1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %%%%%%% FRONT PAGE
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4: % chose the style
5: %
6: %\documentstyle[preprint,tighten,floats,eqsecnum,pra,prb,pre,prl,aps]{revtex}
7: % aps -> general APS format
8: % preprint -> preprint style output (full pages)
9: % without this specification: galley style output
10: %(columns)
11: % tighten -> single-spaced output with preprint style
12: % floats -> floating figures and tables
13: % eqsecnum -> number equations by section
14: % !! use it only if you have sections
15: % prx -> special style for these journals
16: % prb: uses superscripts instead of [ ] for citations
17: % prl: for correct linecount in galley style
18: %
19: % for preprint version:
20: %RP23 \documentstyle[preprint,prl,aps]{revtex}
21: % for journal version:
22: %\documentstyle[twocolumn,prl,aps,epsf]{revtex}
23: %\documentstyle[prb,aps,multicol]{revtex}
24: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
25: \documentstyle[prl,aps,twocolumn,epsf,epsfig]{revtex}
26: %\documentstyle[aps,preprint]{revtex}
27: \begin{document}
28: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname@twocolumnfalse\endcsname
29: \draft
30:
31: \title{Magnetic Phase Diagram of Ca$_{2-x}$Sr$_x$RuO$_4$ Governed by
32: Structural Distortions}
33:
34: \author{Z. Fang$^1$ and K. Terakura$^{2,3}$}
35:
36: \address{ $^1$JRCAT, Angstrom Technology Partnership, 1-1-4 Higashi,
37: Tsukuba, Ibaraki 305-0046, Japan}
38:
39: \address{ $^2$JRCAT, National Institute for Advanced Interdisciplinary
40: Research, 1-1-4 Higashi, Tsukuba, Ibaraki 305-8562, Japan}
41:
42: \address{ $^3$ Tsukuba Advanced Computing Center, 1-1-4 Higashi,
43: Tsukuba, Ibaraki 305-8561, Japan}
44:
45: \date{\today }
46: \maketitle
47:
48:
49: \begin{abstract}
50: We constructed, by the first-principles calculations, a magnetic
51: phase diagram of Sr$_{2}$RuO$_4$ in the space spanned by structural
52: distortions. Our phase diagram can qualitatively explain the
53: experimental one for Ca$_{2-x}$Sr$_x$RuO$_4$. We found that the
54: rotation and the tilting of RuO$_6$ octahedron are responsible for
55: the ferro- and antiferro-magnetism, respectively, while the
56: flattening of RuO$_6$ is the key factor to stabilize those magnetic
57: ground states. Our results imply that the magnetic and the
58: structural instabilities in Sr$_2$RuO$_4$ are closely correlated
59: cooperatively rather than competitively.
60: \end{abstract}
61:
62: \pacs{PACS numbers: 75.30.Kz, 74.70.-b, 71.27.+a}
63: ]
64:
65: \narrowtext
66: %\newpage
67: %\twocolumn
68: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
69: %%%%%%% BODY OF TEXT
70: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
71: % some useful remarks
72: % use \lowercase{#} in \section{} if # should be set in lowercase
73: % in the uppercase heading
74: % use \protect\\ to force a linebreak in heading
75: % use \section*{} etc. to suppress the numbering
76: % use \text{} if you want to insert text in math mode
77: % use Refs.\ \onlinecite{} to pull down superscript reference numbers
78: % use \cite{a,b,c,d,e,f} rather then \cite{a}-\cite{f} and
79: % break a hughe citation sequence as \cite{a,b,c,%
80: % d,e,f} before starting a new line
81: % use \ref{} for referencing to figures, tables, and sections
82: %%%%% possible subdivisions
83: % \section{}
84: % \subsection{}
85: % \subsubsection{}
86: % \paragraph{}
87: % use \label{} immediately after the \section{} etc. commands
88: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
89:
90: Both the magnetic and the structural instabilities are essential
91: issues for the unconventional superconductivity in
92: Sr$_2$RuO$_4$~\cite{nature,NMR,Impurity,time_reversal,flux,Andreev},
93: which is the only example of a noncuprate layered perovskite
94: superconductor. It was first suggested that the Sr$_2$RuO$_4$ is
95: close to the ferromagnetic (FM) instability~\cite{p-wave} with strong
96: FM spin fluctuations, which may naturally lead to a spin-triplet
97: $p$-wave pairing mechanism~\cite{p-wave,Mazin1,Tewordt,p-wave2}.
98: However, the recent observation~\cite{Sidis} of sizable
99: antiferromagnetic (AF) incommensurate spin fluctuation, due to the
100: Fermi surface nesting~\cite{Mazin2}, indicates that more careful
101: studies are needed. As for the structural aspect, it was pointed out
102: by experiments that Sr$_2$RuO$_4$ is very close to the structural
103: instability with respect to the RuO$_6$ rotation~\cite{phonon}. With
104: such a situation, one may consider that three kinds of instabilities,
105: superconducting, magnetic and structural ones, may compete.
106: Nevertheless, the correlation among those instabilities has not been
107: fully discussed. It was found recently~\cite{plummer} that the
108: cleaved surface of this material is reconstructed to form the c(2x2)
109: structure which can be regarded as the frozen RuO$_6$ rotation
110: mentioned above. Furthermore, the density-functional calculation
111: predicts that the surface ferromagnetism is strongly stabilized by the
112: structural reconstruction~\cite{plummer}. This prediction suggests
113: that the structural and magnetic instabilities cooperate rather than
114: compete, although the surface ferromagnetism has not been
115: experimentally confirmed up to now.
116:
117: On the other hand, the recent studies on the Ca$_{2-x}$Sr$_x$RuO$_4$
118: suggest strongly the cooperative feature of the structural and
119: magnetic instabilities in the bulk. Moreover, the system shows a very
120: rich phase diagram and provides us with an opportunity to analyze the
121: correlation between the magnetism and the structure more extensively.
122: Below is a brief description of the experimental observation for
123: Ca$_{2-x}$Sr$_x$RuO$_4$ by Nakatsuji {\it et al}.~\cite{Nakatsuji}.
124: With the Ca substitution for Sr, the system is successively driven
125: from the non-magnetic (NM) 2-dimensional (2D) Fermi liquid
126: ($x\sim2.0$) to a nearly FM metal ($x\sim0.5$), an
127: antiferromagnetically correlated metal ($0.2<x<0.5$), and finally an
128: AF insulator ($x<0.2$). Since the substitution is isovalent, the
129: dominant effects are the structural modifications due to the reduced
130: ionic size of Ca compared with Sr. Evidence~\cite{Friedt} has been
131: presented by neutron scattering that the structural distortions caused
132: by the Ca substitution correlate with the changes in the magnetic and
133: the electronic properties.
134:
135: The main aim of this letter is to study how and why the magnetism of
136: Ca$_{2-x}$Sr$_x$RuO$_4$ is affected by structural distortions. In
137: order to extract essential aspects, we assume that for a given crystal
138: structure, the electronic structure is not affected by the relative
139: content of Ca and Sr. Therefore, in the following, we study the
140: stable magnetic phases of Sr$_2$RuO$_4$ for given structural
141: distortions. Three types of structure distortions, i.e. RuO$_6$
142: octahedron {\it rotation} about the $c$-axis, RuO$_6$ {\it tilting}
143: around an axis parallel to the edge of octahedron basal plane and the
144: {\it flattening} of RuO$_6$ along the $c$-axis are identified from
145: experiments~\cite{Friedt}. Our phase diagram can qualitatively explain
146: the experimental phase diagram of Ca$_{2-x}$Sr$_x$RuO$_4$,
147: demonstrating the crucial roles of structural distortions for the
148: tuning of electronic and magnetic properties, and further supporting
149: our previous prediction for the surface . In particular, we found that
150: the RuO$_6$ rotation can enhance the FM instability significantly,
151: while the combination of tilting and rotation of RuO$_6$ is
152: responsible for the enhancement of AF instability. Furthermore, we
153: point out that the flattening of RuO$_6$ is a key factor to stabilize
154: the magnetic (both FM and AF) ground states. The basic physics
155: governing the phase diagram can be understood in terms of the strong
156: coupling between the lattice and the magnetism through the orbital
157: degrees of freedom. Our results strongly suggest that, in
158: Sr$_2$RuO$_4$, the magnetic fluctuations can be significantly enhanced
159: by the structural fluctuations, implying the necessity of
160: reconsidering the coupling mechanism in the bulk superconductivity.
161:
162: The calculations were performed with the first-principles plane-wave
163: basis pseudopotential method based on the local density approximation
164: (LDA). The validity of LDA treatment for ruthenates was demonstrated
165: in Ref.~\cite{Mazin1,Mazin3}. The $2p$ states of oxygen and $4d$
166: states of Ru are treated by the Vanderbilt ultrasoft
167: pseudopotential~\cite{PP}, while the norm-conserving scheme~\cite{NC}
168: is used for other states. The cutoff energy for the wave function
169: expansion is 30~Ry. The $k$-point sampling of the Brillouin zone (BZ)
170: was well checked to provide enough precision in the calculated total
171: energies. The theoretically optimized lattice parameters
172: $a=3.84$~\AA~and $c=12.70$~\AA~ for the bulk Sr$_2$RuO$_4$ are in good
173: agreement with the experimental ones $a=3.86$~\AA~ and
174: $c=12.73$~\AA. The degree of flattening of RuO$_6$ octahedron $\lambda
175: $ is defined by $\lambda = d_c / d_{ab}$ with $d_c$ ($d_{ab}$)
176: denoting the Ru-O bond length along the $c$-axis (in the $ab$-plane)
177: with the RuO$_6$ volume fixed. Rotation and tilting of RuO$_6$
178: octahedron are operated with the Ru-O bond lengths fixed. In order to
179: construct the magnetic phase diagram, the lowest energy magnetic phase
180: for each crystal structure is searched for among different (NM, FM and
181: AF) phases. In the present work, we focus our attention only on
182: phases described within the c(2x2) unit cell. The soft phonon mode of
183: $\Sigma _3$ at the zone boundary in Sr$_2$RuO$_4$~\cite{phonon} and
184: the AF state of Ca$_2$RuO$_4$~\cite{CRO} are in this category.
185:
186: Figure 1 shows the calculated phase diagram~\cite{comment1}, while the
187: Table 1 summarizes the calculated total energies and magnetic moments
188: for some particular points in the phase diagram. Hereafter, $\phi$
189: and $\theta$ denote the rotation angle and the tilting angle,
190: respectively. The apical oxygen and the oxygen in the $ab$-plane are
191: called O(2) and O(1) respectively. From right to left of the phase
192: diagram, first the RuO$_6$ starts to rotate along the $c$-axis by up
193: to 12$^{\circ}$, and then with the 12$^{\circ}$ rotation being fixed,
194: the RuO$_6$ starts to tilt up to 12$^{\circ}$. The structural
195: analysis by the neutron scattering~\cite{Friedt} allows us to make a
196: one-to-one correspondence between the structural changes, i.e. the
197: horizontal axis of our phase diagram, and the doping level $x$ in
198: Ca$_{2-x}$Sr$_x$RuO$_4$. For $x=2.0$ (Sr$_2$RuO$_4$), the system has
199: $I4/mmm$ symmetry with $\phi=\theta=0^\circ$, corresponding to the
200: right end of our phase diagram. With reduction of $x$, RuO$_6$ starts
201: to rotate and the symmetry is reduced to $I4_1/acd$ until $x=0.5$
202: (Ca$_{1.5}$Sr$_{0.5}$RuO$_4$), where $\phi=12.78^\circ$ and
203: $\theta=0^\circ$ at 10 K. With further reduction of $x$, RuO$_6$
204: starts to tilt and the symmetry is further reduced to $Pbca$ until
205: $x=0.0$ (Ca$_2$RuO$_4$) where $\phi=11.93^\circ$ and
206: $\theta\sim12^\circ$ at low temperature, corresponding to the left end
207: of our phase diagram. It was also pointed out by the
208: experiment~\cite{Friedt} that, from $x=2.0$ to $x=0.5$, the degree of
209: flattening $\lambda$ remains almost constant ($\sim 1.07$), while from
210: $x=0.5$ to $x=0.0$, the rotation angle $\phi$ is almost unchanged
211: ($\sim 12^\circ$). Three representative experimental points are shown
212: in our phase diagram by triangles. Now, the basic tendency suggested
213: by our phase diagram is that the RuO$_6$ rotation will drive the
214: system from a NM state to a FM state, while the subsequent tilting
215: plus the flattening of RuO$_6$ will push the system to an AF region.
216: This general tendency is quite consistent with the experimental
217: results. It should be noted that the rich experimental phase diagram
218: can be simply understood in terms of the close coupling between
219: structural distortions and magnetism. Although the real long range FM
220: ordering in Ca$_{1.5}$Sr$_{0.5}$RuO$_4$ is not confirmed yet, the
221: significant enhancement of spin susceptibility in this doping level
222: undoubtedly implies the strengthening of FM instability. Another
223: important aspect in our phase diagram is that the flattening of
224: RuO$_6$ is so important not only for the AF state but also for the FM
225: state. This suggests that {\it simply by uniaxial pressure, the
226: Sr$_2$RuO$_4$ can be driven from the NM state to a FM state}.
227:
228: The basic questions concerning our phase diagram are: 1) Why are the
229: RuO$_6$ rotation and tilting correlated with the tendency to the FM
230: and AF states? 2) Why is the RuO$_6$ flattening so important for the
231: magnetic solutions? Before answering these questions, let us discuss
232: the role of each $4d$ orbital in the electronic properties of
233: Sr$_2$RuO$_4$, which is essential to our later discussions. The three
234: Ru $t_{2g}$ orbitals ($d_{xy}$, $d_{yz}$, $d_{zx}$) hybridize with
235: each other only very weakly in tetragonal Sr$_2$RuO$_4$. Therefore
236: each orbital plays distinct roles. The projected density of states
237: (DOS) shown in Fig.2(a) indicates that the $d_{xy}$ orbital
238: contributes dominantly to the well-known van Hove singularity (VHS)
239: just above the Fermi level. The $\gamma$ Fermi surface has the
240: character of $d_{xy}$. It is mostly responsible to FM spin
241: fluctuation due to the high DOS around the Fermi level. On the other
242: hand, $d_{yz}$ and $d_{zx}$ orbitals contribute to the $\alpha$ and
243: $\beta$ Fermi surfaces and produce the incommensurate spin fluctuation
244: coming from the strong nesting effect due to the quasi-one-dimensional
245: nature of those states. The calculated bare spin susceptibility shown
246: in Fig. 3(a) for the undistorted compound has the incommensurate peak
247: at {\bf Q}=($2\pi/3a, 2\pi/3a$), being consistent with previous
248: calculations~\cite{Mazin2}. In real materials, the two factors,
249: i.e. the FM instability due to the high DOS at the Fermi level
250: combined with the $q$ dependent Stoner factor~\cite{Mazin1,Mazin2} and
251: the AF instability due to the nesting effect~\cite{Mazin2}, will
252: compete.
253:
254:
255: The RuO$_6$ rotation couples mostly with the $d_{xy}$ orbital but not
256: with the $d_{yz}$, $d_{zx}$ orbitals because the $pd \pi$ type
257: hybridization between the O(1)-$2p$ and the $d_{xy}$ states will be
258: significantly reduced by the RuO$_6$ rotation, but those between the
259: O-$2p$ and the $d_{yz}$, $d_{zx}$ states are not affected so much. The
260: direct results of this reduced $pd \pi$ type hybridization between the
261: O(1)-$2p$ and the $d_{xy}$ states are, first the narrowing of $d_{xy}$
262: band width and second the downward shift of $d_{xy}$ band, as shown in
263: Fig. 2(b) (about 0.4 eV narrowing and 0.1 eV downward shift of
264: $d_{xy}$ band for $\phi=12^{\circ}, \lambda=1.07$). As the latter
265: brings the VHS closer to the Fermi level, both of the two results will
266: enhance the DOS at the Fermi level. Another effect coming from the
267: downward shift of the $d_{xy}$ states is the population reduction in
268: the $d_{yz}$, $d_{zx}$ states, which may shift the Fermi surface
269: nesting vector closer to the zone boundary. However, the increase of
270: the DOS at the Fermi level is the dominant effect and the tendency
271: towards FM instability is enhanced by the RuO$_6$ rotation. Once
272: tilting is additionally introduced, all of the $t_{2g}$ bands will
273: become narrower. This will enhance the nesting effect and enhance the
274: AF instability. The discussion so far can answer the first question.
275:
276: Now let us discuss the effects of RuO$_6$ flattening. There are two
277: factors also. First, with the flattening of RuO$_6$ octahedron, the
278: Ru-O(1) bond length will increase, while the Ru-O(1)-Ru angle remains
279: $180^{\circ}$. The increased bond length will reduce all the $pd \pi$
280: type hybridizations between the O(1)-$2p$ and the $d_{xy}$, $d_{yz}$,
281: $d_{zx}$ states. Therefore, width of all the bands of three
282: Ru-$t_{2g}$ states is reduced as shown in Fig.2(c) (about 0.4 eV for
283: $d_{yz,zx}$ bands and 0.3 eV for $d_{xy}$ band for $\lambda=0.96$),
284: making the DOS at the Fermi level higher. This will favor the FM
285: solution. Another very important results of flattening is the orbital
286: polarization. It is obvious that the tetragonal distortion by the
287: flattening will populate the $d_{xy}$ states and depopulate the
288: $d_{yz}$, $d_{zx}$ states (about 0.2 eV downward shift of $d_{xy}$
289: band for $\lambda=0.96$). The effect is similar to the RuO$_{6}$
290: rotation. The orbital polarization due to flattening will also shift
291: the nesting vector to the zone boundary as shown in the susceptibility
292: calculations (Fig.3). This will favor the commensurate AF state of the
293: system. The net effect by rotation, tilting and flattening of the
294: RuO$_6$ will depend on the competition among them and the phase
295: diagram of Fig.1 demonstrates the situation in a space spanned by
296: those distortion modes. Figure 2(f) shows the calculated DOS for the
297: AF state with $\phi=\theta=12^\circ$, i.e. almost the experimental
298: structure of Ca$_2$RuO$_4$. It is clear in this case that the occupied
299: minority spin states mostly come from the $d_{xy}$ orbital due to
300: flattening of RuO$_6$. Therefore, the strong superexchange interaction
301: between the occupied majority-spin and unoccupied minority-spin
302: $d_{yz}$, $d_{zx}$ orbitals will stabilize the AF ground
303: state~\cite{comment2}.
304:
305:
306: In summary, by constructing a phase diagram of Sr$_2$RuO$_4$ with
307: structural distortions, we find the strong coupling between the
308: lattice and the magnetism. Our phase diagram can qualitatively explain
309: the experimental phase diagram of Ca$_{2-x}$Sr$_x$RuO$_4$. We
310: demonstrate that the RuO$_6$ rotation will enhance the FM instability
311: in the system, while the tilting plus the flattening of RuO$_6$ make
312: the system AF. We pointed out that the flattening of RuO$_6$ is so
313: important not only for the AF state but also for the FM state. An
314: important implication of our results is that the magnetic and the
315: structural instabilities in Sr$_2$RuO$_4$ should be strongly
316: correlated. The structure fluctuation and the magnetic fluctuation
317: cooperate. Actually the phonon mode corresponding to the RuO$_6$
318: rotation is quite soft~\cite{phonon} in the bulk, and this rotation
319: will enhance the FM instability. All these results imply the necessity
320: of reconsidering the coupling mechanism for the unconventional
321: superconductivity. In this context, we propose a possible way to
322: identify experimentally the relationship between the FM fluctuation
323: and the superconducting state. As the uniaxial compression of
324: Sr$_2$RuO$_4$ will enhance the FM fluctuation without introducing the
325: disorder, the variation of superconducting transition temperature
326: against uniaxial compression may provide important information.
327:
328: The authors thank Dr. S. Nakatsuji, Prof. E. W. Plummer,
329: Prof. Y. Tokura, Dr. R. Matzdorf, and Dr. J. D. Zhang for valuable
330: discussions and for providing us with their experimental data. One of
331: the anthors (Z. Fang) also thanks Dr. David Singh for valuable
332: discussions. The present work is partly supported by NEDO.
333:
334:
335: \begin{thebibliography}{50}
336: \bibitem{nature} Y.~Maeno, H.~Hashimoto, K.~Yoshida, {\it et al.},
337: Nature (London) {\bf 372}, 532 (1994).
338:
339: \bibitem{NMR} T.~Imai, A.~W.~Hunt, K.~R.~Thurber and F.~C.~Chou, Phys.
340: Rev. Lett. {\bf 81}, 3006 (1998).
341:
342: \bibitem{Impurity} A.~P.~Mackenzie {\it et al.}, Phys. Rev. Lett. {\bf
343: 80}, 161 and 3890(1998).
344:
345: \bibitem{time_reversal} G. M. Luke, Y. Fudamoto, K. M. Kojima and {\it
346: et al.}, Nature {\bf 394}, 558(1998).
347:
348: \bibitem{flux} T. M. Riseman, P. G. Kealey, E. M. Forgan and {\it et
349: al.}, Nature {\bf 396}, 242(1998).
350:
351: \bibitem{Andreev} F.~Laube, G.~Goll, H.~v.~L{\"o}hneysen and {\it et
352: al.}, Phys. Rev. Lett. {\bf 84}, 1595(2000).
353:
354: \bibitem{p-wave} T.~M.~Rice and M.~Sigrist, J. Phys. Condens. Matter
355: {\bf 7}, L643 (1995).
356:
357: \bibitem{Mazin1} I.~I.~Mazin and D.~J.~Singh, Phys. Rev. Lett. {\bf
358: 79}, 733 (1997).
359:
360: \bibitem{Tewordt} L.~Tewordt, Phys. Rev. Lett. {\bf 83}, 1007 (1999).
361:
362: \bibitem{p-wave2} K.~Ishida, H.~Mukuda, Y.~Kitaoka and {\it et al.},
363: Nature {\bf 396}, 658(1998).
364:
365: \bibitem{Sidis} Y.~Sidis, M. Braden, P.~Bourges and {\it et al.},
366: Phys. Rev. Lett. {\bf 83}, 3320 (1999).
367:
368: \bibitem{Mazin2} I.~I.~Mazin and D.~J.~Singh, Phys. Rev. Lett. {\bf
369: 82}, 4324 (1999).
370:
371: \bibitem{phonon} M.~Braden and W.~Reichardt, Phys. Rev. B {\bf 57},
372: 1236(1998).
373:
374: \bibitem{plummer} R.~Matzdorf, Z. Fang, X. Ismail and {\it et al.},
375: Science {\bf 289}, 746 (2000).
376:
377: \bibitem{Nakatsuji} S. Nakatsuji and Y. Maeno, Phys. Rev. Lett. {\bf
378: 84}, 2666(2000); {\it ibid}, Phys. Rev. B {\bf 62}, 6458 (2000).
379:
380: \bibitem{Friedt} O. Friedt, M. Braden, G. Andr\'{e} and {\it et al.},
381: cond-mat/0007218.
382:
383: \bibitem{Mazin3} I.~I.~Mazin and D.~J.~Singh, Phys. Rev. B {\bf 56},
384: 2556 (1997).
385:
386: \bibitem{PP} D. Vanderbilt, Phys. Rev. B {\bf 41}, 7892 (1990).
387:
388: \bibitem{NC} N.~Troullier and J.~L.~Martins, Phys. Rev. B {\bf 43},
389: 1993(1991).
390:
391: \bibitem{CRO} M. Braden, G. Andr\'{e}, S. Nakatsuji and Y. Maeno,
392: Phys. Rev. B {\bf 58}, 847 (1998).
393:
394: \bibitem{comment1} The experimental information [16] was adopted for
395: the structural changes in the phase diagram. However, the real
396: alloy system Ca$_{2-x}$Sr$_x$RuO$_4$ can be treated by the virtual
397: crystal approximation (VCA) [Z. Fang {\it et al.}, Phys. Rev. Lett.
398: {\bf 84}, 3169 (2000)]. Our test calculations in the LDA for
399: Ca$_{1.5}$Sr$_{0.5}$RuO$_4$ with experimental lattice parameters
400: give the FM ground state with the optimized RuO$_6$ rotation angle
401: $\phi$=9.5$^{\circ}$ and bond length ratio $\lambda$=1.076.
402:
403: \bibitem{comment2} For Ca$_2$RuO$_4$, the generalized gradient
404: approximation (GGA) may be better than LDA due to the narrowing of
405: bands. However, our test calculations for Ca$_2$RuO$_4$
406: ($\phi=\theta=12^{\circ}, \lambda=0.96$) suggested that the total
407: energy difference between the AF and the FM solutions changes only
408: slightly (from -17 meV in LDA to -20 meV in GGA), although a tiny
409: gap can be obtained in GGA.
410:
411: %\bibitem{gga} J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev.
412: %Lett. {\bf 77}, 3865 (1996).
413:
414:
415:
416:
417:
418: \end{thebibliography}
419:
420:
421:
422:
423: %\newpage
424: \begin{figure}
425: \centering
426: \epsfig{file=Fig1.ps,width=75mm}
427: \caption{The calculated magnetic phase diagram of Sr$_2$RuO$_4$
428: with structural distortions. When the tilting of RuO$_6$
429: octahedron is conducted, 12 degrees of RuO$_6$ rotation is
430: reserved (see the text for detailed description). The solid
431: bold lines are calculated phase boundaries, while the triangles
432: linked by dashed line correspond to experimental data. }
433: \end{figure}
434:
435:
436: %\newpage
437: \begin{figure}
438: \centering \psfull
439: \epsfig{file=Fig2.ps,width=75mm}
440: \caption{The calculated electronic densities of states (DOS) for
441: some particular points in our phase diagram, i.e. (a)
442: $\phi$=$\theta$=$0^\circ$, $\lambda=1.07$, NM state;
443: (b)$\phi$=$12^\circ$, $\theta=0^\circ$, $\lambda=1.07$, NM
444: state; (c)$\phi$=$\theta$=$0^\circ$, $\lambda=0.96$, NM state;
445: (d)$\phi$=$\theta$=$0^\circ$, $\lambda=0.96$, FM state;
446: (e)$\phi$=$12^\circ$, $\theta$=$0^\circ$, $\lambda=1.07$, FM
447: state; (f)$\phi$=$12^\circ$, $\theta$=$12^\circ$,
448: $\lambda=0.96$, AF state. The bold solid lines show the total
449: DOS (in (f), the local DOS is shown), while the thin solid and
450: dashed lines give the projected DOS for the $d_{xy}$ and
451: $d_{yx}$, $d_{zx}$ orbitals respectively. Only the regions (-2 eV
452: $\sim$ 1 eV), where Ru-$t_{2g}$ states dominate, are shown. The
453: Fermi levels are located at the energy zero. }
454: \end{figure}
455:
456:
457: %\newpage
458: \begin{figure}
459: \centering
460: % \epsfig{file=Fig3.ps,width=75mm}
461: \leavevmode \epsfxsize=75mm \epsfbox[50 300 650 800]{Fig3.ps}
462: \caption{A contour plot of the calculated bare spin susceptibility for
463: (a)$\phi$=$\theta$=$0^\circ$, $\lambda=1.07$;
464: (b)$\phi$=$\theta$=$0^\circ$, $\lambda=0.96$. The dashed lines
465: are guide to the eye for the nesting vectors. The red color
466: denote the higher intensity.}
467: \end{figure}
468:
469: \newpage
470: \begin{table}
471: \caption{The calculated total energies and magnetic moments
472: for some particular points in our phase diagram.}
473: \begin{tabular}{c|c|c|c}
474: &NM &FM &AF \\ \hline
475: $\lambda=0.96$ &0 meV &-40 meV &-8 meV \\
476: $\phi=\theta=0^\circ$ & &1.26 $\mu_B$/Ru &0.7 $\mu_B$/Ru \\
477: \hline
478: $\lambda=1.07$ &0 meV &-25 meV & \\
479: $\phi=12^\circ$, $\theta=0^\circ$ & &0.74 $\mu_B$/Ru & \\
480: \hline
481: $\lambda=0.96$ &0 meV &-100 meV &-117 meV \\
482: $\phi=12^\circ$, $\theta=12^\circ$ & &1.13 $\mu_B$/Ru &0.93
483: $\mu_B$/Ru \\
484: \end{tabular}
485: \end{table}
486:
487:
488:
489:
490: \end{document}
491:
492: