Structural and magnetic phase transitions in ${\text{Na}}_{1-\delta }\text{FeAs}$

We use neutron scattering to study the spin and lattice structures of single crystal and powder samples of ${\text{Na}}_{1-\delta }\text{FeAs}$ $\left(T_c=23\text{K}\right)$. Upon cooling from room temperature, the system goes through a series of phase transitions: first changing the crystal symmetry from tetragonal to orthorhombic at 49 K, then ordering antiferromagnetically with a spin structure similar to that of LaFeAsO and a small moment $\left(0.09\pm 0.04{\mu }_B\right)$, and finally becoming superconducting below about 23 K. These results confirm that antiferromagnetic order is ubiquitous for the parent compounds of the iron arsenide superconductors and suggest that the separated structural and magnetic phase-transition temperatures are due to the reduction in the $c$-axis exchange coupling of the system.

Figure 1

(Color online) (a) Nuclear and (b) magnetic structures of the ideal NaFeAs. (a) includes two orthorhombic nuclear unit cells for comparison with the magnetic unit cell in (b). (c) and (d) show the $\left[H,0,1\right]$ and $\left[1,0,L\right]$ scans around ${\left(1,0,1\right)}_M$ magnetic Bragg peak at 5 and 70 K. Both peaks disappear at 70 K, indicating their magnetic nature.

Figure 2

(Color online) (a) Temperature dependence of the $2\theta$ scans around the ${\left(2,2,1\right)}_T$ Bragg peak in powder-diffraction data across the tetragonal to orthorhombic phase transition. Clear splitting of the peaks is seen below 33 K. We collected full powder-diffraction spectra at several temperatures, and refinement results gave the temperature dependence of several key parameters: (b) lattice constants $a$, $b$, and $c$; (c) bond lengths of Fe-Fe and Fe-As; and (d) bond angles of Fe-As-Fe. In all panels, left and right axes have the same scale.

Figure 3

(Color online) (a) Temperature dependence of the FWHM of $\theta \text{−}2\theta$ scan at the nuclear peak ${\left(1,1,0\right)}_T$ position using $\lambda /2$ scattering by removing the PG filter. The peak width clearly increases below about 50 K as marked by the vertical line. (b) Temperature dependence of the peak intensity at the AF peak ${\left(1,0,3\right)}_M$ position suggesting a Néel temperature of 39 K. The black line is the fitted result for the whole temperature range by using a simple Ising model described in the text, while the blue line only focuses on temperatures above 20 K which shows strong deviation from the experimental data below 20 K. No anomaly is seen across $T_c$, suggesting that superconductivity is filamentary and not a bulk phenomenon.

Figure 4

(Color online) (a) Comparison of the $\theta \text{−}2\theta$ scans at ${\left(1,0,1\right)}_M$ with and without PG filter, measuring magnetic and nuclear peaks, respectively. (b) Calculated and experimental intensities of magnetic peaks ${\left(1,0,L\right)}_M$ $\left(L=1,3,5,7\right)$. In the calculation, the moment is assumed to be along the $a_O$ axis.