Electronic Evidence of Unusual Magnetic Ordering in a Parent Compound of FeAs-Based Superconductors
High resolution angle-resolved photoemission measurements have been carried out on BaFeAs, a parent compound of the FeAs-based superconductors. In the magnetic ordering state, there is no gap opening observed on the Fermi surface. Instead, dramatic band structure reorganization occurs across the magnetic transition. The appearance of the singular Fermi spots near (,) is the most prominent signature of magnetic ordering. These observations provide direct evidence that the magnetic ordering state of BaFeAs is distinct from the conventional spin-density-wave state. They reflect the electronic complexity in this multiple-orbital system and necessity in involving the local magnetic moment in describing the underlying electron structure.
pacs:74.70.-b, 74.25.Jb, 79.60.-i, 71.20.-b
The interplay between magnetism and superconductivity is a prominent topic in the newly-discovered FeAs-based superconductorsKamihara (); ZARenSm (); RotterSC (). The ground state of the parent compound of the FeAs-based superconductors has a collinear anti-ferromagnetic orderingDongSDW (); PCDai (); BFSNeutron3 (). Superconductivity is achieved from suppressing the magnetic ordering by chemical dopingKamihara (); ZARenSm (); RotterSC () or application of high pressurePressureTc (). Magnetism has been proposed to play an important role in generating high temperature superconductivity in this new class of high temperature superconductorsFeSCMagnetic (); Triplet (). However, the nature of the magnetic ordering and its origin remain under hot debateDongSDW (); J1J2Picture (). It is proposed that the magnetic ordering is associated with the spin-density-wave (SDW) formation driven by the Fermi surface nestingDongSDW (). Recent opticalWZHuOpt () and quantum oscillationOscillations () measurements reported gap opening related with the magnetic transition which is in favor of the SDW picture. It becomes critical to pin down whether there is really a gap opening and, if there is, where the gap opens in the momentum space. In principle, angle-resolved photoemission (ARPES) measurements can provide direct information on these issues but so far the ARPES results remain inconclusiveDLFBFe2As2 (); CLiuBFeAs (); HYLiu (); LZhao (); HDing (); Borisenko (); DLFSFA (); ZHasanSFS (); MYi ().
In this paper we report angle-resolved photoemission evidence to show the unusual nature of magnetic ordering in a parent compound of FeAs-based superconductors, BaFeAsRotterParent (). Unlike the conventional SDW systemCrSDW () where magnetic ordering is associated with Fermi surface nesting and a gap openingCrARPES (), the electronic structure of BaFeAs shows no sign of gap opening on the Fermi surface in the magnetically-ordered state. Instead, it shows a dramatic band reorganization and particularly an emergence of singular strong Fermi spots. These observations provide direct and clear-cut evidences that the magnetic ordering state of BaFeAs is distinct from the conventional SDW state. They also point to the importance and necessity in involving the local magnetic moment in the underlying electron structure. These results and their understanding will have important implications on understanding the magnetism and superconductivity in this new class of superconductors.
The ARPES measurements are carried out on our photoemission system equipped with Scienta R4000 electron energy analyzer and Helium discharge lamp which gives a photon energy of h= 21.218 eVGDLiu (). The light on the sample is partially polarized. The energy resolution was set at 10 meV and the angular resolution is 0.3 degree. The Fermi level is referenced by measuring on the Fermi edge of a clean polycrystalline gold that is electrically connected to the sample. The BaFeAs single crystals were grown using flux methodWZHuOpt () which show a magneticstructural transition at T138 KRotterParent (); WZHuOpt (). The crystals were cleaved in situ and measured in vacuum with a base pressure better than 510 Torr. For ARPES measurements, we found that the cleaved surface of BaFeAs shows a change with time. To reduce the influence of this aging effect, we took many independent measurements on fresh surfaces. All the Fermi surface mapping measurements reported here were taken within four hours after cleaving (Fig. 1), and for the typical momentum cuts (Fig. 2), the measurements were carried out within two hours after cleaving.
In the magnetic ordering state, the Fermi surface of BaFeAs near the point shows three hole-like sheets (Fig. 1a), as evidenced from Fig. 2a and Fig. 2d where three Fermi crossings are clearly resolved in the bands taken along the M1(,) direction (Cut 1 in Fig. 1a. For the convenience of description, we denote the four M points as M1(,), M2(-,), M3(-,-) and M4(,-)). The corresponding Fermi momenta are 0.06, 0.20 and 0.35/a, respectively. Additional measurements along X(,0) and other directions also give three Fermi crossings with similar Fermi momenta.
The most prominent features in the Fermi surface topology of BaFeAs below T are the strong Fermi spots near M points (Fig. 1a), labeled as S1 and S2 near the M2(-,) point, and S3 and S4 near the M3(-,-) point. In these two independent measurements near M2 and M3 (Fig.1a) under different geometries, the strong spots are only obvious along the M direction, with a distance between them (S1 and S2, or S3 and S4) at 0.65 /a. To check whether this could be affected by the matrix element effect associated with the particular electric vector, we took another measurement with a different polarization geometry and observed similar pattern (Fig. 1b), suggesting a weak polarization dependence. These strong spots disappear above the magnetic transition temperature, T, (Fig. 1c), and therefore are intimately related with the magnetic ordering state.
In addition to the strong spots, there are also intensity patches on both sides of the -M2(M3) line(s) (Fig. 1a). Band structure measurements (Fig. 2c and 2f for the Cut 3 in Fig. 1a) indicate they originate from the electronlike bands. The Fermi surface topology near the M point below T is therefore composed of two electronlike Fermi surface sheets and two strong Fermi spots, as schematically shown in the bottom-right quadrant of Fig. 1a. Note that the strong Fermi spots are not part of the electronlike Fermi surface sheets because they originate from different bands (Fig. 2).
It is wellknown that the conventional SDW formation like in CrCrSDW () is associated with a gap opening along the nested section of Fermi surfaceCrARPES (). It remains under debate whether similar mechanism operates in the magnetic ordering state in the parent compound of FeAs-based superconductorsDongSDW (); J1J2Picture (). The detailed Fermi surface and band structure results allow us to directly examine on this issue. Considering the nesting vector in BaFeAs close to (0,0)-(,)DongSDW (), the most likely locations to search for possible gap opening are close to the Fermi surface sections near the -M line (Cut 2 and its equivalent cuts in Fig. 1a). It is clear from Fig. 2g that, near the point, all photoemission spectra at the Fermi momenta cross the Fermi level, showing no gap opening along -M direction. Measurements along other sections of three Fermi surface sheets near see no signature of gap opening either. Furthermore, there is no gap opening seen for the strong Fermi spots(Fig. 2h) and for the two electronlike Fermi surface sheets(Fig. 2i). These results provide a direct evidence that the magnetic ordering state of BaFeAs does not involve gap opening which is fundamentally different from the conventional SDW systemCrSDW (); CrARPES ().
The band structure of BaFeAs experiences a dramatic reorganization across T=138 K near the M point, in terms of both the number of bands and their energy position(Fig. 3). Above T, only one major branch of band is resolved (Fig. 3a,b,e and f). As plotted in a more quantitative manner in Fig. 3j, it appears to be composed of a holelike band () and an electronlike band (). Below T, four bands emerge with inverseparabolic shape (Fig. 3c,d,g,h and Fig. 3j): two of them ( and in Fig. 3j) crossing the Fermi level while the other two below the Fermi level ( and in Fig. 3j). Such a band structure change is remarkable considering its occurrence within a narrow temperature window between 140 K (Fig. 3b and f) and 130 K (Fig. 3c and g). This also indicates that ARPES signal is sensitive to the magnetic transition at a temperature that is consistent with the bulk BaFeAs material.
The striking difference of the band structure of BaFeAs above T (Fig. 4a) and below T (Fig. 4b) clearly indicates that the magnetic ordering state does not inherit its electronic structure simply from its high-temperature non-magnetic counterpart. A natural question is whether such a strong band reorganization can be attributed to the band-folding effect associated with the magneticstructural transition. Such a transition renders the initial and M points equivalent in the new shrunk Brillouin zone causing the bands at and M to fold to each other. In principle, such a band folding effect is always present regardless of the origin of the magnetic ordering. In real experiment, whether the folded bands can be resolved depends on the interaction potential and matrix element effect: the intensity of the folded bands can be much weaker than the primary onesVoit (); Brouet (). Such a folding effect was previously proposed to account for the magnetic transition in SrFeAsZHasanSFS (). But our observations indicate it is not the main contributor to the dramatic band reorganization (Fig. 4). The two hole-like bands below the Fermi level near M ( and in Fig. 4b) definitely can not be due to the band folding effect because there are no corresponding bands near . For the same reason, the inner-most band near ( in Fig. 4b) can not be due to such a band folding from M either. As for the two hole-like bands crossing the Fermi level near M ( and in Fig. 4b), they appear similar to the bands near and one may wonder whether they may originate from the folding effect. However, from the bands in Fig. 3j where the bands near and above E are clearly visible (Fig. 3), the two hole-like bands cross each other near E. This observation makes them unlikely to originate from the two bands near which are separate. The band folding is also expected to produce similar Fermi surface topology near and M. While the strong Fermi spots are obvious near M, no such replica can be identified near point(Fig. 1a). This further indicates that the folding effect in BaFeAs is weak although in principle it is present.
The strong Fermi spots are most prominent electronic signature of magnetic ordering in BaFeAs as they only appear below T (Fig. 1a and c). One natural way to form such a singular Fermi spots is to have two bands crossing near the Fermi level giving rise to a Dirac cone-like structure. In FeAs-based compounds with multiple bands, below T, the folding between the hole-like bands near and the electron-like bands near M can give rise to such Dirac nodes along -M direction which are gaplessYRan (). These characteristics seem to be in qualitative agreement with our observations and similar folding picture was also used for a previous ARPES experimentZHasanSFS (). However, some of our experimental observations are at variance with this picture. In the folding picture, the Dirac nodes are produced from one hole-like band folded from to M and one electron-like band near MYRan (); ZHasanSFS (). However, our results indicate that the spectral weight of the strong spots are mainly from the hole-like bands (Fig. 2b and 2e). Fig. 3j further suggests that the strong spots are more likely formed by the two hole-like bands ( and ) crossing each other near the Fermi level. The contribution to the strong spots formation from the electron-like bands is weak because no trace of electron-like bands is resolved along the -M direction(Figs. 2b, 2e and Fig. 3) probably due to matrix element effect.
We note that the strong Fermi spots we have observed (Fig. 1a) show some resemblance to the “propeller-shaped” pattern reported before for a given polarization in both BaFeAs and (BaK)FeAs superconductor which was attributed to “hidden” (,) excitations that exist all the way to room temperatureBorisenko (). Our results clearly indicate that the strong Fermi spots can not originate from the same “hidden” (,) order because they appear only below T (Fig. 1 and Fig. 3). We also observed strong Fermi spots near M in (BaK)FeAs superconductorLZhao (). In terms of the Fermi surface topology, our observation of strong Fermi spots in both magneticallyordered BaFeAs and superconducting (BaK)FeAs seems to be consistent with the previous experiment showing similar patternsBorisenko (). However, the underlying band structure we have revealed between them is remarkably differentLZhao (). This indicates that the seemingly similar strong Fermi spots in the parent (Fig. 1) and K-doped BaFeAsLZhao () are most likely from different origins, rather than from the same (,) excitations proposed beforeBorisenko ().
The dramatic band reorganization across T and the lack of gap opening provide clear-cut evidences that the magnetic ordering state of BaFeAs is distinct from the usual SDW stateCrSDW (); CrARPES (). These observations may provide a new approach to understand previous measurementsOscillations (); WZHuOpt (). For example, the multiple bands developed below T (Fig. 4b) may provide new channels for inter-band transitions, offering an alternative interpretation of gap opening in terms of SDW picture that was proposed in the optical spectroscopyWZHuOpt (). Our results show that the multiple-band structure (Fig. 4b) and the appearance of strong Fermi spots (Fig. 1a) below T are neither inherited from its high temperature non-magnetic state, nor from the band folding effect associated with the magneticstructural transition. These indicate that they are electronic fingerprints inherent to the magnetic ordered state. These characteristics reflect the electronic complexity in this multiple-orbital system and point to the importance and necessity in involving the local magnetic moment in the underlying electron structure. How to describe the peculiar electronic features of the magnetically ordered state on a quantitative level presents a challenge to further theoretical work. Our observations and their understanding will shed new light on understanding the magnetism and high temperature superconductivity in FeAs-based superconductors.
We thank A. V. Balatsky, J. P. Hu, D. H. Lee, Z. Y. Lu, J. L. Luo, E. W. Plummer, N. L. Wang and T. Xiang for helpful discussions. This work is supported by the NSFC, the MOST of China (973 project No: 2006CB601002, 2006CB921302).
Corresponding author: XJZhou@aphy.iphy.ac.cn
- (1) Y. Kamihara et al., J. Am. Chem. Soc. 130, 3296 (2008).
- (2) Z. A. Ren et al., Chin. Phys. Lett. 25, 2215 (2008).
- (3) M. Rotter et al., Phys. Rev. Lett. 101, 107006(2008).
- (4) J. Dong et al., Europhys. Lett. 83, 27006 (2008).
- (5) C. de la Cruz et al., Nature (London) 453, 899(2008).
- (6) Q. Huang et al., Phys. Rev. Lett. 101, 257003 (2008).
- (7) M. S. Torikachvili et al., Phys. Rev. Lett. 101, 057006 (2008); T. Park et al., J. Phys.: Condens. Matter 20, 322204 (2008); P. L. Alireza et al., J. Phys.: Condens. Matter 21, 012208 (2009).
- (8) I. I. Mazin et al., Phy. Rev. Lett. 101, 057003(2008); K. Kuroki et al., Phys. Rev. Lett. 101, 087004 (2008); F. Wang et al., Phys. Rev. Lett. 102, 047005(2009); A. V. Chubukov et al., Phys. Rev. B 78, 134512(2008); V. Stanev et al., Phys. Rev. B 78, 184509(2008); F. Wang et al., Europhys. Lett. 85, 37005 (2009); I. I.Mazin and M. D. Johannes, Nature Phys. 5, 141(2009).
- (9) X. Dai et al., Phys. Rev. Lett. 101, 057008(2008); P. A. Lee and X. G. Wen, Phys. Rev. B 78, 144517 (2008).
- (10) T. Yildirim, Phys. Rev. Lett. 101, 057010(2008); Q. M. Si and E. Abrahams, Phys. Rev. Lett. 101, 076401(2008); J. S. Wu et al., Phys. Rev. Lett. 101, 126401(2008); C. Fang et al., Phys. Rev. B 77, 224509 (2008); C. K. Xu et al., Phys. Rev. B 78, 020501(R) (2008); S. P. Kou et al., arXiv:cond-mat/0811.4111.
- (11) W. Z. Hu et al., Phys. Rev. Lett. 101, 257005(2008).
- (12) S. E Sebastian1 et al., J. Phys.: Condens. Matter 20, 422203 (2008).
- (13) L. X. Yang et al., arXiv:cond-mat/0806.2627.
- (14) C. Liu et al., Phys. Rev. Lett. 101, 177005 (2008).
- (15) H. Y. Liu et al., Phys. Rev. B 78, 184514(2008).
- (16) L. Zhao et al., Chin. Phys. Lett. 25, 4402(2008).
- (17) H. Ding et al., Europhys. Lett. 83, 47001 (2008).
- (18) V. B. Zabolotnyy et al., Nature (London) 457, 569 (2009).
- (19) Y. Zhang et al., arXiv:cod-mat/0808.2738.
- (20) D. Hsieh et al., arXiv:cond-mat/0812.2289.
- (21) M. Yi et al., arXiv:cond-mat/0902.2628.
- (22) M. Rotter et al., Phys. Rev. B 78, 020503(R)(2008).
- (23) E. Fawcett, Rev. Mod. Phys. 60, 209 (1988).
- (24) J. Schaefer et al., Phys. Rev. Lett 83, 2069 (1999).
- (25) G. D Liu et al., Rev. Sci. Instruments 79, 023105 (2008).
- (26) J. Voit et al., Science 290, 501 (2000).
- (27) V. Brouet et al., Phys. Rev. B 77, 235104 (2008).
- (28) Y. Ran et al., Phys. Rev. B 79, 014505 (2009).
- (29) I.A. Nekrasov et al., JETP Lett., 88, 144(2008).
- (30) F. J. Ma et al., arXiv:cond-mat/0806.3526V2.
- (31) D. J. Singh, Phys. Rev. B 78, 094511 (2008).