1: \documentclass[twocolumn,floats,showpacs,preprintnumbers,amsmath,amssymb,epsf,prb]{revtex4}
2: % Some other (several out of many) possibilities
3: %\documentclass[preprint,aps]{revtex4}
4: %\documentclass[preprint,aps,draft]{revtex4}
5: %\documentclass[prb]{revtex4}% Physical Review B
6: % Include figure files
7: % Align table columns on decimal point
8: % bold math
9: %\nofiles
10: %\documentclass[aps,twocolumn,showpacs,preprintnumbers,epsf]{revtex4}
11: %\documentclass[prb,aps,floating,twocolumn,showpacs]{revtex4}
12: %\documentclass[prb,aps,multicol]{revtex4}
13: %\documentclass[twocolumn,showpacs,floats,prb]{revtex4}
14: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
15: \usepackage{graphicx}
16: \usepackage{dcolumn}
17: \usepackage{bm}
18: %TCIDATA{OutputFilter=LATEX.DLL}
19: %TCIDATA{LastRevised=Mon Jan 12 21:27:34 2004}
20: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
21: %TCIDATA{Language=American English}
22: %\input{tcilatex}
23: \begin{document}
24: \title{Low-temperature phases in PbZr$_{0.52}$Ti$_{0.48}$O$_{3}$ : A neutron powder diffraction
25: study}
26: \author{D. E. Cox}
27: \affiliation{Physics Department, Brookhaven National Laboratory, Upton, NY
28: 11973, USA}
29: \author{B. Noheda}
30: \altaffiliation[Present address: ]{Materials Science Center, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands} \affiliation{Physics Department, Brookhaven
31: National Laboratory, Upton, NY 11973, USA}
32: \author{G. Shirane}
33: \affiliation{Physics Department, Brookhaven National Laboratory, Upton, NY
34: 11973, USA}
35:
36: \date{July 21, 2004}
37: \begin{abstract}
38: A neutron powder diffraction study has been carried out on PbZr$_{0.52}$Ti$_{0.48}$O$_{3}$
39: in order to resolve an ongoing controversy about the nature of the
40: low-temperature structure of this strongly-piezoelectric and
41: technologically-important material. The results of a detailed and systematic
42: Rietveld analysis at 20 K are consistent with the coexistence of two
43: monoclinic phases having space groups $Cm$ and $Ic$ respectively, in the
44: approximate ratio 4:1, and thus support the findings of a recent electron
45: diffraction study by Noheda {\it et al.} [Phys. Rev. B {\bf 66}, 060103 (2002)]. The
46: results are compared to those of two recent conflicting neutron powder
47: diffraction studies of materials of the same nominal composition by Hatch {\it et
48: al.} [Phys. Rev. B {\bf 65}, 212101 (2002)] and Frantti {\it et al.} [Phys. Rev. B {\bf 66},
49: 064108 (2002)].
50: \end{abstract}
51: \pacs{61.12.Ld, 61.50.Ks, 61.66.Fn, 77.84.Bw}
52: %PACS, the Physics and Astronomy Classification Scheme.
53: %\keywords{Suggested keywords}%Use showkeys class option if keyword
54: %display desired
55: \maketitle
56: %\preprint{cond-mat/######}
57:
58:
59: \section{Introduction}
60: The strongly piezoelectric system PbZr$_{1-x}$Ti$_{x}$O$_{3}$ (PZT) has long
61: been known to have a perovskite-type structure with regions of rhombohedral
62: and tetragonal symmetry below the ferroelectric Curie point separated by an
63: almost vertical line at $x \approx 0.5$ in the temperature-composition phase
64: diagram, which is known as the morphotropic phase boundary (MPB).\cite{Jaffe}%
65: Following the recent discovery of a narrow region with monoclinic $Cm$
66: symmetry in the vicinity of the MPB,\cite{Noheda1} numerous experimental
67: and theoretical studies of PZT and related systems have been undertaken in
68: order to clarify the relationships between the structural features and the
69: piezoelectric properties. As previously discussed,\cite{Noheda2} the
70: ferroelectric polarization in the new phase is no longer constrained by
71: symmetry to lie along a symmetry axis, but instead is free to rotate within
72: the symmetry plane. Furthermore, because of the near-degeneracy of the free
73: energies of the various phases, rotation of the polarization axis away from
74: the polar axes of the rhombohedral and tetragonal phases can be accomplished
75: with an applied electric field, resulting in an induced monoclinic phase and
76: a large electromechanical response.\cite{Bellaiche,Cohen}
77:
78: The phase diagram of the PZT system around the MPB as reported in a recent
79: paper by Noheda {\it et al.}\cite{Noheda3} is shown in Fig. 1. Above the Curie
80: temperature, the structure is cubic over the entire range of composition,
81: with space group $Pm3m$ and lattice parameter $a_{0}\approx ${} 4 \AA {}. The
82: rhombohedral region is characterized by high- and low-temperature phases ($\rm R%
83: _{HT}$ and $\rm R_{LT}$) in which there are polar shifts of the
84: atoms along the pseudocubic [111] axis.\cite{Michel,Corker} $\rm
85: R_{HT}$ has space group symmetry $R3m$, with lattice parameters
86: $a_{R} \approx a_{0}$, and $\alpha $ slightly less than 90$^{\circ}$
87: (hexagonal values $a_{H} \approx a_{0}\sqrt{2}, {} c_{H} \approx
88: a_{0}\sqrt{3}$). In $\rm R_{LT}$ there are additional displacements
89: of the oxygen atoms superimposed on the ferroelectric shifts due
90: to antiphase tilting of the oxygen octahedra about the [111] axis,
91: corresponding to an R-point instability. As a consequence, the
92: unit cell is doubled and the mirror plane is destroyed, resulting
93: in the appearance of superlattice peaks in the diffraction
94: pattern. The new space group symmetry is $R3c$, with hexagonal
95: lattice parameters $a_{H} \approx a_{0}\sqrt{2}$ and $c_{H}\approx
96: 2a_{0}\sqrt{3}$. It should be emphasized that these must be
97: regarded as \lq\lq average'' long-range structures, since the
98: presence of short-range order due to local displacements has been
99: clearly demonstrated by the appearance of other types of
100: superlattice peaks in electron diffraction
101: studies\cite{Viehland1,Viehland2,Ricote} not observed in x-ray or
102: neutron diffraction patterns.\cite{Michel,Corker} Significant
103: deviations of the local atomic structure from the crystallographic
104: long-range structure have also been found from pair-distribution
105: function (PDF) analysis of time-of-flight neutron
106: data.\cite{Teslic1,Teslic2,Dmowski}
107:
108: In the tetragonal region of the phase diagram, the space group is
109: $P4mm$ and the polar shifts lie along the [001] axis ($a_{T}
110: \approx c_{T} \approx a_{0},{} c_{T}/a_{T} > 1)$. Nevertheless,
111: the time-of-flight neutron data show that this too should be
112: viewed as an \lq\lq average'' long-range structure. In addition,
113: Raman scattering studies have revealed the presence of local
114: displacements of lower symmetry, which are also reflected in a
115: broadening of some of the x-ray diffraction
116: peaks.\cite{Frantti1,Frantti2} The nature of the local structure
117: has been revealed in more detail from the PDF analysis described
118: in Ref. 14, which shows that there are only gradual changes
119: through the MPB, and suggests that the local environment of each
120: element remains relatively invariant of composition. It is
121: furthermore proposed that the population of local Pb displacements
122: changes between the pseudocubic $<$100$>$ and $<$110$>$ directions
123: as a function of the Ti/Zr ratio. This model is supported by
124: recent theoretical calculations in which the Pb distortions are
125: identified as the determining factor for the average structure of
126: the system.\cite{Grinberg}
127:
128: In the original x-ray
129: study by Noheda {\it et al.}\cite{Noheda1} the unit cell of the low-temperature
130: monoclinic phase (now usually designated $\rm M_{A}$\cite{Cohen}) was found to be doubled
131: with respect to the primitive cell, with the monoclinic $a$ and $b$ axes
132: directed along the [110] and [1$\overline{1}$0] axes of the latter ($a%
133: _{M} \approx b_{M} \approx a_{0}\surd 2, c_{M} \approx a_{0}$,
134: space group $Cm$). Based upon the atomic positions determined from
135: Rietveld analysis of the synchrotron x-ray data from PbZr
136: $_{0.52}$Ti$_{0.48}$O$_{3}$, it was concluded\cite{Noheda2} that
137: at 20 K the polar axis was tilted about 24$^{\circ}$ from the [001] axis
138: towards the pseudocubic [111] axis. The structure can be regarded
139: as a condensation of either the local displacements present in the
140: tetragonal $P4mm$ phase along one of the $\left\langle
141: 110\right\rangle $ directions, or alternatively those present in
142: the rhombohedral $R3m$ phase along one of the $\left\langle
143: 100\right\rangle $ directions, as inferred by Corker {\it et
144: al.}\cite{Corker}
145:
146: However, it is clear that there is a missing
147: ingredient in this simple picture, for in a neutron powder
148: diffraction study of the same sample, Noheda {\it et al.} reported
149: the presence of one very weak superlattice peak at 20 K
150: corresponding to a doubling of the $c$ axis of this monoclinic
151: cell, but did not identify the nature of this additional
152: distortion.\cite{Noheda3} A similar cell-doubled phase was also
153: observed for $x = 0.48$ by Ragini and coworkers in electron
154: diffraction patterns below 200 K, but not in their low-temperature
155: x-ray patterns.\cite{Ragini} Based on a subsequent Rietveld
156: analysis of neutron powder data collected at 10 K, the structure
157: of this new phase was reported by Ranjan {\it et al.} to be
158: monoclinic, with space group $Pc$.\cite{Rajan} The appearance of
159: the weak superlattice reflections was attributed to antiphase
160: tilting of the oxygen octahedra about the [001] direction,
161: corresponding to an R-point instability in the cubic Brillouin
162: zone. It was later reported that the correct space group for this
163: proposed model was in fact $Cc$,\cite{Hatch} and a modified set of
164: refined structural parameters was presented.
165:
166: The $x = 0.48$ composition has
167: also been the subject of a recent low-temperature neutron powder study by
168: Frantti and colleagues.\cite{Frantti3} They, too, note the presence of
169: similar superlattice reflections, but reach very different conclusions;
170: namely that these reflections are attributable to a minority rhombohedral
171: phase with $R3c$ symmetry in coexistence with the monoclinic $Cm$ phase, a model
172: that was not considered by Ranjan {\it et al.}\cite{Rajan} or Hatch {\it et al.}\cite
173: {Hatch} In a footnote to their paper, Frantti {\it et al.} comment that the
174: monoclinic $ Pc$ and $Cc$ models proposed by the latter authors predict peaks
175: that are not observed experimentally, and that the observed superlattice
176: peaks can be accounted for by the $R3c$ phase. However, this conclusion was
177: not supported by the results obtained by Noheda {\it et al.}\cite{Noheda4} in an
178: electron diffraction study of the same $x = 0.48$ sample used in the earlier
179: x-ray study,\cite{Noheda2} which showed no evidence for a rhombohedral
180: phase, but instead the monoclinic $Cm$ phase in coexistence with nanoregions
181: of a minority $Cc$ phase ranging in size from 30-100 \AA {}. These conclusions
182: have been questioned by Frantti {\it et al.}, who comment that their neutron data
183: provide no evidence of a $Cc$ phase, and argue that since electron diffraction
184: probes only small volumes of the sample, it is generally not suitable for
185: the determination of average symmetry, and furthermore that the ion-milling
186: technique used for sample thinning is a very violent one which can easily
187: generate significant defects.
188:
189: In the light of these different interpretations, we have undertaken a
190: detailed Rietveld analysis of the neutron data cited by Noheda {\it et al.}\cite
191: {Noheda3} in an attempt to discriminate between the three models described
192: above. Plausible results were obtained in all three cases, illustrating how
193: difficult it is to identify the correct structural model in complex systems
194: of this type simply on the basis of the standard goodness-of-fit criteria.
195: Nevertheless, we conclude that, taken in conjunction with the electron
196: diffraction data, the results point strongly towards the coexistence model
197: of $Cm$ and minority $Cc$ phases.
198: \section{Experimental}
199: The sample consisted of about 4 g of sintered pellets roughly 1 cm in
200: diameter and 1 mm thick from the same batch of material used in the previous
201: x-ray study. \cite{Noheda2} Long-range fluctuations in the composition of
202: the x-ray sample, $\Delta x$, were estimated to be less than $\pm $ 0.003
203: based upon an analysis of the peak widths. The pellets were loaded into a
204: thin-walled vanadium can and mounted in a closed-cycle helium cryostat. Data
205: were collected at the NIST reactor on the powder diffractometer BT1 with a
206: Cu monochromator set for a wavelength of 1.54 \AA {}, collimation of 15$^{\prime}$ and
207: 40$^{\prime}$ before and after the monochromator, and 10$^{\prime}$ in front of each of the 32 $%
208: ^{3}$He detectors. With this configuration, the best angular
209: resolution attained is about 0.2$^{\circ}$ at 2$\theta \approx$ 80$^{\circ}$
210: ($\Delta d/d
211: \approx 2 \times 10^{-3}$).
212:
213: Extended data sets were collected at 2$%
214: \theta $ step intervals of 0.05$^{\circ}$ in the monoclinic region at 20
215: K, in the vicinity of the monoclinic-tetragonal transition at 325
216: K, and in the tetragonal region at 550 K. Analysis of the data was
217: carried out with the FULLPROF program,\cite{FULLPROF} using the
218: pseudo-Voigt peak-shape function with appropriate corrections for
219: instrumental asymmetric broadening, \cite {Finger} and linear
220: interpolation between background points. Particular attention was
221: paid to the problem of anisotropic peak-broadening, which reflects
222: the fact that closely-adjacent peaks may have markedly different
223: widths arising from local strains or compositional fluctuations,
224: for example, as previously noted for PZT and related piezoelectric
225: systems.\cite {Noheda1,Noheda3,Frantti2,Kiat} In standard
226: Rietveld analysis the peak widths are assumed to vary smoothly as
227: a function of scattering angle, and it is important to note that
228: anisotropic peak-broadening due to microstructural effects can be
229: mistakenly interpreted as a symmetry-lowering distortion of the
230: unit cell of the average long-range structure. With the rapidly
231: increasing use of high-resolution x-ray and neutron techniques, it
232: is becoming clear that anisotropic peak-broadening is a common
233: feature of powder diffraction patterns, and should be allowed for
234: as appropriate. One convenient way to do this is provided by the
235: phenomenological model recently proposed by
236: Stephens,\cite{Stephens} in which the broadening is represented by
237: a series of coefficients $\sum_{HKL}S_{HKL}$$h^{H}k^{K}l^{L}$
238: $(H+K+L=4)$, which take into acccount the Laue symmetry of the
239: space group and are incorporated as refinable parameters in the
240: Rietveld program. For tetragonal $4mm$ and monoclinic $2/m$
241: symmetry, there are respectively 4 and 9 such coefficients.
242:
243: The data analysis is now described in detail for the tetragonal phase at 550
244: K, the monoclinic phase at 20 K, and the intermediate region at 325 K.
245: \subsection{550 K}
246: All the peaks could be unambiguously indexed in terms of a tetragonal cell
247: with $a = 4.060, c = 4.100$ \AA {}, except for two very weak peaks
248: attributable to the vanadium sample holder, which were excluded from the
249: analysis. Rietveld refinement was carried out with individual isotropic
250: temperature factors assigned and the atoms placed in the following positions
251: of space group $P4mm$: Zr/Ti and O(1) in 1(a) sites at $0, 0, z$; O(2) in 2(c)
252: sites at $0.5, 0, z$; and Pb statistically distributed among the 4(d) sites at
253: $x, x, 0$. The Pb positions correspond to random displacements in the
254: $\left\langle 110\right\rangle$ directions away from the origin, as noted
255: in the previous synchrotron x-ray study.\cite{Noheda2} The
256: refinement converged rapidly and smoothly to a goodness-of-fit $\chi ^{2}$
257: value of 1.30. However, as in the x-ray study, an examination
258: of the observed and calculated peak profiles revealed a number of systematic
259: discrepancies indicative of anisotropic peak broadening, and additional
260: refinements were carried out in which various combinations of the four
261: possible coefficients were allowed to vary. A definite improvement was
262: obtained when $S_{004}$ was varied alone ($\chi ^{2}$ = 1.20), but the
263: results obtained with additional coefficients were judged to be of dubious
264: significance, and these coefficients were accordingly set to zero. This
265: result most likely reflects the sensitivity of the $c$ lattice parameter to
266: the presence of long-range compositional fluctuations. The final refinement
267: results based on this model are listed in Table I, and the profile and
268: difference plots are shown in Fig. 2.
269: \subsection{20 K}
270: A series of refinements was carried out for each of the three models
271: described in the Introduction, namely: single-phase $Cc$, two-phase $Cm/Cc$,
272: and two-phase $Cm/R3c$. However, instead of $Cc$, the non-conventional space
273: group setting $Ic$ was chosen, which has the distinct advantage of having a
274: near-orthogonal unit cell closely related to the $Cm$ cell, in which the
275: mirror plane is replaced by a $c$-glide plane and the $c$ axis is doubled. The
276: unit-cell axes are related via the transformation {\bf a}$_I$ = -{\bf c}$_C$,
277: {\bf b}$_I$ = {\bf b}$_C$, {\bf c}$_I$ = {\bf a}$_C$ + {\bf c}$_C$, where the
278: subscripts refer to the unit cells of the $Ic$ and $Cc$ space groups respectively. In
279: this setting, it is much easier to visualize the small displacements from
280: the ideal $Cm$ atomic positions. In the $Ic$ cell, the Pb atom was chosen to lie
281: at the origin, with Zr/Ti and three inequivalent O atoms in fourfold general
282: positions at $x, y, z$ and $x, -y, 1/2+z$, and at the related body-center sites.
283: The Zr/Ti and O(1) atoms are in positions similar to those in the $Cm$
284: structure at $x, 0, z$, the main difference being that they are no longer
285: required to lie on a mirror plane at $y = 0$. The O(2) and O(3) atoms are in
286: two sets of positions derived from the $x, y, z$ sites and the symmetry-equivalent
287: mirror plane sites at $x, -y, z$ in the $Cm$ structure.
288: \subsubsection {Single-phase $Ic$ model}
289: In the first series of refinements the
290: atoms were intially assigned the positions found in our previous
291: x-ray study\cite {Noheda2} with the exception of the O(2) and O(3)
292: atoms, which were displaced from the ideal $Cm$ positions by small
293: shifts corresponding to antiphase tilting of the oxygen octahedra
294: about the [001] axis, as assumed by Ranjan {\it et
295: al.}\cite{Rajan} The corresponding positions chosen for O(2) and
296: O(3) in the $Ic$ structure were $x - \delta, y - \delta, z/2$ and
297: $1/2 + x + \delta, 1/2 + y + \delta, z/2$, where $\delta$ is the
298: shift in the $x$ and $y$ directions due to tilting. It is
299: important to note that with such a constrained-tilt,
300: rigid-octahedron model, the $x$ and $y$ values assumed for the
301: O(2) and O(3) positions (in this case the $Cm$ values found in the
302: previous x-ray study) do not change in the course of the
303: refinement. The resulting fit was reasonably good ($\chi ^{2}$ =
304: 1.97), but inspection of the individual peak profiles once again
305: revealed some significant discrepancies due to anisotropic peak
306: broadening. As before, a distinct improvement was obtained when
307: the anisotropy coefficient $S_{004}$ was refined ($\chi ^{2}$ =
308: 1.78), but further refinements with various combinations of the
309: other eight anisotropy coefficients gave only minimally improved
310: fits, and the one-parameter anisotropy model was accordingly
311: adopted for subsequent refinements.
312: At this point, the constraints on the Zr/Ti and O(1) $y$ parameters were
313: relaxed, but the shifts from the ideal positions at $y$ = 0 and the
314: improvement in the overall fit were judged to be insignificant. A similar
315: result was obtained when the Zr/Ti compositional parameter $x$ was varied.
316:
317: Further refinements were performed, first with constrained models
318: corresponding to rigid-octahedron tilting about the [111] and [110] axes
319: respectively, and finally, with all the constraints on the O(2) and O(3)
320: positions removed. In the latter case, the refinement proceeded smoothly and
321: converged rapidly to a set of positions which were much closer to those of
322: the [001]-tilt model than the other tilt models independent of which tilt
323: model was used to provide the initial values of the positions. However, in
324: none of these cases did the overall fit appear to be significantly improved,
325: and we therefore conclude that the [001]-tilt model is a reasonable choice,
326: although it is clearly not possible to rule out the other models on the
327: basis of the present data. The final refinement was therefore carried out
328: for the constrained [001]-tilt model, but with the $x$ and $y$ values for O(2)
329: and O(3) derived from the results for the unconstrained model. This
330: refinement yielded a $\chi ^{2}$ value of 1.52, with refined parameters as
331: listed in Table II (column 1). Also shown are the values reported by Hatch
332: {\it et al.}\cite{Hatch} transformed from $Cc$ to $Ic$ symmetry (column 3). From a
333: comparison of the two sets of atomic positions, it appears that the
334: constraints applied by Hatch {\it et al.} do not in fact correspond to an
335: [001]-tilt model, but instead to a simpler model in which only the $y$
336: parameters of the O(2) and O(3) atoms are displaced from their ideal $Cm$
337: positions. An additional refinement based on such a $y$-shift model yielded
338: results which are seen to be in excellent agreement with those of Hatch {\it et
339: al.} (column 2 of Table II), although the fit is somewhat inferior to that
340: given by the [001]-tilt model ($\chi ^{2}$ = 1.62). It is also worth noting
341: that the values of axial ratio, $c_{0}/a_{0}$, and the pseudocubic cell volume, $V_{0}$,
342: obtained by the latter authors suggest a slightly higher Zr content (%
343: $\approx 0.5$\%) relative to the present sample.\cite{Noheda3,Frantti3}
344:
345: From the atomic positions listed in the first column of Table II
346: the octahedral tilt angle is calculated as about 3$^{\circ}$. The
347: polar displacements of the Zr/Ti and Pb atoms with respect to the
348: respective polyhedra centers are -0.08 and -0.22 \AA {} along
349: monoclinic [100], and 0.18 and 0.44 \AA {} along [001],
350: corresponding to a rotation of the polar axis towards pseudocubic
351: [111] of roughly 25$^{\circ}$. However, these values are
352: representative only of the average long-range structure, since
353: they do not allow for the local distortions revealed in the PDF
354: analysis cited earlier.\cite{Dmowski}
355: \subsubsection{Two-phase $Cm/R3c$ model}
356: The next set of refinements was performed for the two-phase
357: $Cm/R3c$ model favored by Frantti {\it et al.}\cite{Frantti3}
358: Significantly better peak profiles were obtained with an
359: anisotropic-broadening model for the $Cm$ phase in which $S_{004}$
360: was allowed to vary, together with an isotropic particle-size
361: broadening coefficient for the $R3c$ phase. The refinement
362: converged rapidly to a $\chi ^{2}$ value of 1.27 with the final
363: parameters as listed in Table III. The latter are in close
364: agreement with those reported by Frantti {\it et al.}, including
365: the respective weight fractions of the two phases. Compared to the
366: single-phase $Ic$ model, the overall fit is considerably better
367: ($\chi ^{2}$ = 1.27 vs. 1.52), but because several additional
368: variable parameters are involved, it is difficult to judge the
369: true significance of this result.
370: Inspection of the results listed in Table III reveals that in both samples, $V_{0}$
371: for the rhombohedral phase is larger by about 0.3 \AA $^{3}$, which
372: would imply a significantly higher Zr content of some 3-4\%.\cite
373: {Noheda3,Frantti3} For the present sample, at least, such a conclusion
374: would be inconsistent with the previously estimated long-range compositional
375: fluctuations.\cite{Noheda2}
376: \subsubsection{Two-phase $Cm/Ic$ model}
377: The final set of refinements was carried out for the two-phase
378: $Cm/Ic$ model deduced by Noheda {\it et al.} from the results of
379: an electron diffraction study.\cite{Noheda4} Since we did not
380: anticipate that a meaningful result would be obtained for an
381: unconstrained refinement of two such closely-related structures, a
382: highly constrained model was used; namely, the atomic positions in
383: the $Cm$ and $Ic$ phases were constrained to be equivalent except
384: for one additional parameter $\delta $ for the latter representing
385: the displacement along the $x$ and $y$ axes for the idealized
386: [001]-tilt model described above. The peak-shape model, including
387: an
388: $S_{004}$ anisotropy coefficient, was also constrained to be equivalent for
389: both phases, except for an isotropic particle-size broadening coefficient
390: which was included for the $Ic$ phase.
391: The refined values for the two phases
392: are listed side-by-side in Table IV for easy comparison, and the profile fit
393: and difference plot are shown in Fig. 3. The relative proportions of the $Cm$
394: and $Ic$ phases are approximately 4:1 and thus consistent with the electron
395: diffraction results, but the estimated particle size derived from the
396: broadening coefficient is much larger, about 1000 \AA {} compared to
397: 100 \AA {}. As pointed out by Frantti {\it et al.},\cite{Frantti3} this
398: discrepancy could arise because of the ion-milling techniques used to thin
399: the electron diffraction sample, which can generate significant numbers of
400: defects.
401:
402: Detailed comparison of the results in Tables II, III and IV
403: reveals that a better fit is obtained with the two-phase $Cm/Ic$
404: model ($\chi ^{2}$ = 1.16, $R_{wp} = 0.064)$ than with the
405: $Cm/R3c$ model ($\chi ^{2}$ = 1.27, $R_{wp} = 0.067$) or the
406: single-phase $Ic$ model ($\chi ^{2}$ = 1.52, $R_{wp} = 0.073$),
407: but it would nevertheless be premature to conclude that the former
408: must therefore be correct, since there are no generally-accepted
409: statistical tests to judge the true significance of the results.
410: However, although the extended profile and difference plots for
411: the two latter models are hardly distinguishable by eye from those
412: shown in Fig. 3, there are significant differences in some of the
413: individual peak profiles which provide additional insight, as
414: shown in Figs. 4 and 5 respectively. Fig. 4 shows the region
415: around the strongest superlattice peak at 2$\theta \approx
416: 36.8^{\circ}$, from which it is evident that a much better fit is
417: obtained with the $Ic$ and $Cm/Ic$ models than with the $Cm/R3c$
418: model. On the other hand, the fit shown in Fig. 5 in the
419: pseudocubic (200) region reveals serious deficiencies for the $Ic$
420: model compared to the $Cm/R3c$ and particularly the $Cm/Ic$ model,
421: which accounts much better for the asymmetry of the profiles in
422: the central region.
423:
424: We note also that further analysis of the synchrotron x-ray data reported in
425: Ref. 3 shows that the $Cm/Ic$ coexistence model is superior to
426: the two-phase $Cm/Pm3m$ model previously used, with $\chi ^{2}$ values of 7.3
427: and 12.8 respectively. The results are in reasonable agreement with those of
428: the neutron study; in particular, the ratio of the two phases is found to
429: be about 4:1, very similar to the value listed in Table IV. The failure to
430: detect any superlattice peaks analogous to the one in the neutron pattern
431: can be explained by the relatively much weaker x-ray scattering power of
432: oxygen compared to Pb and Zr/Ti, resulting in calculated intensities that
433: are insignificant compared to the background signal.
434:
435: From the atomic positions listed in Table IV, the octahedral tilt
436: angle about the [001] axis is calculated as about 7$^{\circ}$. The
437: polar shifts of the Zr/Ti and Pb atoms with respect to the
438: polyhedra centers are essentially the same as those obtained for
439: the single-phase $Ic$ refinement. It is also seen that the values
440: of the cell volume $V_0$
441: and the axial ratio $c_0/a_0$ for the $Ic$ phase are respectively slightly larger and smaller
442: than those for $Cm$, and thus suggestive of a slightly higher Zr content ($\approx 0.5\%$)
443: for the former.\cite{Noheda3,Frantti3}
444: \subsection{325 K}
445: Refinement was first carried out based on a model similar to that used for
446: the 550 K data; namely, a single-phase tetragonal structure with $P4mm$
447: symmetry, Pb atoms statistically distributed among the 4(d) sites at $x, x,
448: 0$, and a single anisotropy-broadening coefficient, $S_{004}$. However, the
449: overall fit was only mediocre ( $\chi ^{2}$ = 2.94, $R_{wp} = 0.072$), and a
450: detailed inspection of the individual peak profiles revealed asymmetries
451: consistent with the presence of a monoclinic component. Such a coexistence
452: model of monoclinic and tetragonal phases for $x = 0.48$ at room temperature
453: was proposed in an earlier neutron study by Frantti {\it et al.},\cite{Frantti4}
454: and in a more recent x-ray study by Ragini {\it et al.}\cite{Ragini2} Further
455: refinements based on this model gave a markedly improved fit ( $\chi ^{2}$ = 1.76,
456: $R_{wp}$ = 0.056), but some residual diffuse scattering was clearly
457: present between some of the peaks. This scattering is probably associated
458: with locally disordered regions in the vicinity of domain walls and can be
459: modeled in a simple, albeit rather artificial, way by the addition of a
460: cubic phase with $Pm3m$ symmetry, as assumed in our previous x-ray study.\cite
461: {Noheda2} Such a three-phase model yielded a reasonably satisfactory fit
462: ($\chi ^{2}$ = 1.47, $R_{wp} = 0.051$), with weight fractions of tetragonal,
463: monoclinic and cubic phases in the ratio 0.61:0.33:0.06.
464:
465: The refined parameters are listed in Table V, and the profile fit and difference plot
466: are shown in Fig. 6. Also listed are the parameters reported by Ragini {\it et
467: al.}\cite{Ragini} and Frantti {\it et al.}\cite{Frantti3} In the latter case, it
468: is rather surprising in the light of the results reported in Ref. 6
469: that the lattice strain $c_{0}/a_{0}$ was found to be
470: significantly larger for the monoclinic phase than for the tetragonal one, since one would not expect
471: rotation of the polarization direction away from [001] in the monoclinic
472: phase to increase this strain. Other than this, the three sets of parameters
473: are in reasonable agreement except that the fraction of $Cm$ phase in the room
474: temperature studies is considerably larger than at 325 K, as would be
475: expected.
476:
477: Further analysis of the earlier x-ray data\cite{Noheda2} revealed that this
478: three-phase model gives a noticeably better profile fit than that obtained
479: with the two-phase $P4mm/Pm3m$ model previously used, with $\chi ^{2}$ values
480: of 7.5 and 9.9 respectively. The weight fractions of the three phases were
481: in the ratio 0.55:0.40:0.05, comparable to the neutron values listed in
482: Table V. We emphasize, however, that these results should be regarded as
483: representative only of an average long-range structure, since the
484: true nature of the material in the transition region is surely far more
485: complex than implied by a simple three-phase model. It is more likely
486: in this temperature interval that small fluctuations in
487: composition lead to coexistence of the tetragonal phase with
488: locally-ordered monoclinic regions of widely varying sizes and possibly some
489: disordered regions.
490: \section{Discussion}
491: In summary, the results obtained in the present neutron investigation are
492: consistent with the coexistence of majority $Cm$ and minority $Ic$ phases in PbZr%
493: $_{0.52}$Ti$_{0.48}$O$_{3}$, in agreement with the results of a recent
494: electron diffraction study of the same sample. The $Cm$ ($\rm M_{A}$) phase, which
495: plays a key role in the piezoelectric and ferroelectric behavior of PZT and
496: related systems is the majority phase at low temperature. The structure of
497: the minority $Ic$ phase is readily visualized as the superposition of an
498: antiphase octahedral-tilt system on the parent $Cm$ structure. Furthermore,
499: the close agreement between the refinement results for the alternative $Ic$
500: and $Cm/R3c$ models and those in Refs. 20 and 21
501: respectively suggests that the two-phase $Cm/Ic$ model is worth consideration
502: in those cases as well. It is possible that the coexistence of $Cm$ and $Ic$
503: phases in the present $x = 0.48$ sample reflects the existence of a narrow
504: thermodynamically-stable region with $Ic$ symmetry at low temperature
505: somewhere between 0.45$ < x <$ 0.48.
506: In this case, the coexistence of $Ic$ and $Cm$ phases could plausibly be
507: attributed to the presence of long-range compositional fluctuations, as suggested
508: by the values of the lattice parameters for the two-phase refinement in Table V.
509: In this context, it is interesting to note that recent neutron data obtained by Frantti {\it et al.}
510: for a sample with $x = 0.46$ show clear evidence of a superlattice peak at 4 K.\cite{Frantti5}
511: However, the authors interpret this as evidence for the coexistence of $Cm$ and $R3c$
512: phases, and did not consider the possibility of $Cm$ and $Ic$ phase coexistence.
513: Alternatively, the $Ic$ phase could be a metastable one resulting from the
514: presence of local strains at domain-wall boundaries, for example. Indeed, it
515: is noteworthy that first-principles calculations by Fornari and Singh have
516: shown that local stress fields may lead to the coexistence of both
517: ferroelectric and rotational instabilities near the MPB.\cite{Fornari} In
518: any case, it is clear that a very careful high-resolution x-ray, neutron and
519: electron diffraction study of extremely well-characterized samples would be
520: required in order to throw further light on these issues.
521:
522: On a final note of
523: caution, the present study also demonstrates that the interpretation of the
524: results of Rietveld analysis may be very tricky for complex systems such as
525: this one in which allowance must be made for the possible coexistence of
526: closely-related phases and the presence of anisotropic peak broadening. The
527: choice of any particular model should take into account not only the quality
528: of the refinement as judged by the agreement factors and goodness-of-fit,
529: but also the diffraction profiles of alternative models in selected key
530: regions of the pattern, and, if feasible, data from complimentary structural
531: techniques such as electron diffraction.
532: \begin{center}
533: \textbf{ACKNOWLEDGMENTS}
534: \end{center}
535: We would like to acknowledge the
536: support of the NIST Center for Neutron Research, U. S. Department
537: of Commerce, and to thank B. H. Toby for experimental assistance
538: at beamline BT1. Financial support from the
539: U.S. Department of Energy, Division of Materials Sciences, under contract No. DE-AC02-98CH10886
540: is also acknowledged.
541:
542: \begin{references}
543: \bibitem{Jaffe} B. Jaffe, W. R. Cook, and H. Jaffe, Piezoelectric Ceramics (Academic, London, 1971).
544: \bibitem{Noheda1} B. Noheda, D.E. Cox, G. Shirane, J.A. Gonzalo, L.E. Cross, and S-E. Park, Appl. Phys. Lett. {\bf 74}, 2059 (1999).
545: \bibitem{Noheda2} B. Noheda, J.A. Gonzalo, L.E. Cross, R. Guo, S-E. Park, D.E. Cox, and G. Shirane, Phys. Rev. B {\bf 61}, 8687 (2000).
546: \bibitem{Bellaiche} L. Bellaiche, A. Garcia, and D. Vanderbilt. Phys. Rev. Lett. {\bf 84}, 5427 (2000).
547: \bibitem{Cohen} D. Vanderbilt and M.H. Cohen, Phys. Rev. B {\bf 63}, 094108 (2001).
548: \bibitem{Noheda3} B. Noheda, D.E. Cox, G. Shirane, R. Guo, B. Jones, and L.E. Cross, Phys. Rev. B {\bf 63}, 014103 (2001).
549: \bibitem{Michel} C. Michel, J. M. Moreau, G. D. Achenbach, R. Gerson, and W. J. James, Solid State Commun. {\bf 7}, 865 (1969).
550: \bibitem{Corker} D. L. Corker, A.M. Glazer, R.W. Whatmore, A. Stallard, and F. Fauth, J. Phys.: Condensed Matter {\bf 10}, 6251 (1998).
551: \bibitem{Viehland1} D. Viehland, Phys. Rev. B {\bf 52}, 778 (1995).
552: \bibitem{Viehland2} D. Viehland, J-F. Li, X. Dai, and Z. Xu, J. Phys. Chem. Solids {\bf 57}, 1545 (1996).
553: \bibitem{Ricote} J. Ricote, D. L. Corker, R.W. Whatmore, S. A. Impey, A.M. Glazer, J. Dec, and K. Roleder, J. Phys.: Condensed Matter {\bf 10}, 1767 (1998).
554: \bibitem{Teslic1} S. Teslic, T. Egami, and D. Viehland, J. Phys. Chem. Solids {\bf 57}, 1537 (1996).
555: \bibitem{Teslic2} S. Teslic, T. Egami, and D. Viehland, Ferroelectrics {\bf 194}, 271 (1997).
556: \bibitem{Dmowski} W. Dmowski, T. Egami, L. Farber, and P. K. Davies, in {\it Fundamental Physics of Ferroelectrics - Eleventh Williamsburg Ferroelectrics Workshop}, edited by H. Krakauer, AIP Conf. Proc. No. 582 (AIP, Woodbury, NY, 2001), p. 33.
557: \bibitem{Frantti1} J. Frantti, J. Lappalainen, V. Lantto, S. Nishio, and M. Kakihana, Jpn. J. Appl. Phys. {\bf 38}, 5679 (1999).
558: \bibitem{Frantti2} J. Frantti, S. Ivanov, J. Lappalainen, S. Eriksson, V. Lantto, S. Nishio, M. Kakihana, and H. Rundl\"{o}f, Ferroelectrics {\bf 266}, 73 (2002).
559: \bibitem{Grinberg} I. Grinberg, V. R. Cooper, and A. M. Rappe, Nature {\bf 419}, 909 (2002).
560: \bibitem{Ragini} Ragini, S. K. Mishra, D. Pandey, H. Lemmens, and G. Van Tendeloo, Phys. Rev. B {\bf 64}, 054101 (2001).
561: \bibitem{Rajan} R. Ranjan, Ragini, S. K. Mishra, D. Pandey, and B. J. Kennedy, Phys. Rev. B {\bf 65}, 060102 (2002).
562: \bibitem{Hatch} D. M. Hatch, H. T. Stokes, R. Ranjan, Ragini, S. K. Mishra, D. Pandey, and B. J. Kennedy, Phys. Rev. B {\bf 65}, 212101 (2002).
563: \bibitem{Frantti3} J. Frantti, S. Ivanov, S. Eriksson, H. Rundl\"{o}f, V. Lantto, J. Lappalainen, and M. Kakihana, Phys. Rev. B {\bf 66}, 064108 (2002).
564: \bibitem{Noheda4} B. Noheda, L. Wu, and Y. Zhu, Phys. Rev. B {\bf 66}, 060103 (2002).
565: \bibitem{FULLPROF} J. Rodriguez-Carvajal, Program FULLPROF, version 1.9c (2001). Available by anonymous ftp from www.charybde.saclay.cea.fr/pub/divers/.
566: \bibitem{Finger} L. W. Finger, D. E. Cox, and A. W. Jephcoat, J. Appl. Cryst. {\bf 27}, 892 (1994).
567: \bibitem{Kiat} J-M. Kiat, Y. Uesu, B. Dkhil, M. Matsuda, C. Malibert, and G. Calvarin, Phys. Rev. B {\bf 65}, 064106 (2002).
568: \bibitem{Stephens} P. W. Stephens, J. Appl. Cryst. {\bf 32}, 281 (1999).
569: \bibitem{Frantti4} J. Frantti, J. Lappalainen, S. Eriksson, V. Lantto, S. Nishio, M. Kakihana, S. Ivanov, and H. Rundl\"{o}f, Jpn. J. Appl. Phys. {\bf 39}, 5697 (2000).
570: \bibitem{Ragini2} Ragini, R. Ranjan, S. K. Mishra, and D. Pandey, J. Appl. Phys. {\bf 92}, 3266 (2002).
571: \bibitem{Frantti5} J. Frantti, S. Eriksson, S. Hull, V. Lantto, H. Rundl\"{o}f, and M. Kakihana, J. Phys.: Condensed Matter {\bf 15}, 6031 (2003).
572: \bibitem{Fornari} M. Fornari and D. J. Singh, Phys. Rev. B {\bf 63}, 092101 (2001).
573: \end{references}
574:
575: \newpage
576:
577: \begin{table}[p]
578: \centering \caption{{\protect\small Refined structural parameters
579: for tetragonal PbZr$ _{0.52}$Ti$_{0.48}$O$_{3}$
580: at 550 K, space group $P4mm$, lattice parameters $a = 4.0596(1), c = 4.0999(1)$ \AA . The refinement was based on a model with the Pb atoms statistically distributed among 4(d) sites at
581: $x, x, 0,$ corresponding to local displacements along
582: $\left\langle 110\right\rangle $ directions. Figures in
583: parentheses denote standard errors referred to the least
584: significant digit. $R_{wp}$,
585: $R_{B}$ and $\protect\chi ^{2}$
586: are agreement factors as defined in Ref. \protect\onlinecite{FULLPROF}. }}
587: \begin{tabular}{lcccc}\\
588: \hline\hline
589: & \small $x$ & \small $y$ & \small $z$ & {\small$U$(\AA $^{2}$)} \\
590: \hline
591: \small Pb & 0.033(1) & 0.033(1) & 0.0 & 0.028(1) \\
592: \small Zr/Ti & 0.5 & 0.5 & 0.450(2) & 0.005(1) \\
593: \small O(1) & 0.5 & 0.5 & -0.061(1) & 0.027(1) \\
594: \small O(2) & 0.5 & 0.0 & 0.427(1) & 0.027(1) \\
595: & & & & \\
596: \small $R_{wp}$ & \multicolumn{4}{c}{0.048} \\
597: \small $R_{B}$ & \multicolumn{4}{c}{0.034} \\
598: \small $\chi 2$ & \multicolumn{4}{c}{1.20} \\
599: \hline\hline
600: \end{tabular}
601: \end{table}
602: \begin{table}[p]
603: \centering \caption{{\protect\small Refined structural parameters
604: for monoclinic PbZr$_{0.52}$Ti$_{0.48}$O$_{3}$
605: at 20 K, single-phase model with space group $Ic$, for
606: the [001]-tilt and $y$-shift models described in the text. The Pb
607: atom was fixed at the origin, and the O(2) and O(3) temperature
608: factors were constrained to be equal. Also listed are the
609: parameters recently reported by Hatch $et~ al.$
610: \protect\cite{Hatch}, but with the values transformed from $Cc$
611: to $Ic$ symmetry. $V_{0}$ and $c_{0}/a_{0}$ represent respectively
612: the volume and axial ratio of the primitive pseudocubic cell, with
613: $c_{0} =c/2$ and $a_{0}=(a+b)/2\surd 2$. }}
614: \begin{tabular}{lccc}\\
615: \hline\hline
616: & \multicolumn{2}{c}{\small Present study} & {\small Hatch \it{et al.}}\\
617: & {\small $[001]-$tilt} & {\small $y-$shift} & {\small $y-$shift}
618: \\ \hline
619: {\small $a($\AA $)$} & 5.7131(1) & 5.7131(1) & 5.7312(7) \\
620: {\small $b($\AA $)$} & 5.7000(1) & 5.7001(1) & 5.7093(6) \\
621: {\small $c($\AA $)$} & 8.2679(2) & 8.2683(3) & 8.2363(7) \\
622: {\small $\beta (^{o})$} & 90.475(2) & 90.473(2) & 90.50(1) \\
623: {\small $V_{o}($\AA $^{3})$} & 67.31 & 67.31 & 67.37 \\
624: {\small $c_{o}/a_{o}$} & 1.0245 & 1.0246 & 1.0181 \\
625: & & & \\
626: {\small Pb: \ \ $U($\AA $^{2})$} & 0.013(1) & 0.012(1) & 0.013(1) \\
627: {\small Zr/Ti: \ \ \ \ $x$} & 0.524(2) & 0.524(2) & 0.519(5) \\
628: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ $z$} & 0.219(1) & 0.218(1) & 0.216(2) \\
629: {\small \ \ \ \ \ \ \ \ $U($\AA $^{2})$} & 0.002(1) & 0.003(2) & 0.006(4) \\
630: {\small O(1): \ \ \ \ \ \ $x$} & 0.542(1) & 0.543(1) & 0.548(3) \\
631: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ $z$} & -0.046(1) & -0.046(1) & -0.044(1) \\
632: {\small \ \ \ \ \ \ \ \ $U($\AA $^{2})$} & 0.011(1) & 0.011(1) & 0.011(3) \\
633: {\small O(2): \ \ \ \ \ \ $x$} & 0.275(1) & 0.287(1) & 0.289(2) \\
634: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ $y$} & 0.243(1) & 0.233(1) & 0.233(1) \\
635: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ $z$} & 0.193(1) & 0.194(1) & 0.196(1) \\
636: {\small \ \ \ \ \ \ \ \ $U($\AA $^{2})$} & 0.010(1) & 0.011(1) & 0.009(1) \\
637: {\small O(3): \ \ \ \ \ \ $x$} & 0.801(1) & 0.787(1) & 0.789(2) \\
638: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ $y$} & 0.768(1) & 0.767(1) & 0.767(1) \\
639: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ $z$} & 0.193(1) & 0.194(1) & 0.196(1) \\
640: {\small \ \ \ \ \ \ \ \ $U($\AA $^{2})$} & 0.010(1) & 0.011(1) & 0.009(1) \\
641: & & & \\
642: \small $R_{wp}$ & 0.073 & 0.076 & 0.086 \\
643: \small $R_{B}$ & 0.041 & 0.047 & 0.040 \\
644: \small $\chi ^{2}$ & 1.52 & 1.62 & 1.21\\
645: \hline\hline
646: \end{tabular}
647: \end{table}
648:
649: \begin{table}[p]
650: \centering \caption{{\protect\small Refined structural parameters
651: for PbZr$_{0.52}$Ti$_{0.48}$O$_{3}$ at 20 K, two-phase model with
652: space groups $Cm$ and $R3c$. The Pb and Zr/Ti atoms
653: were fixed at the origin for the $Cm$ and $R3c$ refinements respectively,
654: and the temperature factors for the separate atoms were
655: constrained to be the same in both structures. Also listed are the
656: 10 K parameters recently reported by Frantti {\it et
657: al.}\protect\cite{Frantti3}}}
658: \begin{tabular}{lcccc}\\
659: \hline\hline
660: & \multicolumn{2}{c}{\small Present study} & \multicolumn{2}{c}{\small %
661: Frantii {\it et al.}} \\
662: & {\small $Cm$} & {\small $R3c$} & {\small $Cm$} & {\small $R3c$} \\
663: \hline \\
664: {\small $a($ \AA $)$} & 5.7120(1) & 5.7415(6) & 5.7097(7) & 5.744(2) \\
665: {\small $b($ \AA $)$} & 5.6988(1) & - & 5.6984(7) & - \\
666: {\small $c($ \AA $)$} & 4.1353(1) & 14.208(3) & 4.1367(3) & 14.212(8) \\
667: {\small $\beta (^{o})$} & 90.479(2) & - & 90.449(8) & - \\
668: {\small $V_{o}($ \AA$^{3})$} & 67.32 & 67.60 & 67.29 & 67.68 \\
669: {\small $c_{o}/a_{o}$} & 1.0257 & 1.0 & 1.0256 & 1.0 \\
670: & & & & \\
671: {\small Pb: \ \ \ \ \ \ \ \ \ \ \ $z$} & - & 0.283(4) & - &
672: 0.282(5) \\
673: \ \ \ \ \ \ \ \ \ {{\small $U$(\AA }$^{2})$} & 0.012(1) & 0.012(1)
674: & 0.009(1)
675: & 0.004(6) \\
676: {\small Zr/Ti: \ \ \ \ \ \ \ $x$} & 0.531(2) & - &
677: 0.539(3) & - \\
678: {\small \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ $z$} & 0.441(2) & - & 0.441(3) & - \\
679: \ \ \ \ \ \ \ \ \ {{\small $U$(\AA }$^{2})$} & 0.004(1) & 0.004(1)
680: & 0.001(2)
681: & 0.001(2) \\
682: {\small O(1): \ \ \ \ \ \ \ \ $x$} & 0.543(1) & - &
683: 0.540(1) & - \\
684: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $ z$} & -0.090(1) & - & -0.092(2) & - \\
685: \ \ \ \ \ \ \ \ \ {{\small $U$(\AA }$^{2})$} & 0.008(1) & - & 0.011(2) & - \\
686: {\small O(2): \ \ \ \ \ \ \ \ $x$} & 0.288(1) & 0.137(3) & 0.283(1) & 0.148(3) \\
687: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $ y$} & 0.254(1) & 0.347(3) & 0.253(1) & 0.354(3) \\
688: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $ z$} & 0.389(10 & 0.081(4) & 0.388(1) & 0.081(6) \\
689: \ \ \ \ \ \ \ \ \ {{\small $U$(\AA }$^{2})$} & 0.015(1) & 0.015(1) & 0.013(1) & 0.007(5) \\
690: & & & & \\
691: {$f${\small (wt fraction)}} & 0.89(1) & 0.11(1) & 0.87 & 0.13 \\
692: & & & & \\
693: {\small $R_{wp}$} & \multicolumn{2}{c}{0.067} &
694: \multicolumn{2}{c}{0.058} \\
695: {\small $R_{B}$} & 0.039 & 0.067 & \multicolumn{2}{c}{0.042} \\
696: {\small $\chi ^{2}$} & \multicolumn{2}{c}{1.27} &
697: \multicolumn{2}{c}{2.28}\\
698: \hline\hline
699: \end{tabular}
700: \end{table}
701: \begin{table}[p]
702: \centering \caption{{\protect\small Refined structural parameters
703: for monoclinic PbZr$ _{0.52}$Ti$_{0.48}$O$_{3}$ at 20 K, two-phase
704: model with space groups $Cm$ and $Ic$. The atomic positions for
705: $Ic$ symmetry were based upon the [001]-tilt model described in
706: text and constrained to be equivalent to those for $Cm$ except for
707: one additional parameter $\protect\delta$ corresponding to O(2)
708: and O(3) displacements in the $x$ and $y$ directions due to
709: tilting. For comparison with the $Ic$ structure, the O(2) and O(3)
710: positions for the $Cm$ structure are shown separately, although in
711: fact they are symmetry-equivalent. The temperature factors for the
712: separate atoms were constrained to be the same in both structures.
713: }}
714: \begin{tabular}{lcc}\\
715: \hline\hline
716: & {\small $Cm$} &{\small $Ic$} \\
717: $a(${\small \AA}$)$ & 5.7097(1) & 5.7401(7) \\
718: $b(${\small \AA}$)$ & 5.6988(1) & 5.7188(8) \\
719: $c(${\small \AA}$)$ & 4.1373(1) & 8.2098(11) \\
720: $\beta (^{o})$ & 90.473(2) & 90.550(10) \\
721: $V_{o}(${\small \AA }$^{3})$ & 67.31 & 67.37 \\
722: $c_{o}/a_{o}$ & 1.0257 & 1.0127 \\
723: & & \\
724: {\small Pb: \ \ \ \ $U$(\AA$^{2})$} & 0.012(1)
725: & 0.012(1) \\
726: {\small Zr/Ti: \ \ \ \ \ \ \ $x$} & 0.530(2) & 0.530(2)-
727: \\
728: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & 0.437(1) & 0.218(1) \\
729: \ \ \ \ \ \ \ \ \ {\small $U$(\AA$^{2})$} & 0.003(1) & 0.003(1) \\
730: {\small O(1): \ \ \ \ \ \ \ \ $x$} & 0.541(1) & 0.541(1)
731: \\
732: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & -0.089(1) & -0.045(1) \\
733: \ \ \ \ \ \ \ \ \ {\small $U$(\AA$^{2})$} & 0.011(1) & 0.011(1) \\
734: {\small O(2): \ \ \ \ \ \ \ \ $x$} & 0.286(1) & 0.257(1)
735: \\
736: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $y$} & 0.254(1) & 0.225(1) \\
737: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & 0.390(1) & 0.195(1) \\
738: \ \ \ \ \ \ \ \ \ {\small $U$(\AA$^{2})$} & 0.015(1) & 0.015(1) \\
739: {\small \ O(3): \ \ \ \ \ \ \ $x$} & 0.786(1) & 0.814(1)
740: \\
741: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $y$} & 0.754(1) & 0.783(1) \\
742: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & 0.390(1) & 0.195(1) \\
743: \ \ \ \ \ \ \ \ \ {\small $U$(\AA$^{2})$} & 0.015(1) & 0.015(1) \\
744: & & \\
745: $f${\small (wt fraction)} & 0.78(2) & 0.22(2) \\
746: & & \\
747: {\small $R_{wp}$} & \multicolumn{2}{c}{0.064} \\
748: {\small $R_{B}$} & 0.039 & 0.053 \\
749: $\chi ^{2}$ & \multicolumn{2}{c}{1.16}\\
750: \hline\hline
751: \end{tabular}
752: \end{table}
753:
754: \begin{table}[p]
755: \begin{centering}
756: \caption{{\protect\small Refined structural parameters for the
757: $Cm$
758: and $P4mm$ phases in PbZr$_{0.52}$Ti$_{0.48}$O$_{3}$ at 325
759: K with the three-phase model described in text. The temperature
760: factors for the separate atoms were constrained to be the same in
761: both structures. Also listed are the room-temperature parameters
762: recently reported in an x-ray study by Ragini
763: $et~al.$\protect\cite{Ragini2} and in a neutron study by Frantti
764: $et~al.$\protect\cite{Frantti4} The weight fraction of the cubic
765: $Pm3m$ phase was determined as 0.06. }}
766: \begin{tabular}{lcccccc}\\
767: \hline\hline
768: & \multicolumn{2}{c}{\small Present study} & \multicolumn{2}{c}{\small %
769: Ragini $et~al.$} & \multicolumn{2}{c}{\small Frantti $et~al.$} \\
770: & {\small $Cm$} & {\small $P4mm$} & {\small $Cm$} & {\small $P4mm$} & {\small $Cm$} &
771: {\small $P4mm$} \\
772: \hline {\small $a($\AA$)$} & 5.7268(3) & 4.0393(1) & 5.7520(1) &
773: 4.0429(2) &
774: 5.7129(3) & 4.0550(4) \\
775: {\small $b($\AA$)$} & 5.7187(3) & - & 5.7431(2) & - & 5.7073(3) & - \\
776: {\small $c($\AA$)$} & 4.1230(2) & 4.1388(1) & 4.0912(4) &
777: 4.1318(3) &
778: 4.1436(1) & 4.1097(6) \\
779: {\small $\beta (^{o})$} & 90.393(5) & - & 90.48(1) & - & 90.199(3) & - \\
780: {\small $V_{o}($\AA$^{3})$} & 67.51 & 67.53 & 67.57 & 67.53 &
781: 67.55 & 67.58
782: \\
783: {\small $c_{o}/a_{o}$} & 1.0189 & 1.0246 & 1.0067 & 1.0219 & 1.0262 & 1.0135 \\
784: & & & & & & \\
785: {\small Pb: \ \ \ \ \ \ \ \ \ $z$} & - & 0.035(2)$^{a}$ & - & - & - & - \\
786: \ \ \ \ \ \ \ \ {\small $U($\AA$^{2})$} & 0.017(1) & 0.017(1) & 0.107$%
787: ^{b} $ & 0.030$^{c}$ & 0.021$^{d}$ & 0.019(1) \\
788: {\small Zr/Ti: \ \ \ \ \ $x$} & 0.530(4) & - & 0.578(3) & - & 0.507(2)/0.494(4)$^{e}$ & - \\
789: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & 0.432(5) & 0.442(2) &
790: 0.473(3) & 0.447(2) &
791: 0.426(1)/0.404(4)$^{e}$ & 0.431(4) \\
792: \ \ \ \ \ \ \ \ {\small $U($\AA$^{2})$} & 0.003(1) & 0.0043(1) & 0.015(1)
793: & 0.005(2) & 0.004(1) & 0.019(1) \\
794: {\small O(1): \ \ \ \ \ \ \ $x$} & 0.540(2) & - & 0.50(1)
795: & - & 0.522(1) & - \\
796: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & -0.080(3) & -0.085(2)
797: & -0.10(1) & -0.109(6)
798: & -0.090(1) & -0.080(2) \\
799: \ \ \ \ \ \ \ \ {\small $U($\AA$^{2})$} & 0.016(1) & 0.016(1) & 0.00(1) &
800: 0.029(1) & 0.013(1) & 0.019(1) \\
801: {\small $O(2):$ \ \ \ \ \ \ $x$}& 0.287(2) & - & 0.36(1)
802: & -- & 0.270(1) & - \\
803: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $y$} & 0.255(1) & - & 0.219(8) & - & 0.252(1) & - \\
804: \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ {\small $z$} & 0.400(2) & 0.0395(1)
805: & 0.404(8) & 0.389(3) &
806: 0.391(1) & 0.400(1) \\
807: \ \ \ \ \ \ \ \ {\small $U($\AA$^{2})$} & 0.021(2) & 0.021(2) & 0.04(1) &
808: 0.029(1) & 0.013(1) & 0.019(1) \\
809: & & & & & & \\
810: {\small $f$ (wt fraction)}& 0.33(1) & 0.61(2) & 0.58 & 0.42 & 0.69 & 0.31 \\
811: & & & & & & \\
812: {\small $R_{wp}$} & \multicolumn{2}{c}{0.051} &
813: \multicolumn{2}{c}{0.128} & \multicolumn{2}{c}{0.021} \\
814: {\small $R_{B}$} & 0.043 & 0.030 & 0.041 & 0.062 & \multicolumn{2}{c}{-} \\
815: {\small $\chi ^{2}$} & \multicolumn{2}{c}{1.47} &
816: \multicolumn{2}{c}{3.39} & \multicolumn{2}{c}{2.69}\\
817: \hline\hline
818: \end{tabular}
819: \end{centering}
820: \begin{flushleft}
821:
822: {\footnotesize{\protect
823:
824: (a) Pb atoms statistically distributed among 4(d) sites at $x, x,
825: 0$.
826:
827: (b) Equivalent isotropic $U$ $(U_{11}= 0.221, U_{22}= 0.027,
828: U_{33}= 0.074, U_{13}= 0.030$ \AA$^{2}).$
829:
830: (c) Equivalent isotropic $U$ $(U_{11}= U_{22}= 0.031, U_{33}=
831: 0.027$ \AA$^{2}$).
832:
833: (d) Equivalent isotropic $U$ $(U_{11}= 0.027, U_{22}= 0.026,
834: U_{33}= 0.011, U_{13}= 0.013$ \AA$^{2}$).
835:
836: (e) Zr and Ti parameters refined independently.}}
837: \end{flushleft}
838:
839: \end{table}
840:
841:
842: \newpage
843:
844: \begin{figure}[tbp]
845: \includegraphics {Figure1.eps}
846: \caption{{\protect\small PZT phase diagram as originally proposed
847: by Jaffe et al. in Ref. \onlinecite{Jaffe} (open circles) with the
848: modifications reported by Noheda et al. in Ref.
849: \onlinecite{Noheda3} (full circles). The
850: various phases described in the text are denoted by C (cubic }$Pm3m$%
851: {\protect\small ), R}$_{\rm HT}${\protect\small \ (rhombohedral }$R3m$%
852: {\protect\small), R}$_{\rm LT}${\protect\small \ (rhombohedral }$R3c$%
853: {\protect\small ), T (tetragonal }$P4mm${\protect\small ), and M}$_{\rm A}$%
854: {\protect\small \ (monoclinic} $C{m}${\protect\small ). }}
855: \end{figure}
856:
857: \begin{figure}[tbp]
858: \includegraphics{Figure2.eps}
859: \caption{{\protect\small (Color) Observed and calculated
860: diffraction profiles from
861: the Rietveld refinement of PbZr}$_{0.52}${\protect\small Ti}$_{0.48}$%
862: {\protect\small O}$_{3}${\protect\small \ at 550 K, with space group }$P4mm$%
863: {\protect\small . The difference plot is shown below, with short
864: vertical markers denoting the calculated peak positions.}}
865: \end{figure}
866:
867: \begin{figure}[tbp]
868: \includegraphics{Figure3.eps}
869: \caption{{\protect\small (Color) Observed and calculated
870: diffraction profiles from
871: the two-phase Rietveld refinement of PbZr}$_{0.52}${\protect\small Ti}$%
872: _{0.48}${\protect\small O}$_{3}${\protect\small \ at 20 K, with space groups
873: }$Cm${\protect\small \ and }$Ic${\protect\small . The difference plot is
874: shown below, with upper and lower sets of vertical markers denoting
875: the calculated peak positions for }$Cm${\protect\small \ and }$Ic$%
876: {\protect\small \ respectively. The position of the weak superlattice peak at
877: 2$\theta \approx 36.8^{\circ}$ (pseudocubic 3/2 1/2 1/2) is indicated with an asterisk.}}
878: \end{figure}
879:
880: \begin{figure}[tbp]
881: \includegraphics{Figure4.eps}
882: \caption{{\protect\small (Color) Observed and calculated
883: diffraction profiles and difference plots in the region around the
884: strongest superlattice peak from
885: PbZr}$_{0.52}${\protect\small Ti}$_{0.48}${\protect\small O}$_{3}$%
886: {\protect\small \ at 20 K for single-phase }$Ic${\protect\small \ (left
887: panel), two-phase }$Cm+R3c${\protect\small \ (center panel), and two-phase }$%
888: Cm+Ic${\protect\small \ (right panel).}}
889: \end{figure}
890:
891: \begin{figure}[tbp]
892: \includegraphics{Figure5.eps}
893: \caption{{\protect\small (Color) Observed and calculated
894: diffraction profiles and difference plots in the region around the
895: pseudocubic (200) reflection from
896: PbZr}$_{0.52}${\protect\small Ti}$_{0.48}${\protect\small O}$_{3}$%
897: {\protect\small \ at 20 K for single-phase }$Ic${\protect\small \ (left
898: panel), two-phase }$Cm+R3c${\protect\small \ (center panel), and two-phase }$%
899: Cm+Ic${\protect\small \ (right panel).}}
900: \end{figure}
901:
902: \begin{figure}[tbp]
903: \includegraphics{Figure6.eps}
904: \caption{{\protect\small (Color) Observed and calculated
905: diffraction profiles from
906: the three-phase Rietveld refinement of PbZr}$_{0.52}${\protect\small Ti}$%
907: _{0.48}${\protect\small O}$_{3}${\protect\small \ at 325 K, with space
908: groups }$P4mm${\protect\small , }$Cm${\protect\small \ and }$Pm3m$%
909: {\protect\small . The difference plot is shown below, with upper, middle
910: and lower sets of vertical markers denoting the calculated peak
911: positions for }$P4mm${\protect\small , }$Cm${\protect\small \ and }$Pm3m$%
912: {\protect\small \ respectively.}}
913: \end{figure}
914:
915: \end{document}
916: