Incommensurate spin correlations induced by magnetic Fe ions substituted into overdoped Bi1.75Pb0.35Sr1.90CuO6+z
H. Hiraka, Y. Hayashi, S. Wakimoto, M. Takeda, K. Kakurai, T. Adachi, Y. Koike, I. Yamada, M. Miyazaki, M. Hiraishi, S. Takeshita, A. Kohda, R. Kadono, J.M. Tranquada, K. Yamada
aa r X i v : . [ c ond - m a t . s up r- c on ] M a r Incommensurate spin correlations induced by magnetic Fe ions substituted intooverdoped Bi . Pb . Sr . CuO z H. Hiraka , Y. Hayashi , S. Wakimoto , M. Takeda , K. Kakurai , T. Adachi , Y. Koike , I. Yamada ,M. Miyazaki , M. Hiraishi , S. Takeshita , A. Kohda , , R. Kadono , , J. M. Tranquada , K. Yamada Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan Quantum Beam Science Directorate, Japan Atomic Energy Agency, Tokai, Ibaraki 319-1195, Japan Department of Applied Physics, Tohoku University, Sendai 980-8578, Japan Graduate School of Science and Engineering, Ehime University, Matsuyama 790-8577, Japan Department of Materials Structure Science, The Graduate University for Advanced Studies, Tsukuba, Ibaraki 305-0801, Japan Institute for Materials Structure Science, High Energy Accelerator Research Organization, Tsukuba, Ibaraki 3005-0801, Japan Condensed Matter Physics & Materials Science Department,Brookhaven National Laboratory, Upton, New York 11973-5000, USA and WPI Research Center, Advanced Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan (Dated: July 2, 2018)Spin correlations in the overdoped region of Bi . Pb . Sr . CuO z have been explored with Fe-doped single crystals characterized by neutron scattering, muon-spin-rotation ( µ SR) spectroscopy,and magnetic susceptibility measurements. Static incommensurate spin correlations induced by theFe spins are revealed by elastic neutron scattering. The resultant incommensurability δ is unexpect-edly large ( ∼ δ ∼ / − x Sr x CuO .Intriguingly, the large δ in this overdoped region is close to the hole concentration p . This result isreminiscent of the δ ≈ p trend observed in underdoped La − x Sr x CuO ; however, it is inconsistentwith the saturation of δ in the latter compound in the overdoped regime. While our findings in Fe-doped Bi . Pb . Sr . CuO z support the commonality of incommensurate spin correlations inhigh- T c cuprate superconductors, they also suggest that the magnetic response might be dominatedby a distinct mechanism in the overdoped region. I. INTRODUCTION
Neutron scattering studies have provided valuablecharacterizations of the momentum and energy depen-dences of spin correlations in high- T c cuprate super-conductors. Incommensurate spin correlations (ISCs),which had been discovered in the early stage of high- T c research, are one of the salient features of hole-carrier doped cuprates. The energy evolution of the ISCsin La − x Sr x CuO (La214), YBa Cu O y (Y123), and Bi Sr CaCu O δ (Bi2212), shows a commonalityof magnetic excitations in the form of an hourglass-likedispersion. In the underdoped region, the incommensura-bility δ of static, as well as low-energy, ISCs is close to thehole concentration p in La214 and Y123. As spinstripes with δ ≃ / x = 0 . , the linear relationship δ = p strongly suggests the presence of stripe correlations inthe CuO planes. Another linearity between the onset T c and δ , T onc = δ , as well as the development of thespin-gap below T c14,15 evinces an important impact ofstripe correlations on the superconductivity.In contrast, in the overdoped region, observation ofISCs becomes rather difficult because of peak broaden-ing and doping-induced suppression of magnetic inten-sity, both of which probably result from a change in thedegree of itinerancy of the electrons. Only slowly fluctu-ating ISCs are observable and δ remains nearly constantat around 1/8 in overdoped La214. However, be-cause of the difficulty involved in synthesizing large sin-gle crystals, systematic neutron-scattering data on ISCs in the overdoped region have been restricted to the La214system.Recently, we succeeded in growing sizable single crys-tals of Bi . Pb . Sr . CuO z [(Bi,Pb)2201]. Withthese crystals, the overdoped region is easily accessibleover a wide range of p . Here we report a search for spincorrelations in overdoped (Bi,Pb)2201, where no well-defined spin correlations have been observed to date. In addition to the pristine sample, we also study sam-ples with varying amounts of Fe substituted for Cu, asa strong effect of these magnetic dopants on spin corre-lations is expected, based on recent results in the La214system.
We find that quasistatic ISCs are induced bythe large localized spins of the Fe dopants. This obser-vation for Bi-based cuprates supports an intrinsic natureof dynamic ISCs in high- T c cuprates. The obtained δ of ∼ . T c cuprate, falls on the line δ ≈ p , in contrast tooverdoped La214. Differences in the nature of the spincorrelations compared to the underdoped regime seemlikely.The rest of the paper is organized as follows. In thenext section, we describe the crystal growth and charac-terization by magnetic susceptibility, zero-field µ SR, andelastic neutron scattering; the experimental results arealso presented. In Sec. III, the implications for the natureof the magnetic correlations in overdoped (Bi,Pb)2201are discussed. The paper is summarized in Sec. IV.
II. EXPERIMENTS AND RESULTSA. Crystal growth and characterization
Single crystals of Bi . Pb . Sr . Cu − y Fe y O z ( y = 0 , . , . , .
09, and 0.13) were grown in air by thetraveling-solvent floating-zone technique. The amountsof metal elements determined by ICP spectrometry werein agreement with the nominal values to at least 95%accuracy, except for y = 0 . . .
2. The crystal structure analyzed by X-ray-powder diffraction at room temperature was consistentwith the
Pnan orthorhombic symmetry for all samples,and no significant peak broadening was observed even for y = 0 .
13. The c lattice parameter decreased rapidly uponFe doping, and the rate of decrease follows that observedfor the crystalline Bi . Pb . Sr Cu − y Fe y O z , thusevincing the Fe-atom doping from a structural viewpoint.We also confirmed with neutron diffraction that incom-mensurate structural modulations along the b axis char-acteristic to Bi-based cuprates are absent in our Pb . -substitution samples, in accordance with previous STMstudies. T onc was determined from magnetic shielding measure-ments. The pristine samples show superconductivity be-low T onc = 6 K and 23 K for as-grown and Ar-annealed(600 ◦ C × B. Magnetic susceptibility
Figure 1(a) shows χ ab ( T ) of the series of Fe-doped sam-ples, in which a magnetic field of H = 1 T was applied ina direction parallel to the CuO planes. The Curie-Weisslaw for high-temperature susceptibility was used to cal-culate the Curie constant C and the Weiss temperature( | Θ | ∼
10 K). As can be seen in the inset of Fig. 1(a), C linearly increases up to y = 0 .
09. The effective numberof Bohr magnetons p eff was found to be 4.4 using theequation C = ( N µ / k B ) p , by assuming that only theFe spins contribute to C . The deviation of the C - y rela-tion from linearity for y = 0 .
13 indicated that the effectof the additional magnetic interactions was pronouncedat y ≥ .
1. Furthermore, as we will see, the peak widthin the Q scan of magnetic scattering is actually smallerfor y = 0 .
13 than for y = 0 .
09 (see Fig. 4). As a result,crystals with y = 0 .
09 were the main focus of µ SR andneutron scattering measurements to estimate the inher-ent Cu-spin correlations in (Bi,Pb)2201.At low H and at low temperatures, a spin-glass phaseis observable for all the Fe-doped samples. As typicallyshown in Fig. 1(b), a clear spin-glass transition arises in χ at T sg = 9 K for y = 0 .
09 when H is applied along the c -axis. The anisotropy between χ c and χ ab is attributedto the direction of the Fe spins, although this has not FIG. 1: (a) χ ab over a wide range of temperature for H = 1 T.The inset shows C as a function of the Fe content. (b) χ ab and χ c at low temperatures for H = 100 Oe and y = 0 . ρ ab for y = 0 .
09 (as-grown) and y = 0 (as-grown andAr-annealed). been clarified as yet. C. Estimation of hole concentration
The as-grown pristine sample is expected to be over-doped, due to the increase in T onc by hole reduction. Thein-plane electrical resistivity ρ ab , which was measuredby a standard DC four-terminal method, indeed showsa normal metallic transport above T onc [Fig. 1(c)]. Quan-titatively, the doping rate of the pristine as-grown sampleis determined to be p = 0 . This is close to the value (0.27) in Ref. 25 andfairly consistent with the universal dome-shaped super-conducting phase diagram. Moreover, p of the 9% Fe-doped sample was mea-sured by ARPES, resulting in p = 0 . . This re-duction in p caused by Fe doping indicates a formationof Fe charge states, and it is consistent with prelim-inary measurements of Fe M¨ossbauer spectroscopy, and the edge energy of Fe- K absorption in XAFS. Fedoping of 9% is sufficient for destroying the superconduc-tivity in the heavily overdoped phase, and it causes anincrease in the residual resistivity and carrier localizationat low temperatures, as is evident from the upturn in ρ ab [Fig. 1(c)]. FIG. 2: ZF- µ SR time spectra for the y = 0 .
09 sample at lowtemperatures. The single crystals were placed in a He gas-flow cryostat such that the CuO plane faced the muon beam(i.e., muon spin polarization vector k c -axis of the sample). A ( t ) is given by A ( t ) = { F ( t ) − αB ( t ) } / { F ( t )+ αB ( t ) } , where F ( t ) and B ( t ) are the total muon events of a forward and abackward counter, respectively. α is a calibration factor. Thecurved lines are fitted in the manner described in Ref. . D. Zero-field µ SR Zero-field (ZF) µ SR measurements using a positivemuon beam at TRIUMF in Vancouver, Canada, werecarried out at the M15 beamline. Figure 2 shows the tem-perature variation of the asymmetry parameter A ( t ) forthe as-grown single crystals with y = 0 .
09. The Gaussiandecay of the asymmetry found at 30 K and 20 K changesto a much more rapid relaxation at T = 16.5 K and 13 K,which we attribute to the development of quasistatic elec-tronic spin correlations. At the base temperature of 2 K,a short-lived oscillatory signal is observed at t< µ s and A ( t> µ s) remains nearly constant [ ∼ A (0) / A ( t ) ; two of them indicate slowand fast depolarization, and the other expresses a muonspin precession. The fast depolarization starts to appearat T µ SR ∼
20 K, and the internal field at the muon site isestimated to be ∼
100 Oe at 2 K from the precession fre-quency. For reference, these ZF- µ SR measurements werealso carried out using as-grown pristine crystals. Thenearly temperature-independent A ( t ) below 30 K con-firmed that no static magnetic order is present down to2 K, as in the case of the pristine (Bi,La)2201 crystal. E. Elastic neutron scattering
Neutron scattering experiments were performed to elu-cidate the static spin correlations indicated by bulk mag-netic susceptibility and µ SR measurements. The mag-netic scattering was investigated by triple-axis neutron
FIG. 3: Elastic scattering from the y = 0 .
09 sample. (a)Difference plot of Q spectra obtained on TOPAN (neutronenergy E = 14 . q res-olution is represented by the vertical bar. Inset shows thescan trajectory parallel to the Cu-O-Cu bond axis. (b) Thethermal evolution of the incommensurate peak intensity. Thesolid line is drawn as a guide. (c) Temperature variationof Q spectra observed using polarized neutrons at TAS-1( E = 14 . ∼
20 Oe applied along the scattering vector. Heusler(111) reflections were used in the monochromator and ana-lyzer. Contaminations by higher-order incident neutrons weresuppressed by inserting PG filters in the neutron beam path. spectroscopy in the ( h, k,
0) scattering plane, where a ∗ and b ∗ are ∼ − , respectively. Unpo-larized (polarized) neutron-scattering experiments wereperformed on triple-axis spectrometers TOPAN andAKANE (TAS-1) at the research reactor JRR-3 of theJapan Atomic Energy Agency in Tokai, Japan. Fig-ure 3(a) shows that peaks due to distinct incommen-surate elastic scattering occur near (1 , ,
0) at low tem-peratures for y = 0 .
09. The diffuse peaks, which startappearing below T ∗ ∼
40 K [Fig. 3(b)], correspond toantiferromagnetic short-range modulations propagatingalong the Cu-O-Cu bond axes. The direction of thespin modulation is identical to that observed for the su-perconducting La214 and Y123 systems. Further, re-sults of the polarized-neutron analysis performed in aspin-flip channel confirm that the diffuse incommensu-rate peaks appearing below T ∗ are of magnetic origin, asshown in Fig. 3(c). By assuming four Lorentzian peaksat Q = (1 + δ, ± δ,
0) and (1 − δ, ± δ,
0) with a HWHM κ , δ = 0 . κ = 0 . − are extracted byresolution-convoluted fitting to the difference plot shownin Fig. 3(a). The magnetic modulation period and mag-netic correlation length ξ (= 1 /κ ) are found to be ∼ . a tet FIG. 4: Difference plots for y = 0 . , . , and 0 .
03 mea-sured at AKANE ( E = 19 . × × in size, and 2 η < . ◦ for y ≤ .
09 and ∼ ◦ for y = 0 .
13 in mosaicness. The relativeintensity spectra are plotted with offset. The curved lines arefitted to a pair of Lorentzians and a constant. Dependence ofthe q -integrated intensity on y is shown in the inset. and ∼ . a tet , respectively, where a tet ≈ a/ √ ∼ . δ changes little upon Fe dopingwithin this concentration range, but the peak width at y = 0 .
13 becomes smaller than that at y = 0 .
09. Thenormalized q -integrated intensity presented in the insetshows a linear increase up to y = 0 .
09 but an additionalincrease at y = 0 .
13, thereby indicating a similarity withthe non-linear dependence of C on y . III. DISCUSSIONSA. Spin clusters around Fe
Summarizing the data, we can obtain an overall pictureof the Fe-spin-induced ISCs in overdoped (Bi,Pb)2201.Since the average Fe-Fe separation ( ∼ a tet / √ y ) nearlycorresponds to ξ when y = 0 .
09, small clusters with diam-eter ∼ ξ are formed around Fe. The linear dependencesof both C and the neutron scattering intensity on y be-low y = 0 .
09, together with the weak dependence of δ on y , suggest that the clusters do not interact much witheach other. However, Fe doping beyond y = 0 . C and scatteringintensity on y to deviate from linearity.The onset temperature of the induced ISCs depends onthe nature of probe; T ∗ ∼
40 K determined by neutronscattering is considerably higher than T sg and T µ SR de- FIG. 5: Comparison of δ between Fe-doped (Bi,Pb)2201,pristine La214 , and Y123 . Dynamical data are plot-ted for La214 ( ω = 3–6 meV at T ∼ T c ) and Y123 ( ω ≪ (resonance energy) at T < T c ). The p range of the super-conducting phase is indicated by a thick bar at the base line.The solid line is a guide for La214 and Y123 data. termined by magnetization and µ SR measurements, re-spectively. Therefore, the induced ISCs are quasistaticin nature similar to the case of the spin-glass phase ofLa214 . Our recent reinvestigation of magnetic suscep-tibility found that χ ab ( T ) of Fig. 1(a) starts to deviatefrom the Curie-Weiss law below ∼ T ∗ . This phenomenais also seen in La214 and it might be a precursor of thespin-glass transition. Although the local charge state of Fe is most likelyFe , the obtained p eff is much smaller than expectedfor high-spin S = 5 / p eff = 5 . p eff = 3 . . To explain the reduced p eff , we speculate that the Fe spin strongly couples withneighbor Cu spins and/or ligand hole spins in an anti-parallel fashion. Note that a similar reduction in theeffective spin value is observed in Ni-doped La214. B. ISCs in overdoped (Bi,Pb)2201
The highlight of the current study is the observationof quasi-static ISCs induced by Fe dopants in overdoped(Bi,Pb)2201. To the best of our knowledge, this is thefirst observation of ISCs in the (Bi,Pb)2201 system, al-though the aid of large Fe spins was needed to see them.Hence, our finding supports the idea that the ISCs are acommon feature in high- T c cuprate superconductors. Atthe same time, we can provide neutron scattering datafor the overdoped region of a system other than La214.The ISCs in overdoped (Bi,Pb)2201 possess the largestvalue of δ among the high- T c cuprates studied so far.Figure 5 shows δ , determined from low-energy modula-tions along the Cu-O-Cu bond axes in La214 andY123. Intriguingly, δ of the Fe-doped (Bi,Pb)2201 stud-ied here appears to follow an extrapolation of the linearrelationship δ = p established in La214 for p . / p in pristine overdoped (Bi,Pb)2201. Inthis picture, the Fe-spin injection may contribute to car-rier localization [Fig. 1(c)], reducing the characteristicfrequency of dynamic ISCs locally, pinning stripe corre-lations around Fe sites, and thus resulting in the spinclusters. This is essentially the same story as has beenproposed for the impact of Zn doping on underdopedLa214 ; however, there are some problems in apply-ing it to the overdoped regime. For one thing, the Zndopants that cause pinning of stripes in La214 also causea reduction in the spin correlation length, in contrast tothe present behavior. Even more significant is the ques-tion of whether spin stripes (or at least spin stripes in-duced by charge stripes) with a period as short as 5 a tet could be energetically stable, even locally. To further testthe possibility of Fe-induced stripes and to compare withstripes in La214, an investigation of magnetic-field effectson neutron scattering and electric resistivity in Fe-doped(Bi,Pb)2201 is in progress .An alternative possibility is that the magnetic corre-lations are determined by the electronic Fermi surface.A simple Fermi-surface nesting scenario, as in Cr, seemsunlikely as one would not expect nesting behavior to beestablished or amplified by magnetic impurity ions, whichshould cause considerable scattering of quasiparticles [asindicated by the increased in-plane resistivity, Fig. 1(c)].On the other hand, the presence of the Fe ions does sup-press the superconductivity and presumably closes thesuperconducting gap, which would increase the densityof states available for magnetic interactions. We notethat a crossover in the electronic response to magneticimpurities between under- and over-doped regimes hasbeen observed in a recent study of Ni-doped La214. Perhaps the magnetism involves an RKKY-type of cou-pling between Fe moments via the conduction electrons,as in dilute Cu-Mn alloys. Such an analogy would en-compass both the spin-glass behavior and the ISCs. Inany case, more work is required to understand the natureof the (induced) magnetism in overdoped (Bi,Pb)2201.
IV. SUMMARY
Quasi-static spin correlations induced by Fe dopants inoverdoped (Bi,Pb)2201 were studied by magnetization, µ SR, and neutron-scattering measurements throughFe doping. The magnetization measurements indicatespin-glass-like ordering below 10 K for a crystal with y = 0 .
09, and the µ SR measurements on the samesample confirm the presence of static magnetic orderat 2 K. ISCs propagating along the Cu-O-Cu bondingaxes as in the superconducting phase of other high- T c cuprates are detected by elastic neutron scattering attemperatures below 40 K, with the intensity saturatingbelow 10 K. The large value of δ ( ∼ p are unexpected and quite differentfrom the saturation of δ in overdoped La214. An expla-nation in terms of stripes seems unlikely. An alternativepossibility involves the coupling between Fe momentsthrough conduction electrons. Further experimentalwork is required in order to come up with a properunderstanding of the dopant-induced static magnetismin overdoped (Bi,Pb)2201. Acknowledgments
We are grateful to K. Kudo, H. Kobayashi, D. Mat-sumura, and T. Sato for their helpful discussions. Wealso thank K. Nemoto and M. Sakurai for their assis-tance in the neutron-scattering and crystal-growth ex-periments, respectively. This study was carried out un-der the Common-Use Facility Program of JAEA, andthe Quantum Beam Technology Program of JST. Thestudy performed at Tohoku University was supported bya Grant-In-Aid for Science Research C (19540358) andB (19340090) from the MEXT. JMT is supported by theU.S. Department of Energy, Office of Basic Energy Sci-ences, Division of Materials Sciences and Engineering,under Contract No. DE-AC02-98CH110886. H. Yoshizawa, S. Mitsuda, H. Kitazawa, K. Katsumata, J.Phys. Soc. Jpn. , 3686 (1988). R.J. Birgeneau, Y. Endoh, Y. Hidaka, K. Kakurai,M.A. Kastner, T. Murakami, G. Shirane, T.R. Thurston,K. Yamada, Phys. Rev. B , 2868 (1989). J.M. Tranquada, H. Woo, T.G. Perring, H. Goka, G.D. Gu,G. Xu, M. Fujita, K. Yamada, Nature , 534 (2004). B. Vignolle, S.M. Hayden, D.F. McMorrow, H.M. Rønnow, B. Lake, C.D. Frost, T.G. Perring, Nature Phys. , 163(2007). C. Stock, W.J.L. Buyers, R.A. Cowley, P.S. Clegg,R. Coldea, C.D. Frost, R. Liang, D. Peets, D. Bonn,W.N. Hardy, R.J. Birgeneau, Phys. Rev. B , 024522(2005). V. Hinkov, P. Bourges, S. Pailh`es, Y. Sidis, A. Ivanov,C.D. Frost, T.G. Perring, C.T. Lin, D.P. Chen, B. Keimer,
Nature Phys. , 780 (2007). B. Fauque, Y. Sidis, L. Capogna, A. Ivanov, K. Hradil,C. Ulrich, A. Rykov, B. Keimer, P. Bourges, Phys. Rev. B , 214512 (2007). Guangyong Xu, G.D. Gu, M. Hucker, B. Fauque, T.G. Per-ring, L.P. Regnault, J.M. Tranquada, Nature Phys. , 642(2009). K. Yamada, C.H. Lee, K. Kurahashi, J. Wada, S. Waki-moto, S. Ueki, H. Kimura, Y. Endoh, S. Hosoya, G. Shi-rane, R.J. Birgeneau, M. Greven, M.A. Kastner, Y.J. Kim,Phys. Rev. B , 6165 (1998). M. Matsuda, M. Fujita, K. Yamada, R.J. Birgeneau,M.A. Kastner, H. Hiraka, Y. Endoh, S. Wakimoto, G. Shi-rane, Phys. Rev. B , 9148 (2000). M. Fujita, K. Yamada, H. Hiraka, P.M. Gehring, S.-H. Lee,S. Wakimoto, G. Shirane, Phys. Rev. B , 064505 (2002). P. Dai, H.A. Mook, R.D. Hunt, F. Doˇgan, Phys. Rev. B , 054525 (2001). J.M. Tranquada, B.J. Sternlieb, J.D. Axe, Y. Nakamura,S. Uchida, Nature , 561 (1995). K. Yamada, S. Wakimoto, G. Shirane, C.H. Lee,M.A. Kastner, S. Hosoya, M. Greven, Y. Endoh, R.J. Bir-geneau, Phys. Rev. Lett. , 1626 (1995). B. Lake, G. Aeppli, T. E. Mason, A. Schr¨oder, D. F. Mc-Morrow, K. Lefmann, M. Isshiki, M. Nohara, H. Takagi,and S. M. Hayden, Nature (London) , 43 (1999). C.H. Lee, K. Yamada, H. Hiraka, C.R.V. Rao, Y. Endoh,Phys. Rev. B , 134521 (2003). S. Wakimoto, H. Zhang, K. Yamada, I. Swainson, H. Kim,R.J. Birgeneau, Phys. Rev. Lett. , 217004 (2004). P.L. Russo, C.R. Wiebe, Y.J. Uemura, A.T. Savici,G.J. MacDougall, J. Rodriguez, G.M. Luke, N. Kaneko,H. Eisaki, M. Greven, O.P. Vajk, S. Ono, Y. Ando, K. Fu-jita, K.M. Kojima, S. Uchida, Phys. Rev. B , 054511(2007). M. Fujita, M. Enoki, K. Yamada, J. Phys. Chem. Solids , 3167 (2008). M. Fujita, M. Enoki, S. Iikubo, K. Kudo, N. Kobayashi,and K. Yamada, arXiv:0903.5391. C.C. Torardi, E.M. McCarron, P.L. Gai, J.B. Parise,J. Ghoroghchian, D.B. Kang, M.-H. Whangbo, J.C. Barry,Physica C , 347 (1991). G. Xu, Q. Pu, Z. Ding, Y. Zhang, Physica C , 178(2000). T. Nishizaki, K. Kudo, N. Okumura, N. Kobayashi, Phys-ica C , 156 (2007). T. Sato, private communications. A. Maeda, M. Hase, I. Tsukada, K. Noda, S. Takebayashi,K. Uchinokura, Phys. Rev. B , 6418 (1990). J.L. Tallon, C. Bernhard, H. Shaked, R.L. Hitterman,J.D. Jorgensen, Phys. Rev. B , 12911 (1995). H. Kobayashi, private communications. D. Matsumura, private communications. I. Watanabe, T. Adachi, K. Takahashi, S. Yairi, Y. Koike,K. Nagamine, Phys. Rev. B , 180516 (2002). S. Wakimoto, S. Ueki, Y. Endoh, K. Yamada, Phys. Rev.B , 3547 (2000). S. Wakimoto, H. Hiraka, K. Kudo, D. Okamoto, T.Nishizaki, K. Kakurai, J. M. Tranquada, A. Zheludev, N.Kobayashi, and K. Yamada (in preparation). A. Maeda, T. Yabe, S. Takebayashi, M. Hase, K. Uchi-nokura, Phys. Rev. B , 4112 (1990). H. Hiraka, D. Matsumura, Y. Nishihata, J. Mizuki, K. Ya-mada, Phys. Rev. Lett. , 037002 (2009). K. Hirota, Physica C , 61 (2001). Y. Tanabe, K. Suzuki, T. Adachi, Y. Koike, T. Kawamata,Risdiana, T. Suzuki, and I.Watanabe, J. Phys. Soc. Jpn. , 023706 (2010). F. J. Lamelas, S. A. Werner, S. M. Shapiro, and J. A.Mydosh, Phys. Rev. B51