cond-mat0602138/cond.tex
1: \documentclass{iopart}
2: %\documentclass[12pt]{iopart}
3: \usepackage{epsfig}
4: 
5: \begin{document}
6: 
7: \letter%
8: {Magnetic ordering in Gd$_2$Sn$_2$O$_7$:  the archetypal 
9: Heisenberg pyrochlore antiferromagnet}
10: 
11: 
12: \author{A~S Wills,$^{1,2}$ M E Zhitomirsky,$^3$ B Canals,$^4$ 
13: J~P~Sanchez,$^3$ P~Bonville,$^5$ P Dalmas de R\'eotier$^3$ 
14: and A Yaouanc$^3$}
15: 
16: \address{$^1$Department of Chemistry, University College London,
17: 20 Gordon Street, London, WC1H 0AJ, UK}
18: \address{$^2$Davy-Faraday Research Laboratory, The Royal Institution of 
19: Great Britain, London W1S 4BS, UK}
20: \address{$^3$% 
21: Commissariat \`a l'Energie Atomique, DSM/DRFMC/SPSMS, 38054 Grenoble, France}
22: \address{$^4$% 
23: Laboratoire Louis N\'eel, CNRS, BP-166, 38042 Grenoble,  France}
24: \address{$^5$% 
25: Commissariat \`a l'Energie Atomique, DSM/SPEC, 
26: 91191 Gif-sur-Yvette, France}
27: 
28:  
29: 
30: \begin{abstract}
31: Low-temperature powder neutron diffraction measurements are 
32: performed in the ordered magnetic state of the
33: pyrochlore antiferromagnet Gd$_2$Sn$_2$O$_7$.
34: Symmetry analysis of the diffraction data
35: indicates that this compound has
36: the ground state predicted theoretically for a Heisenberg
37: pyrochlore antiferromagnet with dipolar interactions. 
38: The difference in magnetic structures of Gd$_2$Sn$_2$O$_7$ 
39: and of nominally analogous Gd$_2$Ti$_2$O$_7$ is found to be 
40: determined by a specific type of third-neighbor  
41: superexchange interaction on the pyrochlore lattice between spins across 
42: empty hexagons.
43: \end{abstract}
44: 
45: 
46: Frustration or inability to simultaneously satisfy all 
47: independent interactions \cite{Anderson} has become an important
48: theme in condensed matter research, coupling at the fundamental
49: level a wide range of phenomena, such as high-$T_{\rm c}$
50: superconductivity, the folding of proteins and neural networks.
51: Magnetic crystals provide one of the simplest stages
52: within which to explore the influence of frustration,
53: particularly when it arises as a consequence of lattice
54: geometry, rather than due to disorder. For this reason,
55: geometrically frustrated magnetic materials have been the objects of
56: intense scrutiny for  over 20
57: years \cite{Villain}. Particular interest has been focussed on
58: {\it kagom\'e} and \textit{pyrochlore} 
59: (see Fig.~\ref{pyrolattice})
60: geometries of vertex-sharing triangles and tetrahedra
61: respectively. Model materials with their structures 
62: display a wide range of exotic low-temperature
63: physics, such as spin ice \cite{Bramwell_spin_ice}, spin
64: liquids \cite{YMn2}, topological spin glasses \cite{H3O_Fe}, heavy
65: fermion \cite{LiV2O4}, and co-operative paramagnetic ground 
66: states \cite{Tb2Ti2O7}. Research into these
67: systems was spawned from studies of the archetypal geometrically
68: frustrated system---the Heisenberg pyrochlore antiferromagnet, which 
69: in the classical limit was shown theoretically to
70: possess a disordered ground state. Raju and co-workers found that 
71: the Heisenberg pyrochlore antiferromagnet with dipolar interactions 
72: has an infinite number of degenerate spin configurations near 
73: the mean-field transition temperature, which are described by 
74: propagation vectors $[hhh]$ \cite{Raju}. Later, Palmer and Chalker 
75: showed that quartic terms in the free energy lift this degeneracy 
76: and stabilize a four-sublattice state with the ordering vector 
77: ${\bf k}=(0\,0\,0)$ (the PC state) \cite{Palmer_Chalker}. 
78: 
79: \begin{figure}[t]
80: \centerline{\hbox{\epsfig{figure=pyro.eps,width=0.4\columnwidth}}}
81: \caption{Pyrochlore lattice of vertex sharing tetrahedra.
82: Next-neighbor exchanges are shown by long-dashed line.}
83: \label{pyrolattice}
84: \end{figure}
85: 
86: Among various pyrochlore materials 
87: Gd$_2$Ti$_2$O$_7$ and Gd$_2$Sn$_2$O$_7$ are believed to be 
88: good realizations of Heisenberg antiferromagnets.
89: Indeed, the Gd$^{3+}$ ion has a half-filled $4f$-shell with nominally 
90: no orbital moment.
91: A strong {\it intrashell} spin-orbit coupling
92: mixes, however,  $^8S_{7/2}$ and $^6P_{7/2}$ states leading to a 
93: sizable crystal-field splitting. Recent ESR measurements on dilute systems
94: gave comparable ratios of the single-ion anisotropy constant $D>0$
95: to the nearest-neighbor exchange $J$ for the two compounds:
96: $D/J\sim 0.7$ \cite{Glazkov}. 
97: This corresponds to a planar anisotropy for the ground state.
98: Magnetic properties of the stannate and the titanate 
99: would be expected, therefore, to be very similar.
100: In this light, the contrasts between the low-temperature behavior of
101: Gd$_2$Sn$_2$O$_7$ and that of the analogous titanate are 
102: remarkable. While the titanate displays two magnetic transitions, 
103: at $\sim 0.7$ and 1~K, to structures with the ordering vector 
104: ${\bf k}=\left(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2}\right)$  
105: \cite{Champion,Stewart},
106: the stannate undergoes a
107: first-order transition into an ordered state near 
108: 1~K \cite{Bonville}. Furthermore, 
109: M\"ossbauer measurements indicate
110: that in the latter the correlated 
111: Gd$^{3+}$ moments still 
112: fluctuate as $T\rightarrow 0$~K \cite{Bonville2}.
113: 
114: 
115: In this article we demonstrate that Gd$_2$Sn$_2$O$_7$ orders with
116: the PC ground state, evidencing it as an experimental realization
117: of a Heisenberg pyrochlore antiferromagnet with dipolar
118: interactions. We also note that the magnetic structure observed in
119: Gd$_2$Sn$_2$O$_7$ differs from those observed in the closely
120: related Gd$_2$Ti$_2$O$_7$, indicating that an extra interaction  is at play in the latter which leads to the observed differences.
121: 
122: 
123: \begin{figure}[t]
124: \centerline{\hbox{\epsfig{figure=Refinement.eps,width=0.6\columnwidth}}}
125: \caption{
126: Fit to the magnetic diffraction pattern of Gd$_2$Sn$_2$O$_7$ obtained 
127: from the $\psi_6$ basis state by Rietveld refinement. The dots correspond
128: to experimental data obtained by subtraction of that measured in
129: the paramagnetic phase (1.4~K) from that in the magnetically ordered
130: phase (0.1~K). The solid line corresponds to the theoretical
131: prediction and the line below to the difference. Positions for the
132: magnetic reflections are indicated by vertical markers.}
133:     \label{refinement fig}
134: \end{figure}
135: 
136: In order to reduce the absorption of neutrons, a 500~mg sample of
137: Gd$_2$Sn$_2$O$_7$ enriched with $^{160}$Gd was prepared following
138: conditions given in Ref.~\cite{Bonville}. 
139: Powder neutron diffraction spectra
140: were collected with neutrons of wavelength 2.4 \AA\ using the D20
141: diffractometer of the ILL at two temperatures below
142: $T_{\rm N}$=1~K (0.1~K and 0.7~K), as well as one temperature in the
143: paramagnetic phase above $T_{\rm N}$ (1.4~K). Due to the high
144: residual absorption of the Gd sample, extended counting times of 5 hours per temperature were required. The magnetic contribution to diffraction at 0.1~K
145: could be well isolated from scattering by the cryostat walls and
146: dilution insert by subtraction of the spectrum at 1.4~K. Symmetry calculations were made using SARA\textit{h} \cite{Sarah} and Rietveld refinement
147: of the diffraction data  using
148: Fullprof \cite{Fullprof}  together with
149: SARA\textit{h}.
150: 
151: The magnetic diffraction peaks, Fig.~\ref{refinement fig}, can be 
152: indexed with the propagation vector ${\bf k}=(0\,0\,0)$.  The 
153: different types of magnetic structure can be classified in terms of  
154: the irreducible corepresentations of the reducible magnetic 
155: corepresentation, $c\Gamma_{mag}$. In the case of Gd$_2$Sn$_2$O$_7$ (space
156: group $Fd\bar{3}m$ with ${\bf k}=(0\,0\,0)$ and Gd$^{3+}$ ion at the $16d$
157: crystallographic position) this can be written: 
158: $c\Gamma_{mag}= 1c\Gamma_{3+} + 1c\Gamma_{5+}
159: + 1c\Gamma_{7+} + 2c\Gamma_{9+}$
160: \cite{Kovalev,corep numbering note}.
161: Their associated basis vectors are represented in Fig.~\ref{corep fig}. 
162: Inspection reveals 
163: that $c\Gamma_{3+}$ 
164: corresponds to the antiferromagnetic structure observed in
165: FeF$_3$ \cite{Greedan_FeF3}, 
166: $c\Gamma_{5+}$ to the linear combination observed in the
167: model $XY$ pyrochlore antiferromagnet
168: Er$_2$Ti$_2$O$_7$ \cite{Champion}, $c\Gamma_{7+}$ to the manifold of
169: states proposed as the ground states for the Heisenberg pyrochlore
170: antiferromagnet with dipolar terms (the PC ground state), and 
171: $c\Gamma_{9+}$ to a spin-ice like manifold observed in 
172: the non-collinear ferromagnetic pyrochlores such as Dy$_2$Ti$_2$O$_7$
173: \cite{Bramwell_spin_ice}. 
174: While the phase transition in Gd$_2$Sn$_2$O$_7$ has been shown to be 
175: first order which allows ordering according to several 
176: irreducible corepresentations, it is commonly found that 
177: the terms which drive the transition to being first order are relatively 
178: weak and cause only minor perturbation to the resultant 
179: magnetic structure. Following this, we examined whether 
180: the models detailed above could fit the observed magnetic neutron 
181: diffraction spectrum. The goodness of fit parameter, $\chi^2$,  for  the fit to models characterised by each irreducible corepresentation are:  $c\Gamma_{3+}$ (69.0), $c\Gamma_{5+}$ (35.6),  $c\Gamma_{7+}$ (5.18),  $c\Gamma_{9+}$ (13.6). We find that the magnetic scattering can only be well 
182: modeled by $c\Gamma_{7+}$, the PC state in which the  moments of 
183: a given Gd tetrahedron are parallel to the tetrahedron's edges. 
184: In this state each moment is fixed to be perpendicular to the 
185: local 3-fold  axis of each tetrahedron,  consistent with
186: M\"ossbauer data \cite{Bonville,Bonville2}. 
187: Powder averaging leads to the structures ascribed to $\psi_4$,
188: $\psi_5$ and $\psi_6$ being indistinguishable
189: by neutron diffraction and prevents contributions of the individual basis vectors from
190: being refined. For this reason only $\psi_6$ was used
191: in the refinement and the final fit is presented in Figure \ref{refinement fig}. While the  value of the ordered moment,  6(1) $\mu_B$/Gd$^{3+}$, obtained by scaling the
192: magnetic and nuclear peaks is
193: imprecise due to the uncertainty over the isotopic
194: composition of the Gd and the concomitant neutron absorption, it is 
195: consistent with the free-ion value (7~$\mu_B$) and that measured by M\"ossbauer spectroscopy \cite{Bonville}.
196: 
197: 
198: \begin{figure}[t]
199: \begin{center}
200: \includegraphics[width=0.8\columnwidth]{basis.eps}
201: \end{center}
202: %\centerline{\hbox{\epsfig{figure=basis_vector_pictures.eps,width=10cm}}}
203:  \caption{The magnetic structure basis vectors labelled according to the 
204: different irreducible corepresentations for Gd$_2$Sn$_2$O$_7$.}
205:     \label{corep fig}
206: \end{figure}
207: 
208: 
209: Realization of the PC state in Gd$_2$Sn$_2$O$_7$, 
210: but not in Gd$_2$Ti$_2$O$_7$, indicates that the magnetic 
211: Hamiltonian of the titanate contains additional terms.
212: C\'epas and Shastry \cite{Cepas_Shastry} 
213: have suggested that next-neighbor exchange 
214: may stabilize magnetic ordering at
215: ${\bf k}=(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$, though
216: the corresponding region in the parameter space
217: was found to be tiny. 
218: Also, possible exchange paths were not investigated 
219: in their work  as
220: both types of third-neighbor exchange (Fig.~1) were
221: assumed to be equal.
222: 
223: The pyrochlore A$_2$B$_2$O$_7$ structure
224: has two inequivalent oxygen sites: 
225: O1 at $(x,\frac{1}{8},\frac{1}{8})$ and 
226: O2  at $(\frac{3}{8},\frac{3}{8},\frac{3}{8})$.
227: The oxygen parameter 
228: is  $x=0.335$ and $0.326$  for
229: the stannate and the titanate, respectively \cite{structure}.
230: Using this information we have determined
231: that the nearest-neighbor gadolinium ions
232: are connected with short Gd--O1(2)--Gd bonds.
233: The second-neighbor exchange $J_2$ is produced by two 
234: distinct Gd--O1--O1--Gd bridges. 
235: The O1--O1 distance in the first path is $2.63$~\AA\
236: with two equal bond angles of  $118^\circ$. 
237: In the second path, $|$O1--O1$|=3.04$~\AA, while the 
238: angles are $148^\circ$ and $98^\circ$.
239: (Distances and angles are given 
240: for Gd$_2$Ti$_2$O$_7$.) 
241: 
242: There are two types of 
243: third-neighbour pairs in the pyrochlore lattice
244: indicated in Fig.~1 as $J_{31}$ and $J_{32}$, which  
245: correspond to three- and two-step Manhattan (city-block) distances, 
246: respectively. 
247: The superexchange $J_{31}$ is determined by two equivalent Gd--O1--O1--Gd
248: paths with $|$O1--O1$|=3.04$~\AA\ and two equal angles  $148^\circ$.
249: The superexchange $J_{32}$ is produced by
250: two Gd--O1--O2--Gd bridges with
251: a significantly larger interoxygen  distance $|$O1--O2$|=3.62$~\AA\ and 
252: bond angles of $152^\circ$ and $143^\circ$. 
253: As a result, the two exchange constants for third-neighbor pairs
254: have to be different with $J_{32}\ll J_{31}$.
255: The Goodenough-Kanamori-Anderson rules also suggest that 
256: the second-neighbor constant
257: $J_2$ must be smaller than $J_{31}$ since 
258: bond angles in the corresponding superexchange paths 
259: are significantly closer to $90^\circ$.
260: Similar relations should hold for next-neighbor
261: exchange constants in Gd$_2$Sn$_2$O$_7$ with, perhaps,
262: a somewhat larger ratio $J_2/J_{31}$ due to a larger angle $126^\circ$
263: in the second-neighbor superexchange path.
264: The overall effect of further neighbor exchange is, however,
265: reduced in the stannate because of a larger lattice constant
266: $a=10.45$~\AA\ compared to $a=10.17$~\AA\ in the titanate \cite{structure}.
267: 
268: \begin{figure}[t]
269: \begin{center}
270: \includegraphics[width=0.6\columnwidth]{diagram.eps}
271: \end{center}
272: \caption{Instability wave-vectors for different values of second-
273: and third-neighbor exchange constants for
274: a Heisenberg pyrochlore antiferromagnet with dipolar interactions.
275: Incommensurate states
276: are indicated by nonzero components of the wave-vectors.
277: All transition lines are of the first-order.
278: \label{diagram}}
279: \end{figure}
280: 
281: Next, we consider the following Hamiltonian
282: 
283: 
284: 
285: \begin{eqnarray}
286: \fl \hat{\mathcal H}  =  \sum_{\langle i,j\rangle}
287: J_{ij}{\bf S}_i\cdot{\bf S}_j + D \sum_i ({\bf n}_i\cdot {\bf S}_i)^2 \label{Hamiltonian} 
288: + (g\mu_B)^2 \sum_{\langle i,j\rangle}
289: \Bigl[ \frac{{\bf S}_i\cdot{\bf S}_j}{r_{ij}^3}
290: -\frac{3({\bf S}_i\cdot{\bf r}_{ij})({\bf S}_j\cdot{\bf r}_{ij})}
291: {r_{ij}^5} \Bigr] \ ,
292: \nonumber
293: \end{eqnarray}
294: 
295: 
296: \noindent
297: where the superexchange $J_{ij}$ extends up to the third-neighbor pairs of
298: spins and $D>0$ is a single-ion anisotropy.
299: The strength of the dipolar interaction between nearest-neighbor
300: spins  $E_{dd} = (g\mu_B)^2/(a\sqrt{2}/4)^3$ is estimated as
301: $E_{dd}/J\approx 0.2$ for the titanate \cite{Raju}, where $J$ is
302: the nearest-neighbor exchange constant (in the following $J\equiv 1$). 
303: Applying mean-field theory \cite{Raju,Cepas_Shastry} 
304: and evaluating dipolar sums via the Ewald's summation
305: we have determined the instability wave-vector for different
306: values of second- and third-neighbor exchange constants.
307: Results are essentially independent of the anisotropy constant $D>0$ since the dipolar interaction already selects spins to be orthogonal
308: to the local trigonal axes ${\bf n}_i$
309: for the eigenstates with the highest transition temperature.
310: 
311: If only the nearest-neighbor exchange is present, in agreement with 
312: previous works \cite{Raju,Cepas_Shastry} we find an approximate degeneracy 
313: of modes along the cube
314: diagonal with a very shallow minimum $\sim 0.5$\% at
315: ${\bf k}=(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$. 
316: In such a case, a fluctuation driven
317: first-order transition is expected to the PC state 
318: \cite{Palmer_Chalker,first_order}.
319: The diagram of possible ordering wave-vectors for a restricted
320: range of $J_{31}$  and $J_2$ are presented in Fig.~4. 
321: It contains two commensurate states with ${\bf k}=(0\,0\,0)$ and 
322: ${\bf k}=(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$ and three 
323: incommensurate phases.
324: Remarkably, already weak {\it antiferromagnetic} $J_{31}$
325: robustly stabilizes the ${\bf k}=(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$ magnetic
326: structure, which exists in a wide range
327: $0<J_{31}<0.335J$. In contrast, a small {\it ferromagnetic}
328: $J_2$ within a narrow window $-0.04J<J_2<0$is needed to obtain the same ordering without $J_{31}$.
329: In the whole range of parameters, the eigenstate with
330: ${\bf k}=(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$ 
331: corresponds to 120$^\circ$ spin structure with $q=0$ in transverse kagom\'e plane 
332: with zero ordered moment on interstitial sites.
333: The only remaining degeneracy corresponds to a four-fold orbit of
334: ${\bf k}=(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$ vector.
335: 
336: We have also verified that the second type of third-neighbor exchange
337: $J_{32}$ does not lead to further stabilization of 
338: the  $(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$ spin structure. 
339: Based on these results and the above analysis of the exchange paths
340: we conclude that it is the third-neighbor exchange across
341: empty hexagons $J_{31}$ (Fig.~1), which is responsible for
342: the magnetic structure observed in 
343: Gd$_2$Ti$_2$O$_7$ \cite{Champion,Stewart}.
344: In Gd$_2$Sn$_2$O$_7$ further neighbor exchanges play a less
345: significant role due to a larger lattice constant
346: and, in addition, the ratio $J_2/J_{31}$ might be enhanced
347: due to somewhat different bond angles such that
348: it lies closer to the transition boundary between
349: ${\bf k}=(0\,0\,0)$ and $(\frac{1}{2}\,\frac{1}{2}\,\frac{1}{2})$ states.
350: 
351: In conclusion, the occurrence of the PC state in
352: Gd$_2$Sn$_2$O$_7$ but not in
353: Gd$_2$Ti$_2$O$_7$ indicates that the latter possesses additional
354: contributions, which we identify as a type
355: of third--neighbor exchange.
356: Gd$_2$Sn$_2$O$_7$ presents, therefore, the only 
357: accurate realization of the Heisenberg pyrochlore antiferromagnet
358: with dipolar interactions.
359: 
360: We are grateful to A. Forget for preparing the $^{160}$Gd enriched sample and to the ILL for provision of neutron time. ASW would like to thank the Royal Society and EPSRC (grant number EP/C534654) for financial support.
361: 
362: 
363: \section*{References}
364: 
365: \begin{thebibliography}{10}
366: 
367: \bibitem{Anderson}
368: Anderson P~W 1973 {\it Mat. Res. Bull.} {\bf 8} 153
369: 
370: \bibitem{Villain}
371: Villain J 1979 {\it Z. Phys.} B \textbf{33}  31 
372: 
373: \bibitem{Bramwell_spin_ice}
374: Harris M~J, Bramwell S~T, McMorrow D~F, Zeiske T and Godfrey K~W 
375: 1997 {\it Phys. Rev. Lett.} {\bf 79}  2554 
376: 
377: \bibitem{YMn2}
378: Ballou R, Leli\`evre-Berna E and F{\aa}k B 1996 {\it Phys. Rev. Lett.}
379: {\bf 77}  790
380: 
381: \bibitem{H3O_Fe}
382: %A.~S. Wills {\it et al.\/}, 
383: Wills A~S, Depuis V, Vincent E and Calemczuk R 2000 
384: \textit{Phys. Rev.} B {\bf 62}  9264(R) 
385: 
386: 
387: \bibitem{LiV2O4}
388: Urano C, Nohara M, Kondo S, Sakai F, Takagi H, Shiraki T and Okubo T 
389: 2000 \textit{Phys. Rev. Lett.} {\bf 85} 1052
390: 
391: \bibitem{Tb2Ti2O7}
392: %M. J. P. Gingras {\it et al.\/},  
393: Gingras M~J~P, den Hertog B~C, Faucher M, Gardner J~S, Dunsiger S~R, 
394: Chang L~J, Gaulin B~D, Raju N P and Greedan J E 2000
395: {\it Phys. Rev.} B {\bf 62}  6496
396: 
397: 
398: \bibitem{Raju}
399: %Gd2Ti2O7 paper
400: %N.~P. Raju {\it et al.\/}, 
401: Raju N~P, Dion M, Gingras M~J~P, Mason T~E and Greedan J~E
402: 1999 {\it Phys. Rev.} B \textbf{59} 14489
403: 
404: \bibitem{Palmer_Chalker}
405: %pyrochlore AFM with dipolar interactions to quartic level
406: Palmer S~E and Chalker J~T 2000 {\it Phys. Rev.} B {\bf 62}  488
407: 
408: \bibitem{Glazkov}
409: %Gd2Ti2O7 ESR
410: %V. N. Glazkov  {\it et al.\/}, 
411: Glazkov V~N, Zhitomirsky M~E, Smirnov A I, Krug von Nidda H-A, Loidl A,  
412: Marin C and Sanchez J P 
413: 2005 {\it Phys. Rev.} B {\bf 72} 020409(R);
414: Glazkov V~N, Smirnov A I, Sanchez J P, 
415: Forget A, Colson D and  Bonville P 2005 {\it Preprint}
416: cond-mat/0510575 
417:  
418:  
419: \bibitem{Champion}
420: %original Gd2Ti2O7 magnetic structure paper
421: %J.~D.~M. Champion {\it et al.\/}, 
422: Champion J~D~M, Wills A~S, Fennell T, Bramwell S~T, Gardner J~S and 
423: Green M~A 
424: 2001 \textit{Phys. Rev.} B \textbf{64} 140407
425: 
426: \bibitem{Stewart}
427: %4k structure of Gd2Ti2O7
428: %J.~R. Stewart {\it et al.\/}, 
429: Stewart J~R, Ehlers G, Wills A~S, Bramwell S~T and Gardner J~S 
430: 2004 {\it J. Phys.: Condens. Matter} \textbf{16}  L321
431: 
432: \bibitem{Bonville}
433: %P. Bonville  {\it et al.\/}, 
434: Bonville P, Hodges J~A, Ocio M, Sanchez J~P, Vulliet P., Sosin S 
435: and Braithwaite D
436: 2003 \textit{J. Phys. Condens. Matter.} \textbf{15}  7777
437: 
438: \bibitem{Bonville2}
439: %Gd2Sn2O7 mossbauer
440: %E. Bertin  {\it et al.\/}, 
441: Bertin E, Bonville P, Bouchaud J-P, Hodges J~A, Sanchez J~P and Vulliet P 
442: 2002 \textit{Eur. Phys. J.} B \textbf{27}  347
443: 
444: \bibitem{Sarah}
445: Wills A~S 2000 \textit{Physica} B {\bf 276} 680; progam available from ftp.ill.fr/pub/dif/sarah/
446: 
447: \bibitem{Fullprof}
448: Rodriguez-Carvajal J  1993 \textit{Physica B} \textbf{192} 55
449: 
450: \bibitem{Kovalev}
451: Kovalev O~V  1993 {\it Representations of the Crystallographic Space
452: Groups} Edition 2 (Gordon and Breach Science Publishers, Switzerland)
453: 
454: \bibitem{corep numbering note}
455: The corepresentations are real and are labelled according to the notation of 
456: Kovalev \cite{Kovalev} for the parent representation 
457: and whether the antiunitary 
458: halfing group was created according to the 
459: $d(a)=\pm \delta(aa_0^{-1})\beta$, 
460: where $d(a)$ is the matrix representitive of antiunitary 
461: symmetry element $a$, 
462: $\delta(aa_0^{-1})$ is the matrix representitive of unitary symmetry element 
463: $aa_0^{-1}$, $a_0$ is an antiunitatry generating element and $\beta$ is 
464: an unitary matrix. 
465: 
466: \bibitem{Greedan_FeF3}
467: Greedan J~E, O'Reilly A~H and Stager C~V 1987 \textit{Phys. Rev.} B
468: \textbf{35}  8770
469: 
470: \bibitem{Cepas_Shastry}
471: %further-neighbour stabilised phases in pyrochlore AFM
472: C\'epas O. and Shastry B.~M. 2004 \textit{Phys. Rev.} B \textbf{69} 184402
473: 
474: 
475: \bibitem{structure}
476: Kennedy B J,  Hunter B~A and Howard C J
477: 1997 \textit{J. Solid State Chem.} {\bf 130} 58;
478: Helean K B, Ushakov S V, Brown C E, Navrotsky A, Lian J, Ewing R C,
479: Farmer J M and Boatner L A 2004  {\it ibid\/}.\ {\bf 177}  1858
480: 
481: \bibitem{first_order}
482: Brazovskii S A 1975 Zh. \'Eksp. Teor. Fiz. {\bf 68}, 175
483: [Sov. Phys. JETP {\bf 41} 85];
484: Cepas O,  Young A P and Shastry B S  2005
485: Phys. Rev. B {\bf 72} 184408 
486: 
487: \end{thebibliography}
488: \end{document}
489: 
490: 
491: %%%%%%%%%%%%%%%%%
492: