1: %%%%%%%%%%%%%%%%%%% file aplop.tex %%%%%%%%%%%%%%%%%%%%
2: % Copyright (c) Optical Society of America, 1992. %
3: %%%%%%%%%%%%%%%%%% October 20, 1992 %%%%%%%%%%%%%%%%%%%
4: % Substitute rm (Roman) font.
5: % add backslash char to \tt font
6: %\documentstyle[osa,aplop, manuscript,]{revtex}
7:
8:
9: \documentstyle[prl,aps,twocolumn,epsfig]{revtex}
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11: %TCIDATA{OutputFilter=Latex.dll}
12: %TCIDATA{LastRevised=Thu May 31 14:13:19 2001}
13: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
14: %TCIDATA{Language=American English}
15: %TCIDATA{CSTFile=revtex.cst}
16:
17: \newcommand{\MF}{{\large{\manual META}\-{\manual FONT}}}
18: \newcommand{\manual}{rm}
19: \newcommand\bs{\char '134 }
20:
21: \begin{document}
22:
23: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname @twocolumnfalse\endcsname
24:
25: \title{Diffuse Neutron Scattering Study of a Disordered Complex Perovskite Pb(Zn$_{1/3}$Nb$_{2/3}$)O$_{3}$ Crystal}
26: \author{D. La-Orauttapong,$^{1}$ J. Toulouse,$^{1}$ J. L. Robertson,$^{2}$ and Z.-G. Ye$^{3}$}
27: \address{$^{1}$Physics Department, Lehigh University, Bethlehem, PA 18015-3182}
28: \address{$^{2}$Oak Ridge National Laboratory, Solid State Division, Oak Ridge, TN 37831-6393}
29: \address{$^{3}$Department of Chemistry, Simon Fraser University, Burnaby, BC, V5A 1S6, Canada}
30: \maketitle
31:
32: \begin{abstract}
33: Diffuse scattering around the (110) reciprocal lattice point has been
34: investigated by elastic neutron scattering in the paraelectric and the
35: relaxor phases of the disordered complex perovskite crystal-Pb(Zn$_{1/3}$Nb$_{2/3}$)O$_{3}$\ (PZN). \
36: The appearance of a diffuse intensity peak
37: indicates the formation of polar nanoregions at $\ $temperature $T^{\ast }$,
38: approximately 40K above $T_{c}$=413K. \ The analysis of this diffuse
39: scattering indicates that these regions are in the shape of ellipsoids, more
40: extended in the $<111>$ direction than in the $<001>$ direction. \ The
41: quantitative analysis provides an estimate of the correlation length, $\xi $,
42: or size of the regions and shows that $\xi _{\left< 111\right> }\sim
43: 1.2\xi _{\left< 001\right> }$, consistent with the primary or dominant
44: displacement of Pb leading to the low temperature rhombohedral phase. \ Both
45: the appearance of the polar regions at $T^{\ast }$ and the structural
46: transition at $T_{c}$ are marked by kinks in the $\xi _{\left< 111\right> }$
47: curve but not in the $\xi _{\left< 001\right> }$ one, also indicating that
48: the primary changes take place in a $<111>$ direction at both temperatures.
49:
50: PACS Numbers: 77.84.Dy,61.12.Ld,77.80.-e
51: \end{abstract}
52: \vskip1pc]
53:
54: \narrowtext
55:
56: Many of the relaxor ferroelectrics known today are lead-based compounds with
57: the perovskite structure. \ In addition to the characteristic frequency
58: dispersion of their dielectric constant, several of them exhibit remarkable
59: piezoelectric or electrostrictive properties that are
60: finding important applications, e.g. as transducers and actuators. \ The Pb$%
61: ^{2+}$-containing relaxor perovskites such as Pb$($R$_{1/3}$Nb$_{2/3})$O$%
62: _{3} $ (R=Mg$^{2+}$, Zn$^{2+}$) have a common ABO$_{3}$ cubic perovskite
63: structure in which the B-site can be occupied by $\frac{1}{3}$R$^{2+}$ and
64: $\frac{2}{3}$Nb$^{5+}$. \ Because of the different atomic radii and valences
65: of the B-site cation, PMN and PZN exhibit short-range chemical ordering\cite
66: {Yokomizo et al_1278_1970,Randall_Bhalla_327_1990}. \ Over the years,
67: these relaxors and their properties have been described in a variety of ways,
68: most often in terms of the formation of polar micro- or nanoregions
69: \cite{Smolensky_26_1970,Cross_241_1987}. \ However, these models are primarily
70: based on indirect experimental evidence for such regions and more direct evidence is necessary \
71: in order to elucidate the true origin of the relaxor behavior. For this purpose, neutron and
72: X-ray techniques are most suitable, as they can provide evidence of local
73: structural ordering. \ Obtaining such evidence is crucial if the formation
74: of polar regions is indeed responsible for the relaxor behavior. \
75:
76: PZN is a prototype relaxor ferroelectric. \
77: It was found to exhibit a large dielectric dispersion and a broad dielectric maximum that
78: depended on both frequency and temperature. \ Earlier studies
79: also reported a structural phase transition from cubic to rhombohedral symmetry
80: near 413K\cite{Khuchua et al_194_1968,Kuwata et al_863_1979}, which falls in
81: the temperature region of the maximum of the dielectric peak. \ More recent
82: studies only mention the coexistence of cubic and rhombohedral phases, with
83: polar nanodomains growing into polar microdomains \cite{Mulvihill et al_1462_1997}
84: and their volume fraction increasing with decreasing temperature. \ Strain appears
85: to play an important role in a nanodomain-to-microdomain phase transition, which
86: therefore resembles a martensitic phase transformation\cite{Mulvihill et
87: al_1462_1997}. \ In the last few years, PZN has also been intensively
88: investigated in the form of a solid solution with PbTiO$_{3}$\cite
89: {Kuwata et al_1298_1982,Liu et al_2810_1999,Gehring et al_5216_2000},
90: because of its enhanced dielectric and piezoelectric properties. \
91: PZN is already a disordered perovskite in the cubic phase\cite{Iwase et al_1419_1999},
92: with individual Pb atoms shifted from the A-site position along one of eight
93: equivalent $\left\langle 111\right\rangle $ directions and O atoms forming a \
94: ring-like distribution within $\left\{ 100\right\} $ planes. \
95: In contrast, the Zn/Nb atoms remain located at the normal B-site position. \
96: In the rhombohedral phase, the displacements of Pb and O
97: atoms become correlated, both along a specific [111] direction, with $\delta r_{\text{%
98: Pb}}$ = 0.31 \AA\ and $\delta r_{\text{O}}$ = 0.17 \AA\ at 295K. \ More
99: specifically, a recent X-ray study \cite{Takesue et al_2001} of the related
100: system PMN suggests that the three oxygen atoms may be displaced along
101: three converging [110] directions, all three pointing towards the same A
102: position, such that the average oxygen displacement or the net oxygen dipole
103: moment in the unit cell is indeed along the same [111] direction as that of Pb. \
104: It is important to note that the displacement of Pb \cite{Iwase et al_1419_1999}
105: as well as its neutron scattering length are almost twice those of the O atom. \
106:
107: In the present paper, we report the results of a diffuse neutron scattering
108: study of PZN, in order to identify and directly monitor the formation of
109: polar nanoregions resulting from the correlated displacements of the Pb and
110: O atoms. \ This study follows a similar study of another mixed perovskite
111: relaxor, K$_{1-x}$Li$_{x}$TaO$_{3}$ (KLT), which provided the correlation
112: length or size of the polar nanoregions as a function of temperature \cite
113: {Yong et al_14736_2000}. \ In contrast to KLT, PZN exhibits not only
114: substitutional chemical disorder but also charge disorder. \ Moreover, due
115: to the lone electron pair effects, Pb$^{2+}$ions form partially covalent
116: bonds with the neighboring oxygen atoms, which results in the off-centering
117: of lead and in an additional local electric dipole moment \cite{Egami et
118: al_S935_1998}. \ Our purpose here is to provide direct evidence for these
119: regions and to characterize them in a single crystal of PZN.
120:
121: In complex perovskites such as PZN, the diffuse scattering has two possible
122: origins \cite{Krivoglaz_NewYork_1969}: i) different scattering factors of
123: the two possible atoms (B$_{1}$,B$_{2}$) occupying a given B site ($j$), or
124: ii) static displacements away from this site. \ Both factors arise from
125: concentration inhomogeneities in the\ compound and may simultaneously
126: contribute to the diffuse scattering. \ It is well known that the static
127: structure factor is given by \cite{Schmatz_NewYork_London_1973}:
128:
129: \begin{equation}
130: F(\text{{\bf Q}})={\sum_{j}}b_{j}e^{i\text{{\bf Q{.r}}}_{j}}e^{-W_{j}}
131: \end{equation}
132:
133: \noindent where $b_{j}$\ is the neutron scattering length, $e^{-W_{j}}$ is
134: the Debye-Waller factor for the $j$th atom with coordinate ${\bf r}_{j}$ in
135: the unit cell and the summation is carried out over all positions occupied
136: in the unit cell. \ The scattering vector {\bf Q} is defined as
137: ${\bf k}-{\bf k}^{{\acute{}}}$. \ In the presence case, small displacements,
138: ${\bf \delta r}$, of the atoms from
139: their normal positions distort the local cubic symmetry such that \
140: ${\bf r}_{j}$=$({\bf r}+ {\bf \delta r})_{j}$. At high temperature, these displacements
141: vary both in magnitude and direction. \ The scattered intensity is proportional to
142: $F^{2}=F.F^{\ast }$ so that, experimentally, it is possible to determine the
143: amplitude $F$ of the structure factor but not its phase. \ For the sake of
144: argument we omit the Debye-Waller factor in the following calculation. \ In accordance with
145: Ref.\cite{Iwase et al_1419_1999}, which reported the Zn/Nb atoms at their
146: ideal perovskite position at all temperatures, no displacement is included
147: in eq. (1) for the atoms at the B-site. With the origin chosen at the Pb site, eq. (1) \
148: and for small displacements of the Pb and O atoms, the neutron scattering structure factor $F$ can then be evaluated at
149: the (110) reciprocal lattice point:
150:
151: \begin{eqnarray}
152: F(\text{{\bf Q}})&=&[{\underbrace{b_{\text{Pb}}+b_{\text{Zn/Nb}}-b_{\text{O }%
153: }}_{\text{Bragg}}}] \nonumber \\
154: & &+
155: \begin{array}{c}
156: {\underbrace{i\text{{\bf Q.}}\left[ \delta \text{{\bf {r}}}_{\text{Pb}}b_{%
157: \text{Pb}}+\left( \delta \text{{\bf {r}}}_{\text{O}1}-\delta \text{{\bf {r}}}%
158: _{\text{O}2}-\delta \text{{\bf {r}}}_{\text{O}3}\right) b_{\text{O}}\right] }%
159: _{\text{Diffuse}}}
160: \end{array}
161: \end{eqnarray}
162:
163: \noindent where $b_{\text{Pb}}$, $b_{\text{Zn/Nb}}$, and $b_{\text{O}}$ are
164: the scattering lengths of Pb, Zn/Nb, and O atoms respectively. \ The
165: scattering lengths for the Zn/Nb atoms can be found from $b_{\text{Zn/Nb}}=%
166: \frac{1}{3}b_{\text{Zn}}$ +$\frac{2}{3}b_{\text{Nb}}$. \ In the structure factor above,
167: we have separated the Bragg scattering
168: contribution which contains information about the average lattice from the
169: diffuse scattering which arises from local deviations from the ideal
170: structure. \ These deviations result from correlated displacements of atoms. \
171: In the presence case, the diffuse scattering is primarily associated with
172: the Pb atom, which is strongly off-centered in the PbO$_{12}$ unit and has
173: almost twice the neutron scattering length of the O atom. \ In addition, because \
174: of the Pb displacement along the [111] direction, ${\bf \delta }r_{\text{Pb}}$is \
175: expected to make a greater contribution in that direction than along the [001] and [110] \
176: directions. \
177:
178: The neutron scattering experiments were performed on the HB-1 triple-axis
179: spectrometer of the High Flux Isotope Reactor at Oak Ridge National
180: Laboratory. \ Pyrolytic graphite (002) crystals were used for both the monochromator and
181: analyzer, to suppress harmonic contamination. \ The collimation was
182: 48$^{{\acute{}}}$-20$^{{\acute{}}}$-20$^{{\acute{}}}$-70$^{{\acute{}}}$
183: and the incident neutron energy was 13.6 meV ($\lambda $=2.453\AA ). \
184: Initially, the scattering around the (110) reciprocal lattice point in the [110]-[001]
185: scattering plane was studied because the (110) reflection has the largest
186: structure factor \cite{Yokomizo et al_1278_1970}. In each scan, the stepsize was
187: $q=$0.01 reciprocal lattice units (rlu) or 0.016 \AA $^{-1}$. \
188: The yellowish single crystal of Pb(Zn$_{1/3}$Nb$_{2/3}$)O$_{3}$
189: used in the experiment was grown by a high temperature solution growth
190: technique \cite{Zhang et al_96_2000}. \ The measurements were performed upon
191: cooling from 550 to 295K and no external electric field was applied.
192: \begin{figure}[tbp]
193: \epsfig{width=0.8 \linewidth, figure=Fig1.eps }
194: \caption{Elastic diffuse scattering of neutrons in the [001] and [111]
195: directions near the (110) reciprocal lattice point at T=550, 500, 450, 400,
196: and 350K. Inset is scattering intensities at $T_{c}$=413K in the [001]and [111] directions.}
197: \end{figure}
198: The narrow Bragg peak as well as the relatively broad diffuse scattering
199: peak at the (110)reciprocal lattice point are presented in Fig.1 for several temperatures. \
200: As expected, the Bragg scattering is equally strong in both the [001] and [111]
201: scans(inset of Fig.1). \ However, the diffuse scattering is more extended in the [001]
202: than in the [111] direction, having the form of a
203: rod along a cubic direction. \ It is important to note that the diffuse
204: scattering is more extended in a direction in which the correlation length $%
205: \xi $ is shorter. \ We will discuss this point later. \ In order to separate the Bragg
206: and diffuse scattering contributions, the intensity, $I(q)$, was fitted as the sum \
207: of a Gaussian and a Lorentzian, using a least-squares method(Fig.2). \
208: In the inset of Fig.2, the Bragg intensity is seen to increase slowly, starting at \
209: approximately 500K, then more rapidly below 400K. The latter rapid increase \
210: of the Bragg intensity can be explained by i) a relief
211: of extinction due to an increase in mosaicity below the transition or ii) an
212: unusual behavior of the Debye-Waller factor due to freezing of certain
213: atomic motion at lower temperatures.
214: \ A broadening of the Bragg peak is normally observed to accompany the
215: freezing process but is not observed here. \ This would rule out the effect
216: of the Debye-Waller factor in the present case and favor the first
217: explanation. \ In fact, a similar behavior was observed in a single crystal
218: of PMN by Gosula et al. \cite{Gosula et al_221_2000} using X-ray diffuse
219: scattering. \ They speculated that the sharp increase in intensity arose
220: from the anti-ferrodistortive ordering that developed at low temperatures. \
221: This anti-ferrodistortive ordering grows to a few hundred angstroms with
222: decreasing temperature but never reaches truly long-range order.
223: \begin{figure}[tbp]
224: \epsfig{width=0.8 \linewidth, figure=Fig2.eps }
225: \caption{Elastic neutron scattering in the [001] direction at T=550 and 350K.
226: The solid curve shows the scattering profile fitted by the sum of a Gaussian and a
227: Lorentzian and the two components (dash and dot curves)are shown separately.
228: Inset is temperature dependence of Bragg peak intensities in the [001] and [111] directions.}
229: \end{figure}
230: Turning to the diffuse scattering component, in Fig.2, the $q$ or $l$%
231: -dependence of the scattered intensity is presented on a semilogarithmic
232: plot at 550K and 350K. \ At 550 K the diffuse scattering intensity is slightly
233: asymmetric which could be related to rather large deformations of the crystal lattice
234: giving rise to Huang scattering\cite{Krivoglaz_NewYork_1969}. \ The Lorentzian function
235: used in the fit of the experimental $I(q)$ curves, yielded a peak intensity and
236: a full width at half- maximum $\left( \Delta q_{_{FWHM}}\right) $. \ Such a Lorentzian
237: lineshape is predicted by the Ornstein-Zernike model \cite{Stanley_1971}:
238: $I\simeq \frac{1}{q^{2}+\xi ^{-2}}$ where $q=l$ is the momentum transfer relative to the
239: {\bf Q}=110 Bragg reflection and $\xi $ is the correlation
240: length. \ The Ornstein-Zernike model assumes a correlation function of the form
241: $\frac{e^{-\frac{r}{\xi }}}{r}$ for the polarization. \ Therefore the diffuse
242: scattering width at half maximum, $\Delta q_{_{FWHM}}$, provides an estimate of $\xi $ or,
243: equivalently, of the size of the polar nanoregions:$\xi =\frac{2}{\Delta q_{_{FWHM}}}$
244: where $a$ is the lattice parameter. \ The Lorentzian fit parameters,
245: $I\left( q\right) _{\max }$, $\Delta q_{_{FWHM}}^{2}$, and $\xi $, are shown in Fig.3 (a)
246: and (b) as a function of temperature. \ In both the $[001]$, and $[111]$ directions,
247: the results indicate that:
248: \begin{figure}[tbp]
249: \epsfig{width=0.8 \linewidth, figure=Fig3.eps }
250: \caption{Temperature dependence of (a).the Lorentzian Peak Intensity and
251: square of the FWHM,$\Delta q_{_{FWHM}}^{2}$, (b).the
252: correlation length,$\protect\xi$ in the [001] and [111] directions.}
253: \end{figure}
254: i) the diffuse scattering peak intensity, $I\left( q\right) _{\max }$,
255: increases as the temperature decreases. \ It is interesting however to note
256: that the peak intensity, as measured in the $[111]$ direction, goes through
257: a plateau around the transition temperature $(T_{c}=413$K$)$ before
258: increasing again below 400K. \ This feature is much less visible in the $%
259: [001]$ direction, for which the intensity increases continuously. \ Fig.3
260: (a) also shows that the strongest diffuse scattering is observed in the $%
261: [111]$ direction, an observation that is consistent with the displacement
262: of the Pb atoms producing the local polarization in the
263: rhombohedral phase. \ The diffuse scattering intensity is in fact generally
264: proportional to $\left\langle P_{local}^{2}\right\rangle $ \cite{Nomura et
265: al_39_1999}.\ \ Consequently, an increase of the diffuse scattering
266: intensity with decreasing temperature reflects the growth of the polar
267: clusters or the net correlated displacements of lead and oxygen atoms in a $%
268: [111]$ direction as reported in Ref.\cite{Iwase et al_1419_1999} for PZN. \
269: It is also in agreement with the structure factor calculation $F$ developed
270: above, which predicts a greater magnitude of $F.F^{\star \text{ }}$along the
271: $[111]$ direction than along the others.
272:
273: ii) the width squared, $\Delta q_{_{FWHM}}^{2}$, initially decreases almost
274: linearly with temperature, but then deviates from a straight line at
275: approximately 450K, or 40 degrees above the transition $(T_{c}=413$K$)$ and
276: subsequently levels off at low temperature.\ The initial linear dependence
277: is\ the manifestation of a Curie-Weiss law. \ This should be expected in a
278: temperature range in which the polarization fluctuations are dynamic, since
279: the correlation length squared, $\xi ^{2}$, is proportional to the
280: dielectric constant, $\epsilon $, \cite
281: {Jona_Shirane_NewYorK_1962} which itself is expected to follow a Curie-Weiss
282: law initially:
283:
284: \begin{equation}
285: \frac{1}{\Delta q_{_{FWHM}}^{2}}\sim \epsilon =\frac{C}{T-T_{c}}
286: \end{equation}
287:
288: \noindent where $C$ is the Curie constant. \ The departure of $\Delta
289: q_{_{FWHM}}^{2}$from a linear dependence, seen in Fig.3 (a), marks the
290: appearance of long-lived polar fluctuations or permanent polar regions in
291: the crystal, accompanied by local strain fields. \ This also explains the
292: simultaneous increase in Bragg intensity noted above, due to an increase in
293: mosaicity. \ It is worth noting that a very similar evolution has been
294: reported by Iwase et al. \cite{Iwase et al_1419_1999} for the lattice
295: constant, but with a departure from a linear dependence at approximately
296: 550K instead of 450K. \ However, their measurements were made on a powder
297: obtained from a ground crystal and it is well known that relaxors are very
298: susceptible to internal strains. \ The correlation length, $\xi $, shown in Fig.3 (b)
299: increases continuously with decreasing temperature and saturates at low
300: temperatures. \ This is uncommon for conventional ferroelectrics, in which the
301: correlation length decreases below the transition. \ The present behavior
302: indicates that\ the low temperature state is not a homogeneous ferroelectric
303: state but, rather, a mesoscopic one, presumably made of large polar
304: regions. \ At $T_{c}=\allowbreak 413$K, the finite size of the polar
305: regions\ is $\sim $79\AA\ (19 unit cells)\ and $\sim $114\AA\ (28 unit
306: cells) respectively in the $[001]$ and $[111]$ directions. \ The longer
307: correlation length or effective size of the polar regions in the $[111]$
308: direction is found to be $\xi _{\left[ 111\right] }$ $\sim 1.2\xi _{\left[
309: 001\right] }$ (see dotted line in Fig.3b), consistent with the larger Pb
310: displacement component in the $%
311: [111]$ direction. \ \ This result suggests that the polar regions are in the
312: form of ellipsoids, preferentially oriented in $[111]$ directions.\ \
313: Another noticeable difference between the two directions is the appearance
314: of two kinks in the $\xi _{\lbrack 111]}$ curve, at $\sim $ 450K and $\sim $
315: 413K, that are not visible in the $\xi _{\lbrack 001]}$ curve. \ As
316: discussed above, the first one, at 450K, reveals the relatively sudden
317: formation of permanent polar regions, and the second one corresponds to the
318: transition at $T_{c}$. \ The absence of these kinks in the $\xi _{\lbrack
319: 001]}$ curve\ indicates that the growth of the correlation length in that
320: direction is only indirectly affected by the formation of the polar regions
321: at 450K and by their collective realignment at the transition at 413K. \
322: This result is also consistent with the primary net displacements of Pb and
323: three O atoms in one of eight possible $[111]$ directions.
324:
325: In conclusion, thermal diffuse neutron scattering measurements in the
326: paraelectric and relaxor\ temperature regions of the PZN crystal clearly
327: reveal the relatively abrupt formation of polar nanoregions at 450K,
328: approximately 40 degrees above the transition at $T_{c}=\allowbreak 413$K. \
329: With decreasing temperature, these polar regions grow and so does their
330: volume fraction. \ The correlation lengths in the two directions studied are
331: found to be in the ratio $\xi _{\left[ 111\right] }\sim 1.2\xi _{\left[ 001%
332: \right] }$ over the whole temperature range, except for two kinks in the $%
333: \xi _{\left[ 111\right] }$ curve at 450K and 413K. \ The fact that the
334: correlation lengths remain high below the transition indicates that, at low
335: temperature, the polar regions subsist at low temperature in what can be
336: called a mesoscopic ferroelectric phase. Diffuse scattering results at other
337: reciprocal lattice points are presently being analyzed and will be reported
338: in a subsequent paper.
339:
340: We wish to thank R.K. Pattnaik and G. Shirane for helpful discussions. \
341: This work was supported by DOE Grant No. DE-FG02-00ER45842 and ONR Grant No
342: N00014-99-1-0738 as well as the Oak Ridge National Laboratory at the High
343: Flux Isotope Reactor. \
344: \begin{references}
345: \bibitem{Yokomizo et al_1278_1970} Y.Yokomizo, T. Takahashi, and S.
346: Nomura, J. Phys. Soc. Jpn.\ {\bf 28}, 1278 (1970).
347:
348: \bibitem{Randall_Bhalla_327_1990} C.A. Randall and A.S. Bhalla, Jpn. J.
349: Appl. Phys. {\bf 29}, 327 (1990).
350:
351: \bibitem{Smolensky_26_1970} G.A. Smolensky, J. Phys. Soc. Jpn. {\bf 28},
352: Suppl. 26 (1970).
353:
354: \bibitem{Cross_241_1987} L.E. Cross, Ferroelectrics {\bf 76}, 241 (1987).
355:
356: \bibitem{Khuchua et al_194_1968} N.P. Khuchua, V.A. Bokov, and I.E. Myl\'{}nikova,
357: Sov. Phys.-Solid State. {\bf 10}, 194 (1968).
358:
359: \bibitem{Kuwata et al_863_1979} J. Kuwata, K. Uchino,and S. Nomura,
360: Ferroelectrics {\bf 22}, 863 (1979).
361:
362: \bibitem{Mulvihill et al_1462_1997} M. Mulvihill, L.E. Cross, W. Cao, and
363: K. Uchino, J. Am. Ceram. Soc. {\bf 80}, 1462 (1997).
364:
365: \bibitem{Kuwata et al_1298_1982} J. Kuwata, K. Uchino, and S. Nomura, Jpn.
366: J. Appl. Phys., Part1 {\bf 21}, 1298 (1982).
367:
368: \bibitem{Liu et al_2810_1999} S-F. Liu, S-E. Park, T.R. Shrout, and L.E.
369: Cross, J. Appl. Phys. {\bf 85}, 2810 (1999).
370:
371: \bibitem{Gehring et al_5216_2000} P.M. Gehring, S.-E. Park, and G. Shirane,
372: Phys. Rev. Lett. {\bf 84}, 5216 (2000).
373:
374: \bibitem{Iwase et al_1419_1999} T. Iwase, H. Tazawa, K. Fujishiro, Y. Uesu,
375: and Y. Yamada, J. Phys. Chem. Solids. {\bf 60}, 1419 (1999).
376:
377: \bibitem{Takesue et al_2001} N. Takesue, Y. Fujii,and H. You, (unpublished).
378:
379: \bibitem{Yong et al_14736_2000} G. Yong, J. Toulouse, R. Erwin, S. Shapiro,
380: and B. Hennion, Phys. Rev. B {\bf 62}, 14736 (2000).
381:
382: \bibitem{Egami et al_S935_1998} T. Egami, S. Teslic, W. Dmowski, P.K.
383: Davies, and I.-W. Chen, J. Korean Phys. Soc. {\bf 32}, S935 (1998).
384:
385: \bibitem{Krivoglaz_NewYork_1969} M.A. Krivoglaz, Theory of X-Ray and
386: Thermal-Neutron Scattering by Real Crystals, Plenum Press, New York (1969).
387:
388: \bibitem{Schmatz_NewYork_London_1973} W. Schmatz, in X-ray and Neutron
389: Scattering Studies on Disordered Crystals (Edited by Herbert Herman), p.105,
390: New York and London (1973).
391:
392: \bibitem{Zhang et al_96_2000} L. Zhang, M. Dong and Z.-G. Ye, Mater. Sci.
393: Eng. B. {\bf 78}, 96 (2000).
394:
395: \bibitem{Gosula et al_221_2000} V. Gosula, A. Tkachuk, K. Chung, and H.
396: Chen, J. Phys. Chem. Solids. {\bf 61}, 221 (2000).
397:
398: \bibitem{Stanley_1971} H.E. Stanley, Introduction to Phase Transitions and
399: Critical Phenomena, Oxford University Press, (1971).
400:
401: \bibitem{Nomura et al_39_1999} K. Nomura, T. Shingai, S.Ishino, H.
402: Terauchi, N. Yasuda, and H. Ohwa, J. Phys. Soc. Jpn.\ {\bf 68}, 39 (1999).
403:
404: \bibitem{Jona_Shirane_NewYorK_1962} F. Jona. and G. Shirane, Ferroelectric
405: Crystals, Dover Publications, New York (1962).
406: \end{references}
407:
408: \end{document}
409: