1: \documentstyle[epsf,12pt]{mrsproc}
2: \textheight 23.0cm
3: \textwidth 16.45cm
4: \voffset 1.3cm
5: \hoffset -0.1cm
6: %\voffset=0.6cm
7: \begin{document}
8: \begin{center}
9: {\bf Enhanced polarization in strained BaTiO$_3$ from first principles}
10: \end{center}
11:
12: \vspace{0.2cm}
13: \noindent
14: {J. B. Neaton, C.-L. Hsueh, and K. M. Rabe} \\
15: {Department of Physics and Astronomy, Rutgers University\\
16: Piscataway, NJ 08854-8019}
17: \begin{abstract}
18: The structure, polarization, and zone-center phonons of
19: bulk tetragonal BaTiO$_3$ under compressive epitaxial stress are
20: calculated using density functional theory within the local density approximation.
21: The polarization, computed using the Berry-phase formalism, increases
22: with increasing tetragonality and is found to be enhanced by
23: nearly 70\% for the largest compressive misfit strain considered (-2.28\%).
24: The results are expected to be useful for the analysis
25: of coherent epitaxial BaTiO$_3$ thin films and heterostructures
26: grown on perovskite substrates having a smaller lattice constant, such
27: as SrTiO$_3$.
28: \end{abstract}
29:
30:
31: %
32: \section{INTRODUCTION}
33: %
34:
35:
36: With recent advances in oxide epitaxy, coherent growth of
37: nearly perfect ultrathin perovskite
38: films and heterostructures is now possible \cite{schlom1,schlom2,ahn}.
39: In these systems, individual constituent layers
40: are under mixed mechanical boundary conditions: constrained to the
41: substrate lattice constant in-plane, the films are free to relax in the
42: normal direction. The strains achieved in this way can be quite significant, and
43: can be controlled to tune structure, ferroelectric, and dielectric
44: properties of these nanostructured materials \cite{sup1,sup2}.
45: For example, the polarization could be increased over that
46: of known bulk compounds \cite{schimizu,pertsev1}, improving the performance of ferroelectric
47: field-effect transistors and nonvolatile memories.
48:
49: When the interfaces in lattice-matched heterostructures are between
50: sufficiently similar materials, even the thinnest layers can be well described by considering
51: the behavior of the bulk under the same boundary conditions.
52: Thus, for example, many properties of
53: BaTiO$_3$/SrTiO$_3$ superlattices can be understood by
54: considering the effects of large strains
55: on bulk BaTiO$_3$. This has led to great interest in information about highly strained
56: bulk and film states in the perovskites,
57: for example as presented previously in a phenomenological
58: Landau-Devonshire framework \cite{pertsev1,pertsev2,pertsev3}.
59:
60: Here we perform first-principles calculations and obtain information
61: about the structure, spontaneous polarization, and zone-center phonons
62: of bulk tetragonal BaTiO$_3$ with misfit strains down to about -2.3\%.
63: While our results correspond to zero temperature, previous work \cite{pertsev1} suggests
64: that in this high-strain regime,
65: the tetragonal phase is stable with increasing temperature and thus
66: our analysis is relevant to the study of ultrathin BaTiO$_3$
67: films and BaTiO$_3$/SrTiO$_3$ superlattices at room temperature \cite{us}.
68:
69: %
70: \section{METHODOLOGY}
71: %
72:
73: To predict the ground state structure of strained BaTiO$_3$ and compute the associated
74: polarization, we use density functional theory (DFT) \cite{hk} within the local density
75: approximation (LDA) \cite{ks}, as implemented with a plane-wave basis in the Vienna {\it ab initio}
76: Simulations Package (VASP) \cite{kresse}. Results are obtained using
77: projector-augmented wave (PAW) \cite{paw} potentials provided with VASP.
78: The PAW potentials explicitly treat 10 valence electrons
79: for Sr ($4s^24p^65s^2$), 12 for Ti ($3s^23p^63d^24s^2$), and 6 for oxygen ($2s^22p^6$).
80: The ions are steadily relaxed toward equilibrium
81: until the Hellmann-Feynman forces are less than 10$^{-3}$ eV/{\AA}.
82: Brillouin zone integrations are performed with a
83: Gaussian broadening of 0.1 eV during all relaxations.
84: All calculations are performed with a 6$\times$6$\times$6
85: Monkhorst-Pack {\bf k}-point mesh and a 50 Ry plane-wave cutoff.
86: Once ground-state structures are obtained, polarizations are determined from
87: first principles using the Berry-phase formalism \cite{ksv}.
88: The force-constant matrices are then obtained with a frozen-phonon approach.
89: Specifically, a series of calculations is performed in which each ion in turn is displaced along
90: $\hat z$ by 0.1\% of the lattice parameter and the Hellmann-Feynman forces on all
91: the atoms computed; the phonon frequencies and eigenvectors are then acquired through
92: diagonalization of the corresponding dynamical matrix.
93:
94: \section{RESULTS}
95:
96: We begin by calculating the equilibrium structural parameters of
97: BaTiO$_3$ under zero stress. The observed phases of bulk BaTiO$_3$ are
98: closely related to the cubic perovskite structure $Pm\bar{3}m$ \cite{lg},
99: which is the structure of the high-temperature phase (T$>$403 K).
100: The simple cubic unit cell of this structure can be chosen so the Ti cation
101: occupies the body-centered site 1b ($1 \over 2$,$1 \over 2$,$1 \over 2$)
102: and is octahedrally coordinated by oxygen anions at 3c ($1 \over 2$,$0$,$1 \over 2$);
103: the Ba cation is then at Wyckoff position 1a ($0$,$0$,$0$), and twelve-fold
104: coordinated by oxygens.
105:
106: The rhombohedral $P3m1$ phase is computed to be the ground state of BaTiO$_3$,
107: consistent with observations \cite{lg}; however,
108: it is just slightly lower in energy (only $\sim$1 meV per unit cell)
109: than the optimized tetragonal $P4mm$ structure \cite{jorge},
110: the observed room temperature phase \cite{lg}.
111: In the tetragonal phase, the simple cubic lattice distorts, resulting in two
112: independent lattice parameters $a$ (in-plane) and $c$ (normal). The lower symmetry
113: also permits relative displacement of the atomic sublattices along $\hat z$, with
114: Wyckoff positions 1a ($0$,$0$,$0$) for Ba and
115: 1b ($1 \over 2$,$1 \over 2$,$1 \over 2$+$\Delta_{\rm Ti}$) for Ti,
116: while two Wyckoff positions are occupied by oxygens, designated
117: O$_{\rm I}$, and O$_{\rm II}$:
118: 1b ($1 \over 2$,$1 \over 2$,$\Delta_{\rm O_{\rm I}}$)
119: and 2c ($1 \over 2$,$0$,$1 \over 2$+$\Delta_{\rm O_{\rm II}}$).
120: These displacements result in a nonzero polarization along [001].
121:
122: The computed values of the structural parameters are given in the first line of Table I.
123: Our calculated lattice parameters
124: $a$=3.945~{\AA} and $c$=3.988~{\AA} are slightly
125: smaller than those found experimentally ($a$=3.994~{\AA}
126: and $c$=4.036~{\AA} \cite{shirane}),
127: the size of the underestimate being typical of that found in LDA
128: studies of a wide variety of insulators and semiconductors.
129: The smaller volume is known, however, to suppress the
130: polar instability \cite{volume}, and indeed the
131: computed displacements (in reduced units of $c$)
132: of 0.0122, -0.0192, and -0.0124 for $\Delta_{\rm Ti}$,
133: $\Delta_{\rm O_{\rm I}}$, and $\Delta_{\rm O_{\rm II}}$, respectively,
134: are smaller than the measured values \cite{shirane},
135: which are 0.015, -~0.023, and -0.014.
136: The polarization of the optimized tetragonal phase ($P_0$=24.97~$\mu$C/cm$^2$)
137: is thus also significantly less than that computed with the
138: experimental structural parameters (32.2~$\mu$C/cm$^2$).
139: As a consequence, in the following we define misfit strains to be
140: relative to the calculated lattice constants, and define
141: polarization enhancement relative to the computed zero-stress polarization $P_0$.
142:
143: In coherent epitaxial thin films and [001] superlattice heterostructures,
144: the two lattice vectors in the plane of the interface are constrained to
145: match the substrate while the third lattice vector is free to relax.
146: Here, we consider tetragonal BaTiO$_3$ ($P4mm$) under
147: mechanical boundary conditions corresponding to coherent epitaxy of a $c$-axis
148: oriented phase on simple
149: cubic substrates, such as SrTiO$_3$, having lattice constants slightly
150: smaller than that of stress-free bulk tetragonal BaTiO$_3$.
151: We perform a series of constrained relaxations in which
152: the in-plane lattice constant $a$ of bulk tetragonal BaTiO$_3$ is
153: compressed (e.g., to the value of a hypothetical substrate) and the length of
154: the normal axis and atomic positions are optimized. We consider misfit strains
155: ranging from -0.28\% to -2.28\%.
156: The absolute stabilities of the resulting structures require independent investigation.
157: For example, it may be energetically favorable for the polarization vector and c-axis to
158: tilt toward the [110] direction
159: (this has been referred to as the $r$ phase in previous work \cite{pertsev1}),
160: lowering the point symmetry to monoclinic.
161: We have, however, verified that for misfit strains of -0.76\% and larger, lowering the symmetry to monoclinic
162: does not result in a lower energy, and we can therefore tentatively conclude that the $P4mm$ structure
163: is relevant under the boundary conditions considered in this work.
164:
165: The optimized structural parameters of strained $c$-phase BaTiO$_3$ are summarized
166: in Table I. A tendency to conserve volume is manifested as an
167: increase in $c$ with increasing magnitude of the misfit strain, through this increase
168: does not result in a full compensation of the in-plane compression and the volume actually decreases moderately.
169: The atomic displacements relative to the centrosymmetrically-strained cubic perovskite structure exhibit
170: a significant, nearly linear, increase with
171: increasing misfit strain. Both the average and deviation (or splitting) of the two Ti-O$_{\rm I}$ distances
172: (along the $\hat z$-oriented Ti-O chain) increase with misfit strain, from
173: 1.994 $\pm$ 0.125 {\AA} at zero misfit to 2.066 $\pm$ 0.233 {\AA} at -2.28\%. Thus,
174: the smaller Ti-O$_{\rm I}$ distance actually decreases slightly, from 1.87 to 1.83 {\AA},
175: while the larger distance increases by 8\% over the same range.
176: The shortest Ba-O$_{\rm II}$ distance also declines modestly from 2.77 to 2.75 {\AA}.
177: The shortest Ba-O$_{\rm I}$ distance, however,
178: decreases more rapidly with increasing strain and, for the largest
179: strain considered (2.28\%), this separation is smaller than the minimum
180: Ba-O$_{\rm II}$ contact.
181:
182: These structural trends are accompanied by a significant change in polarization $P_s$.
183: Specifically, the increased atomic displacements associated with the increase of tetragonality ($c/a$)
184: tend to increase the polarization; for 2.28\% strain, the polarization
185: {\it increases} by nearly a factor of two. These results are consistent with a previous first-principles
186: study, in which a strong dependence of the polarization on the $c$ parameter
187: was observed for BaTiO$_3$~\cite{schimizu}. We also report the increase in energy per unit cell,
188: $\Delta E$, which reflects the energy cost of nonzero misfit strain. It varies
189: quadratically with misfit strain over the range shown, with an estimated
190: effective elastic constant of 135 eV/unit cell.
191:
192: The computed frequencies for the $\hat z$ polarized zone-center phonons for each
193: optimized structure are given in Table II. As there are five atoms in each primitive cell,
194: 12 zone-center optical phonons are expected (plus three acoustic modes).
195: At the zone center these are labeled by the irreducible representations of the C$_{4v}$ point group.
196: There are 4 A$_1$ modes (Raman and IR-active),
197: 1 B$_1$ mode (Raman active), and 5 doubly-degenerate E modes (Raman and IR-active),
198: of which the A$_1$ and B$_1$ modes have atomic displacements and polarization
199: (if nonzero) only along $\hat z$.
200: The frequencies of the three A$_1$
201: modes are 503, 252, and 178~cm$^{-1}$, respectively. These agree quite well with
202: a room-temperature experiment, which reports 512, 276, and 178~cm$^{-1}$ for the
203: same TO modes \cite{porto}. An analysis of the corresponding eigenvectors
204: indicates that the two highest-frequency modes involves primarily Ti and O motion;
205: in the lowest the cations move opposite to the anions, with each ion in the primitive cell
206: participating almost equally.
207: The B$_1$ mode, for which the computed frequency is 294~cm$^{-1}$, is Raman active and involves
208: alternating motion of the two O$_{\rm II}$ ions. As for the A$_1$ modes, our
209: computed frequency agrees well with the value of 308~cm$^{-1}$
210: measured at room temperature \cite{porto}.
211: We are presently unaware of previous first-principles
212: calculations of zone-center phonon frequencies in tetragonal BaTiO$_3$,
213: although we note previous calculations have been performed using an empirical
214: rigid-ion model \cite{freire} and shell-model potentials \cite{khalal}.
215:
216: In the remainder of Table II, we report the effect of
217: in-plane compression on the frequencies of these four modes.
218: We find that the two highest A$_1$ modes {\it increase} with
219: increasing misfit strain, while the frequency of the lowest-lying A$_1$ mode
220: is nearly independent of misfit strain down to about -2.3\%.
221: In contrast, the computed Raman-active B$_1$ mode frequency is found to
222: {\it decrease} with increasing misfit strain.
223: \begin{table}
224: \begin{center}
225: \caption{\small Structural parameters of BaTiO$_3$ under in-plane compression,
226: computed from first principles.
227: The misfit strain (\%) is measured relative to the calculated
228: lattice parameter of the computed tetragonal phase, reported in the table as 0.0\% strain. Displacements
229: are given relative to the centrosymmetrically-strained cubic-perovskite structure with the Ba
230: position fixed at (0,0,0)
231: in reduced coordinates (units of $c$). P$_s$ is the
232: Berry-phase polarization in $\mu$C/cm$^2$, with P$_0$ being the polarization of the equilibrium tetragonal structure. $\Delta E$ is the increase in
233: energy in meV per unit cell relative to the equilibrium tetragonal structure.}
234: \vspace{0.15cm}
235: \begin{tabular}{c c c c c c c c c c c}
236: \% & $a$ (\AA) & $c$ (\AA) & $c/a$ & V (\AA$^3$)& $\Delta_{\rm Ti}$
237: & $\Delta_{\rm O_{\rm I}}$ & $\Delta_{\rm O_{\rm II}}$ & P$_s$
238: & P$_s$/P$_0$ & $\Delta$E \\\hline
239: 0.0 & 3.945 & 3.988 & 1.0110 & 62.06 & 0.0122 & -0.0192 & -0.0124 & 24.97 & 1 & 0.0\\
240: -0.28 & 3.934 & 4.0097 & 1.0192 & 62.06 & 0.0134 & -0.0227 & -0.0143 & 28.08 & 1.13 & 1.08\\
241: -0.76 & 3.915 & 4.0353 & 1.0309 & 61.83 & 0.0148 & -0.0256 & -0.0165 & 31.36 & 1.26 & 8.48\\
242: -1.26 & 3.895 & 4.0647 & 1.0437 & 61.66 & 0.0160 & -0.0295 & -0.0191 & 34.87 & 1.40 & 22.67\\
243: -1.77 & 3.875 & 4.0961 & 1.0571 & 61.51 & 0.0170 & -0.0336 & -0.0220 & 38.37 & 1.54 & 43.63\\
244: -2.28 & 3.855 & 4.1322 & 1.0719 & 61.42 & 0.0179 & -0.0384 & -0.0253 & 42.12 & 1.69 & 71.26\\
245: \end{tabular}
246: \end{center}
247: \end{table}
248:
249:
250: \begin{table}
251: \begin{center}
252: \caption{\small Computed zone-center phonon frequencies
253: of BaTiO$_3$ as a function of misfit strain (\%) for $\hat z$-polarized phonons
254: labelled by irreducible representations A$_1$ and B$_1$.}
255: \vspace{0.15cm}
256: \begin{tabular}{c c c c c}
257: \multicolumn{1}{c}{\%} &
258: \multicolumn{3}{c}{$\omega_{\rm A_1}$ (cm$^{-1}$)} &
259: \multicolumn{1}{c}{$\omega_{\rm B_1}$ (cm$^{-1}$)}\\
260: \hline
261: 0.0 & 503 & 252 & 178 & 294 \\
262: -0.28 & 508 & 275 & 179 & 292 \\
263: -0.76 & 514 & 296 & 180 & 290 \\
264: -1.26 & 524 & 316 & 180 & 286 \\
265: -1.77 & 536 & 332 & 180 & 282 \\
266: -2.28 & 549 & 347 & 180 & 278 \\
267: \end{tabular}
268: \end{center}
269: \end{table}
270:
271: %
272: \section{DISCUSSION}
273: %
274:
275: The trends witnessed here in the computed properties of strained BaTiO$_3$
276: are quite general and reflect the underlying physics of ferroelectric perovskites under
277: the mixed-mechanical boundary conditions associated with epitaxy.
278: In-plane compression resulting from epitaxial stress favors expansion
279: of $c$ normal to the substrate and, by a mechanism analogous
280: to that responsible for the enhancement of the ferroelectric instability with increasing volume,
281: the polar distortion and associated polarization along $\hat z$ is expected to increase,
282: as observed.
283:
284: While larger polarization enhancements may in principle be achieved through
285: larger misfit strains, the role of misfit strain in materials design is
286: limited by a rapidly increasing
287: strain energy $\Delta E$. Above a critical thickness, misfit dislocations
288: will become energetically favorable, and the film or heterostructure will
289: relax to its bulk lattice constant, relieving stress and decreasing the
290: polarization.
291: However, coherent epitaxy of BaTiO$_3$-based heterostructures of useful thicknesses
292: has already been demonstrated \cite{schlom2,sup2}, and the present results show that
293: the misfit strain of BaTiO$_3$ on a SrTiO$_3$
294: substrate ($\sim$2\%) is large enough to achieve significant enhancement of the polarization.
295:
296: Our computed phonon frequencies are of interest chiefly for two reasons.
297: First, combined with the computed Born effective charges and electronic dielectric constant
298: (to be reported in a separate publication), they
299: determine the phonon contribution to the static dielectric response $\epsilon_{0}$. While clearly an
300: important property in itself, $\epsilon_0$ should also be relevant for understanding
301: the effects of changing electrical boundary conditions on films and superlattices.
302: Second, the sensitivity to strain of the mode frequencies, particularly the
303: intermediate frequency A$_1$ mode, can provide an experimental probe of
304: the strain state of a BaTiO$_3$ layer.
305: Indeed, for the closely related material PbTiO$_3$, mode hardening has been observed in
306: films under compressive epitaxial stress \cite{sun}.
307:
308: Although our focus here centers on the effects of strain,
309: macroscopic electric fields, interfacial effects, and finite
310: size effects must also play a fundamental role in determining the properties
311: of coherent epitaxial thin films and superlattices.
312: However, knowledge of strain effects has already proved to be invaluable
313: for analyzing experimental and first-principles
314: results \cite{us}, and for separating out the contributions of other intrinsic
315: as well as extrinsic effects. Therefore, a broader study of BaTiO$_3$,
316: which includes calculation of the
317: dielectric tensor and related quantities, is currently in progress.
318: A similar study of other perovskites, such as PbTiO$_3$, PbZrO$_3$, SrTiO$_3$,
319: KNbO$_3$ and KTaO$_3$, would also be pertinent.
320:
321: \section{ACKNOWLEDGMENTS}
322: This work was supported by NSF-NIRT and ONR N00014-00-1-0261.
323: We thank C. H. Ahn, M. H. Cohen, J. \'I\~niguez, X. Q. Pan, D. G. Schlom,
324: and D. Vanderbilt for valuable discussions.
325:
326: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
327: % REFERENCES
328: {\small
329: \begin{thebibliography}{0}
330:
331: \vspace{-0.27cm}\bibitem{schlom1}
332: J. C. Jiang, X. Q. Pan, W. Tian, C. D. Theis, and D. G. Schlom,
333: Appl. Phys. Lett. {\bf 74}, 2851 (1999).
334:
335: \vspace{-0.27cm}\bibitem{schlom2}
336: W. Tian, X. Q. Pan, J. H. Haeni, and D. G. Schlom, preprint.
337:
338: \vspace{-0.27cm}\bibitem{ahn}
339: T. Tybell, C. H. Ahn and J. M. Triscone, Appl. Phys. Lett. {\bf 75},
340: 856 (1999).
341:
342: \vspace{-0.27cm}\bibitem{sup1}
343: T. Tsurnumi, T. Ichikawa, T. Harigai, H. Kakemoto, S. Wada,
344: J. Appl. Phys. {\bf 91}, 2284 (2002).
345:
346: \vspace{-0.27cm}\bibitem{sup2}
347: T. Shimuta, O. Nakagawara, T. Makino, S. Arai, H. Tabata,
348: and T. Kawai, J. Appl. Phys. {\bf 91}, 2290 (2002).
349:
350: \vspace{-0.27cm}\bibitem{schimizu}
351: T. Schimizu and T. Kawakubo, Jpn. J. Appl. Phys. {\bf 37}, L235 (1998).
352:
353: \vspace{-0.27cm}\bibitem{pertsev1}
354: N. A. Pertsev, A. G. Zembligotov, and A. K. Tagantsev, Phys. Rev. Lett. {\bf 80},
355: 1988 (1998).
356:
357: \vspace{-0.27cm}\bibitem{pertsev2}
358: N. A. Pertsev, A. G. Zembligotov, S. Hoffmann, R. Waser, and A. K. Tagantsev,
359: J. Appl. Phys. {\bf 85}, 1698 (1999).
360:
361: \vspace{-0.27cm}\bibitem{pertsev3}
362: N. A. Pertsev, A. K. Tagantsev, and N. Setter, Phys. Rev. B {\bf 61}, R825 (2000).
363:
364: \vspace{-0.27cm}\bibitem{us}
365: J. B. Neaton and K. M. Rabe, in preparation.
366:
367: \vspace{-0.27cm}\bibitem{hk}
368: P. Hohenberg and W. Kohn, Phys. Rev. {\bf 136}, 864B (1964).
369:
370: \vspace{-0.27cm}\bibitem{ks}
371: W. Kohn and L. J. Sham, Phys. Rev. {\bf 140}, 1133A (1965).
372:
373: \vspace{-0.27cm}\bibitem{kresse}
374: G. Kresse and J. Hafner, Phys. Rev. B {\bf 47}, R558 (1993);
375: G. Kresse and J. Furthm\"uller, Phys. Rev. B {\bf 54}, 11169 (1996).
376:
377: \vspace{-0.27cm}\bibitem{paw} P. Bl{\"o}chl, Phys. Rev. B {\bf 50}, 17953 (1994); G. Kresse
378: and D. Joubert, Phys. Rev. B {\bf 59}, 1758 (1999).
379:
380: \vspace{-0.27cm}\bibitem{ksv}
381: D. King-Smith and D. Vanderbilt, Phys. Rev. B. {\bf 47}, 1651 (1993).
382:
383: \vspace{-0.27cm}\bibitem{lg}
384: M. E. Lines and A. M. Glass, {\it Principles and Applications of
385: Ferroelectrics and Related Materials} (Clarendon Press, Oxford, 1977).
386:
387: \vspace{-0.27cm}\bibitem{jorge}
388: J. \'I\~niguez, J. B. Neaton, and D. Vanderbilt,
389: {\it Fundamental Physics of Ferroelectrics 2002}, R. E. Cohen
390: and T. Egami, eds., (AIP, Melville, New York, 2002), in press.
391:
392: \vspace{-0.27cm}\bibitem{volume}
393: R. E. Cohen and H. Krakauer, Phys. Rev. B {\bf 42}, 6416 (1990).
394:
395: \vspace{-0.27cm}\bibitem{shirane}
396: G. Shirane, H. Danner, and P. Pepinsky, Phys. Rev. {\bf 105}, 856 (1957).
397:
398: \vspace{-0.27cm}\bibitem{porto}
399: A. Scalabrin, A. S. Chaves, D. S. Shim, and S. P. S. Porto, Phys.
400: Stat. Sol. B {\bf 79}, 731 (1977).
401:
402: \vspace{-0.27cm}\bibitem{freire}
403: J. D. Freire and R. S. Katiyar, Phys. Rev. B {\bf 37}, 2074 (1988).
404:
405: \vspace{-0.27cm}\bibitem{khalal}
406: A. Khalal, D. Khatib, and B. Jannot, Ann. Ch. Sci. Mat. {\bf 24}, 471 (1999).
407:
408: \vspace{-0.27cm}\bibitem{sun}
409: L. Sun {\it et al.}, Phys. Rev. B {\bf 55}, 12218 (1997).
410:
411: \end{thebibliography} }
412: % END REFERENCES
413: %
414: \end{document}
415: