AA Large Iron Isotope Effect in SmFeAsO F x and Ba K x Fe As R. H. Liu , T. Wu , G. Wu , H. Chen , X. F. Wang , Y. L. Xie , J. J. Yin , Y. J. Yan , Q. J. Li , B. C. Shi , W. S. Chu , Z. Y. Wu , X. H. Chen Hefei National Laboratory for Physical Sciences at Microscale and Department of Physics, University of Science and Technology of China, Hefei, Anhui 230026, China Beijing Synchrotron Radiation Facility, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049, China and National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei 230026, China
The recent discovery of superconductivity in oxypnictides with the critical temperature (T C ) higher than McMillan limit of 39 K (the theoretical maximum predicted by Bardeen-Cooper-Schrieffer (BCS) theory) has generated great excitement . Theoretical calculations indicate that the electron-phonon interaction is not strong enough to give rise to such high transition temperatures , while strong ferromagnetic/antiferromagnetic fluctuations have been proposed to be responsible . However, superconductivity and magnetism in pnictide superconductors show a strong sensitivity to the lattice, suggesting a possibility of unconventional electron-phonon coupling. Here we report the effect of oxygen and iron isotopic mass on T C and the spin-density wave (SDW) transition temperature (T SDW ) in SmFeAsO F x and Ba K x Fe As systems. The results show that oxygen isotope effect on T C and T SDW is very little, while the iron isotope exponent α C =-d lnT C /d lnM is about 0.35, being comparable to 0.5 for the full isotope effect. Surprisingly, the iron isotope exchange shows the same effect on T SDW as T C . These results indicate that electron-phonon interaction lays some role in the superconducting mechanism, but simple electron-phonon coupling mechanism seems to be rather unlikely because a strong magnon-phonon coupling is included. Sorting out the interplay between the lattice and magnetic degrees of freedom is a key challenge for understanding the mechanism of high-T C superconductivity. Recent inelastic neutron scattering measurements on Ba K x Fe As (x=0 and 0.4) provide evidence for presence of magnetic excitations . It suggests that spin fluctuation may play an important role for the mechanism of superconductivity. However, it is demonstrated that phonons couple selectively to the spin system . The structural transition from tetragonal to orthorhombic is driven by the antiferromagnetic SDW order , and the antiferromagnetic SDW exists only in the orthorhombic structure . The pressure coefficient of T C , dT C / dP , changes from positive to negative with a crossover from orthorhombic to tetragonal symmetry for the superconducting phase . The superconductivity and spin-density wave coexist in the orthorhombic structure . These results indicate remarkable sensitivity of superconductivity and magnetism to the lattice. The isotopically substituted polycrystalline samples with nominal compositions SmFeAsO F x (x=0, 0.15) and Ba K x Fe As (x=0, 0.4) were synthesized by conventional solid state reaction described in Ref.5 and Ref.17, respectively. Figure 1 shows the Raman spectra for the samples SmFeAsO F x by replacing O with O, and for the samples SmFeAsO F x and Ba K x Fe As by replacing Fe with the isotope Fe. The frequency shift of 4.2% and 4.5% for the E g mode of oxygen around 420 cm -1 suggests about 71% and 77% O substitution for O for x=0 and 0.15 samples, respectively. Raman shift of about 1.7% observed in the four samples for the B mode of ron indicates almost 100% Fe substitution for Fe for the two systems. These data are listed in Table 1. The temperature dependence of resistivity ( ρ ) and its derivative (d ρ /dT) for typical samples SmFeAsO F x on replacing O with the isotope O are shown in Fig. 2. T C and T SDW are listed in Table 2 for all samples from different batches. Based on isotope exponent α C =-d lnT C /d lnM , the α C is deduced to be -0.06(1) for the superconducting transition. To quantitatively compare the isotope effect on T SDW with on T C , we also define an isotope exponent α SDW =-d lnT
SDW /d lnM for SDW transition although no theory is established for isotope effect on magnetic phase transition yet. α SDW =-0.05(1) is obtained. These results indicate that oxygen isotope effect on T C and T SDW is very little. Temperature dependence of resistivity and its derivative for typical samples SmFeAsO F x and Ba K x Fe As by substitution of Fe for Fe are shown in Fig.3. An increase in T C is clearly observed in resistivity measurements, and d ρ /dT clearly shows an increase of SDW transition by substitution of Fe for Fe. The average results for several different samples are listed in Table 2. The average isotope component α SDW for several samples of SmFeAsO and BaFe As is 0.39(2) and 0.36(2), and α C is found to be 0.34(3) and 0.37(3) averaged over several samples of SmFeAsO F and Ba K Fe As , being comparable to 0.5 for full isotope effect in the framework of BCS theory. It indicates a strong iron isotope effect on T C and T SDW . It implies that electron-phonon interaction should play an important role for the superconducting mechanism. It is striking that the iron isotope exponents ( α C and α SDW ) on Tc and T
SDW are almost the same for he two systems, and much larger than the oxygen isotope exponents. Isotope effect studies require well-characterized samples with reproducible crystal chemistry properties. An important experimental point has to be addressed about the sample processing. It is found that T
SDW is insensitive to the sample processing for the parent compounds. T C of the sample SmFeAsO F is sensitive to the sample processing because the F content is not easy to control. A detailed description of the synthesis procedure used to ensure the same F content is given in the Supplementary Information. No difference in the lattice constants (see Supplementary Fig. S1) provides strong evidence for the same F content for isotope exchange. To confirm that the observed results is intrinsic instead of impurity effect, we checked the difference of T C and T SDW for the samples Ba K x Fe As with natural abundance iron ( n Fe ) with purity of 99.9% and Fe with purity of 99.78%. T
SDW and T C are nearly the same for the samples with n Fe and Fe (see Supplementary Fig. S3). We synthesized the samples Ba K Ba As using n Fe with purity of 98% and 99.9% to check impurity effect on Tc. The difference of Tc for the two samples is 0.07 K (see Supplementary Fig. S4). It indicates that the effect of impurity on Tc is very little, and does not affect the intrinsic isotope effect observed in Table 2. It should be emphasized that iron isotope exchange has a strong effect on spin-density-wave state. Substitution of Fe for Fe leads to a remarkable decrease in resistivity below the SDW ordering temperatures with a large α SDW for the two systems. It suggests a strong magnon-phonon coupling. A giant oxygen isotope effect has been observed in magnetoresistive La Ca x MnO , and the isotope exponent α FM for ferromagnetic transition is as high as 0.85 . Such a large isotope shift is elieved to arise from coupling of the charge carriers to Jahn-Teller lattice distortions . In pnictide superconductors, the strong sensitivity of superconductivity and magnetism to the lattice may be responsible for the large isotope effect. These results definitely indicate that electron-phonon interaction plays an important role for the superconducting mechanism, but the strong magnon-phonon coupling has to be considered. The iron isotope effect on T SDW and T C is much larger than the oxygen isotope effect in pnictide superconductors. The reason could be that the iron-arsenide plane is conducting layer and responsible for the superconductivity, and SDW ordering originates from Fe moment. For MgB superconductor, no magnetic correlation is included and the superconductivity can be understood with BCS theory with α c =0.32 . In cuprates, the isotope effect on T C is sensitive to doping level. The effect is vanishing at optimum doping, but increases systematically with decreasing doping level to be maximum at the border to the antiferromagnetic state . It seems that the isotope effect is somewhat related to magnetic fluctuation. Such unconventional isotope effects demonstrate that the electron–phonon interaction also plays an important role in the physics of cuprates. Sorting out the interplay between the lattice and magnetic degrees of freedom is a key challenge for understanding the mechanism of high-T C superconductivity. . Kamihara, Y. et al. Iron-based layered superconductor LaO F x FeAs (x=0.05–0.12) with T C = 26 K. J. Am. Chem. Soc. , 3296–3297(2008). 2. Chen, X. H. et al. Superconductivity at 43 K in SmFeAsO F x . Nature , 761-762(2008). 3. Chen, G. F. et al. Superconductivity at 41 K and its competition with spin-density-wave instability in layered CeO F x FeAs.
Phys. Rev. Lett. , 247002(2008). 4. Ren, Z. A. et al. Superconductivity in iron-based F-doped layered quaternary compound NdO F x FeAs.
Europhys. Lett . , 57002(2008). 5. Liu, R. H. et al. Anomalous transport properties and phase diagram of the FeAs-based SmFeAsO F x superconductors. Phys. Rev. Lett. , 087001(2008). 6. Boeri, L. et al. Is LaFeAsO F x an electron-phonon superconductor? Phys. Rev. Lett . , 026403(2008). 7. Cao, C. et al. Proximity of antiferromagnetism and superconductivity in LaFeAsO − x F x : Effective Hamiltonian from ab initio studies. Phys. Rev. B , 220506(R)(2008). 8. Dai, X. et al. Even parity, orbital singlet, and spin triplet pairing for superconducting LaFeAsO F x . Phys. Rev. Lett. , 057008(2008). 9. Ma, F. et al. Iron-based layered compound LaFeAsO is an antiferromagnetic semimetal.
Phys. Rev. B , 033111(2008). 10. Ewings, R. A. et al. High energy spin excitations in BaFe As . Phys. Rev. B , 220501(R) (2008). 11. Christianson, A. D. et al. Resonant spin excitation in the high temperature superconductor Ba K Fe As . Nature , 930-932(2008). 12. Zbiri, M. et al. Combined ab initio lattice dynamics simulations and improved inelastic eutron scattering spectra for studying phonons in BaFe As : Effect of structural phase transition, structural relaxation and magnetic ordering. arXiv:0807.4429(2008). 13. Fang, C. et al., Theory of electron nematic order in LaOFeAs. Phys. Rev. B , 224509(2008).. 14. Margadonna, S. et al., Crystal structure and phase transitions across the metal-superconductor boundary in the SmFeAsO F x (0 ≤ x ≤ Phys. Rev. B , 014503(2009). 15. Zhao, J. et al. Structural and magnetic phase diagram of CeFeAsO F x and its relation to high-temperature superconductivity. Nature Materials , 953-959(2008) 16. Drew, A. J. et al., Coexistence of static magnetism and superconductivity in SmFeAsO F x as revealed by muon spin rotation. arXiv:0807.4876(2008) Nature Materials (in press) (2009). 17. Chen, H et al., Coexistence of the spin-density-wave and superconductivity in the (Ba,K)Fe As . Europhys. Lett. , 17006(2009). 18. Zhao, S. C. et al., Raman spectra in iron-based quaternary CeO F x FeAs and LaO F x FeAs,
Supercond. Sci. Technol. , 015017(2009). 19. Cruz, C. et al., Magnetic order close to superconductivity in the iron-based layered LaO F x FeAs systems.
Nature , 899-902(2008). 20. Huang, Q. et al. Magnetic order in BaFe As , the parent compound of the FeAs based superconductors in a new structural family. Phys. Rev. Lett. , Ca x MnO , Nature ,676-678(1996). 22. Hinks, D. G., Claus, H., and Jorgensen, J. D., The complex nature of superconductivity in gB as revealed by the reduced total isotope effect. Nature , 457-460(2001). 23. Franck, J.P.: In: Ginsberg, D.M. (ed.) Physical Properties of High Temperature Superconductors, p. 189.World Scientific, Singapore (1994) 24. Keller, H.: In: Müller, K.A., Bussmann-Holder, A. (eds.) Structure and Bonding, vol. 114, p. 143. Springer, Berlin (2005).
Supplementary Information
Acknowledgments:
This work is supported by the Nature Science Foundation of China, and by the Ministry of Science and Technology of China and Chinese Academy of Sciences. We acknowledge Prof. Z. X. Shen for helpful discussion and encouragement, and Profs. D. L Feng and S. Y. Li for discussions.
Author Contributions:
X.H.C. designed and coordinated the whole experiment, and analysed the data and wrote the paper. R.H.L. and T.W. did the main experiments including sample preparation and analysed the data. G.W., X.F.W. and B.C.S synthesized the samples. H.C. and Y.L.X. partially measured the resistivity. J.J.Y. measured the susceptibility. Y.J.Y. and Q.J.L. did X-ray powder diffraction measurements. W.S.C. and Z.Y.W. provided the iron isotope Fe.
Author information:
The authors declare no competing financial interests. Correspondence and requests for materials should be addressed to X. H. Chen ([email protected]). Table 1:
Raman shifts of the E g mode of oxygen ( ω O ) and the B mode of iron ( ω Fe ) in the samples SmFeAsO F x (x=0, 0.15) with natural abundance oxygen n O and O, and in the samples SmFeAsO F x (x=0, 0.15) and Ba K x Fe As (x=0, 0.4) with Fe and Fe, respectively. O and Fe isotopic contents are deduced on the Raman shift being proportional to 1-(16/M') for the substitution of O for n O and proportional to 1-(55.93/M') for the substitution of Fe for Fe, respectively . In the calculation the molar mass for isotope exchange used is 53.93 g for Fe and 55.93 g for Fe, and 16 g for n O and 18 g for O. ω O (cm -1 ) n O O △ ω O O% SmFeAsO 418.6 401.2 -17.4 71% SmFeAsO F ω Fe (cm -1 ) Fe Fe △ ω Fe Fe%
SmFeAsO 208.0 211.5 3.5 92% SmFeAsO F As K Fe As Spin-density-wave and superconducting transition temperature (T
SDW and T C ), and their shifts ( △ T SDW and △ T C ) and isotope exponents ( α SDW and α C ) for the samples of SmFeAsO F x with natural abundance oxygen n O and O, and Fe and Fe, and for the samples Ba K x Fe As with Fe and Fe, respectively.
Isotope exponent is deduced by α =-d lnT /d lnM , where M is the atomic mass. To guarantee that one measures the intrinsic isotope effect, isotope back-exchange is usually used: the transition should come back to that before isotope exchange. In our case there is no way to carry out the isotope back-exchange experiment. To get intrinsic isotope effect, we prepared several batches of samples to check the results reproducible from sample to sample. Although T C slightly varies from batch to batch for SmFeAsO F , the shift of T C caused by isotope exchange is nearly the same for all batches. T C of the sample Ba K Fe As is very stable and nearly independent of batches. In the calculation the molar masses used are the same as used in Table 1 except for O which is reduced to 17.42 g and 17.54 g because of only 71% and 77% exchange. The error bar for the temperature determination is equal to the step temperature in the temperature sweep that is 0.02 K for T C and 0.05 K for T SDW in the resistivity measurements, respectively. T SDW n O O △ T SDW α SDW
SmFeAsO 130.00(5) 130.80(5) 0.80(10) -0.07(1) 130.10(5) 130.60(5) T SDW 56 Fe Fe △ T SDW α SDW
SmFeAsO 130.00(5) 131.90(5) 1.90(10) 0.41(2) 130.10(5) 131.80(5) 1.70(10) 0.37(2) Average 0.39(2) BaFe As T C n O O △ T C α C SmFeAsO F T C 56 Fe Fe △ T C α C SmFeAsO F K Fe As Figure legends:
Figure 1:
Raman spectra at room temperature for the samples SmFeAsO F x (x=0 and 0.15) and Ba K x Fe As (x=0 and 0.4) . SmFeAsO F x (x=0 and 0.15), (a): with n O and O, (b): with Fe and Fe; (c): Ba K x Fe As (x=0 and 0.4) with Fe and Fe. The mode around 420 cm -1 is ascribed to the E g mode of oxygen and the mode around 210 cm -1 is assigned to the B mode of Fe . Figure 2: Temperature dependence of resistivity ( ρ ) and its derivative (d ρ /dT) for the samples SmFeAsO F x with n O and O. The peak in the derivative is considered as the transition temperature. The peak temperature in d ρ /dT corresponds to mid-transition temperature for superconducting transition. An anomalous peak at about 147 K in resistivity is associated with the structural transition for the x=0 sample, and the peak temperature in d ρ /dT is very close to the SDW ordering temperature observed by neutron scattering . Figure 3: Temperature dependence of resistivity ( ρ ) and its derivative (d ρ /dT) for the samples SmFeAsO F x and Ba K x Fe As isotopically substituted with Fe and Fe. (a): SmFeAsO F x with x=0; (b): SmFeAsO F x with x=0.15; (c): Ba K x Fe As with x=0; (d): Ba K x Fe As with x=0.4. In order to accurately determine the transition temperature, we took the derivative of the resistivity, and the peak in the derivative is considered as the transition temperature. The peak temperature in d ρ /dT for the SDW is very close to the ordering temperature bserved by neutron scattering . The α C and α SDW are deduced to be 0.34(3) and 0.37(2) for the samples SmFeAsO F x with x=0.15 and 0, respectively. For a typical sample of BaFe As , the resistivity anomaly around 145 K arises from the structural and SDW transition . For a typical superconducting sample of Ba K x Fe As with x=0.4, the onset transition occurs around 39 K. The α C and α SDW are found to be 0.38(3) and 0.39(2) for the samples Ba K x Fe As with x=0.4 and 0, respectively. To confirm the results from resistivity measurements, the susceptibility for the same samples is measured (see Supplementary Figure S2). The shift of T C and T SDW determined from susceptibility measurements due to iron isotope exchange is consistent with that obtained from the resistivity measurements. It should be pointed out that a remarkable difference in resistivity is observed below SDW transition between the x=0 samples with Fe and Fe for the SmFeAsO F x and Ba K x Fe As systems. As shown in the inset of Fig.3a and 3c, the resistivity decreases by about 28% for SmFeAsO sample and by about 35% for BaFe As sample at 10 K due to iron isotope exchange. It suggests that iron isotope exchange has strong effect on spin-density-wave state. upplementary Information The isotopically substituted polycrystalline samples with nominal compositions SmFeAsO F x (x=0, 0.15) were synthesized by conventional solid state reaction using SmAs, SmF , As, Fe and Fe O /Fe O as starting materials for oxygen isotope exchange, while using SmAs, SmF , As, Fe/ Fe and Fe O / Fe O as starting materials for iron isotope exchange. Fe O and Fe O were obtained by sintering Fe powder in natural abundance oxygen n O and O with an 80% enrichment, respectively. The Fe O and Fe O were prepared by sintering Fe and Fe powder in flowing natural abundance oxygen atmosphere, respectively. The iron isotope polycrystalline samples of Ba K x Fe As (x=0, 0.4) were synthesized using BaAs, KAs, and Fe As and Fe As as starting materials, respectively. Fe As and Fe As were synthesized by reacting Fe and Fe powder with As powder in evacuated quartz tubes at 650 o C for 24 h, respectively. To keep the same condition, the isotope exchange samples with the same composition were sealed in the same evacuated quartz tube for annealing. To mitigate the difference of F content in isotope exchange samples SmFeAsO F , SmAs, SmF and As were first mixed according to stoichiometric ratio and grounded, then mixture of SmAs, SmF and As was equally separated into two parts. Finally, Fe/ Fe and Fe O / Fe O as starting materials were weighed and put into the separated mixture of SmAs, SmF and As for iron isotope exchange. It guarantees the same F content for isotope exchange in the beginning of sample process. They were loaded into the same quartz tube for annealing. Natural abundance iron was used for the oxygen isotope exchange experiments and obtained from lfa Aesar, while the Fe and Fe enriched isotopes from ISOFLEX (San Francisco, USA) were used for the iron isotope experiments. The purity of the natural abundance iron n Fe is 99.9%, while the purity of the Fe (enrichment: 97%+) is 99.78%, and the purity of the Fe (enrichment: 96%+) is 99.86%. Natural abundance oxygen was used for the iron isotope exchange experiments for the system SmFeAsO F x . The purity of the natural abundance oxygen n O is 99.99%. The purity of the O (enrichment: 80%) is 99.99%. Raman spectra were obtained on a LABRAM-HR Confocal Laser MicroRaman Spectrometer using the 514.5 nm line from an argon-ion laser with in-plane light polarization. X-ray diffraction (XRD) was performed by MAC MXPAHF X-Ray diffractometer with graphite monochromated CuK α radiation ( λ =0.15406 nm) at room temperature. The superconducting transition temperatures and the SDW transition temperatures were determined by standard four-probe resistance and susceptibility measurements. The resistance was measured by an AC resistance bridge (LR-700, Linear Research). Magnetic susceptibility measurements were performed with a superconducting quantum interference device magnetometer (Quantum Design MPMS-7). Figure S1:
X-ray diffraction patterns at room temperature for the samples SmFeAsO F x and Ba K x Ba As with iron isotope exchange. As shown in the Figure, XRD patterns look like the same for isotope exchange. The XRD patterns are fitted by Rietveld analysis. The obtained lattice constants are: a=0.3961(1) nm, c=1.3015(4) nm for BaFe As with Fe, a=0.3962(1) nm, c=1.3017(4) nm for BaFe As with Fe; a=0.3918(2) nm, c=1.3289(7) nm for Ba K Fe As with Fe, a=0.3917(2) nm, c=1.3288(7) nm for Ba K Fe As with Fe; a=0.39415(6) nm, c=0.8502(3) nm for SmFeAsO with Fe, a=0.39410(6) nm, c=0.8501(3) nm for SmFeAsO with Fe; a=0.39320(5) nm, c=0.8490(2) nm for SmFeAsO F with Fe, a=0.39326(5) nm, c=0.8491(2) nm for SmFeAs O F with Fe, respectively. The lattice constants are the same within experimental error. Figure S2:
Temperature dependence of susceptibility for the typical samples of SmFeAsO F and Ba K x Fe As isotopically substituted with Fe and Fe. (a): SmFeAsO F ; (b): Ba K x Fe As with x=0; (c): Ba K x Fe As with x=0.4. To confirm the results from resistivity measurements, the susceptibility of the same samples used in Figure 3 is measured. The shift of T C and T SDW determined from susceptibility measurements is consistent with that obtained from the resistivity measurements as shown in Figure 3. The susceptibility of superconducting samples SmFeAsO F and Ba K Fe As is measured in the zero-field cooling process under the magnetic field of 10 Oe. The susceptibility of parent compound BaFe As is measured in the field-cooled process under the magnetic field of 50000 Oe. Figure S3:
Temperature dependence of resistivity ( ρ ) and its derivative (d ρ /dT) for the samples Ba K x Fe As (x=0 and 0.4) with nature abundance iron n Fe and Fe .
The natural abundance iron n Fe with a 91.8% enrichment of Fe and purity of 99.9%; the enrich iron Fe with a 97%+ enrichment and purity of 99.78%. It clearly shows that no apparent change in T
SDW and T C is observed in the samples with nature abundance iron n Fe and enriched iron Fe. Figure S4: