1: % sw5.tex (04/27/01)
2: % This is final
3:
4: %%%% prb format (below) %%%%%%%%%%%%%%%%%%%%%%%%%%%%
5: \documentstyle[aps,prb]{revtex}
6: %
7: \begin{document}
8: %
9: \input epsf.sty
10: \twocolumn[
11: \hsize\textwidth\columnwidth\hsize\csname %
12: @twocolumnfalse\endcsname
13: %%
14: %% \draft command makes pacs numbers print
15: %%
16: %% \draft
17: %
18: \widetext
19: %%%%%%%% prb (above) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
20:
21: %%%%%%%%% draft format (below) %%%%%%%%%%%%%%%%%%%%%
22: %\documentstyle[aps,preprint]{revtex}
23: %%
24: %\begin{document}
25: %%
26: %\draft
27: %%%%%%%% draft format (above) %%%%%%%%%%%%%%%%%%%%%%
28:
29: \title{Spin dynamical properties and orbital states of the layered perovskite
30: La$_{2-2x}$Sr$_{1+2x}$Mn$_{2}$O$_{7}$ $(0.3 \le x < 0.5)$}
31: %%
32: \author{K. Hirota}
33: \address{Department of Physics, Tohoku University, Sendai 980-8578, Japan}
34: %%
35: \author{S. Ishihara\cite{S.Ishihara}}
36: \address{Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan}
37: %%
38: \author{H. Fujioka\cite{H.Fujioka}}
39: \address{Department of Physics, Tohoku University, Sendai 980-8578, Japan}
40: %%
41: \author{M. Kubota\cite{M.Kubota} and H. Yoshizawa}
42: \address{Neutron Scattering Laboratory, Institute for Solid State Physics, University of
43: Tokyo, Tokai 319-1106, Japan}
44: %%
45: \author{Y. Moritomo}
46: \address{Center for Integrated Research in Science and Engineering, Nagoya University,
47: Nagoya 464-8601, Japan}
48: %%by
49: \author{Y. Endoh and S. Maekawa}
50: \address{CREST, Institute for Materials Research, Tohoku University, Sendai 980-8577,
51: Japan}
52: %%
53: \date{\today}
54: \maketitle
55:
56: %%
57: %% insert abstract here
58: %%
59:
60: \begin{abstract}
61: %
62: % What we did in the preset paper.
63: %
64: Low-temperature spin dynamics of the double-layered perovskite
65: La$_{2-2x}$Sr$_{1+2x}$Mn$_{2}$O$_{7}$ (LSMO327) was systematically studied in a wide
66: hole concentration range ($0.3 \le x < 0.5$).
67: %
68: % What we found in the experiment.
69: %
70: The spin-wave dispersion, which is almost perfectly two-dimensional, has two branches
71: due to a coupling between layers within a double-layer. Each branch exhibits a
72: characteristic intensity oscillation along the out-of-plane direction. We found that
73: the in-plane spin stiffness constant and the gap between the two branches strongly
74: depend on $x$.
75: %
76: % What we analyzed.
77: %
78: By fitting to calculated dispersion relations and cross sections assuming Heisenberg
79: models, we have obtained the in-plane $(J_{\parallel})$, {\em intra}-bilayer
80: $(J_{\perp})$ and {\em inter}-bilayer $(J')$ exchange interactions at each $x$. At
81: $x=0.30$, $J_{\parallel}=-4$~meV and $J_{\perp}=-5$~meV, namely almost isotropic and
82: ferromagnetic. Upon increasing $x$,
83: $J_{\perp}$ rapidly approaches zero while $|J_{\parallel}|$ increases slightly,
84: indicating an enhancement of the planar magnetic anisotropy. At $x=0.48$,
85: $J_{\parallel}$ reaches
86: $-9$~meV, while $J_{\perp}$ turns to $+1$~meV indicating an antiferromagnetic
87: interaction.
88: %
89: % Consideration in terms of orbital degrees of freedom.
90: %
91: Such a drastic change of the exchange interactions can be ascribed to the change of the
92: relative stability of the $d_{x^{2}-y^{2}}$ and $d_{3z^{2}-r^{2}}$ orbital states upon
93: doping. However, a simple linear combination of the two states results in an orbital
94: state with an orthorhombic symmetry, which is inconsistent with the $I4/mmm$ tetragonal
95: symmetry of the crystal structure. We thus propose that an ``orbital liquid'' state
96: realizes in LSMO327, where the charge distribution symmetry is kept tetragonal
97: around each Mn site.
98: %
99: %%% added by Ishihara (12/Jan./01)
100: %
101: Orbital liquid states are formulated in a theoretical model which takes into account
102: strong electron correlation. The calculated results satisfactorily explain the
103: systematic changes of the exchange interactions in LSMO327 observed in the experiments.
104: \end{abstract}
105:
106: % insert suggested PACS numbers in braces on next line
107:
108: %\vskip .5in
109: \pacs{75.30.Ds,75.30.Et,61.12.-q,75.10.-b}
110:
111: % body of paper here
112:
113: %%%%%% prb format (below) %%%%%%%%%%%%%%%%%%%%%%%%%%%%
114: \phantom{.}
115: %
116: ]
117: %
118: \narrowtext
119: %%%%%%%%% prb (above) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
120:
121:
122: \section{Introduction}
123: \label{Introduction}
124:
125: %
126: % General introduction of Mn perovskite oxides and orbital degrees of freedom
127: %
128:
129: Rare-earth doped Mn perovskite oxide R$_{1-x}$A$_{x}$MnO$_{3}$ (R: rare-earth ion, A:
130: alkaline-earth ion) is a prototype of the colossal magnetoresistance (CMR) materials.
131: Considerable amounts of efforts have been made to clarify the magnetic, electrical
132: and structural properties of these
133: systems.\cite{Kusters_89,Chabara_93,Jin_94,Urushibara_95,Hwang_95} It has become
134: recognized that the complex and delicate balance among the internal degrees of freedom of
135: electrons, i.e., charge, spin and orbital, is a key to understand the physics of these
136: materials.\cite{Tokura_00a,Tokura_00b} In the Mn perovskite oxides, the orbital degree
137: of freedom arises from an electron in the doubly degenerated $e_{g}$ states of a
138: Mn$^{3+}$ ion in a MnO$_{6}$ octahedron. Comparing with the charge and spin degrees of
139: freedom, however, the orbital degree of freedom has been much less explored, partially
140: due to lack of experimental techniques to {\it directly} measure the ordering
141: processes and the states, with the exception of a pioneering work by Akimitsu and Ito
142: who have established the orbital ordered state in K$_{2}$CuF$_{4}$ by measuring the
143: anisotropy of the magnetic form factor using polarized neutrons.\cite{Akimitsu_76} Recently,
144: resonant x-ray scattering techniques were successfully applied to the detection of the
145: orbital ordering process in La$_{0.5}$Sr$_{1.5}$MnO$_{4}$,\cite{Murakami_98a} which was a
146: significant step toward the understanding of the ``third'' degree of freedom of
147: electrons. In the case of LaMnO$_{3}$, the resonant x-ray scattering techniques have
148: provided direct evidence of the orbital ordering. \cite{Murakami_98b} Two-dimensional (2D)
149: planar ferromagnetic coupling found in this three-dimensional (3D) lattice can be naturally
150: explained by this orbital ordering.\cite{Hirota_96,Moussa_96,Solovyev_96,Ishihara_96}
151:
152: %
153: % Some difficulties in LSMO113
154: %
155:
156: R$_{1-x}$A$_{x}$MnO$_{3}$ consists of MnO$_{6}$ octahedra which are three dimensionally
157: connected with sharing the corners. Due to mismatch of ionic radii of Mn and
158: (R$_{1-x}$A$_{x}$) ions, the Mn-O-Mn bond angle deviates from 180$^{\circ}$, which is
159: called buckling.\cite{Goodenough_70} Since doping holes inevitably alter the average
160: ionic radius of (R$_{1-x}$A$_{x}$) ion, the amount of buckling also changes, resulting in
161: a variety of crystal structures.\cite{Kawano_96} Structural phase transitions are also
162: observed with changing temperature,\cite{Urushibara_95,Caravajal_98} and can be induced
163: by an external magnetic field.\cite{Asamitsu_95} The complexity of structural properties
164: of R$_{1-x}$A$_{x}$MnO$_{3}$ is an interesting issue,\cite{Cox_01} but makes it difficult
165: to study the role and significance of the orbital degree of freedom in affecting the
166: magnetic and transport properties, because orbitals are strongly affected by surrounding
167: structural environment.
168:
169: %
170: % Introduction of the LSMO327 system
171: %
172:
173: Mn perovskite oxides are generally represented by the Ruddlesden-Popper notation
174: (R,A)$_{n+1}$Mn$_{n}$O$_{3n+1}$, where the effective dimensionality
175: can be adjusted by changing the number of MnO$_{2}$ sheets, $n$, blocked with
176: (R,A)$_{2}$O$_{2}$ layers. As for the double layered Mn perovskite $(n=2)$, Moritomo
177: {\it et al.}\cite{Moritomo_96} have found an extremely large magnetoresistance around
178: Curie temperature $T_{c}$ in a single crystal of La$_{1.2}$Sr$_{1.8}$Mn$_{2}$O$_{7}$,
179: which was followed by intensive studies of La$_{2-2x}$Sr$_{1+2x}$Mn$_{2}$O$_{7}$
180: (LSMO327) and related compounds.\cite{Kimura_00} Figure~\ref{Fig:Structure} schematically
181: shows the structure and the magnetically ordered state at $x=0.40$.\cite{Hirota_98} A
182: comprehensive magnetic and structural phase diagram of LSMO327 has been established by
183: Kubota {\it et al.}\cite{Kubota_99a,Kubota_00} in a wide range of $x$ $(0.30 \le x \le
184: 0.50)$ as shown in Fig.~\ref{Fig:Phase_diagram}(a) through systematic powder
185: neutron-diffraction studies combined with the Rietveld analysis. They have found that
186: LSMO327 exhibits a planar ferromagnetic (FM) structure FM-I in the range $0.32 \le x \le
187: 0.38$ at low temperatures and that a finite canting angle between neighboring layers
188: starts appearing around $x \sim 0.39$ and reaches 180$^{\circ}$ (AFM-I, i.e., A-type AFM)
189: for $x \ge 0.48$. They also found that the magnetic moments are aligned parallel to the
190: $c$-axis at $x=0.30$, indicating a phase boundary between $x=0.30$ and 0.32. At $x=0.50$,
191: the magnetic structure exhibits complicated temperature dependence due to charge
192: ordering.\cite{Kubota_99b,Argyriou_00}
193:
194: %
195: % Structural properties of LSMO327
196: %
197:
198: In contrast to the rich magnetic phase diagram, the structure of LSMO327 is fairly
199: simple. Although anomalous structural behaviors were reported around $T_{c}$
200: suggesting a strong coupling among charge, spin and
201: lattice,\cite{Mitchell_97,Argyriou_99} there is only a single tetragonal ($I4/mmm$) phase
202: in the entire hole concentration ($0.30 \le x \le 1.0$) and temperature ranges ($T \le
203: 400$~K) studied so far, except a recently discovered orthorhombic
204: ($Immm$) phase which
205: exists in a limited concentration range ($0.75 < x < 0.95$).\cite{Ling_00} This is
206: most likely due to the layered structure which absorbs the changes of the Mn-O bond
207: lengths and the average ionic radius of La and Sr ions upon doping.
208: %
209: %%%(12/Jan/01)
210: %
211: It was also found that the Jahn-Teller (JT) type lattice distortion
212: $\Delta_{JT}$ of Mn-O$_{6}$ octahedra monotonically changes with increasing
213: $x$ as shown in Fig.~\ref{Fig:Phase_diagram}(c).\cite{Kubota_00} Note that $\Delta_{JT}$
214: is defined by the ratio of the averaged apical Mn-O bond length to the equatorial Mn-O
215: bond length, i.e., $\Delta_{JT} \equiv ( d_{{\rm Mn-O(1)}}+d_{\rm Mn-O(2)} )/ 2 d_{\rm
216: Mn-O(3)}$, where $d_{\rm Mn-O}$ is a distance between nearest neighbor (NN) Mn and O
217: ions. The positions of the O ions are depicted in Fig.~\ref{Fig:Phase_diagram}(c). The
218: results are in good agreement with x-ray diffraction
219: measurements.\cite{Moritomo_98,Okuda_99} The JT distortion in LSMO327 stabilizes either
220: $d_{x^2-y^2}$ or $d_{3z^2-r^2}$ state. The contraction of MnO$_{6}$ octahedron upon
221: doping implies the stabilization of $d_{x^{2}-y^{2}}$ state, i.e., a pseudo-2D
222: $e_{g}$ band, in a heavily doped region.
223: %
224: As discussed later, such a structural change itself cannot account for the systematic
225: changes of the exchange interactions within the conventional double exchange scenario.
226: The dominance of the A-type AFM structure with the decrease of
227: $\Delta_{JT}$ is ascribed to the change in the $e_{g}$ orbital state from
228: $d_{3z^{2}-r^{2}}$ to $d_{x^{2}-y^{2}}$.\cite{Hirota_98,Kubota_00} The importance of the
229: $e_{g}$ orbital state was also pointed out in striction measurements by Kimura {\it et
230: al.}\cite{Kimura_98} as well as Argyriou {\it et al.} \cite{Argyriou_99} and Medarde {\it
231: et al.}\cite{Medarde_99}
232:
233: %
234: % Spin dynamics of LSMO327
235: %
236:
237: The dynamical magnetic properties of LSMO327 with $x=0.4$ were measured by Fujioka {\it
238: et al.}\cite{Fujioka_99} They found that the spin wave dispersion is almost perfectly
239: 2D with the in-plane spin stiffness constant $D \sim 151$~meV\AA. This
240: value is similar to that of La$_{1-x}$Sr$_{x}$MnO$_{3}$ (LSMO113) with $x \sim 0.3$,
241: though $T_{C}$ is three times higher. They found that there exist two branches due to
242: a coupling between layers {\em within} a double-layer. They have analyzed the
243: spin-wave dispersion and the differential scattering cross section by applying the
244: Holstein-Primakoff transformation to a Heisenberg Hamiltonian with in-plane
245: $(J_{\parallel})$ and {\em intra}-bilayer $(J_{\perp})$ interactions (See
246: Fig.~\ref{Fig:Structure}). They have estimated that the {\em intra}-bilayer coupling is
247: $\sim 30$~\% of the in-plane coupling, which is contrary to the fact that the Mn-O bond
248: lengths are similar. They speculated that $d_{x^{2}-y^{2}}$ orbital is dominant at
249: $x=0.40$, which enhances the double-exchange, i.e., ferromagnetic, interaction within
250: the planes. This interpretation is consistent with the conclusion
251: drawn by previous structural studies\cite{Kubota_99a,Kubota_00} The $l$-dependence of the
252: scattering intensity due to the spin wave show an excellent agreement with the theoretical
253: calculation of the differential scattering cross section. Similar inelastic
254: neutron-scattering experiments were independently performed by Chatterji {\it et
255: al.},\cite{Chatterji_99a,Chatterji_99b} which gives consistent results with that of
256: Fujioka {\it et al.}\cite{Fujioka_99}
257:
258: %
259: % Studying the orbital state of LSMO327 system.
260: %
261:
262: The importance of the $e_{g}$ orbital state in determining the magnetic and transport
263: properties of LSMO327 is clear. Moreover, its simple structure makes LSMO327
264: more favorable platform to study the roles of orbital degrees of freedom than LSMO113.
265: The resonant x-ray scattering technique, however, is not directly applicable to LSMO327.
266: %
267: %%% (12/Jan./01)
268: %
269: This is because the orbital state is presumably not antiferro-type long-range ordering as
270: seen in LaMnO$_{3}$, where the two types of the orbital are alternately aligned, thus the
271: superlattice reflections due to the orbital ordering do not appear in the resonant x-ray
272: scattering. Instead, it is necessary to determine the $e_{g}$ orbital polarization. In
273: the present study, we have carried out a series of inelastic neutron-scattering
274: measurements on single crystals of the layered perovskite
275: La$_{2-2x}$Sr$_{1+2x}$Mn$_{2}$O$_{7}$ (LSMO327) at $x=0.30, 0.35, 0.40,$ and $0.48$. To
276: quantitatively determine the magnetic interactions in LSMO327, we have calculated the
277: spin-wave dispersion and the differential scattering cross section numerically by
278: applying the Holstein-Primakoff transformation\cite{Holstein_40} and the Bogoliubov
279: transformation to a Heisenberg model with the in-plane $(J_{\parallel})$, {\em
280: intra}-bilayer $(J_{\perp})$ and {\em inter}-bilayer $(J')$ interactions. We found that
281: the exchange interactions systematically changes with changing $x$. Such an $x$
282: dependence of the exchange interactions is well explained by an orbital liquid picture.
283:
284: %
285: %%%
286: % Format of the present paper
287: %
288:
289: %\newpage
290: \section{Experimental Procedures}
291: \label{Experimental}
292:
293: %
294: % Sample preparation
295: %
296:
297: LSMO327 powder was prepared by solid-state reaction using prescribed amount of pre-dried
298: La$_{2}$O$_{3}$ (99.9~\%), Mn$_{3}$O$_{4}$ (99.9~\%) and SrCO$_{3}$ (99.99~\%). The
299: powder mixture was calcined in the air for 3 days at 1250 --1400$^{\circ}$C with
300: frequent grinding. The calcined powder was then pressed into a rod and heated at
301: 1450$^{\circ}$C for 24~h. Single crystals were melt-grown in flowing 100~\% O$_{2}$ in
302: a floating zone optical image furnace with a travelling speed of 15~mm/h. We powderized
303: a part of single crystals and performed x-ray diffraction, which shows no indication of
304: impurities. Some of the crystals were also checked with electron probe microanalysis
305: (EPMA), which indicates no particular spatial inhomogeneity within the instrumental
306: error. The $x=0.30$ sample was examined by inductively coupled plasma (ICP) analysis,
307: which revealed that the ratio of La, Sr and Mn is 28.6 : 32.2: 39.2, which is in good
308: agreement with the ideal ratio, 28.0: 32.0: 40.0. These analyses indicate that our
309: samples are sufficiently stoichiometric and homogeneous. All the single crystals
310: studied are domain samples with mosaic spread of 0.3 -- 0.8$^{\circ}$
311: full-width-at-half-maximum (FWHM). The samples have a cylindrical shape, which size is
312: typically $5~\phi \times 20-30$~mm. Similarly grown samples in the range $0.3 \le x \le
313: 0.5$ were powderized and studied in detail by powder neutron diffraction techniques and
314: the Rietveld analysis. The results were already published in
315: Refs.\onlinecite{Kubota_99a,Kubota_00}. The transport properties of the samples were also
316: measured, part of which were published in Refs.\onlinecite{Kubota_00b,Moritomo_00}. The
317: results are consistent with previously reported data.\cite{Moritomo_96}
318:
319: %
320: % Neutron scattering
321: %
322:
323: Neutron scattering measurements were carried out using the Tohoku University triple-axis
324: spectrometer TOPAN located in the JRR-3M reactor of the Japan Atomic Energy Research
325: Institute (JAERI). The spectrometer was set up in the standard triple-axis mode with
326: the fixed final energy at 14.7~meV and the horizontal collimation of
327: Blank-60$'$-S-60$'$-Blank. The $(0\ 0\ 2)$ reflection of pyrolytic graphite (PG) was
328: used to monochromate and analyze the neutron beam, together with a PG filter to
329: eliminate higher order contamination. The sample was mounted in an Al can so as to give
330: the $(h\ 0\ l)$ zone in the tetragonal $I4/mmm$ notation. The Al can was then attached
331: to the cold finger of a closed-cycle He gas refrigerator. All the data were taken at
332: 10~K.
333:
334: %
335: % Characterization of the sample measured
336: %
337:
338: \section{Results}
339: \label{Results}
340:
341: %
342: % Spin wave dispersion
343: %
344:
345: As shown by Fujioka {\it et al.}\cite{Fujioka_99} and Chatterji {\it et
346: al.},\cite{Chatterji_99a,Chatterji_99b} the spin wave dispersion of LSMO327 with $x=0.4$
347: should have two modes, i.e., acoustic (A) and optical (O) branches, due to a coupling
348: between layers within a double-layer. It was theoretically shown that the A-branch has
349: maximum intensity at $l=5n$ ($n$: integer), while the phase of the O-branch is shifted by
350: $\pi$ in the double layered system.\cite{Fujioka_99} We thus measured the spin-wave
351: dispersions along $[h\ 0\ 0]$ around $(1\ 0\ 0)$ and $(1\ 0\ 5)$ for the A-branch and
352: around $(1\ 0\ 2.5)$ and $(1\ 0\ 7.5)$ for the O-branch. To study the differential cross
353: sections of spin waves, we have also measured the
354: $l$-dependence of the spin-wave intensities of A and O branches at a fixed transfer
355: energy $\Delta E=E_{i}-E_{f}$.
356:
357: %
358: % x=0.30 FM-II
359: %
360:
361:
362: Figure~\ref{Fig:Dispersion}(a) shows the dispersion relations of spin waves at 10~K for
363: $x=0.30$. Error bars correspond to the full-width at half-maximum (FWHM) of peak
364: profiles including the instrumental resolution. Spin waves of the A-branch are well
365: defined in the low $q$ and low energy region. However, the O-branch exhibits a large
366: broadening even at the magnetic zone center. The $l$-dependence of the constant energy
367: scan at $\Delta E=20$~meV are shown in Fig.~\ref{Fig:Cross_section}(a). As expected, the
368: A and O branches exhibit intensity maxima at $l=5n$ and $l=\frac{5}{2} (2n+1)$. Solid and
369: dotted curves are fitting to theoretical calculations, which is described in the next
370: section.
371:
372: %
373: % x=0.35 FM-I
374: %
375:
376: Figure~\ref{Fig:Dispersion}(b) shows the dispersion relations for $x=0.35$. The
377: dispersion curves of both the A and O branches become slightly steeper than those of
378: $x=0.30$, indicating that the in-plane magnetic interaction $J_{\parallel}$ increases
379: only gradually. However, the gap between the A and O branches becomes almost half of
380: that at $x=0.30$. Since the gap corresponds to the out-of-plane magnetic interaction
381: $J_{\perp}$, this result indicates that $J_{\perp}$ decreases considerably. More
382: quantitative analysis will be made in the following sections. We have also noticed that
383: spin waves are fairly well defined below 20~meV, and that they becomes significantly
384: broad above 20~meV, even in the constant energy scans. The $l$-dependence of the
385: constant energy scan at $\Delta E=11$~meV are shown in Fig.~\ref{Fig:Cross_section}(b).
386:
387: %
388: % x=0.40 FM-I (canted -- how to take into account)
389: %
390:
391: Figure~\ref{Fig:Dispersion}(c) shows the dispersion relations for $x=0.40$. Part of the
392: data has been already reported.\cite{Fujioka_99} The $l$-dependence of the constant
393: energy scan at $\Delta E=5$~meV are shown in Fig.~\ref{Fig:Cross_section}(c). Following
394: the tendency between $x=0.30$ and 0.35, the dispersion curves become steeper and the gap
395: becomes smaller. As shown in Fig.~\ref{Fig:Dispersion}(c), the energy-width shows
396: anomalous broadening near the zone boundary, which was found by Fujioka {\it et
397: al.}\cite{Fujioka_99} Furukawa and Hirota\cite{Furukawa_00} investigated this broadening
398: from both theoretical and experimental point of view, and ascribed it to a strong
399: magnon-phonon coupling. Let us consider a dispersionless optical phonon branch at
400: $\hbar\Omega_{0}$, and a spin wave dispersion $\hbar\omega(q)$. When a magnon with
401: momentum $q$ has energy $\omega(q) > \Omega_{0}$, it is possible to find an inelastic
402: channel to decay into a magnon-phonon pair with momentum $q'$ and $q-q'$, respectively,
403: which satisfies the energy conservation law, $\omega(q)=\omega(q')+\Omega_{0}$. This
404: decay channel gives rise to an abrupt broadening of the line-width of the spin wave
405: branch which crosses the optical phonon.
406:
407: %
408: % x=0.48 AFM-I
409: %
410:
411: Figure~\ref{Fig:Dispersion}(d) shows the dispersion relations for
412: $x=0.48$. The $l$-dependence of the constant energy scan at $\Delta E=5$~meV are shown
413: in Fig.~\ref{Fig:Cross_section}(d).
414:
415: %
416: %%%(12/Jan./01)
417: %
418: The two branches measured at (1\ 0\ 5) and (1\ 0\ 2.5) nearly degenerate.
419: %%%
420: Unlike $x=0.30$, 0.35 and 0.40, $x=0.48$ has the A-type AFM (AFM-I) structure, resulting
421: in a fundamental difference in the spin wave dispersion. We will discuss this difference
422: in detail in the next section.
423: %
424: %%%
425: %
426: % Overall
427: %
428:
429: %
430: % Note for peak profiles
431: %
432:
433: As pointed out by Furukawa and Hirota, there exists an optical phonon branch around
434: $\Delta E=20$~meV for $x=0.40$, which we have also confirmed for the other compositions we
435: studied in the present work. We have noticed that the line-width of spin waves above
436: this characteristic energy of 20~meV become significantly broaden, which is consistent
437: with the strong magnon-phonon coupling model mentioned above. Khaliullin and
438: Kilian\cite{Khaliullin_00} considered an orbitally degenerate double-exchange system
439: coupled to Jahn-Teller active phonons, which explains the softening of spin waves at the
440: zone boundary found in various ferromagnetic manganese oxides. Their model could be
441: applicable to the anomalous broadening of the spin waves of LSMO327 near the zone
442: boundary. However, the large line-width of the O branch near the {\em zone center} may
443: not be accounted for because their theoretical model does not affect the small-momentum
444: spin dynamics. We thus believe that there exists a significantly strong magnon-phonon
445: coupling as suggested. In the present paper, we have combined constant $q$ and energy
446: scans to efficiently measure the dispersion relations, which is our principle target of
447: the present work. Constant energy scans are particularly useful to avoid contamination
448: from dispersionless optical phonon branches. To further investigate this issue, however,
449: it is necessary to measure the energy widths at various $q$ utilizing constant
450: $q$ scans, which we plan to carry out in the next step.
451:
452: \section{Analysis}
453: \label{Analysis}
454:
455: In order to analyze the experimental results of the spin-wave dispersion relation and
456: the scattering cross section in LSMO327, we start from the Heisenberg model where $3d$
457: electrons in a Mn ion are treated as localized spins. In the FM (A-type AFM) structure
458: for LSMO327 with $x=$0.3, 0.35 and 0.4 ($x=$0.48), a magnetic unit cell includes two
459: (four) Mn ions termed $A$ and $B$ ($A$, $B$, $C$, and $D$). Between Mn ions, three kinds
460: of the exchange interactions, i.e. the in-plane ($J_{\parallel}$), intra-bilayer
461: ($J_{\perp}$) and inter-bilayer ($J'$) exchange interactions are introduced. The
462: schematic picture is shown in Fig.~\ref{Fig:Structure}. The Hamiltonian is given by
463: %%
464: %%
465: \begin{eqnarray}
466: {\cal H} & = & {1 \over 2} \sum_{i l} \vec S^{l}(\vec r_{il}) \biggl \{
467: J_{\parallel} \sum_{\delta_{\parallel}} \vec S^l (\vec r_{il}+\vec \delta_{\parallel})
468: \nonumber \\ & & +
469: J_{\perp} \sum_{\delta_{\perp}} \vec S^{l_{\perp}} (\vec r_{il}+\vec \delta_{\perp})+
470: J' \sum_{\delta'} \vec S^{l'} (\vec r_{il}+\vec \delta')
471: \biggr \} ,
472: \label{eq:hamiltonian}
473: \end{eqnarray}
474: %%
475: %%
476: where $\vec S^l(\vec r_{il})$ is the spin operator at a Mn ion $l$ in the $i$-th unit
477: cell and $\vec r_{il}$ is a position of the ion. Spin quantum number is assumed to be
478: $S=2(1-x)+{3 \over 2}x$ with a hole concentration $x$.
479: $l_{\perp}=l'=(B,A)$ for $l=(A,B)$ in the FM structure and $l_{\perp}=(B,D,A,C)$ and
480: $l'=(D,C,B,A)$ for $l=(A,B,C,D)$ in the A-type AFM one.
481: $\vec \delta_{\parallel}$, $\vec \delta_{\perp}$ and $\vec \delta'$ indicate the vectors
482: connecting the nearest neighboring (NN) Mn ions;
483: $\vec \delta_{\parallel}=(\pm a,0,0)$ and $(0, \pm a,0)$ where $a$ is a distance between
484: NN Mn ions in the $ab$ plane.
485: $\vec \delta_{\perp}=(0,0, c_{\perp})$ and $\vec \delta'=(\pm a/2,\pm a/2,-c')$ for $l=B$
486: and $D$, and
487: $\vec \delta_{\perp}=(0,0,-c_{\perp})$ and $\vec \delta'=(\pm a/2,\pm a/2, c')$ for $l=A$
488: and $C$, where $c_{\perp}$ and $c'$ are distances between NN MnO$_2$ layers and NN
489: bilayers, respectively.
490: %
491: By applying the Holstein-Primakoff transformation to Eq.~(\ref{eq:hamiltonian}), the
492: Hamiltonian is rewritten as follows,
493: %%
494: %%
495: \begin{equation}
496: {\cal H}=\sum_k \psi^\dagger (\vec k) \varepsilon (\vec k) \psi(\vec k) .
497: \label{eq:dis}
498: \end{equation}
499: %%
500: %%
501: Here, $\psi(\vec k)=(a_k, b_k)$ for the FM structure and $\psi(\vec k)=(a_k,
502: b_k^\dagger, c_k , d_k^\dagger)$ for the A-type AFM structure.
503: $a_k$, $b_k$, $c_k$ and $d_k$ are the boson operators for the spin operators $\vec S^A$,
504: $\vec S^B$, $\vec S^C$ and $\vec S^D$, respectively.
505: $\varepsilon (\vec k)$ is given by
506: %%
507: %%
508: \begin{equation}
509: \varepsilon(\vec k)= \pmatrix{
510: {\rm x} & {\rm y} \cr
511: {\rm y}^\ast & {\rm x} \cr
512: } ,
513: \end{equation}
514: %%
515: %%
516: with
517: %%
518: %%
519: \begin{eqnarray}
520: {\rm x} & = & -4J_{\parallel}S \biggl \{1-{1 \over 2}(\cos ak_{\rm x}+\cos ak_{\rm y})
521: \biggr\} \nonumber \\ & &
522: -J_{\perp}S-4J'S ,
523: \end{eqnarray}
524: %%
525: %%
526: and
527: %%
528: %%
529: \begin{equation}
530: {\rm y}=J_{\perp}Se^{i k_{\rm z} c_{\perp}}+4J' S\cos \biggl( {ak_{\rm x} \over 2} \biggr)
531: \cos \biggl( {ak_{\rm y} \over 2} \biggr) e^{i k_{\rm z} c'} ,
532: \end{equation}
533: %%
534: %%
535: for the FM structure,
536: and
537: %%
538: %%
539: \begin{equation}
540: \varepsilon(\vec k)= \pmatrix{
541: X & Y^\ast& Z & \cr
542: Y & X & & Z^\ast \cr
543: Z^\ast& & X & Y \cr
544: & Z & Y^\ast & X \cr
545: } ,
546: \end{equation}
547: %%
548: %%
549: with
550: %%
551: %%
552: \begin{eqnarray}
553: X & = &-4J_{\parallel} S \biggl \{ 1-{1 \over 2}(\cos ak_{\rm x}+\cos ak_{\rm y}) \biggr \}
554: \nonumber \\ & &
555: +J_{\perp}S-4J'S ,
556: \end{eqnarray}
557: %%
558: %%
559: \begin{equation}
560: Y=J_{\perp}Se^{i k_{\rm z} c_{\perp}} ,
561: \end{equation}
562: %%
563: %%
564: and
565: %%
566: %%
567: \begin{equation}
568: Z=4J'S \cos \biggl( {ak_{\rm x} \over 2} \biggr)
569: \cos \biggl( {ak_{\rm y} \over 2} \biggr) e^{i k_{\rm z} c'} ,
570: \end{equation}
571: %%
572: %%
573: for the A-type AFM structure. By utilizing the canonical transformation of the
574: Hamiltonian Eq.~(\ref{eq:dis}), the dispersion relations of the spin waves are
575: obtained. As for the FM structure, in particular, the dispersion relations are given
576: analytically as $\omega_k={\rm x}\pm | {\rm y} |$. On an equal footing, the differential
577: scattering cross section for the inelastic-neutron scattering from spin wave is given by
578: %%
579: \begin{eqnarray}
580: \lefteqn{
581: {d^2 \sigma \over d \Omega d \omega'} =
582: {\gamma e^2 \over m c^2}
583: \biggl ( {1 \over 2} g F(\vec Q) \biggr )^2
584: {k' \over k} e^{-2W(\vec Q)} N {S \over 8}
585: } & &\nonumber \\
586: & \times &
587: \sum_{ll'm} \sum_{q G} \biggl \{
588: \delta (\omega-\omega_q^m) \delta(\vec Q-\vec G-\vec q) (1+n_q^m)
589: U_{lm}^\dagger (\vec q) U_{ml'} (\vec q) \nonumber \\
590: & + &
591: \delta (\omega+\omega_q^m) \delta(\vec Q-\vec G+\vec q) n_q^m
592: U_{lm} (\vec q) U_{ml'}^\dagger (\vec q) \biggr \} ,
593: \label{eq:cs}
594: \end{eqnarray}
595: %%
596: %%
597: where $\vec q$ and $\omega_q^m$ are the momentum and energy of spin wave of the mode $m$,
598: respectively, and $n_q^m=1/(e^{\beta \omega_q^m}-1)$ is a Bose factor with temperature
599: $T=1/\beta$. $U(\vec q)$ is a matrix introduced in the canonical transformation, $\vec
600: Q=\vec k_{i}-\vec k_{f}$ is the momentum transfer with $\vec k_{i}$ ($\vec k_{f}$) being
601: the momentum of the incident (scattered) neutron, and $\vec G$ is a reciprocal lattice
602: vector. $F(Q)$ and $W(Q)$ are the magnetic structure factor and the Debye-Waller factor,
603: respectively.
604:
605: The experimental results of the dispersion relation and the differential scattering
606: cross section in LSMO327 are fitted by utilizing the least-squares method. The
607: calculated results are shown in Figs.~\ref{Fig:Dispersion} and \ref{Fig:Cross_section}
608: together with the experimental data. Note that the dispersion relation along
609: $[h\ 0\ 0]$ and the cross section along $[0\ 0\ l]$ are sensitive to $J_{\parallel}$
610: and $J_{\perp}$, and $J_{\perp}$ and $J'$, respectively. In $x=0.3$, 0.35 and 0.4, the
611: A and O branches are well separated. These two correspond to the in-phase and out-of
612: phase motions of spins in NN MnO$_2$ layers. The energy separation between the two
613: branches at the point $\Gamma$ and the stiffness constant of the A-branch in the $ab$
614: plane are given by $-2S(J_{\perp}+4J')$ and
615: $D=-SJ_{\parallel}$, respectively. An intensity oscillation along $[0\ 0\ l]$ is
616: factorized by the functions $1+\cos(c_{\perp}q_z)$ and $1-\cos(c_{\perp}q_z)$ for the A
617: and O branches, respectively, where $c_{\perp} \sim c/5$ with $c$ being the lattice
618: constant in the $c$ axis. This is attributed to the spin correlation between NN MnO$_2$
619: layers controlled by $J_{\perp}$. Additional fine structures in the intensity are
620: caused by $J'$. On the contrary, in the A-type AFM structure, four modes of the spin
621: wave exist and separate into the A and O branches corresponding to the in-phase and
622: out-of phase motions of spins between NN bilayers. Here, the each branch is doubly
623: degenerate and the energy separation between the two is of the order of $J'$. Since $J'
624: \sim J_{\parallel}/1000$ at $x=0.48$, as mentioned later, these cannot be observed
625: separately by the experiments. An intensity oscillation along [0\ 0\ l] is factorized
626: by a function $1-\cos(c_{\perp}q_z)$ originating from the antiferromagnetic spin
627: alignment between the NN MnO$_2$ layers.
628:
629:
630: The $x$ dependence of the exchange interactions is shown in Fig.~\ref{Fig:Exchange}.
631: All interactions systematically change with $x$; with increasing $x$ from 0.3,
632: $|J_{\parallel}|$ increases,
633: $J_{\perp}$ rapidly approaches to zero and changes its sign from negative to positive.
634: $|J'|$ decreases with $x$, although its value is 10$\sim$1000 times smaller than
635: $|J_{\parallel}|$ and $|J_{\perp}|$. The systematic change of the interactions
636: correlates with that of the lattice distortion in a MnO$_6$ octahedron represented by
637: $\Delta_{JT}$ (Fig.~\ref{Fig:Phase_diagram}(c)). However, this is an opposite direction
638: predicted by the conventional double-exchange scenario, where $|J_{\parallel}|$ is
639: reduced with increasing $x$ because magnitude of the double-exchange interaction is
640: proportional to the hopping integral between Mn ions in the strong Hund-coupling
641: limit. Therefore, a structural change itself cannot account for that of the exchange
642: interactions. Let us taking into account the orbital degree of freedom in a Mn ion. In
643: LSMO327 with hole concentration $x$, $1-x$ electrons occupy the two $e_g$ orbitals.
644: Character of the occupied orbital controls the anisotropy of the hopping integral of
645: electrons, i.e., that of the ferromagnetic double exchange interaction. The systematic
646: change of the exchange interactions can be explained by assuming that the $3d_{x^2-y^2}$
647: orbital is relatively stabilized with increasing $x$. The exchange interaction between
648: Mn ions is sum of the ferromagnetic double-exchange interaction and the
649: antiferromagnetic superexchange one $J_{AFM}$ acting between $t_{2g}$ spins. The more the
650: $d_{x^2-y^2}$ orbital is stabilized, the more the ferromagnetic interaction in $ab$
651: plane (along the
652: $c$ axis) becomes strong (weak). Then $J_{AFM}$ overcomes the ferromagnetic interaction
653: along the $c$ axis as shown in the region of $x=0.4-0.48$. This assumption for the
654: orbital stability is consistent with the $x$ dependent $\Delta_{JT}$
655: \cite{Kubota_00,Kimura_98,Medarde_99} and is supported by the previous theoretical work
656: where the stability of the orbitals is examined by the Madelung potential calculation.
657: \cite{Akimoto_99,Okamoto_01}
658:
659: \section{Summary and Discussion}
660:
661: We discuss possible orbital states in LSMO327 with $0.3 \le x<0.5$ and its relation to
662: the anisotropy of the ferromagnetic interaction in more detail. The orbital state at
663: each Mn ion is represented by the pseudo-spin operator defined by
664: %%
665: %%
666: \begin{equation}
667: T_{i \mu} = {1 \over 2} \sum_{s \gamma \gamma'}
668: d_{i \gamma s}^\dagger \sigma_\mu d_{i \gamma' s} ,
669: \label{eq:ps}
670: \end{equation}
671: %%
672: %%
673: for $\mu=(x,z)$. $d_{i \gamma s}$ is the annihilation operator for the $e_g$ electron at
674: site $i$ with spin $s$ and orbital $\gamma$, and $\sigma_\mu$ are the Pauli matrices.
675: In the eigen state of $T_{i z}=+(-)1/2$, an electron occupies the $d_{3z^2-r^2}$
676: ($d_{x^2-y^2}$) orbital at site $i$. $T_{iz}$ ($T_{ix}$) describes the charge
677: quadrupole moment with tetragonal (orthorhombic) symmetries and couples with the lattice
678: distortion with the same symmetry;
679: %%
680: %%
681: \begin{equation}
682: {\cal H}_{JT}=-g\sum_{i \mu=x,z}T_{i \mu }Q_{i \mu} ,
683: \end{equation}
684: %%
685: %%
686: where $Q_{i z}$ ($Q_{i x}$) describes the O ion distortions in a MnO$_6$ octahedron. The
687: orbital ordered state is characterized by a magnitude and an angle of this operator,
688: i.e. $|\langle \vec T \rangle|=\sqrt{\langle T_x \rangle^2+\langle T_z \rangle^2}$ and
689: $\Theta=\tan^{-1}(\langle T_x \rangle/\langle T_z \rangle)$ where $\langle \cdots
690: \rangle$ is the thermal average. For example, in the $(d_{3x^2-r^2},d_{3y^2-r^2})$-type
691: orbital ordered state observed in LaMnO$_3$,
692: $\Theta=2\pi/3$ and $-2\pi/3$ for the Mn sites where the $d_{3x^2-r^2}$ and
693: $d_{3y^2-r^2}$ orbitals are occupied, respectively.
694: %%
695: \par
696: %%
697: As mentioned in the previous section, a relative weight of the occupied $d_{x^2-y^2}$
698: orbital increases continuously with increasing $x$ from 0.3 to 0.48 in LSMO327 where
699: the crystal structure remains to be tetragonal ($I4/mmm$). That is, $\langle T_z (\vec
700: k=0) \rangle$ is gradually reduced with keeping the condition $\langle T_x (\vec
701: k=0)\rangle=0$ where $\vec T(\vec k)=1/N \sum_i e^{i \vec k \cdot \vec r_i} \vec T_i$
702: with the number of the Mn ion $N$ and the position of the $i$-th ion $\vec r_i$. This
703: cannot be satisfied by the uniform orbital ordered state where one kind of orbital
704: characterized by $\Theta$ is occupied at all Mn sites. This is because the change of the
705: orbital state is represented by the rotation of $\langle \vec T \rangle$ in the
706: $\langle T_z \rangle$-$\langle T_x \rangle$ plane. One may think that the
707: antiferro-type orbital ordered state explains the experiments, when the condition
708: $\Theta_A=-\Theta_B$, with $\Theta_{A(B)}$ being the angle in the orbital space for the
709: $A$ $(B)$ sublattice, is satisfied.
710: %%
711: %% Is the following statement correct?
712: %%
713: However, this is ruled out by the experimental fact
714: that the expected superlattice reflection was not reported in $0.3 \le x < 0.5$ by the x-ray and
715: electron diffractions.\cite{Li_99,Wakabayashi_00}
716:
717:
718: One of the possible orbital states realized in LSMO327 is an {\em orbital liquid} state.
719: This state was originally proposed in the ferromagnetic metallic state in LSMO113 by
720: Ref.~\onlinecite{Ishihara_97}, where the orderings of both $T_z $ and $T_x $ are
721: suppressed by the low dimensional character of the orbital fluctuation. In the case of
722: LSMO327, $\langle T_z \rangle$ is finite due to the layered crystal structure. On the
723: other hand, $T_{ix}$ does not show ordering and symmetry of the charge distribution
724: remains to be tetragonal at each Mn site. In order to formulate this orbital state,
725: let us start from the Hamiltonian where the intra-site Coulomb interactions in Mn ions are
726: taken into account;
727: %%
728: %%
729: \begin{eqnarray}
730: {\cal H} & = & \sum_{\langle ij \rangle \sigma} (t_{ij}^{ \gamma \gamma'}
731: \widetilde d_{i \gamma \sigma}^\dagger \widetilde d_{j \gamma' \sigma} + H.c.)
732: \nonumber \\
733: & & - J_H \sum_i \vec S_i \cdot \vec S_{t i}+\Delta \sum_i T_{i z} ,
734: \label{eq:ham}
735: \end{eqnarray}
736: %%
737: %%
738: where $\widetilde d_{i\gamma \sigma}=d_{i \gamma \sigma} (1-n_{i \gamma {\bar
739: \sigma}})(1-n_{i {\bar
740: \gamma} \sigma}) (1-n_{i {\bar \gamma} {\bar \sigma}})$ is the annihilation operator of
741: an $e_g$ electron excluding the doubly occupied states of electrons due to the strong
742: Coulomb interaction.
743: $\vec S_i$ is the spin operator for an $e_g$ electron defined by $\vec S_i={1 \over 2}
744: \sum_{s s'
745: \gamma} d_{i \gamma s}^\dagger \vec \sigma_{s s'} d_{i \gamma s'}$ and $\vec S_{t i}$ is
746: the spin operator for
747: $t_{2g}$ electrons with $S=3/2$. The first and second terms in Eq.~(\ref{eq:ham})
748: represent the hopping of $e_g$ electrons between NN Mn sites and the Hund coupling
749: between $e_g$ and
750: $t_{2g}$ spins, respectively. In the third term, $\Delta$ describes the splitting of the energy
751: levels of
752: $d_{3z^2-r^2}$ and $d_{x^2-y^2}$ orbitals due to the tetragonal distortion $\Delta_{JT}$
753: of a MnO$_6$ octahedron. It is shown from the theoretical calculation in
754: Ref.~\onlinecite{Okamoto_01} that $\Delta$
755: monotonically decreases with increasing $x$ for
756: LSMO327 and its maximum value is of the order of 0.5 eV. Instead of the actual crystal
757: structure of LSMO327, a pair of the 2D sheets where a squared lattice
758: consists of the Mn ions is introduced because of the weak inter-bilayer exchange
759: interaction.
760: We adopt the slave-boson scheme where
761: $\widetilde d_{i\gamma \sigma}$ is decomposed into a product of operators:
762: $\widetilde d_{i\gamma \sigma}=f_i^\dagger \tau_{i \gamma} s_{i \sigma}$ where $f_i$ and
763: $s_{i \sigma}$ are bosonic operators for charge and spin degrees of freedom,
764: respectively, and $\tau_{i \gamma}$ is a fermionic one for orbital associated with the
765: constraint of
766: $\sum_\sigma s_{i \sigma}^\dagger s_{i \sigma}=\sum_\gamma \tau_{i \gamma}^\dagger
767: \tau_{i \gamma}$ and
768: $f_i^\dagger f_i+\sum_\sigma s_{i \sigma}^\dagger s_{i \sigma}=1$ at each site. The mean
769: field approximation is introduced; $\langle f_i^\dagger f_j \rangle=x$ and
770: $\sum_{\sigma} \langle s_{i \sigma}^\dagger s_{j \sigma} \rangle=(1-x)\varepsilon_{ij}$
771: with
772: $\varepsilon_{ij}=+(-)1$ for a ferromagnetic (antiferromagnetic) bond in the orbital part
773: of the mean-field Hamiltonian. It is well known that the slave-boson mean-field
774: approximation is suitable to describe the spin liquid state.\cite{Tsvelik_95} The ratio
775: of the ferromagnetic exchange interaction in the $ab$ plane to that in the $c$ axis is
776: given by $R=J_{\parallel}/J_{\perp}=\chi_{ab}/\chi_c$ with
777: $\chi_{l}=\sum_{\gamma \gamma'} \langle \tau_{i \gamma}^\dagger \tau_{i+l \gamma'}
778: \rangle$ for $l=ab$ and $c$. The results are shown in Fig.~\ref{Fig:Ratio}(a) where
779: $J_H$ is assumed to be infinite and $t_0$ is the hopping integral between
780: $d_{3z^2-r^2}$ orbitals in the $c$ axis.
781: $R$ continuously decreases with increasing $\Delta$ implying that $d_{x^2-y^2}$ orbital
782: becomes stable relatively. This feature does not depend on the hole concentration $n_h$
783: in the calculation. Since $\Delta$ is expected to continuously decreases with $x$ in LSMO327,
784: the results in Fig.~\ref{Fig:Ratio}(a) explain the experimental results presented in
785: Fig.~\ref{Fig:Exchange}(b).
786: %
787: $\Delta$ dependence of $\langle T_x \rangle$ are shown in Fig.~\ref{Fig:Ratio}(b).
788: We note that $\langle T_z \rangle$ is zero. It is clearly shown that
789: the continuous change of $R$ is controlled by the
790: character of the occupied orbital. We also present the schematic pictures of the spatial
791: distribution of the electronic charge at a Mn site for the proposed orbital liquid state in
792: Fig.~\ref{Fig:Orbital}. The charge distributions have tetragonal symmetry and are not
793: represented by any linear combination of the atomic wave functions of the $d_{3z^2-r^2}$
794: and $d_{x^2-y^2}$ orbitals.
795: %
796:
797:
798: To summarize, we have systematically studied low-temperature spin dynamics of the
799: double-layered perovskite La$_{2-2x}$Sr$_{1+2x}$Mn$_{2}$O$_{7}$ ($0.3 \le x < 0.5$).
800: The acoustic and optical branch of the 2D spin-wave dispersion relations
801: as well as characteristic intensity oscillations along the out-of-plane direction are
802: successfully explained by theoretical calculations assuming the Heisenberg models with the
803: in-plane $(J_{\parallel})$, {\em intra}-bilayer
804: $(J_{\perp})$ and {\em inter}-bilayer $(J')$ exchange interactions. We have found that
805: the ratio $R=J_{\parallel}/J_{\perp}$ drastically decreases upon doping holes, which
806: indicates that the $d_{x^2-y^2}$ orbital becomes more stable than the $d_{3z^2-r^2}$
807: orbital. Since a simple linear combination of the two states
808: results in an orbital state with an orthorhombic symmetry, inconsistent with the
809: $I4/mmm$ tetragonal symmetry, we have introduced an ``orbital
810: liquid'' state, in which the charge distribution symmetry is kept tetragonal around each
811: Mn site.
812:
813: \begin{acknowledgments}
814: Authors would like to thank S.~Okamoto, G.~Khaliullin, A.~Koizumi, Y.~Murakami and
815: K.~Takahashi for their valuable discussions. This work was supported by the Grant in Aid
816: from Ministry of Education, Science and Culture of Japan, CREST, NEDO, and Science and
817: Technology Special Coordination Fund for Promoting Science and Technology. Part of the
818: numerical calculation was performed in the HITACS-3800/380 supercomputing facilities in IMR,
819: Tohoku University.
820: \end{acknowledgments}
821:
822: %
823: % now the references. delete or change fake bibitem. delete next three
824: % lines and directly read in your .bbl file if you use bibtex.
825:
826:
827: \begin{references}
828:
829: \bibitem[*]{S.Ishihara} Present address: Department of Applied Physics, University
830: of Tokyo. 7-3-1 Hongo, Tokyo 113-8656, Japan.
831:
832: \bibitem[\dag]{H.Fujioka} Present address: NEC corporation.
833:
834: \bibitem[\ddag]{M.Kubota} Present address: Photon Factory, Institute of Materials
835: Structure Science, KEK, Tsukuba 305-0801, Japan
836:
837: \vspace{0.25in}
838:
839: %%
840: %% Introduction (Experiment)
841: %%
842:
843: \bibitem{Kusters_89} R. M. Kusters, D. A. Singleton, R. Mcgreevy, and W. Hayes, Physica B
844: {\bf 155} 362 (1989).
845:
846: \bibitem{Chabara_93} K. Chabara, T. Ohno, M. Kasai, and Y. Kozono, Appl.\ Phys.\ Lett.
847: {\bf 62}, 780 (1993).
848:
849: \bibitem{Jin_94} S. Jin, T. H. Tiefel, M. McCormack, R. A. Fatsnacht, R. Ramesh, and L. H.
850: Chen, Science {\bf 264}, 413 (1994).
851:
852: \bibitem{Urushibara_95} A. Urushibara, Y. Moritomo, T. Arima, A. Asamitsu, G. Kido, and Y.
853: Tokura, Phys.\ Rev.\ B {\bf 51} (1995) 14103.
854:
855: \bibitem{Hwang_95} H. Y. Hwang, S.-W. Cheong, P. G. Radaelli, M. Marezio, and B. Batlogg,
856: Phys.\ Rev.\ Lett. {\bf 75} 914 (1995).
857:
858: \bibitem{Tokura_00a} {\it See, for example,} Colossal Magnetoresistance Oxides, edited
859: by Y.~Tokura (Gordon and Breach, 2000).
860:
861: \bibitem{Tokura_00b} Y.~Tokura and N.~Nagaosa, Science, {\bf 288}, 462 (2000).
862:
863: \bibitem{Akimitsu_76} J. Akimitsu and Y. Ito, J. Phys.\ Soc.\ Japan {\bf 40},
864: 1621 (1976).
865:
866: %%
867: %% Orbital
868: %%
869:
870: \bibitem{Murakami_98a} Y. Murakami, H. Kawada, H. Kawata, M. Tanaka, T. Arima, Y.
871: Moritomo, and Y. Tokura, Phys.\ Rev.\ Lett. {\bf 80}, 1932 (1998).
872:
873: \bibitem{Murakami_98b} Y. Murakami, J. P. Hill, D. Gibbs, M. Blume, I. Koyama, M. Tanaka,
874: H. Kawata, T. Arima, Y. Tokura, K. Hirota, and Y. Endoh, Phys.\ Rev.\ Lett. {\bf 81}, 582
875: (1998).
876:
877: \bibitem{Hirota_96} K. Hirota, N. Kaneko, A. Nishizawa, and Y. Endoh, J. Phys.\ Soc.\
878: Jpn. {\bf 65} 3736 (1996).
879:
880: \bibitem{Moussa_96} F. Moussa, M. Hennion, J. Rodriguez-Caravajal, H. Moudden, L.
881: Pinsard, and A. Revcolevschi, Phys.\ Rev.\ B {\bf 54}, 15149 (1996).
882:
883: \bibitem{Solovyev_96} I. Solovyev, N. Hamada, and K. Terakura, Phys.\ Rev.\ Lett. {\bf
884: 76} 4825 (1996).
885:
886: \bibitem{Ishihara_96}
887: S.~Ishihara, J. Inoue, and S. Maekawa,
888: Physica C {\bf 263}, 130 (1996), and Phys. Rev. B {\bf 55}, 8280 (1997).
889: %%
890: %% Structure of 113
891: %%
892:
893: \bibitem{Goodenough_70} J. B. Goodenough and J. M. Longon, Landolt-B\"{o}rnstein
894: Tabelllen Vol.~III/4a (Springer, Berlin, 1970).
895:
896: \bibitem{Kawano_96} H. Kawano, R. Kajimoto, M. Kubota, and H. Yoshizawa, Phys.\ Rev.\ B
897: {\bf 53}, R14709 (1996).
898:
899: \bibitem{Caravajal_98} J. Rodriguez-Caravajal, M. Hennion, F. Moussa, A. H. Moudden,
900: L. Pinsard, and A. Revcolevschi, Phys.\ Rev.\ B {\bf 57}, R3189 (1998).
901:
902: \bibitem{Asamitsu_95} A. Asamitsu, Y. Moritomo, Y. Tomioka, T. Arima, and Y. Tokura,
903: Nature {\bf 373}, 407 (1995).
904:
905: \bibitem{Cox_01} D. E. Cox, T. Iglesias, E. Moshopoulou, K. Hirota, K. Takahashi, and Y.
906: Endoh, Phys.\ Rev.\ B (to be published, cond-mat/0010339).
907:
908: %%
909: %% Introduction to 327
910: %%
911:
912: \bibitem{Moritomo_96} Y. Moritomo, A. Asamitsu, H. Kuwahara, and Y. Tokura,
913: Nature (London) {\bf 380}, 141 (1996).
914:
915: \bibitem{Kimura_00} {\it For a review,} T. Kimura and Y. Tokura, Annual Review of Material
916: Science {\bf 30}.
917:
918: \bibitem{Hirota_98} K. Hirota, Y. Moritomo, H. Fujioka, M. Kubota, H.
919: Yoshizawa and Y. Endoh, J. Phys.\ Soc.\ Jpn {\bf 67} 3380 (1998); {\bf 68} 1463 (1999).
920:
921: \bibitem{Kubota_99a} M. Kubota, H. Fujioka, K. Ohoyama, K. Hirota, Y.
922: Moritomo, H. Yoshizawa and Y. Endoh, J. Phys.\ Chem.\ Solids {\bf 60} 1161 (1999).
923:
924: \bibitem{Kubota_00} M. Kubota, H. Fujioka, K. Hirota, K. Ohoyama, Y.
925: Moritomo, H. Yoshizawa and Y. Endoh, J. Phys.\ Soc.\ Jpn {\bf 69} 1986 (2000).
926:
927: \bibitem{Kubota_99b} M. Kubota, H. Yoshizawa, Y. Moritomo, H. Fujioka, K.
928: Hirota and Y. Endoh: J. Phys.\ Soc.\ Japan {\bf 68} 2202 (1999).
929:
930: \bibitem{Argyriou_00} D. N. Argyriou, H. N. Bordallo, B. J. Campbell, A. K. Cheetham, D.
931: E. Cox, J. S. Gardner, K. Hanif, A. dos Santos, and G. F. Strouse, Phys.\ Rev.\ B {\bf
932: 61} 15269 (2000).
933:
934: \bibitem{Mitchell_97} J. F. Mitchell, D. N. Argyriou, J. D. Jorgensen, D. G. Hinks, C. D.
935: Potter, and S. D. Bader, Phys.\ Rev.\ B {\bf 55}, 63 (1997).
936:
937: \bibitem{Argyriou_99} D. N. Argyriou, H. N. Bordallo, J. F. Mitchell, J. D. Jorgensen,
938: and G. F. Strouse, Phys.\ Rev.\ B {\bf 60} 6200 (1999).
939:
940: \bibitem{Ling_00} C. D. Ling, J. E. Millburn, J. F. Mitchell, D. N. Argyriou, J. Linton,
941: and H. N. Bordallo, Phys\ Rev.\ B {\bf 62} 15096 (2000).
942:
943: \bibitem{Kimura_98}
944: T. Kimura, Y. Tomioka, A. Asamitsu, and Y. Tokura, Phys.\ Rev.\ Lett. {\bf 81}, 5920
945: (1998).
946:
947: \bibitem{Medarde_99}
948: M. Medarde, J. F. Mitchell, J. E. Millburn, S. Short, and J. D. Jorgensen,
949: Phys.\ Rev.\ Lett. {\bf 83}, 1223 (1999).
950:
951: \bibitem{Moritomo_98} Y. Moritomo, Y. Maruyama, T. Akimoto, and A. Nakamura, J. Phys.\
952: Soc.\ Jpn. {\bf 67} 405 (1998).
953:
954: \bibitem{Okuda_99} T. Okuda, T. Kimura, H. Kuwahara, Y. Tomioka, A. Asamitsu, Y. Okimoto,
955: E. Saitoh, and Y. Tokura, Mater.\ Sci.\ Eng.\ B {\bf 63}, 163 (1999).
956:
957: %%
958: %% Spin Dynamics
959: %%
960:
961: \bibitem{Fujioka_99} H. Fujioka, M. Kubota, K. Hirota, H. Yoshizawa, Y. Moritomo, and
962: Y. Endoh, J. Phys.\ Chem.\ Solids {\bf 60}, 1165 (1999).
963:
964: \bibitem{Chatterji_99a} T. Chatterji, P. Thalmeier, G. J. McIntyre, R. van de Kamp, R.
965: Suryanarayanan, G. Dhalenne, and A. Revcolevschi, Europhys. Lett. {\bf 46}, 801 (1999)
966:
967: \bibitem{Chatterji_99b} T. Chatterji, L. P. Regnault, P. Thalmeier, R. Suryanarayanan, G.
968: Dhalenne, and A. Revcolevschi, Phys.\ Rev.\ B {\bf 60} R6965 (1999).
969:
970: \bibitem{Holstein_40} T. Holstein and H. Primakoff, Phys. Rev. {\bf 58} 1094 (1940).
971:
972: %%
973: %% Sample Preparation
974: %%
975:
976: \bibitem{Kubota_00b} M. Kubota, Y. Oohara, H. Yoshizawa, H. Fujioka, K. Shimizu, K.
977: Hirota, Y. Moritomo, and Y. Endoh, J. Phys.\ Soc.\ Jpn. {\bf 69} 1986 (2000).
978:
979: \bibitem{Moritomo_00} Y. Moritomo, K. Hirota, H. Nakao, T. Kiyama, Y. Murakami, S.
980: Okamoto, S. Ishihara, S. Maekawa, M. Kubota, and H. Yoshizawa, Phys.\ Rev.\ B {\bf 62} 17
981: (2000).
982:
983: %%
984: %% Results
985: %%
986:
987: \bibitem{Furukawa_00} N. Furukawa and K. Hirota, Physica B {\bf 291} 324 (2000).
988:
989: \bibitem{Khaliullin_00} G. Khaliullin and R. Kilian, Phys.\ Rev.\ B {\bf 61} 3494 (2000).
990:
991: \bibitem{Akimoto_99} T. Akimoto, Y. Moritomo, K. Ohoyama, O. Okamoto, S. Ishihara,
992: S. Maekawa, and A. Nakamura, Phys.\ Rev.\ B {\bf 59}, R14153 (1999).
993:
994: \bibitem{Okamoto_01} S. Okamoto, S. Ishihara, and S. Maekawa, Phys.\ Rev.\ B {\bf 63},
995: 104401 (2001).
996:
997: \bibitem{Li_99} J. Q. Li, Y. Matsui, T. Kimura, and Y. Tokura, Phys.\ Rev.\ B {\bf 57},
998: R3205 (1998).
999:
1000: \bibitem{Wakabayashi_00} Y. Wakabayashi, Y. Murakami, I. Koyama, T. Kimura, Y. Tokura,
1001: Y. Moritomo, K. Hirota, and Y. Endoh, J. Phys.\ Soc.\ Jpn. {\bf 69} 2731, (2000).
1002:
1003: \bibitem{Ishihara_97} S. Ishihara, M. Yamanaka, and N. Nagaosa, Phys.\ Rev.\ B {\bf 56},
1004: 686 (1997).
1005:
1006: \bibitem{Tsvelik_95} A. M. Tsvelik, Quantum Field Theory in Condensed Matter Physics
1007: (Cambridge University Press, Cambridge, 1995).
1008:
1009: \end{references}
1010:
1011:
1012:
1013: % tables follow here
1014: %
1015: % Here is an example of the general form of a table:
1016: % Fill in the caption in the braces of the \caption{} command. Put the label
1017: % that you will use with \ref{} command in the braces of the \label{} command.
1018: % Insert the column specifiers (l, r, c, d, etc.) in the empty braces of the
1019: % \begin{tabular}{} command.
1020: %
1021: % \begin{table}
1022: % \caption{}
1023: % \label{}
1024: % \begin{tabular}{}
1025: % \end{tabular}
1026: % \end{table}
1027:
1028:
1029: % figures follow here
1030: %
1031: % Here is an example of the general form of a figure:
1032: % Fill in the caption in the braces of the \caption{} command. Put the label
1033: % that you will use with \ref{} command in the braces of the \label{} command.
1034: %
1035:
1036: %% FIG.1
1037: %%
1038:
1039: \begin{figure}
1040: \epsfxsize=0.6\columnwidth
1041: \centerline{\epsfbox{Figure1.eps}}
1042: \vspace{0.1in}
1043: \caption{Schematic representation of the magnetic spin arrangement on Mn ions in the
1044: $I4/mmm$ tetragonal cell of La$_{1.2}$Sr$_{1.8}$Mn$_{2}$O$_{7}$. Each Mn ions are
1045: surrounded by an O$_{6}$ octahedron. The lattice parameters are $a=b=3.87$ and
1046: $c=20.1$~\AA~at 10~K.\protect\cite{Hirota_98}. Notations are explained in
1047: \S.~\ref{Analysis}.}
1048: \label{Fig:Structure}
1049: \end{figure}
1050:
1051: %% FIG.2
1052: %%
1053:
1054: \begin{figure}
1055: \epsfxsize=0.9\columnwidth
1056: \centerline{\epsfbox{Figure2.eps}}
1057: \vspace{0.1in}
1058: \caption{(a) Structural and magnetic phase diagram of LSMO327 after Kubota {\it et
1059: al.}\protect\cite{Kubota_00}. (b) Several different magnetic structures appearing in the
1060: phase diagram are schematically drawn. (c) Hole concentration dependence of the JT
1061: distortion, which is defined as the ratio of the averaged apical and the equatorial Mn-O
1062: bond lengths, at room temperature and 10~K.}
1063: \label{Fig:Phase_diagram}
1064: \end{figure}
1065:
1066: %% FIG.3
1067: %%
1068:
1069: \begin{figure}
1070: \epsfxsize=0.75\columnwidth
1071: \centerline{\epsffile{Figure3.eps}}
1072: \vspace{0.1in}
1073: \caption{The dispersion relations of spin waves at 10~K for (a) $x=0.30$, (b) $x=0.35$,
1074: (c) $x=0.40$, and (d) $x=0.48$. Error bars correspond to the FWHM of peak profiles.
1075: Solid circles and open circles indicate the acoustic branch and
1076: the optical branch, respectively. Solid and dotted curves are obtained by fitting to
1077: theoretical models described in \S.~\ref{Analysis}.}
1078: \label{Fig:Dispersion}
1079: \end{figure}
1080:
1081: %% FIG.4
1082: %%
1083:
1084: \begin{figure}
1085: \epsfxsize=0.75\columnwidth
1086: \centerline{\epsffile{Figure4.eps}}
1087: \vspace{0.1in}
1088: \caption{Differential cross sections along the out-of-plane direction, which were
1089: obtained from the $l$-dependence of the constant-$E$ scans. The solid and open circles
1090: indicate intensities of the acoustic and optical branches, respectively. Solid and dotted
1091: curves are obtained by fitting to
1092: theoretical models described in \S.~\ref{Analysis}. The acoustic branch is dominant at
1093: (1~0~0) and (1~0~5), and the optical branch is dominant at (1~0~2.5) and (1~0~7.5).}
1094: \label{Fig:Cross_section}
1095: \end{figure}
1096:
1097: %% FIG.5
1098: %%
1099:
1100: \begin{figure}
1101: \epsfxsize=0.75\columnwidth
1102: \centerline{\epsffile{Figure5.eps}}
1103: \vspace{0.1in}
1104: \caption{(a) $x$ dependence of the exchange interactions obtained by analyses of the
1105: dispersion relations and the scattering cross sections in the inelastic neutron scattering
1106: experiments. (b) $x$ dependence of the ratio of the exchange interactions $J_{\perp}$ and
1107: $J_{\parallel}$.}
1108: \label{Fig:Exchange}
1109: \end{figure}
1110:
1111: %% FIG.6
1112: %%
1113:
1114: \begin{figure}
1115: \epsfxsize=0.75\columnwidth
1116: \centerline{\epsffile{Figure6.eps}}
1117: \vspace{0.1in}
1118: \caption{
1119: (a) Theoretical results of the exchange interactions ratio R=$J_{\perp}/J_{\parallel}$.
1120: $J_{\parallel}$ and $J_{\perp}$ are the exchange interactions between NN Mn sites in
1121: the $ab$ plane and along the $c$ direction, respectively.
1122: (b) The relative number of occupied electrons in two $e_g$ orbitals 2$\langle T_z \rangle$.
1123: }
1124: \label{Fig:Ratio}
1125: \end{figure}
1126:
1127: %% FIG.7
1128: %%
1129:
1130: \begin{figure}
1131: \epsfxsize=0.4\columnwidth
1132: \centerline{\epsffile{Figure7.eps}}
1133: \vspace{0.1in}
1134: \caption{Schematic pictures of the charge distribution for $e_g$ electrons in the case of
1135: (a) $\Delta=-0.4$, (b) 0 and (c) 0.4.}
1136: \label{Fig:Orbital}
1137: \end{figure}
1138:
1139:
1140: \end{document}
1141:
1142: %
1143: % ****** End of file template.aps ******
1144:
1145: