Ni-impurity effect in high-Tc cuprates studied by neutron scattering and XAFS spectroscopy

ARTICLE IN PRESS Journal of Physics and Chemistry of Solids 69 (2008) 3136– 3138

Contents lists available at ScienceDirect

Journal of Physics and Chemistry of Solids journal homepage: www.elsevier.com/locate/jpcs

Ni-impurity effect in high-T c cuprates studied by neutron scattering and XAFS spectroscopy H. Hiraka a,, S. Wakimoto b, M. Matsuda b, D. Matsumura c, Y. Nishihata c, J. Mizuki c, K. Yamada a a b c

Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan Quantum Beam Science Directorate, Japan Atomic Energy Agency, Ibaraki 319-1195, Japan Quantum Beam Science Directorate, Japan Atomic Energy Agency, Hyogo 679-5148, Japan

a r t i c l e in f o

Keywords: C. Neutron scattering C. XAFS

a b s t r a c t Neutron scattering experiments using La2x Srx Cu1y Niy O4 clarified that the parallel spin-density modulations (SDMs) in the superconducting phase are susceptible to Ni, in the same way as the diagonal SDMs in the insulating spin-glass phase. Ni substitution reduces the mobile hole concentration from x down to x  y. Polarized XAFS measurements using Ni K-edge probe two types of Ni valence states; Ni2þ and Nið2þaÞþ . It indicates that a strong hole localization occurs around Ni, resulting in an effective spin-12 value at Ni sites. Therefore, a charge impurity nature of Ni is most likely realized when xXy. & 2008 Elsevier Ltd. All rights reserved.

1. Introduction The electron-pairing mechanism of high-T c superconductivity is still controversial and unresolved yet. Historically, investigation by Cu-site replacement with other 3d transition metals has played an important role to limit the possible pairing mechanism [1]. Since the optimal T c in hole-doped cuprates drops by nonmagnetic Zn substitution more rapidly than by Ni substitution [2–4], a pairing formation with s-wave symmetry is invalidated. It contrasts to conventional phonon-mediated BCS superconductors, and electron–phonon interactions have been supposed to be indirect to the origin of high T c . Nonetheless, an interest of the possible role of electron–phonon interactions arises from a research via angle-resolved photoemission spectroscopy (ARPES) recently [5]. It is therefore urgent to re-investigate impurity effects from the microscopic point of view by rapidly advancing measurement techniques [6–8]. Impacts of magnetic impurities on the superconducting (SC) properties are especially intriguing, as many believe that magnetic interactions are essential to create high T c . Due to the quantum spin number of S ¼ 1 (Ni2þ ), Ni atoms have been most widely introduced in the CuO2 plane presumed as a magnetic dopant. Recently, however, a charge-impurity nature of Ni atom is proposed experimentally. Neutron scattering experiments for hole-doped La2x Srx CuO4 (LSCO) reveal that upon Ni doping the Ne´el temperature and the spin-glass transition temperature ðT sg Þ both increase, while the incommensurability of  Corresponding author. Tel.: +81 22 215 2037.

E-mail address: [email protected] (H. Hiraka). 0022-3697/$ - see front matter & 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jpcs.2008.06.043

diagonal spin-density modulations (D-SDMs) decreases in the antiferromagnetic and spin-glass phases [9,10]. These Ni-impurity effects are simply explained with a reduction in the net hole density, which is consistent with the strong hole localization by Ni substitution reported from resistivity measurements [11]. As a result, a Ni3þ impurity state with S ¼ 12 is indirectly proposed in the insulating phase of LSCO. It is therefore crucial to clarify by neutron scattering whether doped Ni ions can function similarly in the SC phase as well, where parallel spin-density modulations (P-SDMs) coexist robustly with superconductivity. Further, we try to settle this issue by means of a charge probe, or by X-rayabsorption-fine-structure (XAFS) spectroscopy, so that one can correctly examine the origin of novel high T c .

2. Experimental procedures Single crystals of La2x Srx Cu1y Niy O4 (LSCNO) in the SC phase (x ¼ 0:06 and 0.07) with Ni doping levels in 0:03pyp0:06 were grown by traveling-solvent-floating-zone techniques. After our standard heat treatments under oxygen flowing gas, the crystals were characterized chemically by ICP measurements and physically using a SQUID magnetometer. Neutron scattering experiments were carried out on coldneutron triple-axis spectrometers LTAS and HER installed in the guide hall of JRR-3 at the Japan Atomic Energy Agency (JAEA), and on SPINS in the research reactor of the National Institute for Standard and Technology in U.S.A. Some parts of sample-quality check determining twin structures were made on AKANE, a thermal-neutron triple-axis spectrometer of Tohoku University

ARTICLE IN PRESS H. Hiraka et al. / Journal of Physics and Chemistry of Solids 69 (2008) 3136–3138

4. XAFS spectroscopy In order to support this intuitive scenario proposed by neutron scattering experiments, we measured fluorescence XAFS spectra near Ni K-edge and characterized the Ni valence state. In case of EkCuO2 plane (Fig. 2), the Ni-rich group ðxeff o0Þ well follows a

P-scan D-scan

P-SDM

κd 0.53 k (r.l.u.)

installed at JAEA. The experimental details are described elsewhere [12]. Polarized XAFS experiments were performed on the beam line 14B1 at the SPring-8, Japan. The energy resolution was about 1 eV by using a Si(111) double-crystal monochromator. Ni K-edge XAFS spectra were measured using single crystals (typical size 5  5  3 mm3 ) at room temperature. A fluorescence-detection technique using a multi solid-state-detector bank of Ge was applied to monitor the absorption spectra. XAFS spectrum of a single crystal La1:95 Sr0:05 NiO4 (marked as ‘‘5–100’’) was taken as the reference of a divalent Ni state. In addition, other reference data were obtained from the insulating antiferromagnetic phase (ðx; yÞ ¼ ð0; 0:01Þ; ð0:01; 0:01Þ and ð0:01; 0:07Þ) and the SC phase (ðx; yÞ ¼ ð0:15; 0:01Þ and ð0:15; 0:03Þ). Two types of experimental geometries of EkCuO2 plane and E ? CuO2 plane were used, where E represents an electrical-field vector of the incident X-ray.

3137

δp

0.50 κp

3. Neutron scattering

δd D-SDM

0.47

≈

resolution ≈

0

0.50 h (r.l.u.)

0.47

0.53

La1.93Sr0.07Cu0.97Ni0.03O4, T = 4 K SPINS, D-scan & P-scan 300

Intensity (counts / min)

We present here the results of Ni-impurity effects in the vicinity of superconductor-insulator boundary (xcri ’ 0:055), where both P-SDMs and D-SDMs coexist at low temperature for pristine LSCO [13]. To evaluate the influence of Ni substitution for each, we carried out two types of Q -scans, i.e., ‘‘parallel scan’’ (P-scan) and ‘‘diagonal scan’’ (D-scan) as depicted in Fig. 1(a). Due to the similar flat-top cross sections along the P-scan and the D-scan (Fig. 1(b)), it is very likely from experiments that Ni impurities give a comparable impact on the incommensurability (dp and dd ) for both types of SDMs, and that dp  dd . Also, since the doped Ni ions will increase charge inhomogeneity in local, the coexistence of two types of SDMs is possible even in the spin-glass LSCNO in consideration of the first-order transition at xcri . Therefore, we quantitatively analyzed the Q-spectra using a cross section that contains parallel and diagonal incommensurate peaks together with an orthorhombic twin structure, as schematically shown in Fig. 1(a). According to the peak profile of Ni-free LSCO, the small incommensurability for Ni-doped sample would be comparable to the peak width (kp and kd ). Hence, the P-scan and the D-scan data were simultaneously fitted to the cross section by taking into account the instrumental resolution. The resultant fit reproduces the observation well as shown by dark lines in Fig. 1(b). Fig. 1(c) shows dd , that is one of the fitting parameters and assumed to be equal to dp , as a function of x  yð¼ xeff Þ. dd in the Ni-doped LSCO follows the data from the Ni-free LSCO [13,15]. It suggests that the mobile hole density decreases from x to x  y by Ni substitution without disturbing the underlying spin framework. We conjecture that Ni and hole couple strongly on the CuO2 plane, so that the net hole density is reduced and a spin-12value is synthesized locally at Ni sites.

Obs, D-scan

Fit, D-scan

Obs, P-scan

Fit, P-scan

200

100 -0.2

0

-0.1

0.1

La2-xSrxCu1-yNiyO4 present 0.09

y>0 Matsuda

(p-SDM)

y=0 δd (r. l. u.)

Fig. 1. (a) Schematic incommensurate-peak configuration of D-SDMs (triangle) and P-SDMs (circle), and scan trajectories together with instrumental resolution. Incommensurate peaks originating only from ð1; 0; 0Þort domain are drawn for convenience. The filled [open] square represents the orthorhombic ð1; 0; 0Þort ½ð0; 1; 0Þort  position in a two-domain structure caused by twin. (b) Superposed plot of P-scan (black) and D-scan (white) for LSCNO with ðx; yÞ ¼ ð0:07; 0:03Þ. The horizontal axis is scaled by a distance from ð12; 12; 0Þ with a unit of A1 , and the instrumental resolution is shown by the thick bar. The results of simultaneous fits of P-scan (broken line) and D-scan (solid line) are drawn by dark lines. For reference, a resolution-convoluted calculation of P-scan for y ¼ 0 is shown by a light broken line using parameters from Ref. [14]. The intensity scale of y ¼ 0 is adjusted so that the total magnetic cross section does not change due to Ni. (c) Plots of incommensurability dd against xeff ð¼ x  yÞ in LSCNO. Filled circles and squares represent data from the SC phase and the SG phase [10], respectively. The data of Ni-free LSCO [13,15] are shown by open circles.

0.2

q (Å-1)

0.06

0.03

Ni 0.03 Ni 0.03 Ni 0.05

0 0

0.03

0.06 xeff (= x - y)

0.09

ARTICLE IN PRESS 3138

H. Hiraka et al. / Journal of Physics and Chemistry of Solids 69 (2008) 3136–3138

5. Summary

O(2) NiO6 octahedron

We can easily group current XAFS data into two series by either xoy or xXy. This simple rule manifests presence of discrete valence states for Ni atom. One group contains a usual Ni2þ state 8 ð3d Þ and the other a new hole-pinning Nið2þaÞþ state, where a is a 8 7 constant of 0oap1 (3d L or 3d ) and measures the degree of hole localization. The attractive force materializing the second case could be the deep Coulomb potential for hole at Ni2þ ions that are randomly distributed in the CuO2 plane. In this scenario, an itinerant hole is very likely trapped at Ni2þ sites until the potential of all Ni2þ ions is buried by other holes. At that time, the net hole concentration is reduced down to xeff naturally. Thus, XAFS data are consistent with the probable scenario from neutron studies. From the point of view of the charge distribution, the Nið2þaÞþ ions give a kind of inhomogeneity in the Cu2þ background. As a result, Ni-substitution in the hole-itinerating CuO2 plane lowers the mobile hole density and it is equivalent to the introduction of charge impurity in the divalent square lattice, when xeff X0.

O(1) Ni

E // CuO2 plane

X-ray La2-xSrxCu1-yNiyO4, Ni K-edge, T = 290 K 15-1 15-3

∼1 eV

Fluorescence intensity (arb. unit)

6-3 6-6 1-1

8340

Acknowledgments We are grateful to T. Tohyama, K. Tsutsui, W. Koshibae, M. Kofu, M. Fujita and Y. Itoh for stimulating discussions. We also thank M. Sakurai for growing the single crystals. The work at Tohoku University was supported by the grants from the Ministry of Education, Culture, Sports, Science and Technology.

8344

0-1 1-7 5-100

References

(Ni2+ standard)

8330

8350

8340

8360

E (eV) Fig. 2. (a) Schematic set up for EkCuO2 plane. (b) XANES spectra nearby Ni K-edge for EkCuO2 plane. The spectra with white (gray) circles represent the data from La2x Srx Cu1y Niy O4 with xXy (xoy). The standard spectrum of Ni2þ is shown by black circles. The inset shows the close-up spectra at the edge.

spectrum of the Ni2þ standard compound, indicating a divalent state of Ni. In contrast, the hole-rich group ðxeff X0Þ shows a partial modification from the divalent spectrum associated with an energy shift of the K-edge by  þ 1 eV. This is the reason why the valence state of Ni for the hole-rich group should be higher than divalent. We note that, for the experimental set-up of E ? CuO2 plane, the concentration dependence is less pronounced (not shown). This quite large polarization dependence indicates that trapped holes around Ni atoms are confined within the CuO2 plane and they occupy hybridized orbitals between 3dx2 y2 of Ni and 2ps of in-plane oxygen, O(1).

[1] For example, Physical Properties of High Temperature Superconductors I, D.M. Ginsberrg (Ed.), World Scientific, Singapore, 1989. [2] Y. Maeno, T. Tomita, M. Kyogoku, S. Awaji, Y. Aoki, K. Hoshino, A. Minami, T. Fujita, Nature 328 (1987) 512. [3] J.M. Tarascon, L.H. Greene, P. Barboux, W.R. McKinnon, G.W. Hull, T.P. Orlando, K.A. Delin, S. Foner, E.J. MacNiff Jr., Phys. Rev. B 36 (1987) 8393. [4] A. Maeda, T. Yabe, S. Takebayashi, M. Hase, K. Uchinokura, Phys. Rev. B 41 (1990) 4112. [5] A. Lanzara, P.V. Bogdanov, X.J. Zhou, S.A. Kellar, D.L. Feng, E.D. Lu, T. Yoshida, H. Eisaki, A. Fujimori, K. Kishio, J.-I. Shimoyama, T. Noda, S. Uchida, Z. Hussain, Z.-X. Shen, Nature 412 (2001) 510. [6] S.H. Pan, E.W. Hudson, K.M. Lang, H. Eisaki, S. Uchida, J.C. Davis, Nature 403 (2000) 746. [7] E.W. Hudson, K.M. Lang, V. Madhavan, S.H. Pan, H. Eisaki, S. Uchida, J.C. Davis, Nature 411 (2001) 920. [8] K. Terashima, H. Matsui, D. Hashimoto, T. Sato, T. Takahashi, H. Ding, T. Yamamoto, K. Kadowaki, Nature Phys. 2 (2006) 27. [9] H. Hiraka, T. Machi, N. Watanabe, Y. Itoh, M. Matsuda, K. Yamada, J. Phys. Soc. Jpn. 74 (2005) 2197. [10] M. Matsuda, M. Fujita, K. Yamada, Phys. Rev. B 73 (2006) 140503(R). [11] N. Watanabe, T. Masui, Y. Itoh, T. Machi, I. Kato, N. Koshizuka, M. Murakami, Physica C 388–389 (2003) 241. [12] H. Hiraka, S. Ohta, S. Wakimoto, M. Matsuda, K. Yamada, J. Phys. Soc. Jpn. 76 (2007) 074703. [13] M. Fujita, K. Yamada, H. Hiraka, P.M. Gehring, S.-H. Lee, S. Wakimoto, G. Shirane, Phys. Rev. B 65 (2002) 064505. [14] H. Hiraka, Y. Endoh, M. Fujita, Y.S. Lee, J. Kulda, A. Ivanov, R.J. Birgeneau, J. Phys. Soc. Jpn. 70 (2001) 853. [15] M. Matsuda, M. Fujita, K. Yamada, R.J. Birgeneau, M.A. Kastner, H. Hiraka, Y. Endoh, S. Wakimoto, G. Shirane, Phys. Rev. B 62 (2000) 9148.