1: \documentclass[twocolumn,aps,amssymb,showpacs,prb]{revtex4}
2: %\documentclass[aps,amssymb,preprint,showpacs,prb]{revtex4}
3: \usepackage{graphicx}
4: \usepackage{textcomp}
5:
6: \hyphenation{sig-nifi-cant-ly}
7:
8: \begin{document}
9:
10: \title{Magnetic-order induced phonon splitting in MnO from far-infrared spectroscopy}
11:
12: \author{T.~Rudolf}
13: \author{Ch.~Kant}
14: \author{F.~Mayr}
15: \author{A. Loidl}
16: \affiliation{Experimental Physics V, Center for Electronic Correlations and
17: Magnetism, University of Augsburg, D-86135~Augsburg, Germany}
18:
19: \date{\today}
20:
21: \begin{abstract}
22: Detailed far-infrared spectra of the optical phonons are reported for
23: antiferromagnetic MnO. Eigenfrequencies, phonon damping and effective plasma
24: frequencies are studied as a function of temperature. Special attention is paid
25: to the phonon splitting at the antiferromagnetic phase transition. The results
26: are compared to recent experimental and theoretical studies of the spin-phonon
27: coupling in frustrated magnets, which are explained in terms of a spin-driven
28: Jahn-Teller effect, and to ab initio and model calculations, which predict
29: phonon splitting induced by magnetic order.
30: \end{abstract}
31:
32: %\doi{}
33:
34: \pacs{63.20.-e, 75.50.Ee, 78.30.-j}
35:
36: \maketitle
37:
38: \section{Introduction}
39:
40: Spin-phonon coupling in magnetic semiconductors has been an active area of
41: research for
42: decades.~\cite{baltensperger,brueesch,lockwood,wakamura,wesselinowa,rudolf72}
43: In correlated matter it has been revived and came into the focus of modern
44: solid state physics again, a fact that had been triggered by the observation of
45: the splitting of phonon modes just at the onset of antiferromagnetic (AFM)
46: order. This was investigated by far-infrared spectroscopy in geometrically
47: frustrated ZnCr$_2$O$_4$ (Ref.~\onlinecite{sushkov}) and CdCr$_2$O$_4$
48: (Ref.~\onlinecite{rudolf9}) as well as in bond frustrated ZnCr$_2$S$_4$
49: (Refs.~\onlinecite{rudolf9,hemberger97}) and ZnCr$_2$Se$_4$
50: (Refs.~\onlinecite{rudolf9,rudolf75}). In the latter compound AFM order can be
51: suppressed by moderate external magnetic fields~\cite{hemberger98} and the
52: phonon splitting fully disappears as the magnetic ordering temperature $T_N$
53: approaches 0~K (Ref.~\onlinecite{rudolf75}).
54:
55: Of course by establishing long-range spin order at $T_N$, these chromium
56: spinels lower their symmetry and reveal slight distortions from cubic symmetry.
57: But these distortions are too small to explain the significant splitting of
58: phonon modes. In local spin-density approximations (LSDA) the non-cubic
59: behavior of the phonon properties can be explained in the absence of any
60: structural distortion and arises due to the anisotropy induced by the magnetic
61: order alone.~\cite{fennie} Furthermore, the structural distortions at the onset
62: of antiferromagnetic order in these and similar spinel compounds have been
63: explained by the concept of a spin-driven Jahn-Teller
64: effect.~\cite{yamashita,tchernyshyov}
65:
66: The idea of a purely magnetic order induced phonon splitting has been put forth
67: by Massidda {\it et al.}~\cite{massidda} for the AFM transition metal
68: monoxides. The splitting of the transverse optical modes has been indeed
69: experimentally documented by Chung {\it et al.}~\cite{chung} in MnO and NiO by
70: inelastic neutron scattering. As the late transition monoxides are prototypical
71: correlated electron systems and benchmark materials for Mott-Hubbard
72: insulators~\cite{anisimov} it seems self-evident to reinvestigate MnO by
73: far-infrared spectroscopy. Despite the fact that a number of far-infrared
74: experiments on MnO have been published,~\cite{plendl,kinney,mochizuki}
75: temperature dependent measurements are absent and to our knowledge the
76: splitting of phonon modes below the AFM phase transition has not been reported
77: so far.
78:
79: In principle it seems fascinating to compare the splitting of phonon modes in a
80: canonical antiferromagnet with strong exchange striction and only weak
81: frustration, with the splitting observed in strongly geometrically frustrated
82: magnets. MnO has a Curie-Weiss temperature of $\Theta_{CW} = -548$~K
83: (Ref.~\onlinecite{tyler}) and hence is characterized by a frustration parameter
84: $f = |\Theta_{CW}|/T_N \sim 5$. This value has to be compared with values in
85: strongly frustrated magnets, like the chromiums spinel ZnCr$_2$O$_4$, with
86: frustration parameters of the order of 20 (Ref.~\onlinecite{ramirez}).
87:
88: At room temperature MnO is paramagnetic and exhibits the cubic NaCl structure
89: ($Fm\bar{3}m$; $a = 0.44457(2)$~nm).~\cite{morosin} The transition into the AFM
90: state at $T_N = 118$~K (Ref.~\onlinecite{morosin}) is accompanied by a small
91: rhombohedral distortion,~\cite{tombs} with $a = 0.44316(3)$~nm and $\alpha =
92: 90.624(8)^\circ$ at liquid helium temperature.~\cite{morosin} Despite the fact
93: that MnO basically is the simplest transition-metal monoxide with Mn$^{2+}$
94: having a half filled $d$-shell and hence a negligible spin-orbit coupling, at
95: the magnetic ordering temperature it exhibits an anomalous large
96: exchange-striction, which arises from the dependence of the exchange energies
97: on inter-ionic distances, and amounts $\Delta a/a \approx 1.1\times10^{-4}$
98: (Ref.~\onlinecite{morosin}). In the AFM phase the magnetic moments are arranged
99: in ferromagnetic (FM) sheets parallel to (111) planes which are coupled
100: antiferromagnetically.~\cite{shull} The moments lie within these FM
101: planes~\cite{shull,shaked} and are oriented parallel and antiparallel to the
102: $\langle11\bar{2}\rangle$ direction.~\cite{goodwin} It has been pointed out
103: that this spin structure is not compatible with the rhombohedral distortion.
104: The true lattice symmetry definitively should be lower.~\cite{shaked}
105:
106: \section{Results and Discussion}
107:
108: A high-quality single crystalline platelet of MnO with one side polished to
109: optical quality was purchased from MaTecK GmbH. The surface was parallel to a
110: (100) plane. The reflectivity measurements were carried out in the far- and
111: mid-infrared range using the Bruker Fourier-transform spectrometers IFS 113v
112: and IFS 66v/S which are equipped with a He bath cryostat. In the far-infrared
113: we measured the phonon spectra with high accuracy in a range from 50 to
114: 700~cm$^{-1}$. Additional measurements up to 8000~cm$^{-1}$ were performed to
115: unambiguously determine $\epsilon_{\infty}$, which is governed by electronic
116: contributions only. The reflectivity spectra of MnO are reported as function of
117: temperature between liquid helium and room temperature.
118:
119: Representative results of the far-infrared reflectivity $R$ at 5~K, 100~K and
120: 305~K are shown in Fig.~\ref{fig1}. At room temperature, as expected for
121: paramagnetic MnO, we find a broad reststrahlen spectrum due to transverse and
122: longitudinal optical phonons of the NaCl structure. At zero wave vector the two
123: transverse optical phonon modes are degenerate and determine the strong
124: increase of the reflectivity close to 250~cm$^{-1}$. The longitudinal optical
125: phonon is responsible for the steep decrease of the reflectivity at
126: 550~cm$^{-1}$. In a harmonic solid, one would ideally expect that the
127: reflectivity in between these two characteristic frequencies, where $\epsilon'$
128: is negative, is close to 1. The observed structure in the reflectivity as
129: documented in Fig.~\ref{fig1}, especially the characteristic hump close to
130: 500~cm$^{-1}$, probably results from two-phonon processes involving zone
131: boundary optical and acoustical modes, which sum up to a zero-wave vector
132: excitation with a dipole moment transferred from the transverse optical phonon
133: as discussed later in the text. Despite this fact and contrary to earlier work
134: we made no attempts to get a better fit involving two
135: eigenmodes,~\cite{plendl,mochizuki} but we will document that a reasonable fit
136: can be obtained with realistic values of one transverse and one longitudinal
137: optical mode.
138:
139: \begin{figure}
140: \includegraphics{fig1.eps}
141: \caption{\label{fig1}(Color online) Reflectivity spectra of MnO at 5~K, 100~K
142: and 305~K. The lines are results of fits based on a generalized oscillator
143: model. The inset shows the splitting of the phonon mode below $T_N$ into at
144: least three modes at 5~K. The fits at 5 and 100~K utilize three modes.}
145: \end{figure}
146:
147: As the temperature is lowered, the reflectivity in the frequency range of the
148: reststrahlen band slightly increases reaching values close to~1 at 5~K. At the
149: same time the hump-like feature becomes weaker while the overall shape of the
150: reflectivity spectrum roughly remains unchanged. Both facts indicate the
151: decreasing influence of anharmonicity when approaching low temperatures.
152: However, a closer inspection reveals that just at $T_N$ a clear dip develops in
153: the restrahlen band close to 280~cm$^{-1}$ which indicates the splitting of
154: modes. This mode splitting is shown in the inset of Fig.~\ref{fig1}, which
155: provides an enlarged view of the reflectivity between 255 and 310~cm$^{-1}$. At
156: 100~K a dip evolves at 280~cm$^{-1}$ and at 5~K it seems that the structure of
157: the dip becomes even more complex indicating a splitting into three modes. This
158: will be analyzed in more detail later.
159:
160: The reflectivity spectrum of MnO has been modeled with a generalized oscillator
161: model,~\cite{gervais} where the factorized form of the complex dielectric
162: constant, $\epsilon = \epsilon' + i \epsilon''$, is determined by transverse
163: and longitudinal eigenfrequencies and damping constants:
164: \begin{equation}
165: \label{equ1} \epsilon(\omega)=\epsilon_{\infty} \prod_j
166: \frac{\omega_{Lj}^2-\omega^2-i\gamma_{Lj}\omega}{\omega_{Tj}^2-\omega^2-i\gamma_{Tj}\omega}
167: \end{equation}
168: Here $\omega_T$, $\omega_L$, $\gamma_T$ and $\gamma_L$ correspond to
169: longitudinal (L) and transversal (T) optical eigenfrequency and damping of the
170: $j$th mode, respectively. $\epsilon_{\infty}$ is the dielectric constant
171: determined by electronic processes alone. In the paramagnetic phase ($T>T_N$)
172: we have a single mode and $j=1$. For temperatures $T<T_N$, in the AFM phase, we
173: had to use $j=3$ to describe the data consistently. At normal incidence the
174: dielectric function is related to the reflectivity via
175: \begin{equation}
176: \label{equ2}
177: R(\omega)=\left|\frac{\sqrt{\epsilon(\omega)}-1}{\sqrt{\epsilon(\omega)}+1}\right|^2.
178: \end{equation}
179: In this formalism the dielectric strength of a given mode is determined by
180: \begin{equation}
181: \label{equ3}
182: \Delta\epsilon=\epsilon_{\infty}\frac{\omega_L^2-\omega_T^2}{\omega_T^2}
183: \end{equation}
184: and the effective plasma frequency, which is related to the effective charges,
185: follows from
186: \begin{equation}
187: \label{equ4}
188: \Omega^2=\epsilon_{\infty}(\omega_L^2-\omega_T^2)=\frac{N}{V}\frac{\epsilon_{\infty}}{\epsilon_{vac}}\frac{(Z^*e)^2}{\mu}.
189: \end{equation}
190: Here $N/V$ is the number of ion pairs per unit cell and $\epsilon_{vac}$ the
191: permittivity of free space. $Z^*e$ is the effective charge, which should be
192: close to~2 in mainly ionic MnO, and $\mu$ is the reduced mass of the ions
193: involved in the lattice vibration.
194:
195: \begin{figure}
196: \includegraphics{fig2.eps}
197: \caption{\label{fig2}(Color online) Eigenfrequencies (upper frame) and damping
198: constants (lower frame) of transverse optical modes of MnO. Closed symbols have
199: been determined from the fits of $R(\omega)$. Empty symbols were read off from
200: the peak maxima in $\epsilon''$. Full circles indicate neutron scattering
201: results (Ref.~\onlinecite{chung}).}
202: \end{figure}
203:
204: The reflectivity spectra have been fitted using Eqs.~(\ref{equ1})
205: and~(\ref{equ2}) utilizing a fit routine which has been developed by A.
206: Kuzmenko.~\cite{kuzmenko} The reflectivity spectrum is fully determined by the
207: transversal and longitudinal eigenfrequencies and damping constants, which are
208: treated as free parameters (four-parameter fit). In the AFM state the
209: dielectric strength and the effective plasma frequency has to be calculated for
210: each mode $j$ separately (see e.g. Ref.~\onlinecite{rudolf2007}). The
211: electronic dielectric constant $\epsilon_{\infty}$ has been determined from the
212: reflectivity measured up to 8000~cm$^{-1}$ ($\sim 1~$eV). $\epsilon_{\infty}$
213: obtained from these fits was found to scatter between 4.7 and 5.0, with no
214: systematic temperature variation. Hence for all further analyses we used an
215: average value of 4.85 for all temperatures from $5~\textrm{K} < T <
216: 305~\textrm{K}$. In addition, because of the strong anharmonic effects we
217: excluded the reflectivity regime between 350 and 480~cm$^{-1}$ in our fit
218: routine at all temperatures and found, that this restriction provides a more
219: reliable fit. The results of the best fits are shown as solid lines in
220: Fig.~\ref{fig1} and provide a good description of the reststrahlen spectrum of
221: MnO. Even in the magnetically ordered phase, where a clear splitting of the
222: modes can be observed, a fit using three eigenmodes yields a convincing
223: description of the experimentally observed reflectivity (see inset in
224: Fig.~\ref{fig1}). From these fits we determined all transverse and longitudinal
225: eigenmodes and damping constants as a function of temperature.
226:
227: Fig.~\ref{fig2}(a) and (b) show eigenfrequencies (upper frame) and dampings
228: (lower frame) of the transverse eigenmodes as a function of temperature. In the
229: paramagnetic phase $\omega_{TO}$ reveals a slight increase on decreasing
230: temperature as expected for an anharmonic solid. But more importantly, as is
231: impressively documented in Fig.~\ref{fig2}(a), a clear splitting of the
232: eigenfrequencies can be observed in the AFM phase. At 5~K we found three
233: transverse eigenmodes with frequencies close to 264, 278 and 292~cm$^{-1}$.
234: Hence the overall splitting amounts to almost 30~cm$^{-1}$.
235:
236: In the paramagnetic phase the temperature dependence of the damping decreases
237: as expected from anharmonic like phonon-phonon interactions
238: [Fig.~\ref{fig2}(b)]. Below $T_N$ the damping of the peak with the lowest
239: frequency continues to decrease, as does the damping of the mode with the
240: highest frequency and at the lowest temperatures the damping amounts about 2\%
241: of the mode eigenfrequency. The damping of the peak in between of these two
242: modes, with a very low dielectric strength, is of the order of 10~cm$^{-1}$.
243: However, one has to be aware that the experimental uncertainties for this mode
244: are rather large. We also would like to mention that at room temperature the
245: longitudinal eigenmode is characterized by an eigenfrequency
246: $\omega_{LO}=556$~cm$^{-1}$ and a damping constant $\gamma_{LO}=21$~cm$^{-1}$.
247:
248: \begin{figure}
249: \includegraphics{fig3.eps}
250: \caption{\label{fig3}(Color online) Dielectric loss of MnO at various
251: temperatures calculated via Kramers-Kronig transformation from the
252: reflectivity. The shift of the main mode and the emergence of new peaks with
253: low intensity below $T_N$ is clearly documented. The upper frame shows the
254: dielectric loss in the paramagnetic phase, the lower frame in the AFM state.}
255: \end{figure}
256:
257: To arrive at an independent check of the quality and accuracy of the
258: temperature dependence and splitting of the mode eigenfrequencies as shown in
259: Fig.~\ref{fig2}(a) we compared these fit results to the maxima of the
260: dielectric loss function. To do so, we calculated the dielectric loss from the
261: measured reflectivity spectra (70--8000~cm$^{-1}$), with a constant
262: extrapolation towards low frequencies and a smooth $\omega^{-h}$ extrapolation
263: towards high frequencies with exponents $h$ ranging from 0.09 to 0.22. In
264: Fig.~\ref{fig3} we show the dielectric loss between 250 and 300~cm$^{-1}$
265: determined by this procedure. For $T > T_N$, in the paramagnetic phase, we
266: detected a single loss peak indicating one well defined transverse optical
267: phonon as expected in the cubic NaCl-type phase. Below $T_N$ this mode reveals
268: significant splitting into two modes at 115~K and finally, at lower
269: temperatures, even into three modes which, at 5~K, are located at 263, 278 and
270: 289~cm$^{-1}$. These eigenfrequencies, as determined from the peak maxima of
271: the loss peaks, are plotted in the upper frame of Fig.~\ref{fig2} and we find
272: excellent agreement between the eigenfrequencies as determined out of
273: $R(\omega)$ by the four-parameter fit routine and those derived from the loss
274: peaks.
275:
276: Our far-infrared results on the splitting of the modes have to be compared with
277: recent neutron scattering results, which determined one mode at 300~K close to
278: 268~cm$^{-1}$ and two modes at 4.3~K, located at 268.6 cm$^{-1}$ and 293.6
279: cm$^{-1}$.~\cite{chung} The eigenfrequencies of these modes are indicated in
280: Fig.~\ref{fig2}(a). The two low-temperature modes obviously correspond to the
281: two more intense modes as documented in Fig.~\ref{fig3}. The small mode in
282: between probably is unobservable in neutron scattering due to the weak
283: intensity and the limited experimental resolution. However, the observation of
284: a splitting into three modes may elucidate the underlying mechanism driving
285: this phonon splitting and may put serious constraints on model calculations.
286:
287: Using Eqs.~(\ref{equ3}) and~(\ref{equ4}) we calculated the static dielectric
288: constant $\epsilon_0$ and the ionic plasma frequency, the results are shown in
289: Fig.~\ref{fig4}. For $T>T_N$, the ionic plasma frequency is almost temperature
290: independent with an average value of $\Omega_p = 1092$~cm$^{-1}$
291: [Fig.~\ref{fig4}(b)]. Below the N\'{e}el temperature the effective plasma frequency
292: exhibits an order parameter like increase and saturates close to 1140~cm$^{-1}$
293: at low temperatures. This increase of the plasma frequency of more than 6\% in
294: the magnetically ordered state indicates significant charge transfer processes.
295: The temperature dependence of the static dielectric constant, as calculated by
296: Eq.~\ref{equ3}, is shown in the upper panel of Fig.~\ref{fig4}. At low
297: temperatures we summed up the dielectric strength of all three eigenmodes.
298: $\epsilon_0$ slightly decreases from 20.75 at room temperature to 20.6 at 5~K,
299: with a dip-like suppression close to $T_N$. The results of the static
300: dielectric constant can be compared with measurements of the dielectric
301: constants at radio frequencies utilizing canonical dielectric
302: spectroscopy.~\cite{seehra} At 270~K, zero external pressure and depending on
303: polarization these authors determined dielectric constants between 18.05 and
304: 18.92, which is significantly lower than our value. In addition they report a
305: slight decrease of the dielectric constant below the magnetic ordering
306: temperature while we find a cusp like suppression close to $T_N$. These
307: discrepancies remain unexplained.
308:
309: Longitudinal and transverse eigenfrequencies, as well as $\epsilon_0$ and
310: $\epsilon_{\infty}$, are in good agreement with earlier
311: results.~\cite{plendl,kinney,mochizuki} However, one has to keep in mind that
312: these earlier results have been obtained by analyzing the reflectivity spectrum
313: using a model involving two damped oscillators. In these earlier
314: investigations, the longitudinal eigenfrequency has been calculated from the
315: Lyddane-Sachs-Teller (LST) relation:
316: \begin{equation}
317: \label{equ7}
318: \frac{\epsilon_0}{\epsilon_{\infty}}=\left(\frac{\omega_L}{\omega_T}\right)^2
319: \end{equation}
320: In our analysis the LST relation automatically is fulfilled. Hence, in the
321: present model a good fit is derived with the use of five parameters [see
322: Eq.~(\ref{equ1})], while the models used in Refs.~\onlinecite{plendl}
323: and~\onlinecite{mochizuki} utilize a set of seven parameters. In addition, it
324: is clear that the second mode is rather unphysical or has to be assigned to a
325: zero-wave-vector two-phonon process, which should be observable throughout the
326: Brillouin zone. At the zone boundary only the density of states is maximum,
327: producing the hump around 470~cm$^{-1}$ as observed in Fig.~\ref{fig1}. Our
328: result of the transverse optical phonon is also in good agreement with early
329: neutron scattering results,~\cite{haywood2,haywood4,wagner1} while it seems
330: that the longitudinal phonon mode has been underestimated by neutron scattering
331: techniques.~\cite{haywood4} In Ref.~\onlinecite{haywood4} the eigenfrequencies
332: of the LO branch at the zone center has been determined as 484~cm$^{-1}$
333: compared to 556~cm$^{-1}$ observed in this work.
334:
335: \begin{figure}
336: \includegraphics{fig4.eps}
337: \caption{\label{fig4}(Color online) Temperature dependence of the static
338: dielectric susceptibility $\epsilon_0$ (upper frame) calculated via the
339: Lyddane-Sachs-Teller relation and a constant average $\epsilon_{\infty}$ of
340: 4.85. The lower frame shows a steep increase in the overall plasma frequency
341: below $T_N$.}
342: \end{figure}
343:
344: An analysis of the ionic plasma frequency also allows clear statements
345: concerning the ionicity of the bonds in MnO. From detailed
346: electron-density-distribution studies utilizing gamma-ray
347: spectroscopy,~\cite{jauch} it has been concluded that the bonding is purely
348: ionic. We can calculate the ionic plasma frequency assuming an ideal ionic
349: valence $Z = \pm 2$ for both ions, which results in an effective plasma
350: frequency of 1873~cm$^{-1}$, which is significantly higher than the
351: experimentally observed value, $\Omega_p = 1077~$cm$^{-1}$ at room temperature.
352: This implies that $Z^*/Z \sim 0.58$, which still signals predominantly ionic
353: bonding, however with distinct covalent contributions. The effective charges
354: are almost independent of temperature for $T > T_N$ but reveal significant
355: charge transfer in the AFM phase.
356:
357: \section{Concluding remarks}
358:
359: In this work we document a detailed far-infrared study of the phonon properties
360: of MnO with special emphasis on the spin-phonon coupling and on the phonon
361: splitting in the AFM phase for $T < 118$~K. MnO reveals a $d^5$ configuration
362: with a spin-only value of $S = 5/2$, it is an electronically strongly
363: correlated material and a prototypical Mott-Hubbard insulator. We provide clear
364: experimental evidence for\\
365: i) a significant influence of spin ordering on eigenfrequencies, damping and
366: effective plasma frequencies. Specifically below the antiferromagnetic phase
367: transition we find an increase of the effective charge by more than 5\%. This
368: fact reflects strong charge transfer processes and/or strong changes in the
369: dynamic polarizability.\\
370: ii) Below $T_N$ we find a splitting of the phonon modes into three modes with
371: an overall splitting of as much as 30~cm$^{-1}$. The observed splitting
372: compares well with neutron scattering results by Chung {\it et
373: al.},~\cite{chung} who, however, observe a splitting into two modes only. In
374: neutron scattering the ratio of the integrated intensities of the two peaks was
375: approximately 2:1 in favor of the lower mode. Also the far-infrared results at
376: 5~K [Fig. 3(b)] show an intense low-frequency peak with approximately 2/3 of
377: the total intensity. It is unclear if the third peak in neutron scattering was
378: unobservable due to a weak intensity or due to a finite resolution. It is
379: possible that this low-intensity peak at 280~cm$^{-1}$ reflects the symmetry in
380: the antiferromagnetic state which should be lower than rhombohedral.
381:
382: It is unclear if the present experiments allow for a critical review on
383: spin-Jahn-Teller transitions. MnO is a strongly correlated material with the
384: spins residing on an fcc lattice. It reveals only weak magnetic frustration $(f
385: \approx 5)$ and exhibits giant exchange striction at $T_N$. One is apt to think
386: that symmetry lowering by exchange striction alone is enough to explain the
387: phonon splitting. However, it has been clearly demonstrated by Massidda {\it et
388: al.}~\cite{massidda} that the structural distortion at $T_N$ only explains
389: phonon splittings of the order of less than 1\%. The calculated magnetic order
390: induced splitting of the optical phonon modes at the zone center was estimated
391: to range model dependent from 3 to 10\%. Experimentally we observed effects
392: larger than 10\%. We conclude that MnO has to be described within the framework
393: of spin-driven Jahn-Teller transitions, developed to describe similar
394: observations in a variety of spinel compounds.
395:
396: \begin{acknowledgements}
397: This research has partly been supported by the Deutsche Forschungsgemeinschaft
398: through the German Research Collaboration SFB~484 (University of Augsburg).
399: \end{acknowledgements}
400:
401: \begin{thebibliography}{10}
402: \bibitem{baltensperger}
403: W. Baltensperger and J. S. Helman, Helv. Phys. Acta~{\bf 41}, 668 (1968); W.
404: Baltensperger, J. Appl. Phys.~{\bf 41}, 1052 (1970).
405: \bibitem{brueesch}
406: P. Br\"{u}esch and F. D'Ambrogio, Phys. Status Solidi~(B)~{\bf 50}, 513 (1972).
407: \bibitem{lockwood}
408: D. J. Lockwood and M. G. Cottam, J. Appl. Phys.~{\bf 64}, 5876 (1988).
409: \bibitem{wakamura}
410: K. Wakamura and T. Arai, J. Appl. Phys.~{\bf 63}, 5824 (1988).
411: \bibitem{wesselinowa}
412: J. M. Wesselinowa and A. T. Apostolov, J. Phys.: Condens. Matter~{\bf 8}, 473
413: (1996).
414: \bibitem{rudolf72}
415: T. Rudolf, K. Pucher, F. Mayr, D. Samusi, V. Tsurkan, R. Tidecks, J.
416: Deisenhofer, and A. Loidl, Phys. Rev. B {\bf 72}, 014450 (2005).
417: \bibitem{sushkov}
418: A. B. Sushkov, O. Tchernyshyov, W. Ratcliff, S. W. Cheong, and H. D. Drew,
419: Phys. Rev. Lett.~{\bf 94}, 137202 (2005).
420: \bibitem{rudolf9}
421: T. Rudolf, Ch. Kant, F. Mayr, J. Hemberger, V. Tsurkan, and A. Loidl, New J.
422: Phys. {\bf 9}, 76 (2007).
423: \bibitem{hemberger97}
424: J.~Hemberger, T. Rudolf, H.-A. Krug von Nidda, F. Mayr, A. Pimenov, V. Tsurkan,
425: and A. Loidl, Phys.~Rev.~Lett.~{\bf 97}, 087204 (2006).
426: \bibitem{rudolf75}
427: T. Rudolf, Ch. Kant, F. Mayr, J. Hemberger, V. Tsurkan, and A. Loidl, Phys.
428: Rev. B {\bf 75}, 052410 (2007).
429: \bibitem{hemberger98}
430: J. Hemberger, H.-A. Krug von Nidda, V. Tsurkan, and A. Loidl, Phys. Rev.
431: Lett.~{\bf 98}, 147203 (2007).
432: \bibitem{fennie}
433: C. J. Fennie and K. M. Rabe, Phys. Rev. Lett.~{\bf 96}, 205505 (2006).
434: \bibitem{yamashita}
435: Y. Yamashita and K. Ueda, Phys. Rev. Lett.~{\bf 85}, 4960 (2000).
436: \bibitem{tchernyshyov}
437: O. Tchernyshyov, R. Moessner, and S. L. Sondhi, Phys. Rev. Lett.~{\bf 88},
438: 067203 (2002); O. Tchernyshyov, R. Moessner, S. L. Sondhi, Phys. Rev. B~{\bf
439: 66}, 064403 (2002).
440: \bibitem{massidda}
441: S. Massidda, M. Posternak, A. Baldereschi, and R. Resta, Phys. Rev. Lett.~{\bf
442: 82}, 430 (1999).
443: \bibitem{chung}
444: E. M. L. Chung, D. McK. Paul, G. Balakrishnan, M. R. Lees, A. Ivanov, and M.
445: Yethiraj, Phys. Rev. B~{\bf 68}, 140406(R) (2003).
446: \bibitem{anisimov}
447: V. I. Anisimov, J. Zaanen, and O. K. Andersen, Phys. Rev. B~{\bf 44}, 943
448: (1991).
449: \bibitem{plendl}
450: J. N. Plendl, L. C. Mansur, S. S. Mitra and I. F. Chang, Solid State
451: Commun.~{\bf 7}, 109 (1969).
452: \bibitem{kinney}
453: T. B. Kinney and M. O'Keeffe, Solid State Commun.~{\bf 7}, 977 (1969).
454: \bibitem{mochizuki}
455: S. Mochizuki, J. Phys.: Condens. Matter~{\bf 1}, 10351 (1989).
456: \bibitem{tyler}
457: R. W. Tyler, Phys. Rev.~{\bf 44}, 776 (1933).
458: \bibitem{ramirez}
459: A. P. Ramirez, {\it Handbook of Magnetic Materials}, edited by K. H. J. Buschow
460: (Elsevier Science, North-Holland, 2001), Vol. 13, p. 423.
461: \bibitem{morosin}
462: B. Morosin, Phys. Rev. B~{\bf 1}, 236 (1970).
463: \bibitem{tombs}
464: N. C. Tombs and H. P. Rooksby, Nature~{\bf 165}, 442 (1950); S. Greenwald and
465: J. S. Smart, Nature~{\bf 166}, 523 (1950).
466: \bibitem{shull}
467: C. G. Shull and J. S. Smart, Phys. Rev.~{\bf 76}, 1256 (1949); C. G. Shull, W.
468: A. Strauser, and E. O. Wollan, Phys. Rev.~{\bf 83}, 333 (1951); W. L. Roth,
469: Phys. Rev.~{\bf 110}, 1333 (1958).
470: \bibitem{shaked}
471: H. Shaked, J. Faber Jr., and R. L. Hitterman, Phys. Rev. B~{\bf 38}, 11901
472: (1988).
473: \bibitem{goodwin}
474: A. L. Goodwin, M. G. Tucker, M. T. Dove, and D. A. Keen, Phys. Rev. Lett.~{\bf
475: 96}, 047209 (2006).
476: \bibitem{gervais}
477: F. Gervais and B. Piriou, J. Phys. C: Solid State Phys. {\bf 7}, 2374 (1974).
478: \bibitem{kuzmenko}
479: RefFIT by A.~Kuzmenko, University of Geneva, Version 1.2.44 (2006),
480: http://optics.unige.ch/alexey/reffit.html.
481: \bibitem{rudolf2007}
482: T. Rudolf, Ch. Kant, F. Mayr, J. Hemberger, V. Tsurkan, and A. Loidl, {\it
483: Polar phonons and spin-phonon coupling in HgCr$_2$S$_4$ and CdCr$_2$S$_4$
484: studied with far-infrared spectroscopy}, Phys. Rev. B, to be published
485: \bibitem{seehra}
486: M. S. Seehra, R. E. Helmick, G. Srinivasan, J. Phys. C~{\bf 19}, 1627 (1986).
487: \bibitem{haywood2}
488: B. C. G. Haywood and M. F. Collins, J. Phys. C~{\bf 2}, 46 (1969).
489: \bibitem{haywood4}
490: B. C. Haywood and M. F. Collins, J. Phys. C~{\bf 4}, 1299 (1971).
491: \bibitem{wagner1}
492: V. Wagner, W. Reichardt, and W. Kress, Proc. Conf. Neutron Scattering {\bf 1},
493: 175 (1976).
494: \bibitem{jauch}
495: W. Jauch, M. Reehuis, Phys. Rev. B~{\bf 67}, 184420 (2003).
496: \end{thebibliography}
497: \end{document}
498: