Robustness of superconducting properties to transition metal substitution and impurity phases in ${\mathrm{Fe}}_{1-x}{\mathrm{V}}_x\mathrm{Se}$

We have performed transverse- and zero-field muon spin rotation/relaxation experiments, as well as magnetometry measurements, on samples of ${\mathrm{Fe}}_{1\text{−}x}{\mathrm{V}}_x\mathrm{Se}$ and their $\text{Li}+{\mathrm{NH}}_3$ intercalates ${\mathrm{Li}}_{0.6}{\left({\mathrm{NH}}_2\right)}_{0.2}{\left({\mathrm{NH}}_3\right)}_{0.8}{\mathrm{Fe}}_{1\text{−}x}{\mathrm{V}}_x\mathrm{Se}$. We examine the low vanadium substitution regime: $x=0.005$, 0.01, and 0.02. The intercalation reaction significantly increases the critical temperature $\left(T_{\mathrm{c}}\right)$ and the superfluid stiffness for all $x$. The nonintercalated samples all exhibit $T_{\mathrm{c}}\approx 8.5$ K while the intercalated samples all show an enhanced $T_{\mathrm{c}}>40$ K. Vanadium substitution has a negligible effect on $T_{\mathrm{c}}$, but seems to suppress the superfluid stiffness for the nonintercalated samples and weakly enhance it for the intercalated materials. The optimal substitution level for the intercalated samples is found to be $x=0.01$, with $T_{\mathrm{c}}\approx 41\mathrm{K}$ and ${\lambda }_{ab}\left(0\right)\approx 0.18\mu \mathrm{m}$. The nonintercalated samples can be modeled with either a single $d$-wave superconducting gap or with an anisotropic gap function based on recent quasiparticle imaging experiments, whereas the intercalates display multigap nodal behavior which can be fitted using $s+d$- or $d+d$-wave models. Magnetism, likely from iron impurities, appears after the intercalation reaction and coexists and competes with the superconductivity. However, it appears that the superconductivity is remarkably robust to the impurity phase, providing an avenue to stably improve the superconducting properties of transition metal substituted FeSe.

Figure 1

(a) Sample TF-$\mu \mathrm{SR}$ spectra above and below $T_{\mathrm{c}}$ for $x=0.01$. Fits as in Eq. (1) are also plotted. (b) Dependence of the nuclear contribution to the superconducting relaxation [described in Eq. (1)] on vanadium substitution for both intercalated $\left(+{\mathrm{NH}}_3\right)$ and nonintercalated (no $+{\mathrm{NH}}_3)$ samples. The temperature dependence of the field width of the superconducting vortex lattice is given in (c) and (d) for the nonintercalated and intercalated samples respectively. The temperature dependence of the inverse square penetration depth for nonintercalated and intercalated samples is shown in (e) and (f) respectively. The data in (e) have been fitted with a single-gap $d$-wave function, and the data in (f) have been fitted with two-gap $d+d$ and $s+d$ models.

Figure 2

Fits of the data for ${\mathrm{Fe}}_{1\text{−}x}{\mathrm{V}}_x\mathrm{Se}$ to a gap function (see inset) based on the results of quasiparticle imaging experiments [29, 30] on the surface of bulk FeSe, as described in the main text.

Figure 3

(a) Temperature dependence of the magnetization for all samples, measured in Bohr magnetons per formula unit. (b) ZF-$\mu \mathrm{SR}$ asymmetry for $x=0.02+{\mathrm{NH}}_3$ at a range of temperatures. The black lines give fits as in Eq. (5). (c) Temperature dependence of relaxation $\lambda$, and initial and baseline asymmetries $(A_0$ and $A_{\mathrm{b}}$ respectively) of the ZF-$\mu \mathrm{SR}$ asymmetry for $x=0.02+{\mathrm{NH}}_3$.

Figure 4

Powder-diffraction data for the $x=0.005$ sample before and after exposure to air. The asterisks indicate new impurity peaks. Notice also the increase in the diffuse background.