Transition from Sign-Reversed to Sign-Preserved Cooper-Pairing Symmetry in Sulfur-Doped Iron Selenide Superconductors

Qisi Wang, J. T. Park, Yu Feng, Yao Shen, Yiqing Hao, Bingying Pan, J. W. Lynn, A. Ivanov, Songxue Chi, M. Matsuda, Huibo Cao, R. J. Birgeneau, D. V. Efremov, and Jun Zhao

Phys. Rev. Lett. 116, 197004 – Published 13 May 2016

Abstract

An essential step toward elucidating the mechanism of superconductivity is to determine the sign or phase of the superconducting order parameter, as it is closely related to the pairing interaction. In conventional superconductors, the electron-phonon interaction induces attraction between electrons near the Fermi energy and results in a sign-preserved $s$ -wave pairing. For high-temperature superconductors, including cuprates and iron-based superconductors, prevalent weak coupling theories suggest that the electron pairing is mediated by spin fluctuations which lead to repulsive interactions, and therefore that a sign-reversed pairing with an $s_{\pm}$ or $d$ -wave symmetry is favored. Here, by using magnetic neutron scattering, a phase sensitive probe of the superconducting gap, we report the observation of a transition from the sign-reversed to sign-preserved Cooper-pairing symmetry with insignificant changes in $T_{c}$ in the S-doped iron selenide superconductors $K_{x} {Fe}_{2 - y} {({Se}_{1 - z} S_{z})}_{2}$ . We show that a rather sharp magnetic resonant mode well below the superconducting gap ( $2 Δ$ ) in the undoped sample ( $z = 0$ ) is replaced by a broad hump structure above $2 Δ$ under 50% S doping. These results cannot be readily explained by simple spin fluctuation-exchange pairing theories and, therefore, multiple pairing channels are required to describe superconductivity in this system. Our findings may also yield a simple explanation for the sometimes contradictory data on the sign of the superconducting order parameter in iron-based materials.

Received 24 November 2015

DOI:https://doi.org/10.1103/PhysRevLett.116.197004

Physics Subject Headings (PhySH)

Chalcogenides Superconductors

Neutron scattering

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

Qisi Wang¹, J. T. Park^2,*, Yu Feng¹, Yao Shen¹, Yiqing Hao¹, Bingying Pan¹, J. W. Lynn³, A. Ivanov⁴, Songxue Chi⁵, M. Matsuda⁵, Huibo Cao⁵, R. J. Birgeneau^6,7, D. V. Efremov⁸, and Jun Zhao^1,9,†

¹State Key Laboratory of Surface Physics and Department of Physics, Fudan University, Shanghai 200433, China
²Heinz Maier-Leibnitz Zentrum (MLZ), Technische Universität München, D-85748 Garching, Germany
³NIST Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, USA
⁴Institut Laue-Langevin, 71 Avenue des Martyrs, 38042 Grenoble Cedex 9, France
⁵Quantum Condensed Matter Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6393, USA
⁶Department of Physics, University of California, Berkeley, California 94720, USA
⁷Department of Materials Science and Engineering, University of California, Berkeley, California 94720, USA
⁸IFW Dresden, Helmholtzstraße 20, 01069 Dresden, Germany
⁹Collaborative Innovation Center of Advanced Microstructures, Fudan University, Shanghai 200433, China

^*jitae.park@frm2.tum.de
^†zhaoj@fudan.edu.cn

Article Text (Subscription Required)

Click to Expand

Supplemental Material (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 116, Iss. 19 — 13 May 2016

Reuse & Permissions

Access Options

Article Available via CHORUS

Download Accepted Manuscript

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
The energy dependence of the magnetic excitations and their temperature evolution in $K_{x} {Fe}_{2 - y} {Se}_{2}$ . (a) Energy scans at $Q = (- 0.5, 0.77, 0)$ and $Q = (- 0.7, 0.77, 0)$ for temperatures below and above $T_{c}$ . A superconductivity-induced resonant mode is clearly observed at 13 meV at $Q = (- 0.5, 0.77, 0)$ , while the scattering at $Q = (- 0.7, 0.77, 0)$ displays little temperature dependence across $T_{c}$ . (b) Background subtracted contour map of change in the magnetic scattering as a function of energy and temperature. The background was measured at $T = 36 K$ . The open circles indicate $2 Δ$ adapted from the ARPES measurements of Ref. [35], which is consistent with the data measured on a single crystal from the same batch used for our neutron scattering measurements (not shown). The dotted line and dashed line are guides to the eye. The error bars indicate 1 standard deviation throughout the Letter.
Reuse & Permissions
Figure 2
Momentum dependence of the magnetic excitations in $K_{x} {Fe}_{2 - y} {({Se}_{1 - z} S_{z})}_{2}$ ( $z = 0$ , 0.25, 0.4, 0.5) below and above $T_{c}$ . (a)–(d) Constant energy scans in the superconducting (black) and the normal state (red) for the $z = 0$ , 0.25, 0.4, and 0.5 samples, respectively. The energies were chosen to correspond to the peak in the extra scattering associated with the superconductivity. The data fit well to Gaussian curves as indicated by solid lines.
Reuse & Permissions
Figure 3
Temperature dependence of the spin excitations and their doping evolution in $K_{x} {Fe}_{2 - y} {({Se}_{1 - z} S_{z})}_{2}$ ( $z = 0$ , 0.25, 0.4, 0.5) below and above $T_{c}$ . (a)–(d) Energy dependence of the intensity difference between the superconducting and the normal states in the vicinity of $Q = (0.5, 0.75, 0)$ or equivalent wave vectors. The dashed lines represent $2 Δ$ determined by ARPES measurements [34]. The shaded area denotes the spectral weight enhancement below $T_{c}$ . The resonance energies and hump (shoulder) energies are marked by black and orange arrows, respectively. The data are normalized to the amplitude of the normal state $Q$ scans at similar energies between different doping levels. The insets show the temperature dependence of dc magnetic susceptibility. The magnetic field ( $H = 10 Oe$ ) is applied parallel to the $a b$ plane. $1 Oe = (1000 / 4 π) A / m$ . (e)–(k) Temperature dependence of the spin excitations measured at various energies. (e) $z = 0$ , $E = 13 meV$ , $Q = (- 0.5, 0.77, 0)$ ; (f) $z = 0.25$ , $E = 13 meV$ , $Q = (0.5, 0.77, 0)$ ; (g) $z = 0.4$ , $E = 13 meV$ , $Q = (0.5, 0.25, 0)$ ; (h) $z = 0.5$ , $E = 20 meV$ , $Q = (0.5, 0.75, 0)$ ; (i) $z = 0.25$ , $E = 13 meV$ , $Q = (0.5, 0.25, 0)$ ; (j) $z = 0.25$ , $E = 21 meV$ , $Q = (0.5, 0.77, 0)$ ; (k) $z = 0.4$ , $E = 21 meV$ , $Q = (0.5, 0.77, 0)$ . The horizontal error bars indicate the temperature range within which data were combined. The solid lines are guides to the eye.
Reuse & Permissions
Figure 4
Doping dependence of $T_{c}$ , superconducting gap ( $2 Δ$ ), and superconductivity induced magnetic peak energy ( $E_{p}$ ). (a) Doping dependence of the $T_{c}$ and the superconducting gap ( $2 Δ$ ) adapted from ARPRES measurements in Ref. [34]. (b) Doping dependence of the superconductivity-induced magnetic peak energy ( $E_{p}$ ) and the ratio of $E_{p} / 2 Δ$ . For $z = 0$ , 0.25, and 0.4 samples, $E_{p}$ denotes the resonance energy. For the $z = 0.5$ sample, $E_{p}$ denotes the energy of the hump structure above $2 Δ$ . The dot-dashed line indicates the empirical universal ratio of $E_{p} / 2 Δ = 0.64$ for magnetic superconductors [37]. The dashed line indicates $E_{p} / 2 Δ = 1$ .
Reuse & Permissions

Physical Review Letters