Control of the Competition between a Magnetic Phase and a Superconducting Phase in Cobalt-Doped and Nickel-Doped NaFeAs Using Electron Count

Using a combination of neutron, muon, and synchrotron techniques we show how the magnetic state in NaFeAs can be tuned into superconductivity by replacing Fe by either Co or Ni. The electron count is the dominant factor, since Ni doping has double the effect of Co doping for the same doping level. We follow the structural, magnetic, and superconducting properties as a function of doping to show how the superconducting state evolves, concluding that the addition of 0.1 electrons per Fe atom is sufficient to traverse the superconducting domain, and that magnetic order coexists with superconductivity at doping levels less than 0.025 electrons per Fe atom.

Figure 1

The evolution at 298 K of (a) As–Fe–As bond angles, (b) Fe–As bond lengths, and (c) lattice parameters of ${\mathrm{NaFe}}_{1-x}{\mathrm{Co}}_x\mathrm{As}$ as a function of Co content. The lines are straight-line fits. Inset: Crystal structure of NaFeAs.

Figure 2

The evolution of the zero-field magnetic susceptibility (measured in an applied magnetic field of 5 mT) of ${\mathrm{Na}}_{1-x}{\mathrm{Co}}_x\mathrm{As}$ samples (left) and the comparison of Ni- and Co-doped samples (right) showing that the susceptibilities are similar for isoelectronic compounds. Doping levels are expressed as percentages of Co unless specified.

Figure 3

(a) Zero-field $\mu \mathrm{SR}$ data in NaFeAs at 1.8 K. (b) Temperature dependence of the three precession frequencies (data for two different samples are shown in open symbols and closed symbols). (c) Relaxation rate for the three frequencies. (d) Zero-field $\mu \mathrm{SR}$ data in ${\mathrm{NaFe}}_{0.99}{\mathrm{Co}}_{0.01}\mathrm{As}$ at 1.8 K and (e) temperature dependence of the precession frequencies (only two frequencies can be extracted in this case).

Figure 4

(a) Rietveld refinements of NaFeAs against 2 K neutron data (HRPD). Data (red crosses), fit (green solid line), and difference plots are shown. Scattering from the sample environment has been excluded. (b),(c) Detail showing the splitting of the tetragonal 112 reflection into the orthorhombic 022 and 202 reflections. (d) Evolution of the orthorhombic $022/202$ reflections in ${\mathrm{NaFe}}_{0.995}{\mathrm{Co}}_{0.005}\mathrm{As}$ measured on ID31. At 46 K the orthorhombic and tetragonal models give fits of comparable quality. (e) Temperature dependence of the distortion for ${\mathrm{NaFe}}_{1-x}{\mathrm{Co}}_x\mathrm{As}$. The error bars lie within the points except when the presence of the distortion is marginal. (f) The size of the distortion at 5 K as a function of composition. Inset: For ${\mathrm{NaFe}}_{0.975}{\mathrm{Co}}_{0.025}\mathrm{As}$ the tetragonal 212 reflection measured on ID31 at 5 K is broader than at room temperature while the tetragonal 104 reflection (orthorhombic 114) does not broaden.

Figure 5

Uemura plot of the superconducting transition temperature $T_c$ versus the low temperature superfluid stiffness. Data obtained here for the doped NaFeAs samples are compared with previously reported data for other iron arsenides and cuprates [20, 21, 22, 23]. The lower dashed line indicates the trend line for electron-doped cuprates, which closely corresponds to the behavior of the LiFeAs system [19], in contrast to the other iron-arsenide superconductors, which sit closer to the trend for hole-doped cuprates. The symbol size represents the typical estimated uncertainty of the data points.

Figure 6

Phase diagram for ${\mathrm{NaFe}}_{1-x}{\mathrm{Co}}_x\mathrm{As}$. The maximum temperature at which the distortion persists is shown by the filled circles and dotted line. The antiferromagnetic (AFM) phase is delineated by $T_N$ (open circles) and superconducting (SC) phase by $T_c$ (filled squares).