cond-mat0304543/cmr.tex
1: %\documentclass[prl,preprint,showpacs]{revtex4}
2: \documentclass[prl,twocolumn,showpacs]{revtex4}
3: \usepackage{graphicx}
4: \catcode`\@=11
5: \def\simle{\mathrel{\mathpalette\@versim<}}   % < over \sim
6: \def\simge{\mathrel{\mathpalette\@versim>}}   % > over \sim
7: \def\@versim#1#2{\lower2.5pt\vbox{\baselineskip0pt \lineskip-.5pt
8:    \ialign{$\m@th#1\hfil##\hfil$\crcr#2\crcr\sim\crcr}}}
9: \newcommand{\mib}[1]{\mbox{\boldmath $#1$}} %\newcommand{\mib}[1]{\mbox{\boldmath $#1$}} 
10: 
11: \begin{document}
12: 
13: \title{
14: Competing Orders and Disorder-induced Insulator to Metal Transition 
15: in Manganites
16: }
17: \author{
18: Yukitoshi Motome$^{1*}$, 
19: Nobuo Furukawa$^2$, and Naoto Nagaosa$^{1,3,4}$
20: }
21: \affiliation{
22: $^{1}$Tokura Spin SuperStructure Project, ERATO,
23: Japan Science and Technology Corporation,
24: c/o National Institute of Advanced Industrial Science and Technology,
25: Tsukuba Central 4, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan
26: \\
27: $^{2}$Department of Physics, Aoyama Gakuin University, 
28: 5-10-1 Fuchinobe, Sagamihara, Kanagawa 229-8558, Japan
29: \\
30: $^{3}$CREST, Department of Applied Physics, University of Tokyo,
31: 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
32: \\
33: $^{4}$Correlated Electron Research Center, AIST,
34: Tsukuba Central 4, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8562, Japan
35: }
36: \date{\today}
37: 
38: \begin{abstract}
39: Effects of disorder on the two competing phases, i.e.,  
40: the ferromagnetic metal and the commensurate charge/lattice 
41: ordered insulator, are studied by Monte Carlo simulation. 
42: The disorder suppresses the charge/lattice 
43: ordering more strongly than the ferromagnetic order, 
44: driving the commensurate insulator to the ferromagnetic metal 
45: near the phase boundary in the pure case. 
46: Above the ferromagnetic transition temperature, 
47: on the contrary, the disorder makes the system more insulating, 
48: which might cause an enhanced colossal magnetoresistance 
49: as observed in the half-doped or Cr-substituted manganites. 
50: No indication of the percolation or the cluster formation is found, 
51: and there remain the charge/lattice fluctuations instead 
52: which are enhanced toward the transition temperature. 
53: \end{abstract}
54: 
55: \pacs{75.47.Gk, 75.47.Lx, 71.10.-w}
56: 
57: \maketitle
58: %%%%%%%%%%%%%%%%%%% Introduction
59: The mechanism of the colossal magnetoresistance (CMR) in manganites 
60: is one of the central issues in the physics of 
61: the strongly correlated electronic systems 
62: \cite{Tokura2000,Dagotto2001}. 
63: By applying the external magnetic field of a few Tesla, 
64: the resistivity changes of the order of $10^4$-$10^6$. 
65: Although the double-exchange (DE) mechanism gives a reasonable description 
66: for the negative magnetoresistance 
67: \cite{Furukawa1999}, 
68: it is not enough to explain such a huge response. 
69: The first proposal for the mechanism of the CMR was based on 
70: the transition from small to large polaron driven by the magnetic field 
71: \cite{Millis1995}. 
72: The idea is that the effective bandwidth controlled 
73: by the DE interaction is enhanced by the spin ferromagnetic 
74: alignment under the external magnetic field, 
75: which leads to the crossover from small to large Jahn-Teller polaron. 
76: However, the change of the effective bandwidth 
77: is rather small, and this mechanism does not lead to such a huge effect. 
78: Thus, this single-particle picture is not satisfactory 
79: to explain the CMR. 
80: 
81: Instead, recently, it has been recognized that the CMR is 
82: a collective phenomenon due to the many-body correlation. 
83: One of the present authors proposed 
84: a new mechanism of CMR based on the multicritical fluctuation 
85: between the ferromagnetic metal (FM) and the charge-ordered insulator 
86: (COI) occurring near $x=0.5$ ($x$: hole concentration) 
87: \cite{MurakamiPREPRINT}. 
88: There, the enhanced fluctuations near the multicritical point 
89: trigger a giant response to the external magnetic field. 
90: An evidence for this multicritical scenario is 
91: the scaling law for the magnetization curve (Arrott plot). 
92: However, this scaling analysis holds only for some class of materials, 
93: while the others do not show a good agreement. 
94: 
95: In the latter class of materials, 
96: the disorder appears to play an important role. 
97: Especially the spatially inhomogeneous structure has been observed 
98: experimentally \cite{Mori1998} and the percolation mechanism of 
99: the CMR is phenomenologically proposed based on this observation 
100: \cite{Moreo1999}. 
101: Namely it is assumed that the resistivity is determined 
102: by the percolating conduction paths, 
103: which are so sensitive to the pattern of 
104: the coexisting metallic and insulating regions 
105: under the influence of the disorder. 
106: The external magnetic field changes this pattern, 
107: and hence causes a large change of the resistivity. 
108: 
109: \begin{figure}
110: \includegraphics[width=6.5cm]{fig1.eps}
111: %\includegraphics[width=12cm]{fig1.eps}
112: \caption{
113: Phase diagram of {\it Ln}$_{1/2}$Ba$_{1/2}$MnO$_3$ 
114: from Ref.~[8]. 
115: $T_{\rm C}$, $T_{\rm CO}$, $T_{\rm N}$, and $T_{\rm SG}$ denote 
116: the ferromagnetic, charge-ordering, antiferromagnetic, and 
117: spin-glass transition temperatures, respectively. 
118: The subscript `(order)' are for the ({\it Ln},Ba)-ordered materials. 
119: The shaded area shows the charge-ordered insulating state 
120: in the ordered materials. 
121: The hatched area represents the disorder-induced ferromagnetic phase. 
122: } 
123: \label{fig:phase_exp} 
124: \end{figure}
125: 
126: Recent experiments in the half-doped manganites 
127: {\it Ln}$_{1/2}$Ba$_{1/2}$MnO$_3$, where {\it Ln} is a rare-earth element, 
128: have elucidated more explicitly the importance of the disorder
129: \cite{AkahoshiPREPRINT,Nakajima2002}. 
130: In these compounds, the strength of disorder due to the solid solution of 
131: ({\it Ln},Ba) ions can be tuned by careful treatment of the synthesis 
132: \cite{Millange1998}. 
133: Figure~\ref{fig:phase_exp} shows the experimental phase diagram 
134: reported in Ref.~\cite{AkahoshiPREPRINT}. 
135: In the `clean' limit where {\it Ln} and Ba ions form 
136: a periodic layered structure,
137: the phase diagram shows a multicritical behavior
138: where FM and COI compete with each other. 
139: On the other hand, in the `disordered' case where 
140: the arrangement of ({\it Ln},Ba) ions is random, 
141: the phase diagram changes in a very asymmetrical manner: 
142: FM is suppressed but survives at finite temperatures, while COI disappears 
143: and instead some glassy state is realized at low temperatures. 
144: This leads to a nontrivial regime where the disorder induces 
145: the transition from COI to FM (hatched area in Fig.~\ref{fig:phase_exp}).
146: The enhanced CMR effect is observed above $T_{\rm C}$ in this regime 
147: \cite{AkahoshiPREPRINT}.
148: Similar phenomena are observed also in 
149: {\it Ln}$_{0.55}$(Ca,Sr)$_{0.45}$MnO$_3$ \cite{TomiokaPREPRINT} and 
150: in {\it Ln}$_{0.5}$Ca$_{0.5}$MnO$_3$ with the Cr substitution 
151: of Mn \cite{Barnabe1997}. 
152: 
153: Therefore, to understand the mechanism of the CMR, 
154: it is crucial to clarify the effects of the quenched disorder on the 
155: competition between the FM and CO, and on 
156: the metal-insulator transition, which we undertake in this letter.
157: Employing the unbiased numerical method to treat large fluctuations properly, 
158: we reveal that these two orders show
159: contrastive responses to the random potential 
160: which indeed leads to the disorder-induced transition from COI to FM. 
161: Through the systematic study of density of states,
162: conductivity, and fluctuations, 
163: we elucidate the mechanism of the enhanced CMR effect in this regime. 
164: 
165: 
166: 
167: %%%%%%%%%%%%%%%%%%% Model
168: We consider a minimal model to describe a competition 
169: between FM of the DE origin and 
170: COI concomitant with the lattice distortion, 
171: whose Hamiltonian reads 
172: \begin{eqnarray}
173: &H& = -t \sum_{\langle ij \rangle \sigma} (c_{i\sigma}^\dagger c_{j\sigma} 
174: + {\rm h.c.}) - J_{\rm H} \sum_i \sigma_i^z S_i 
175: \nonumber \\
176: &-&g \sum_i n_i Q_i 
177: + \frac12 \sum_i Q_i^2 + \frac{\lambda}{2} \sum_{\langle ij \rangle} Q_i Q_j
178: + \sum_i \varepsilon_i n_i.
179: \label{eq:H}
180: \end{eqnarray}
181: Here, the first two terms represent a simplified DE model 
182: which consists of the electrons' nearest-neighbor hopping $t$ and 
183: the Hund's-rule coupling $J_{\rm H}$. 
184: We consider the Ising-type Hund's-rule coupling 
185: in the limit of $J_{\rm H} \rightarrow \infty$ for simplicity, 
186: which retains an essential physics of the DE mechanism 
187: \cite{Motome2001}. 
188: The third term describes the electron-phonon coupling. 
189: Here, we consider the classical phonons 
190: which couple to the electron density (breathing mode). 
191: The next two terms are the elastic energy of phonons. 
192: We take account of the cooperative aspect of the lattice distortion 
193: by the fifth term which reflects the fact that 
194: a displacement of an oxygen affects both MnO$_6$ tetrahedra 
195: connected by the oxygen. 
196: The last term incorporates the random on-site potential energy $\varepsilon_i$ 
197: which couples to the electron density 
198: $n_i = \sum_\sigma c_{i\sigma}^\dagger c_{i\sigma}$. 
199: The alloying effect of {\it A}-site ions in {\it A}MnO$_3$ 
200: and the Cr substitution into the Mn site
201: cause the structural and electrostatic disorder, 
202: which disturbs the on-site potential 
203: through the Madelung energy. 
204: Since large fluctuations in the competition between different orders are crucial 
205: in this study, we employ the Monte Carlo (MC) method in which the 
206: configurations of $\{ S_i \}$ and $\{ Q_i \}$ are stochastically sampled 
207: \cite{Yunoki1998a}. 
208: In the presence of the disorder, we take the random average 
209: for different configurations of $\{ \varepsilon_i \}$. 
210: 
211: We consider model (\ref{eq:H}) on the 2D square lattice 
212: in the half-doped case with $x=0.5$ 
213: ($0.5$ electron per site on average). 
214: We set the half-bandwidth $W=4t$ as an energy unit, and 
215: take $\lambda=0.1$ throughout this paper. 
216: A similar model was studied 
217: in a different context for the low-$x$ regime 
218: in the absence of the disorder 
219: \cite{Verges2002}. 
220: We consider the binary-type distribution of the random potential, 
221: $\varepsilon_i = \pm \Delta$.  
222: We apply the systematic analysis on the finite-size effect 
223: by using the series of $L \times L$-site systems with $L=4,6,8$
224: to distinguish the long-range ordering from short-range correlations. 
225: The numerical details will be reported elsewhere 
226: \cite{MotomePREPARATION}. 
227: 
228: 
229: %%%%%%%%%%%%%%%%%%% Results and Discussions
230: Figure~\ref{fig:phase} (a) shows the phase diagram of model (\ref{eq:H}) 
231: in the absence of the disorder ($\Delta=0$). 
232: Note that the abscissa is reversed for comparison with experimental results. 
233: At $g=0$, the electronic part of model (\ref{eq:H}) 
234: becomes a pure DE model which exhibits 
235: the FM state below $T_{\rm C}^{(0)}$ 
236: (the superscript $(0)$ represents the case of $\Delta=0$). 
237: As $g$ increases, $T_{\rm C}^{(0)}$ decreases and 
238: the checkerboard-type COI concomitant with the cooperative lattice distortion 
239: appears below $T_{\rm CO}^{(0)}$. 
240: In the intermediate region where $T_{\rm C}^{(0)}$ and $T_{\rm CO}^{(0)}$ 
241: intersect, we have the phase with both the ferromagnetism and 
242: the charge/lattice ordering (F+CO phase). 
243: Thus, the phase diagram exhibits the tetracritical behavior
244: where four different phases meet at one point. 
245: The density of states (DOS) shown in Fig.~\ref{fig:phase} (d) indicates that 
246: the phases below $T_{\rm CO}^{(0)}$ including the F+CO phase 
247: have a finite energy gap. 
248: This tetracritical behavior can be turned into 
249: the bicritical one when we include a competing term 
250: between the ferromagnetism and the charge/lattice ordering by hand 
251: \cite{MotomePREPARATION}, 
252: and the following results on the disorder effects are 
253: qualitatively similar in these multicritical phenomena. 
254: 
255: \begin{figure}
256: \includegraphics[width=7.5cm]{fig2.eps}
257: %\includegraphics[width=10cm]{fig2.eps}
258: \caption{
259: (a)-(c) Phase diagram of model (\ref{eq:H}) 
260: for different strength of the random potential
261: (a) $\Delta=0.0$, (b) $\Delta=0.1$, and (c) $\Delta=0.2$. 
262: The lines are guides for the eyes. 
263: The gray data and lines in (b) and (c) are the results in (a) for comparison. 
264: (d)-(g) Density of states at $T=0.016$ for $L=8$ clusters: 
265: (d) $\Delta=0.0$ for various $g$; 
266: (e) $g=0.4$, (f) $g=1.6$, and (g) $g=0.8$ by varying $\Delta$. 
267: The Fermi energy is at $\omega=0$. 
268: Only the typical error is shown for each case. 
269: } 
270: \label{fig:phase} 
271: \end{figure}
272: 
273: The emergence of the F+CO phase is due to the robustness of 
274: the DE ferromagnetic interaction even in the insulating state. 
275: The DE interaction is effective
276: if the electron wavefunction spreads, at least, for a finite spatial range. 
277: We note that this phase is also predicted in more realistic DE model for 
278: manganites \cite{AliagaPREPRINT,Yunoki1998b}, and 
279: is indeed observed in recent experiments \cite{Loudon2002}. 
280: 
281: When we switch on the random potential, the phase diagram shows a drastic 
282: change. 
283: By introducing weak disorder with $\Delta=0.1$ (Fig.~\ref{fig:phase} (b)), 
284: $T_{\rm CO}$ is suppressed remarkably, 
285: whereas $T_{\rm C}$ is lowered only slightly. 
286: Accordingly, the tetracritical point shifts 
287: toward larger $g$ and lower $T$. 
288: For stronger disorder with $\Delta=0.2$ (Fig.~\ref{fig:phase} (c)), 
289: $T_{\rm CO}$ disappears down to the lowest temperatures 
290: in the present MC calculations ($T \simge 0.008$). 
291: This clearly shows the fragility of COI
292: against the random potential compared to the DE ferromagnetism 
293: \cite{AliagaPREPRINT}. 
294: As a result of this asymmetric effect, 
295: there appears the regime where COI 
296: is turned into FM by introducing the random potential, 
297: which is not clearly found in the previous study 
298: \cite{AliagaPREPRINT}. 
299: These features are in good agreement with the experimental results 
300: \cite{AkahoshiPREPRINT,TomiokaPREPRINT}. 
301: 
302: The DOS gives insights on the nature of these disorder effects. 
303: As shown in Fig.~\ref{fig:phase} (e) and (f), 
304: DOS in both the small-$g$ FM regime and the large-$g$ COI regime are
305: not much affected by the random potential. 
306: Especially, Fig.~\ref{fig:phase} (f) indicates that 
307: the energy gap in the large-$g$ regime persists 
308: even when the charge/lattice long-range order is destroyed 
309: by the disorder. 
310: This means that a short-range correlation remains and 
311: the local lattice distortion persists to open the gap. 
312: Therefore, the collapse of the charge/lattice ordering 
313: is not due to the disappearance of 
314: the amplitude of the lattice distortion 
315: but due to the disturbance of the phase of the commensurate ordering
316: with the wavenumber ($\pi,\pi$). 
317: The random potential acts as the `random field' to the CO order parameter, 
318: and its phase coherence is lost.
319: On the contrary, the ferromagnetic order is robust against such pinning 
320: since $T_{\rm C}$ is essentially determined by the kinetic energy of electrons. 
321: These give a comprehensive understanding 
322: of the asymmetrical effect of the disorder on the phase diagram. 
323: 
324: The most remarkable behavior of DOS is obtained near the phase boundary 
325: between the FM state and the F+CO insulating state. 
326: In the F+CO phase but close to the phase boundary at $\Delta=0$, 
327: as shown in Fig.~\ref{fig:phase} (g), 
328: a small but clear gap in DOS below $T_{\rm CO}^{(0)}$ is rapidly collapsed 
329: by introducing the disorder and the system appears to be metallic. 
330: This indicates that the random potential easily destroys COI and induces FM 
331: in the proximity of the metal-insulator transition. 
332: This result also agrees well with experimental results 
333: \cite{AkahoshiPREPRINT,TomiokaPREPRINT}, 
334: which strongly suggests that the simple model (\ref{eq:H}) captures 
335: the essential physics of this competition between FM and CO in manganites. 
336: 
337: \begin{figure}
338: \includegraphics[width=7.5cm]{fig3.eps}
339: %\includegraphics[width=13cm]{fig3.eps}
340: \caption{
341: Optical conductivity for $g=0.8$ at (a) $T=0.02$ and (b) $T=0.06$ ($L=6$). 
342: (c) Average of the square of the lattice distortion ($L=8$). 
343: See the text for details. 
344: (d) Schematic picture of the temperature dependence of 
345: the resistivity.
346: } 
347: \label{fig:conductivity} 
348: \end{figure}
349: 
350: Above $T_{\rm C}$ of the disorder-induced metallic state, 
351: the CMR is found to be much enhanced experimentally
352: \cite{AkahoshiPREPRINT,TomiokaPREPRINT}. 
353: There, the system is insulating and 
354: shows a large drop of the resistivity near $T=T_{\rm C}$
355: toward the low-$T$ metallic phase. 
356: Our results are satisfactorily consistent with this behavior: 
357: Figures~\ref{fig:conductivity} (a) and (b) show the optical conductivity
358: $\sigma(\omega)$ for $T < T_{\rm C}$ and $T > T_{\rm C}$, respectively, 
359: which are obtained by the Kubo formula. 
360: In the low-$T$ regime, $\sigma(\omega)$ shows a disorder-induced 
361: insulator to metal transition which is consistent with 
362: DOS in Fig.~\ref{fig:phase} (g). 
363: On the contrary, at the high-$T$ regime, the disorder substantially reduces
364: the low-energy weight and develops a quasi-gap structure in $\sigma(\omega)$, 
365: namely, it makes the system more insulating.
366: These tendencies are also indicated in the magnitude of the lattice distortion, 
367: $\langle \sum_i Q_i^2 \rangle / L^2$, as shown in Fig.~\ref{fig:conductivity} (c),
368: which correlates with the localization of electrons. 
369: The distortion increases for $T > T_{\rm C}$ 
370: while it decreases for $T < T_{\rm C}$ 
371: when the random potential is introduced. 
372: These contrastive effects of the disorder suggest 
373: the temperature dependence of the resistivity $\rho$ 
374: as schematically drawn in Fig.~\ref{fig:conductivity} (d). 
375: In the absence of the disorder, the system causes the COI transition 
376: at $T_{\rm CO}^{(0)}$, where $\rho$ shows a sharp upturn. 
377: The disorder enhances the insulating nature in the high-$T$ regime, 
378: hence $\rho$ becomes larger 
379: while it does not show any anomaly corresponding to the COI transition. 
380: Below $T_{\rm C}$, the disorder induces FM, therefore
381: $\rho$ should show a large drop at $T \sim T_{\rm C}$ 
382: as shown in the figure. 
383: The large change of $\rho$ potentially leads to a huge response 
384: to the external magnetic field as is indeed observed in the CMR manganites 
385: \cite{AkahoshiPREPRINT,TomiokaPREPRINT}. 
386: Thus, the origin of the enhanced CMR can be attributed to 
387: the contrastive influence of the disorder
388: in low-$T$ and high-$T$ regimes near the metal-insulator phase boundary. 
389: 
390: \begin{figure}
391: \includegraphics[width=8.5cm]{fig4.eps}
392: %\includegraphics[width=14cm]{fig4.eps}
393: \caption{
394: (a) Temperature dependence of the standard deviation of 
395: the square of the local lattice distortion at $\Delta=0.2$. 
396: Transition temperatures for FM (COI) in the absence of the disorder 
397: are indicated by the gray (black) arrows. 
398: (b) Susceptibility for the ($\pi,\pi$) lattice ordering at $g=1.0$. 
399: Both are for $L=8$ clusters.
400: } 
401: \label{fig:rough,kai} 
402: \end{figure}
403: 
404: We discuss here this insulating state for $T > T_{\rm C}$ in more detail. 
405: In the percolation scenario for the CMR \cite{Moreo1999}, 
406: the system is supposed to be a coexisting mixture of 
407: the metallic and the insulating clusters below 
408: the FM or COI transition temperatures in the absence of the disorder, 
409: $T_{\rm C}^{(0)}$ and $T_{\rm CO}^{(0)}$, respectively. 
410: However, our simulation does not show any clear indication of this coexistence. 
411: Figure~\ref{fig:rough,kai} (a) shows the standard deviation of 
412: the square of the lattice distortion 
413: $\langle Q_i^2 \rangle$ among the sites 
414: as a function of the temperature. 
415: The standard deviation changes monotonously 
416: and remains small (compare to the values in Fig.~\ref{fig:conductivity} (c)). 
417: This is contrary to the percolation picture 
418: in which we expect an enhancement for $T < T_{\rm C}^{(0)}$ or 
419: $T_{\rm CO}^{(0)}$  
420: due to the formation of the metallic (smaller $\langle Q_i^2 \rangle$) and 
421: the insulating (larger $\langle Q_i^2 \rangle$) clusters. 
422: Instead of the cluster formation, we find fluctuations or 
423: short-range correlations of the charge/lattice ordering. 
424: Figure~\ref{fig:rough,kai} (b) shows the susceptibility for the ($\pi,\pi$) 
425: lattice ordering. 
426: In the absence of the disorder, the susceptibility shows 
427: a sharp cusp at $T=T_{\rm CO}^{(0)}$, which leads to the divergence 
428: in the thermodynamic limit $L \rightarrow \infty$. 
429: Although the disorder destroys COI and makes the cusp obscure, 
430: the fluctuation remains finite and is enhanced as $T \rightarrow T_{\rm C}$. 
431: The contrastive behavior between above and below $T_{\rm C}$ 
432: in the disordered case is in accord with 
433: this temperature dependence of $\chi$. Namely 
434: the CO fluctuation is enhanced towards $T_{\rm C}$ from above while 
435: it is suppressed at the lowest temperature. This could be regarded as
436: the reminiscence of the multicritical phenomenon in the pure case.
437: This is also in good agreement with the diffuse scattering experiments 
438: which indicate a large fluctuation of the 
439: charge/lattice ordering above $T_{\rm C}$ 
440: \cite{TomiokaPREPRINT,Shimomura1999}. 
441: We note that this fluctuation is suppressed as $\Delta$ increases, 
442: which may account for the deviation from the scaling law 
443: \cite{MurakamiPREPRINT}. 
444: Therefore, our results indicate that the phase above $T_{\rm C}$ is not 
445: likely the inhomogeneous mixture of the static clusters
446: but a rather homogenous state with the thermodynamic
447: charge/lattice fluctuations. 
448: 
449: 
450: %%%%%%%%%%%%%%%%%%% Acknowledgment
451: 
452: The authors acknowledge Y. Tokura, Y. Tomioka, and E. Dagotto 
453: for fruitful discussions. 
454: This work is supported by Grant-in-Aids from the Ministry of Education,  
455: Culture, Sports, Science, and Technology. 
456: 
457: 
458: \noindent
459: $^*$ Present address: RIKEN (The Institute of Physical and Chemical Research), 
460: 2-1 Hirosawa, Wako, Saitama 351-0198, Japan.
461: 
462: 
463: 
464: %%%%%%%%%%%%%%%%%%% References
465: \begin{thebibliography}{99}
466: 
467: 
468: \bibitem{Tokura2000}
469: Y. Tokura and N. Nagaosa, Science {\bf 288}, 462 (2000) 
470: and references therein.
471: 
472: \bibitem{Dagotto2001}
473: E. Dagotto, T. Hotta, and A. Moreo, Phys. Rep. {\bf 344}, 1 (2001) 
474: and references therein. 
475: 
476: \bibitem{Furukawa1999}
477: N. Furukawa, in {\it Physics of Manganites}, 
478: ed. T. Kaplan and S. Mahanti (Plenum Publishing, New York, 1999) 
479: and references therein. 
480: 
481: \bibitem{Millis1995}
482: A. J. Millis, P. B. Littlewood, and B. I. Shraiman, 
483: Phys. Rev. Lett. {\bf 74}, 5144 (1995).
484: 
485: \bibitem{MurakamiPREPRINT}
486: S. Murakami and N. Nagaosa, Phys. Rev. Lett. {\bf 90}, 197201 (2003).
487: 
488: \bibitem{Mori1998}
489: S. Mori, C. H. Chen, and S-W. Cheong, Phys. Rev. Lett. {\bf 81}, 3972 (1998); 
490: M. Uehara {\it et al.}, Nature {\bf 399}, 560 (1999). 
491: 
492: \bibitem{Moreo1999}
493: A. Moreo, S. Yunoki, and E. Dagotto, Science {\bf 283}, 2034 (1999); 
494: J. Burgy {\it et al.}, Phys. Rev. Lett. {\bf 87}, 277202 (2001). 
495: 
496: \bibitem{AkahoshiPREPRINT}
497: D. Akahoshi {\it et al.}, Phys. Rev. Lett. {\bf 90}, 177203 (2003).
498: \label{AkahoshiPRE}
499: 
500: \bibitem{Nakajima2002}
501: T. Nakajima {\it et al.}, J. Phys. Soc. Jpn. {\bf 71}, 2843 (2002). 
502: 
503: \bibitem{Millange1998}
504: F. Millange {\it et al.}, Chem. Mater. {\bf 10}, 1974 (1998).
505: 
506: \bibitem{TomiokaPREPRINT}
507: Y. Tomioka {\it et al.}, Phys. Rev. B, in press.
508: 
509: \bibitem{Barnabe1997}
510: A. Barnabe {\it et al.}, Appl. Phys. Lett. {\bf 71}, 3907 (1997); 
511: T. Kimura {\it et al.}, Phys. Rev. Lett. {\bf 83}, 3940 (1999). 
512: 
513: \bibitem{Motome2001}
514: Y. Motome and N. Furukawa, J. Phys. Soc. Jpn. {\bf 70}, 1487 (2001); 
515: {\it ibid.} {\bf 70}, 2802 (2001). 
516: The Ising symmetry enables us to survey finite-temperature transitions 
517: in two-dimensional systems in which the systematic results can be obtained 
518: in smaller lattice sizes compared to three-dimensional cases. 
519: 
520: \bibitem{Yunoki1998a}
521: S. Yunoki {\it et al.}, Phys. Rev. Lett. {\bf 80}, 845 (1998).
522: 
523: \bibitem{Verges2002}
524: J. A. Verg\'es, V. Mart\'{\i}n-Mayor, and L. Brey, 
525: Phys. Rev. Lett. {\bf 88}, 136401 (2002).
526: 
527: \bibitem{MotomePREPARATION}
528: Y. Motome, N. Furukawa, and N. Nagaosa, preprint (cond-mat/0304542).
529: 
530: \bibitem{AliagaPREPRINT}
531: H. Aliaga {\it et al.}, preprint (cond-mat/0303513).
532: 
533: \bibitem{Yunoki1998b}
534: S. Yunoki, A. Moreo, and E. Dagotto, Phys. Rev. Lett. {\bf 81}, 5612 (1998).
535: 
536: \bibitem{Loudon2002}
537: J. C. Loudon, N. D. Mathur, and P. A. Midgley, Nature {\bf 420}, 797 (2002).
538: 
539: \bibitem{Shimomura1999}
540: S. Shimomura {\it et al.}, Phys. Rev. Lett. {\bf 83}, 4389 (1999);
541: S. Shimomura {\it et al.}, Phys. Rev. B {\bf 62}, 3875 (2000).
542: 
543: \end{thebibliography}
544: 
545: %%%%% FIGURES
546: 
547: 
548: \end{document}
549: