Intrinsic and structural isotope effects in iron-based superconductors

The currently available results of the isotope effect on the superconducting transition temperature Tc in Fe-based high-temperature superconductors (HTSs) are highly controversial. The values of the Fe isotope effect exponent �Fe for various families of Fe-based HTS were found to be as well positive, as negative, or even be exceedingly larger than the BCS value �BCS�0.5. Here we emphasize that the Fe isotope substitution causes small structural modifications which, in turn, affect Tc. Upon correcting the isotope effect exponent for these structural effects, an almost unique value of ��0.35–0.4 is observed for at least three different families of Fe-based HTS. © 2010 The American Physical Society DOI: https://doi.org/10.1103/PhysRevB.82.212505 Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: https://doi.org/10.5167/uzh-45343 Accepted Version Originally published at: Khasanov, R; Bendele, M; Bussmann-Holder, A; Keller, H (2010). Intrinsic and structural isotope effects in iron-based superconductors. Physical Review. B, Condensed Matter and Materials Physics, 82(21):212505 . DOI: https://doi.org/10.1103/PhysRevB.82.212505 ar X iv :1 00 8. 45 40 v1 [ co nd -m at .s up rco n] 2 6 A ug 2 01 0 Intrinsic and structural isotope effects in Fe-based superconductors R. Khasanov, ∗ M. Bendele, 2 A. Bussmann-Holder, and H. Keller Laboratory for Muon Spin Spectroscopy, Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland Physik-Institut der Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland Max-Planck-Institut für Festkörperforschung, Heisenbergstrasse 1, D-70569 Stuttgart, Germany The currently available results of the isotope effect on the superconducting transition temperature Tc in Fe-based high-temperature superconductors (HTS) are highly controversial. The values of the Fe isotope effect (Fe-IE) exponent αFe for various families of Fe-based HTS were found to be as well positive, as negative, or even be exceedingly larger than the BCS value αBCS ≡ 0.5. Here we demonstrate that the Fe isotope substitution causes small structural modifications which, in turn, affect Tc. Upon correcting the isotope effect exponent for these structural effects, an almost unique value of α ∼ 0.35 − 0.4 is observed for at least three different families of Fe-based HTS. PACS numbers: 74.70.Xa, 74.62.Bf, 74.25.Kc The isotope effect on the superconducting transition temperature Tc traditionally plays an important role in identifying the superconducting pairing mechanism. As a rule, an impact of the isotope substitution and, consequently, an involvement of the lattice degrees of freedom into the pairing mechanism are determined by comparing the isotope effect exponent α = −(∆Tc/Tc)/(∆M/M) (M is the atomic mass) with the universal value αBCS ≡ 0.5 as predicted within the framework of BCS theory of electron-phonon mediated superconductivity. In conventional phonon mediated superconductors like simple metals, alloys, etc. α, typically, ranges from 0.2 to 0.5, see, e.g., Ref. 1 and references therein. The only exceptions are Ru and Zr exhibiting zero isotope effect and PdH(D) with αH(D) = −0.25. 2 The negative isotope effect of PdH(D) is explained, however, by the presence of strong lattice anharmonicty caused by the double-well potential in the proton (deuteron) bond distribution. This was confirmed by neutron scattering data where the large zero point motion of H in comparison with that of Deuterium results in 20% change of the lattice force constants. A similar finding exists in organic superconductors where the H(D) isotope effect changes sign as compared, e.g., to S, C, and N isotope replacements, see, e.g., Ref. 5 and references therein. Again, an unusually strong anharmonic lattice dynamics are attributed to this observation. The cuprate high-temperature superconductors (HTS) are characterized by a vanishingly small but positive isotope effect exponent in optimally doped compounds which increases in a monotonic way upon decreasing doping. For the optimally doped cuprate HTS the smallest value of the oxygen-isotope exponent αO ≃ 0.02 was obtained for YBa2Cu3O7−δ and Bi2Sr2Ca2Cu3O10+δ, while it reaches αO ≃ 0.25 for Bi2Sr1.6La0.4CuO6+δ. 7–9,13,14 In addition, it was demonstrated that in underdoped materials αO exceeds substantially the BCS limit αBCS ≡ 0.5. 8,10,14 It is important to note here that the values of both, the oxygen and the copper isotope exponents in cuprate HTS are always positive. Similar tendencies, with the only few above mentioned exceptions, are realized in a case of conventional phonon mediated superconductors. Since the discovery of superconductivity in Fe-based compounds few attempts to measure the isotope effect on Tc in these materials were made. Currently we are aware of four papers reporting, however, rather contradictory results. Liu et al. and Khasanov et al. have found a positive Fe isotope effect (Fe-IE) exponent αFe for Ba0.6K0.4Fe2As2, SmFeAsO0.85F0.15, and FeSe1−x with the corresponding values αFe = 0.34(3), 0.37(3), and 0.81(15), respectively. Note that αFe = 0.81(15) for FeSe1−x exceeds grossly the BCS value. In the other two studies Shirage et al. have reported a negative αFe = −0.18(3) and −0.024(15) for Ba0.6K0.4Fe2As2 16 and SmFeAsO1−y, 17 respectively. These controversial results are unlikely to stem from different pairing mechanisms to be realized in different Fe-based superconductors. Especially, in the case of Ba0.6K0.4Fe2As2, nominally identical samples were isotope replaced with one exhibiting a positive and the other a negative isotope exponent. Note, that the sign reversed isotope exponent seen by Shirage et al. was attributed to multiband superconductivity with different pairing channels, namely a phononic one and an antiferromagnetic (AF) fluctuation dominated one. On the other hand a multiband model cannot exhibit any sign reversed isotope exponent even if solely AF fluctuations were the pairing glue. In the present study we demonstrate that the very controversial results for αFe are caused by small structural changes occurring simultaneously with the Fe isotope exchange. As such, we decompose the Fe-IE exponent into one related to the structural changes α Fe and the genuine (intrinsic) one α Fe to arrive at: αFe = α int Fe + α str Fe . (1) By comparing the c-axis lattice constants for the pairs of isotopically substituted samples we observe that α Fe is negative for Ba0.6K0.4Fe2As2 and SmFeAsO1−y studied by Shirage et al. in Refs. 16 and 17, positive for FeSe1−x from Ref. 18 and close to 0 for Ba0.6K0.4Fe2As2 and SmFeAsO0.85F0.15 measured in Ref. 15. By taking into account the sign of α Fe we arrive at the conclusion that α Fe is positive for all so far studied Fe-based HTS. Our motivation to separate the isotope coefficient into the above mentioned two components, see Eq. (1), stems from the fact that superconductivity in these compounds is intimately related to small structural changes as re-