Distinguishings±ands++electron pairing symmetries by neutron spin resonance in superconducting NaFe

Abstract

A determination of the superconducting (SC) electron pairing symmetry forms the basis for establishing a microscopic mechanism for superconductivity. For iron pnictide superconductors, the s ± -pairing symmetry theory predicts the presence of a sharp neutron spin resonance at an energy below the sum of hole and electron SC gap energies (E 2 ) below T c . On the other hand, the s ++ -pairing symmetry expects a broad spin excitation enhancement at an energy above 2 below T c . Although the resonance has been observed in iron pnictide superconductors at an energy below 2 consistent with the s ± -pairing symmetry, the mode has also been interpreted as arising from the s ++ -pairing symmetry with E 2 due to its broad energy width and the large uncertainty in determining the SC gaps. Here we use inelastic neutron scattering to reveal a sharp resonance at E = 7 meV in SC NaFe 0.935 Co 0.045 As (T c = 18 K). On warming towards T c , the mode energy hardly softens while its energy width increases rapidly. By comparing with calculated spin-excitation spectra within the s ± and s ++ -pairing symmetries, we conclude that the ground-state resonance in NaFe 0.935 Co 0.045 As is only consistent with the s ± pairing, and is inconsistent with the s ++ -pairing symmetry.

Temperature and polarization dependence of low-energy magnetic fluctuations in nearly optimally doped NaFe0.9785Co0.0215As

“…Interestingly, the above analysis indicates that the appearance of the resonance mode in NaFe 0.9785 Co 0.0215 As can be interpreted as removal of damping from an existing mode, a scenario recently proposed for Ce 1−x Yb x CoIn 5 [50]. However, we note that unlike Ce 1−x Yb x CoIn 5 , behaviors of the resonance mode in iron pnictides are also consistent with the spin-exciton scenario [27,51].…”

“…We find intensity of the resonance mode follows the superconducting order parameter, but energy of the mode E r is almost temperature-independent, softening only slightly upon approaching T c . This is different from the behavior in optimally doped BaFe 1.85 Co 0.15 As 2 but similar to other unconventional superconductors [11,[24][25][26][27]. Polarized neutron scattering reveals the resonance to be mostly isotropic, different from electron-, hole-, and isovalently doped BaFe 2 As 2 superconductors near optimal doping [27,36,44], which also exhibit an additional anisotropic component.…”

“…This is different from the behavior in optimally doped BaFe 1.85 Co 0.15 As 2 but similar to other unconventional superconductors [11,[24][25][26][27]. Polarized neutron scattering reveals the resonance to be mostly isotropic, different from electron-, hole-, and isovalently doped BaFe 2 As 2 superconductors near optimal doping [27,36,44], which also exhibit an additional anisotropic component. Significant spin anisotropy is instead present below the resonance mode, in the form of remnant spectral weight inside a partial spin gap induced by superconductivity, similarly to optimally doped YBa 2 Cu 3 O 6.9 [45].…”

“…The energy of the resonance mode was also found to track the superconducting order parameter as a function of temperature in optimally electron-doped BaFe 1.85 Co 0.15 As 2 [23]. However, energies of the resonance modes in YBa 2 Cu 3 O 7 [24], CeCoIn 5 [25], FeTe 0.6 Se 0.4 [26], optimally hole-doped Ba 0.67 K 0.33 Fe 2 As 2 [11], and electron-overdoped NaFe 0.935 Co 0.045 As [27] were found to only slightly soften upon approaching T c .…”

Raman resonance in iron-based superconductors: The magnetic scenario

“…Substituting into (33) we find that the vertex function for holes γ i (k) in the B 1g channel is cos(2θ k ) and has opposite signs for the two hole bands. For electron pockets, the situation is more simple since each electron pocket only has contributions from the d zx or d yz orbital, but not from both.…”