1: %\documentclass[aps,prb,twocolumn]{revtex4}
2: \documentclass[aps,prb,twocolumn]{revtex4}
3: %\documentclass[12pt,aps,prb,preprint]{revtex4}
4: %\documentstyle[12pt]{article}
5: \usepackage[dvips]{epsfig,color}
6: \usepackage{graphicx}
7: \begin{document}
8: \title{Anisotropic Hubbard model on a triangular lattice ---
9: spin dynamics in $\rm Ho Mn O_3$}
10: \author{Saptarshi Ghosh}
11: \email{gsap@iitk.ac.in}
12: \author{Avinash Singh}
13: \email{avinas@iitk.ac.in}
14: \affiliation{Department of Physics, Indian Institute of Technology Kanpur - 208016}
15: \begin{abstract}
16: The recent neutron-scattering data for spin-wave dispersion in
17: $\rm Ho Mn O_3$ are well described by an anisotropic Hubbard model on a
18: triangular lattice with a planar (XY) spin anisotropy.
19: Best fit indicates that magnetic excitations in $\rm Ho Mn O_3$
20: correspond to the strong-coupling limit $U/t > \sim 15$,
21: with planar exchange energy $J=4t^2/U \simeq 2.5$meV and
22: planar anisotropy $\Delta U \simeq 0.35$meV.
23: \end{abstract}
24: \pacs{75.50.Pp,75.30.Ds,75.30.Gw}
25: \maketitle
26: %\section{Introduction}
27: There has been renewed interest in correlated electron systems on triangular lattices,
28: as evidenced by recent studies of antiferromagnetism, superconductivity
29: and metal-insulator transition in the organic systems
30: $\rm \kappa -(BEDT-TTF)_2 X$,\cite{review1,review2}
31: the discovery of superconductivity in $\rm Na_x Co O_2 . y H_2 O$,\cite{watersup}
32: the observation of low-temperature insulating phases
33: in some $\sqrt{3}$-adlayer structures such as K on Si[111],\cite{weitering}
34: and quasi two-dimensional $120^0$ spin ordering and spin-wave excitations
35: in $\rm Rb Fe (MoO_4)_2$ (Refs. 5,6)
36: and the multiferroic materials $\rm Y Mn O_3$ and $\rm Ho Mn O_3$.\cite{sato,holmium}
37:
38: Recent neutron-scattering studies of the multiferroic material $\rm Ho Mn O_3$
39: have revealed a non-collinear $120^0$ antiferromagnetic (AF) ordering below
40: $T_{\rm N} \approx 72$ K of the $S=2$ Mn$^{3+}$ spins arranged in offset layers of
41: two-dimensional (2D) triangular lattice.\cite{holmium}
42: Measurements of the spin wave dispersion were found to be well described by
43: a nearest-neighbour Heisenberg AF with exchange energy $J = 2.44$ meV
44: and a planar anisotropy $D=0.38$ meV at 20 K.
45: No discernible dispersion was observed in the out-of-plane direction,
46: indicating primarily 2D spin dynamics.
47:
48: Recently spin-wave excitations in the $120^0$ AF state of the Hubbard model
49: on a triangular lattice were studied within the random phase approximation (RPA)
50: in the full $U$ range.\cite{tri}
51: The spin wave energy in the large $U$ limit was shown to asymptotically approach
52: the corresponding result for the Quantum Heisenberg antiferromagnet (QHAF),
53: thus providing a continuous interpolation between weak and strong
54: coupling limits. However, competing interactions and frustration
55: were found to significantly modify the dispersion at finite $U$,
56: resulting in vanishing spin stiffness at $U\approx 6$
57: and a magnetic instability at $U \approx 7$ corresponding to
58: vanishing spin-wave energy at wave vector ${\bf q}_M=(\pi/3,\pi/\sqrt{3})$.
59: The sharp fall-off of $\omega_M$ near $U \approx 7$ provides a sensitive indicator
60: of finite-$U$ effects in the AF state.
61: Indeed, recent high-resolution neutron-scattering studies of the spin-wave dispersion
62: in the square-lattice S=1/2 AF $\rm La_2 Cu O_4$
63: have revealed noticeable spin-wave dispersion along the MBZ edge,\cite{spinwave}
64: associated with finite-$U$ double-occupancy effects.\cite{spectrum}
65:
66: In this brief report we extend the spin-wave analysis to include planar spin anisotropy,
67: and show that the neutron-scattering data for $\rm Ho Mn O_3$
68: are also well described by a Hubbard model on a triangular lattice,
69: thus providing a microscopic description of the most essential features
70: of spin dynamics in the $120^0$ AF state of $\rm Ho Mn O_3$,
71: including the spin-wave dispersion and energy scale.
72: We examine the behaviour of spin-wave anisotropy gap with spin anisotropy
73: and also suggest a sensitive measure of finite-$U$, double occupancy effects.
74:
75: However, the Mn spin planar anisotropy is treated here only at a phenomenological level,
76: equivalent to the effective anisotropy $DS_{iz}^2$ included
77: in recent investigations using spin models.\cite{holmium,sato}
78: In a detailed study of single-ion anisotropy and crystal-field effects
79: in the layered rare-earth cuprates $\rm R_2 CuO_4$ (R=Nd,Pr,Sm),
80: the magnetic behaviour (including spin-reorientation transitions)
81: has been attributed to coupling of Cu with the rare-earth magnetic subsystem
82: which exhibits a large single-ion anisotropy resulting in preferential ordering
83: of rare-earth moments along specific lattice directions.\cite{sachi}
84: It has been suggested that the anisotropy of Mn spins,
85: its observed temperature dependence,
86: and the reorientation transitions in $\rm Ho Mn O_3$ also originate from a similar
87: anisotropic exchange coupling with the rare-earth Holmium,\cite{holmium}
88: resulting in magnetic behaviour as seen in layered rare-earth cuprates,
89: where frustrated interlayer coupling allows for weaker, higher-order interactions
90: to control the magnetic structure.
91: Especially relevant for non-collinear ordering,
92: the anisotropic exchange (Dzyaloshinski-Moriya) interaction
93: ${\bf D}.({\bf S}_i \times {\bf S}_j)$
94: originating from spin-orbit coupling has been suggested as responsible for the
95: clamping of ferroelectric and antiferromagnetic order parameters in
96: $\rm Y Mn O_3$.\cite{hanamura}
97:
98: Hund's rule coupling responsible for the $S=2$ spin state of Mn$^{+++}$ ions
99: and crystal-field splitting have also not been
100: realistically incorporated here.
101: However, these realistic details do not qualitatively affect
102: the spin-rotation symmetry and spin dynamics, as discussed below.
103: Hund's rule coupling in the generalized Hubbard model considered here is maximal as
104: inter-orbital Coulomb interaction for parallel spins is dropped completely.
105: Including an inter-orbital density-density interaction $V_0$ and
106: an intra-atomic exchange interaction $F_0$ favouring parallel-spin alignment
107: (Hund's rule coupling),
108: the more realistic orbital Hubbard model\cite{roth,held}
109: \begin{eqnarray}
110: H &=& -t \sum_{i,\delta,\gamma,\sigma} a_{i\gamma\sigma}^\dagger a_{i+\delta,\gamma\sigma}
111: + U\sum_{i\gamma} n_{i\gamma\uparrow} n_{i\gamma\downarrow} \nonumber \\
112: &+&
113: \sum_{i,\gamma<\gamma',\sigma,\sigma'}(V_0 - \delta_{\sigma\sigma'}F_0 )
114: n_{i\gamma\sigma} n_{i\gamma'\sigma'} \nonumber \\
115: &+&
116: F_0 \sum_{i,\gamma<\gamma',\sigma\ne\sigma'}
117: a_{i\gamma\sigma} ^\dagger
118: a_{i\gamma'\sigma}
119: a_{i\gamma'\sigma'} ^\dagger
120: a_{i\gamma\sigma'}
121: \end{eqnarray}
122: remains spin-rotationally invariant under a global rotation of the fermion spin
123: ${\bf S}_{i\gamma}=\Psi^\dagger _{i\gamma} \frac{\mbox{\boldmath $\sigma$}}{2} \Psi_{i\gamma}$
124: (where $\Psi^\dagger _{i\gamma} \equiv (a_{i\gamma\uparrow}^\dagger \; a_{i\gamma\downarrow}^\dagger)$ is the fermion field operator),
125: even if orbitals $\gamma$ are identified with the Mn orbitals $(t_{2g},e_g)$
126: resulting from crystal-field splitting of the atomic 3d orbitals.
127: As the intra-atomic exchange interaction $F_0$ is much larger than the
128: spin excitation energy scale ($\sim \frac{t^2}{U+F_0}$, within a strong-coupling expansion),
129: all fermion spins on a site are effectively coupled,
130: yielding a composite quantum spin $S$
131: and a corresponding multiplication by factor $2S$
132: to obtain the effective spin-wave energy scale.
133: Therefore, orbital multiplicity does not change the Goldstone-mode structure,
134: and the spin-dynamics energy scale in the orbital Hubbard model
135: is essentially determined by $t$ and $U_{\rm eff} = U+F_0$.
136:
137: %\section{Anisotropic Hubbard model}
138: As a simplest extension to phenomenologically include spin-space anisotropy,
139: while retaining only the relevant energy scales $t$ and $U_{\rm eff}$,
140: we consider the generalized $\cal N$-orbital Hubbard model\cite{quantum}
141: \begin{eqnarray}
142: H &=& -t \sum_{i,\delta,\gamma,\sigma}
143: a_{i\gamma\sigma} ^\dagger a_{i+\delta,\gamma,\sigma}
144: + \frac{U_1}{\cal N} \sum_{i,\gamma,\gamma'} a_{i\gamma\uparrow}^\dagger
145: a_{i\gamma\uparrow} a_{i\gamma'\downarrow}^\dagger a_{i\gamma'\downarrow} \nonumber \\
146: &+& \frac{U_2}{\cal N} \sum_{i,\gamma,\gamma'} a_{i\gamma\uparrow}^\dagger
147: a_{i\gamma'\uparrow} a_{i\gamma'\downarrow}^\dagger a_{i\gamma\downarrow}
148: \end{eqnarray}
149: on a triangular lattice with nearest-neighbour (NN) hopping between sites $i$
150: and $i+\delta$.
151: Here $\gamma,\gamma'$ refer to the (fictitious) degenerate $\cal N$ orbitals per site.
152: The factor $\frac{1}{\cal N}$ is included to render the energy density finite
153: in the ${\cal N} \rightarrow \infty$ limit.
154: The two correlation terms involve density-density and exchange interactions
155: with respect to the orbital indices.
156: The Hartree-Fock (HF) approximation and Random Phase Approximation (RPA) are of O(1)
157: whereas quantum fluctuation effects appear at higher order within the
158: inverse-degeneracy expansion and thus $1/\cal N$,
159: in analogy with $1/S$ for quantum spin systems, plays the role of $\hbar$.
160:
161: The key feature of spin-rotation symmetry of the generalized Hubbard model
162: is highlighted by writing the two interaction terms as
163: \begin{equation}
164: H_{\rm int} =
165: - \frac{U_2}{\cal N} \sum_{i} {\bf S}_{i} . {\bf S}_{i}
166: + \frac{U_2 - U_1}{\cal N} \sum_{i} S_{iz} ^2
167: \end{equation}
168: in terms of the total spin operator
169: \begin{equation}
170: {\bf S}_{i} = \sum_\gamma \psi_{i\gamma}^\dagger
171: \frac{\mbox{\boldmath $\sigma$}}{2} \psi_{i\gamma}
172: \equiv \sum_\gamma \frac{\mbox{\boldmath $\sigma$}_{i\gamma}}{2}
173: \end{equation}
174: where $\psi_{i\gamma}^\dagger \equiv (a_{i\gamma\uparrow}^\dagger \;
175: a_{i\gamma\downarrow}^\dagger )$.
176: An Ising (uniaxial) anisotropy is obtained for $U_1 > U_2$,
177: a planar (XY) anisotropy for $U_2 > U_1$,
178: and full spin-rotation symmetry for $U_1 = U_2$.
179:
180: As appropriate for $\rm Ho Mn O_3$,
181: we consider the case $U_2 > U_1$ corresponding to preferential ordering of spins
182: in the $x-y$ plane in spin space and an anisotropy gap for out-of-plane
183: excitations.\cite{holmium}
184: Magnetic excitations were analyzed in terms of a
185: Heisenberg model with a similar anisotropy term in Ref. [7].
186: At the HF level, the interaction term for orbital $\gamma$ then reduces to
187: \begin{equation}
188: H_{\rm int} ^\gamma =
189: - \sum_i \mbox{\boldmath $\sigma$}_{i\gamma} .{\bf \Delta}_i
190: \end{equation}
191: where the self-consistently determined mean field
192: ${\bf \Delta}_i = U_2 \langle {\bf S}_{i\gamma'} \rangle_{\rm HF}$
193: lies in the $x-y$ plane in spin space.
194: The HF sublattice magnetization depends only on $U_2$ and is determined from the
195: self-consistency condition $\langle {\bf S}_{\alpha} \rangle_{\rm HF}=
196: \sum_{{\bf k},l}\langle {\bf k},l| \frac{\mbox{\boldmath $\sigma$}}{2}
197: |{\bf k},l \rangle_\alpha $ in terms of the HF states $|{\bf k},l \rangle$
198: on sublattice $\alpha$, exactly as for the isotropic case.\cite{tri}
199: In the strong coupling limit the sublattice magnetization
200: $\langle {\bf S}_{\alpha} \rangle \approx 1/2$
201: at the HF level, and is reduced by about 50\% (in the isotropic case) due to quantum fluctuations, as found in different calculations cited in Ref. [9].
202:
203: We consider the $120^0$ ordered AF state on the triangular lattice,
204: and examine transverse spin fluctuations about the broken-symmetry state.
205: At the RPA level,
206: the magnon propagator reduces to a sum of all bubble diagrams
207: where the interaction vertices involving
208: $S_{ix}^2$ and $S_{iy}^2$ appear with interaction $U_2$,
209: whereas those involving $S_{iz}^2$ with interaction $U_1$.
210: Introducing a planar spin rotation,
211: so that spins are oriented along the $x'$ direction for all three sublattices,
212: we obtain for the transverse spin-fluctuation propagator
213: \begin{equation}
214: [\chi({\bf q},\omega)]_{\alpha\beta} ^{\mu\nu} =
215: \frac{[\chi^0 ({\bf q},\omega)]}
216: {{\bf 1} - 2[U][\chi^0 ({\bf q},\omega)]}
217: \end{equation}
218: in the $2\otimes 3$ spin-sublattice basis of the two transverse spin directions
219: $\mu,\nu=y',z'$ and the three sublattices $\alpha,\beta=A,B,C$.
220: The sublattice-diagonal interaction matrix
221: \begin{equation}
222: [U] = \left [ \begin{array}{cc} U_2 {\bf 1} & {\bf 0} \\
223: {\bf 0} & U_1 {\bf 1} \end{array} \right ]
224: \end{equation}
225: in the $y',z'$ basis, and the bare particle-hole propagator
226: \begin{eqnarray}
227: & & [\chi^0({\bf q},\omega)]_{\alpha \beta}^{\mu\nu} \nonumber \\
228: &=& \frac{1}{4} \sum_{{\bf k},l,m}
229: \left [ \frac{\langle \sigma_\mu \rangle_\alpha ^{-+}
230: \langle \sigma_\nu \rangle_\beta^{-+*}}
231: {E_{{\bf k-q},m}^+ - E_{{\bf k},l}^- + \omega}
232: + \frac{\langle \sigma_\mu \rangle_\alpha ^{+-}
233: \langle \sigma_\nu \rangle_\beta ^{+-*}}
234: {E_{{\bf k},m}^+ - E_{{\bf k-q},l}^- - \omega} \right ] \nonumber \\
235: \end{eqnarray}
236: involves integrating out the fermions in the broken-symmetry state.
237: In the particle-hole matrix elements
238: \begin{equation}
239: \langle \sigma_\mu \rangle_\alpha ^{-+} \equiv
240: \langle {\bf k-q},m| \sigma_\mu |{\bf k},l\rangle_\alpha
241: \end{equation}
242: of the rotated spins,
243: the spin orientation angles $\phi_\alpha$ in the fermion states
244: $|{\bf k},l\rangle$ are transformed out.
245: The numerical evaluation of $[\chi^0({\bf q},\omega)]$ in terms of the HF-level
246: AF-state energies and amplitudes is exactly as for the isotropic case
247: studied earlier.\cite{tri}
248:
249: The spin-wave energies $\omega_{\bf q}$ are then obtained
250: from the poles $1 - \lambda_{\bf q}(\omega_{\bf q}) = 0$ of Eq. (6)
251: in terms of the eigenvalues $\lambda_{\bf q}(\omega)$
252: of the matrix $2[U][\chi^0({\bf q},\omega)]$.
253: As $\omega_{\bf q}$ corresponds to spin 1/2,
254: it is scaled by the factor $2S$ for arbitrary spin $S$.\cite{magimp}
255: As expected for planar anisotropy, there is only one Goldstone mode corresponding
256: to planar rotation of spins, and the two out-of-plane modes become massive
257: with an anisotropy gap $\omega_{\rm gap}$.
258:
259: %fig 1
260: \begin{figure}
261: \begin{center}
262: \vspace*{-00mm}
263: \hspace*{-5mm}
264: \psfig{figure=holmium2.ps,height=60mm,angle=-90}
265: \vspace{0mm}
266: \end{center}
267: %\caption{.}
268: %\end{figure}
269: %inset
270: \vspace*{-125mm}
271: \hspace*{-40mm}
272: \epsfig{figure=mbz.ps,width=140mm}
273: \vspace{-75mm}
274: \end{figure}
275:
276: %fig 2
277: \begin{figure}
278: \begin{center}
279: \vspace*{-10mm}
280: \hspace*{-5mm}
281: \psfig{figure=holmium1.ps,height=95mm,angle=-90}
282: \vspace{-00mm}
283: \end{center}
284: \caption{(color online) Spin-wave dispersion for the three modes
285: calculated from Eq. (5) with $J=4t^2/U \approx 2.5$meV and $\Delta U \approx 0.35$meV,
286: along with neutron scattering data points for $\rm Ho Mn O_3$ at 20K from Ref. [7].}
287: \end{figure}
288:
289: Spin-wave dispersion $\omega_{\bf q}$ for the three modes is shown in Fig. 1
290: along two symmetry directions in the magnetic Brillouin zone (MBZ),
291: along with the neutron-scattering data for $\rm Ho Mn O_3$ at 20 K from Ref. [7].
292: The anisotropic Hubbard model provides a remarkably good description
293: of the spin dynamics.
294: We find that best fits with the magnetic excitations in $\rm Ho Mn O_3$
295: are only obtained in the strong-coupling limit $U/t > \sim 15$,
296: with a planar exchange energy $J =4t^2/U \simeq 2.5$meV and anisotropy
297: $U_2 - U_1 \equiv \Delta U \simeq 0.35$meV,
298: the individual values of $t$ and $U$ not being resolvable within
299: experimental resolution. To estimate the order of magnitude of the ratio $U/t$,
300: if we nominally take $U=1$eV, we obtain $t=25$meV and $U/t=40$.
301: Spin-wave dispersion calculated in the intermediate coupling regime
302: cannot be fitted with the neutron scattering data,
303: indicating no evidence of finite-$U$ double occupancy effects, as discussed below.
304:
305: %fig3
306: \begin{figure}
307: \begin{center}
308: \vspace*{-05mm}
309: \hspace*{-5mm}
310: \psfig{figure=uvarym.ps,width=80mm}
311: \vspace{-00mm}
312: \end{center}
313: \caption{(color online)
314: The ratio $\omega_1/ \omega_{2,3}$ at wave vector ${\bf q}_M$
315: provides a sensitive indicator of finite-$U$, double-occupancy effects,
316: shown for different values of spin anisotropy $\Delta U$.}
317: \end{figure}
318:
319: %fig4
320: \begin{figure}
321: \begin{center}
322: \vspace*{-5mm}
323: \hspace*{-5mm}
324: \psfig{figure=anisogap.ps,height=80mm,angle=-90}
325: \vspace{0mm}
326: \end{center}
327: \caption{Spin-wave anisotropy gap corresponding to out-of-plane fluctuations
328: shows a $\sim (\Delta U/U)^{1/2}$ behaviour with spin anisotropy. Here $U/t=16$.}
329: \end{figure}
330:
331: For the square-lattice $S=1/2$ AF $\rm La_2 Cu O_4$,
332: finite-$U$, double-occupancy effects associated with higher-order $(t^4/U^3$)
333: spin couplings are manifested in noticeable
334: spin-wave dispersion along the MBZ boundary.\cite{spinwave,spectrum}
335: Similarly, for the isotropic triangular-lattice AF,
336: the ratio $\omega_1/\omega_{2,3}$ of the non-degenerate and
337: degenerate spin-wave energies at wave vector ${\bf q}_M=(\pi/3,\pi/\sqrt{3})$
338: is a sensitive measure of the $U/t$ ratio,\cite{tri}
339: which asymptotically approaches $2/\sqrt{10} = 0.632$ in the strong-coupling
340: $(U/t \rightarrow \infty)$ limit, decreases monotonically with $U/t$,
341: and eventually vanishes at $U/t \approx 7$.\cite{tri}
342: Variation of $\omega_1/ \omega_{2,3}$ with $U/t$ is shown in Fig. 2
343: for different values of anisotropy $\Delta U /t$.
344: Neutron-scattering data for $\rm Ho Mn O_3$ shows that
345: $\omega_1/\omega_{2,3} \approx 11{\rm meV}/16{\rm meV} \approx 0.7$,
346: which yields a lower bound $(\sim 15)$ on the ratio $U/t$ from Fig. 2.
347:
348: The spin-wave anisotropy gap $\omega_{\rm gap}/t$,
349: corresponding to out-of-plane fluctuations with $q=0$,
350: varies as $(\Delta U/U)^{1/2}$ with spin anisotropy (Fig. 3),
351: which translates to $\omega_{\rm gap} \propto \sqrt{J \Delta U}$,
352: the geometric mean of the two energy scales.
353:
354: In conclusion, the anisotropic Hubbard model provides a simple extension to
355: phenomenologically include spin-space anisotropy and study spin excitations in
356: the full range from weak to strong coupling.
357: Comparison of spin-wave dispersion with RPA calculations allows for a quantitative
358: determination of the effective anisotropy, in addition to highlighting
359: any finite-$U$, double occupancy effects as seen in $\rm La_2 Cu O_4$.
360: The recent neutron-scattering data for spin-wave dispersion in
361: $\rm Ho Mn O_3$ are found to be well described by an anisotropic Hubbard model on a
362: triangular lattice with a planar (XY) spin anisotropy.
363: We find that the twin constraints $\omega_{\rm gap}/\omega_{1,2} \approx 1/3$
364: and $\omega_1/\omega_{2,3} \approx 2/3$ in the neutron scattering data
365: cannot be satisfied in the intermediate coupling regime,
366: and best fit indicates that magnetic excitations in $\rm Ho Mn O_3$
367: correspond to the strong-coupling limit $U/t > \sim 15$,
368: with $J \simeq 2.5$meV and $\Delta U \simeq 0.35$meV.
369: The ratio $\omega_1/\omega_{2,3}$ of the non-degenerate and
370: degenerate spin-wave energies at wave vector ${\bf q}_M=(\pi/3,\pi/\sqrt{3})$
371: is suggested as a sensitive measure of finite-$U$, double occupancy effects
372: in a triangular-lattice antiferromagnet.
373: Finally, in view of the formal resemblance with the
374: orbital Hubbard model (1), our RPA calculation provides a step forward towards
375: investigating magnetic excitations using realistic models including Hund's rule
376: coupling, crystal-field splitting etc.
377:
378:
379: \begin{thebibliography}{05}
380:
381: \bibitem{review1}
382: K. Kanoda, Physica C {\bf 282-287}, 299 (1997);
383: K. Kanoda, Hyperfine Interact. {\bf 104}, 235 (1997).
384:
385: \bibitem{review2}
386: R. H. McKenzie, Science, {\bf 278}, 820 (1997).
387:
388: \bibitem{watersup}
389: K. Takada {\em et al.},
390: Nature {\bf 422}, 53 (2003).
391:
392: \bibitem{weitering}
393: H. H. Weitering, X. Shi, P. D. Johnson, J. Chen, N. J. Dinardo, and S. Kempa,
394: Phys. Rev. Lett. {\bf 78}, 1331 (1997).
395:
396: \bibitem{res1}
397: T. Inami, Y. Ajiro, and T. Goto,
398: J. Phys. Soc. Jpn. {\bf 65}, 2374 (1996).
399:
400: \bibitem{res2}
401: L. E. Svistov, A. I. Smirnov, L. A. Prozorova, O. A. Petrenko, L. N. Demianets,
402: and A. Ya. Shapiro,
403: Phys. Rev. B {\bf 67} 094 434 (2003).
404:
405: \bibitem{holmium}
406: O. P. Vajk, M. Kenzelmann, J. W. Lynn, S. B. Kim, and S.-W. Cheong,
407: Phys. Rev. Lett. {\bf 94}, 087601 (2005).
408:
409: \bibitem{sato}
410: T. J. Sato, S. -H. Lee, T. Katsufuji, M. Masaki, S. Park, J. R. D. Copley, and H. Takagi,
411: Phys. Rev. B {\bf 68}, 014432 (2003).
412:
413: \bibitem{tri}
414: A. Singh, Phys. Rev. B {\bf 71}, 214406 (2005).
415:
416: \bibitem{spinwave}
417: R. Coldea, S. M. Hayden, G. Aeppli, T. G. Perring, C. D. Frost,
418: T. E. Mason, S.-W.Cheong, and Z. Fisk,
419: Phys. Rev. Lett. {\bf 86}, 5377 (2001).
420:
421: \bibitem{spectrum}
422: A. Singh and P. Goswami,
423: Phys. Rev. B {\bf 66}, 92402 (2002).
424:
425: \bibitem{sachi}
426: R. Sachidanandam, T. Yildirim, A. B. Harris, A. Aharony, and O. E.-Wohlman,
427: Phys. Rev. B {\bf 56}, 260 (1997).
428:
429: \bibitem{hanamura}
430: E. Hanamura, K. Hagita, and Y. Tanabe,
431: J. Phys.: Condens. Matter {\bf 15}, L103 (2003).
432:
433: \bibitem{roth}
434: L. M. Roth,
435: Phys. Rev. {\bf 149}, 306 (1966).
436:
437: \bibitem{held}
438: K. Held and D. Vollhardt,
439: Eur. Phys. J. B {\bf 5}, 473 (1998).
440:
441: \bibitem{quantum}
442: A. Singh, Phys. Rev. B {\bf 43}, 3617 (1991).
443:
444: \bibitem{magimp}
445: A. Singh and P. Sen,
446: Phys. Rev. B {\bf 57}, 10598 (1998).
447:
448: \end{thebibliography}
449:
450: \end{document}
451: