of the magnetic structure in FeSe single crystals. It is obviously not an AFM in the common sense. It was shown in [1] that for the usual AFM the dependence of the PZM modulus on temperature reproduces the temperature dependence of the sublattice magnetization. But in our case the amplitude in the normal state increases almost linearly under decrease in temperature, and it does not show any tendency to the saturation. It is most likely that in case of FeSe we deal with a short range magnetic order. This conclusion correlates with an increase in AFM fluctuations in FeSe under approaching TC [2]. But we note that in any case such fluctuations should have some preferable orientation of the antiferromagnetic vector.
The signal decreases below the temperature of the superconducting transition, but it does not disappear completely and saturates at the value easily accessible for the measurements. We believe the electromagnetic signal decreases with lowering of temperature due to the change in the London penetration depth. So we can measure the temperature dependence $\lambda^{-2}_{L} (T)$ and the value $\lambda_{L} (0) = 1.65 \mu m$. Our estimate for $\lambda_{L} (0)$ corresponds to the carrier density $n \sim 10^{19} cm^{-3}$.