cond-mat0603307/CCO.tex
1: \documentclass[prb,showpacs,floatfix,preprint]{revtex4}
2: 
3: \usepackage{epsfig} \usepackage{amsmath}
4: 
5: \def\cco{CoCr$_2$O$_4$}  \def\Tc{$T_c$}
6: 
7: \begin{document}
8: 
9: \title{Dielectric anomalies and spiral magnetic order in \cco} 
10: 
11: \author{G.\,Lawes$^1$, B.\,Melot$^2$, K.\,Page,$^2$ C.\,Ederer,$^2$ 
12: M.\,A.\,Hayward,$^3$ Th.\,Proffen$^4$ and  R.\,Seshadri$^2$} 
13: 
14: \affiliation{$^1$  Department of Physics and Astronomy, Wayne State University,
15: Detroit, MI 48201}  \affiliation{$^2$ Materials Department and Materials
16: Research Laboratory\\ University of California, Santa Barbara, CA 93106}
17: \affiliation{$^3$  Department of Chemistry, Inorganic Chemistry Laboratory,\\
18: University of Oxford, South Parks Road, Oxford, OX1 3QR UK} \affiliation{$^4$ 
19: Los Alamos National Laboratory, Manuel Lujan Jr. Neutron Scattering Center\\ 
20: LANSCE-12, MS H805, Los Alamos, NM 87545}
21: 
22: \date{\today}
23: 
24: \begin{abstract}  We have investigated the structural, magnetic, thermodynamic,
25: and dielectric properties of polycrystalline \cco, an insulating spinel
26: exhibiting both ferrimagnetic and spiral magnetic structures.  Below  \Tc =
27: 94\,K the sample develops long-range ferrimagnetic order, and we attribute a
28: sharp phase transition at $T_N$ $\approx$ 25\,K with the onset of long-range
29: spiral  magnetic order. Neutron measurements confirm that while the structure 
30: remains cubic at 80\,K and at 11\,K; there is complex magnetic ordering by 
31: 11\,K. Density functional theory supports the view of a ferrimagnetic 
32: semiconductor with magnetic interactions consistent with non-collinear 
33: ordering. Capacitance measurements on \cco\, show a sharp decrease in the
34: dielectric constant at $T_N$, but also an anomaly showing thermal hysteresis
35: falling between approximately $T$ = 50\,K and $T$ = 57\,K.   We tentatively
36: attribute the appearance of this higher temperature dielectric anomaly to the
37: development of \textit{short-range} spiral magnetic order, and discuss these
38: results in the context of utilizing dielectric spectroscopy to investigate
39: non-collinear short-range magnetic structures. \end{abstract}
40: 
41: \pacs{75.50.Gg, % (ferrimagnets) 
42:       75.80.+q, % (magnetoelectric and magnetostrictive effects)
43:       75.40.Cx  % (static properties-includes heat capacity).
44:      }
45: 
46: \maketitle
47: 
48: \section{Introduction}
49: 
50: In recent years there has been renewed interest in investigating systems in
51: which charge and spin degrees of freedom are strongly coupled.  From colossal
52: magnetoresistive materials\cite{CMR} to diluted magnetic
53: semiconductors\cite{DMS}, studies on  systems exhibiting an interplay between
54: their magnetic and electronic properties have led to a deeper appreciation of
55: the role of spin-charge coupling in determining the behavior of materials.
56: The long-standing problem  of understanding magnetodielectric coupling in
57: insulating magnetic compounds has also been re-visited lately.  Recent results
58: on magnetodielectrics have  highlighted several important features in these
59: systems, including the  observation that incommensurate non-collinear magnetic
60: structures tend to lead to large magnetocapacitive 
61: couplings,\cite{GotoPRL,KimuraNature} as does
62: geometrical magnetic frustration. With this background, it is natural to ask how
63: one can use dielectric  measurements to extract information about incommensurate
64: non-collinear  magnetic structures in frustrated magnets.  
65: 
66: \begin{figure} \epsfig{file=fig1.eps, width=6cm} \caption{Structure projected
67: down [110] showing octahedra around Cr and  tetrahedral Co atoms.} \end{figure}
68: 
69: \cco\, is a spinel ferrimagnet, with Co$^{2+}$ ions on the A sites and Cr$^{3+}$
70: on the B sites of the spinel structure (FIG.\,1).  Both Co$^{2+}$ and Cr$^{3+}$
71: are magnetic; in conjunction with the fact that  coupling between the B sites is
72: also quite strong, a complex magnetic phase  diagram emerges. Below
73: approximately \Tc = 94\,K the system has long-range  ferrimagnetic order.  There
74: remain some discrepancies concerning the  low temperature magnetic phase
75: diagram.  In the original investigation on  powder \cco\, samples, Menyuk {\it
76: et al.} found evidence for short-range  spiral magnetic order below $T$ $\sim$
77: 86\,K, and long-range spiral magnetic order below 
78: $T$ $\sim$ 31\,K.\cite{menyuk} Motivated by
79: suggestions that ferrimagnetic spiral long-range order in \cco\, should be
80: unstable, Tomiyasu {\it et al.} investigated both \cco, and MnCr$_2$O$_4$
81: through extensive  neutron and magnetization measurements on single crystals. 
82: These more recent  studies suggest that the spiral component develops
83: incommensurate short-range  order below approximately $T$ = 50\,K, and that this
84: short-range order  persists to the lowest temperatures with a correlation length
85: of only  3.1\,nm at $T$ = 8\,K. The authors attribute this behavior to weak
86: geometric  magnetic frustration.
87: 
88: These properties suggest that \cco\, would be a promising material for
89: investigations on magnetodielectric coupling in systems with non-collinear
90: magnetic order.  \cco\, is insulating, and has two different types of magnetic 
91: structures
92: (ferrimagnetic and spiral).  Measuring  the dielectric constant in phases with
93: both the ferrimagnetic and spiral components will offer insight into how
94: dielectric properties couple to different magnetic structures within the same
95: material.  
96: 
97: \section{Experimental and computational details}
98: 
99: Ceramic samples (beautiful jade-green pellets) were prepared from cobalt oxalate
100: [Co(C$_2$O$_4$)$_2$$\cdot$2H$_2$O] and chromium oxide (Cr$_2$O$_3$) by heating
101: well-ground mixtures initially at 800$^\circ$C for close to 24\,h followed by
102: regrinding, pelletizing and heating at 1000$^\circ$C for 24\,h. The samples were
103: initially characterized by lab powder x-ray diffraction  (Scintag X-II,
104: CuK$\alpha$ radiation) and then by time-of-flight powder neutron diffraction at
105: room temperatures, 80\,K and 11\,K  at the neutron powder diffractometer
106: (NPDF)\cite{Proffen} at the Lujan Center at Los Alamos National Laboratory. For
107: the measurements, the samples were contained in vanadium cans.
108: 
109: Calculations were performed using the projector augmented wave \cite{paw} method
110: implemented in the Vienna Ab-initio Simulation Package
111: (\textsc{vasp}).\cite{vasp} The cubic spinel structure (space group $Fd\bar 3m$)
112: with the lattice parameters fixed to the experimentally  obtained $a$ =
113: 8.3351\,\AA\/ was used, with an internal structural parameter  for the oxygen of
114: $x$ = 0.264 determined from the diffraction data. Different collinear magnetic
115: structures were calculated in order to extract the Heisenberg exchange
116: constants. The plane wave energy cut-off was 400\,eV and we used a 8 $\times$ 8
117: $\times$ 8 Monkhorst-Pack mesh of $k$-points.  The LDA + $U$ method of Dudarev
118: \textit{et al.}\cite{dudarev} was used and $U_\text{eff}$ was varied
119: independently on both the Co and Cr sites.
120: 
121: \section{Results and discussion}
122: 
123: We measured the temperature and field dependence of the DC magnetization  of
124: \cco\, using a Quantum Design MPMS SQUID magnetometer, operating between  $T$ =
125: 5\,K and $T$ = 350\,K, and measured the AC magnetization of the sample on a
126: Quantum Design Physical Property Measurement System PPMS at  frequencies between
127: $\omega/2\pi$ = 100\,Hz  and $\omega/2\pi$ = 10\,kHz. We measured the specific
128: heat of \cco\, as a function of temperature and magnetic field with a
129: quasi-adiabatic method on the PPMS using a 20\,mg pressed powder pellet.  We
130: found that the internal time constant of the sample was significantly smaller
131: than the other time constants in the fit, suggesting excellent thermal coupling
132: between the sample and calorimeter.  In order to measure the dielectric
133: constant, we pressed approximately 75 mg of \cco\, into a  5\,mm $\times$ 5\,mm
134: $\times$ 1.5\,mm pellet, and fashioned electrodes on  opposite faces of the
135: pellet using conducting silver epoxy. We measured the  dielectric constant
136: between $\omega/2\pi$ = 1\,kHz and $\omega/2\pi$ = 1\,MHz  with an Agilent 4284A
137: LCR meter, using the PPMS for temperature and  magnetic field control. The data
138: at lower frequencies was qualitatively similar to the higher frequency
139: measurements, although they were somewhat noisier and  displayed larger losses,
140: possibly due to conduction at the grain boundaries.
141: 
142: \begin{figure} \epsfig{file=fig2.eps, width=6cm} \caption{Upper panel: 
143: Field-cooled DC magnetization of \cco\, measured in an applied field of H=100
144: Oe. Lower Panel:  Real part of the AC susceptibility of \cco\, measured at H=0
145: and $\omega/2\pi$=10 kHz.  Inset: Lower temperature magnetic anomalies in \cco.}
146: \end{figure}
147: 
148: Figure\,2 shows the magnetization as a function of temperature measured on
149: warming after cooling in a magnetic field (upper panel).  These data are
150: qualitatively similar to the FC data obtained by  Tomiyasu {\it et
151: al.}\cite{tomiyasu}   The magnetization shows a large increase below the onset
152: of ferrimagnetic  long range order at \Tc = 95\,K, and there is an anomaly at
153: $T_S$ = 27\,K associated with spiral magnetic order in the system.  It should be
154: noted that the anomaly at $T_S$ = 27\,K is sharper than that presented in the Tomiyasu
155: {\it et al.} data.\cite{tomiyasu}  
156: 
157: The lower panel of Fig.\,2 plots the AC magnetization measurements at 10 kHz. 
158: The real part of the susceptibility exhibit a very large peak at \Tc. The inset
159: to the lower panel of Fig\,2 shows the low temperature AC susceptibility in more
160: detail.  There is a very clear peak in $\chi'$ at $T_S$ = 27\,K, consistent with
161: the sharp anomaly in the DC magnetization measurements. Furthermore, there is
162: another clear magnetic feature at approximately $T$ = 13\,K, which is  rather
163: more difficult to observe in the DC measurements. Tomiyasu {\it et al} 
164: associate this feature with a saturation of the correlation  length for the
165: spiral component on cooling.\cite{tomiyasu} These bulk  magnetization
166: measurements conclusively show the existence of magnetic  features at \Tc =
167: 95\,K and T$_S$ = 27\,K, associated with the ferrimagnetic  and spiral magnetic
168: structures respectively, and a third magnetic feature at  $T$ = 13\,K.
169: 
170: \begin{figure} \epsfig{file=fig3.eps, width=7cm} \caption{Time-of-flight neutron
171: powder diffraction at room temperature,  80\,K and 11\,K. The symbols in each
172: panel are data, the lines are fits  using the Rietveld method to the nuclear
173: structure, and the traces at the  bottom of each panel are the difference
174: profiles. The peaks in the difference profiles that emerge as the temperature is
175: lowered arise from long-range magnetic ordering.} \end{figure}
176: 
177: In order to get more information about the magnetic transitions in \cco, we also
178: conducted neutron measurements on these samples.  Figure\,3 displays the results
179: of fitting the low-angle time-of-flight  neutron diffraction data from \cco\/
180: powders to the \textit{nuclear\/} structure. At room temperature, the data is
181: almost completely fit except for a small reflection near $d$ = 2.15\,\AA\/ which
182: corresponds to the sample container. At 80\,K, below the first collinear
183: magnetic ordering temperature, we find that some of the peaks whose nuclear
184: contribution is fit, have a  residual intensity seen clearly in the difference
185: spectra, particularly  near $d$ = 1.7\,\AA\/ and 1.9\,\AA, due to magnetic
186: ordering. At 11\,K, the data reveal a number of peaks whose intensities are not
187: correctly fit, as well as new reflections which have come about because of the
188: long-range non-collinear magnetic order, again evident in the difference
189: spectrum. In particular, the low-angle peak at 2.7\,\AA\/ is evident only in the
190: 11\,K data as are peaks near $d$ = 1.7\,\AA\/ and $d$ = 2.2\,\AA. While we have
191: not at this stage been able to completely establish the magnetic structure, we 
192: present these data to strengthen our interpretation of the magnetic and 
193: transport measurements.
194: 
195: The electronic structure of \cco\/ was calculated using the LSDA + $U$
196: methodology. FIG.\,4 displays the total density of states (DOS) and the partial
197: densities of Co $d$ and Cr $d$ states obtained using $U_\text{eff}^\text{Co}$ =
198: $U_\text{eff}^\text{Cr}$ = 0\,eV (LSDA, top panel), as well as
199: $U_\text{eff}^{Co}$ = 4\,eV and $U_\text{eff}^\text{Cr}$ = 2\,eV (lower panel)
200: and a collinear N{\'e}el-type magnetic structure (Co spins antiparallel to Cr
201: spins). The system is an insulating fully spin-polarized ferrimagnetic
202: semiconductor already in LSDA, albeit with a very small gap, less than 0.1\,eV. 
203: Application of $U$ on either of the two magnetic sites enlarges the gap but only
204: application of $U$ on both sites leads to a significant gap of $>$ 1\,eV. The
205: calculations show that this system can be expected to be a robust insulator and
206: therefore promising for magnetodielectric measurements.
207: 
208: \begin{figure} \epsfig{file=fig4.eps, width=6cm} \caption{Total densities of
209: state (gray shaded) and partial densities of Co $d$ (black shaded) and Cr $d$
210: (black line) states obtained for $U_\text{eff}^\text{Co}$ =
211: $U_\text{eff}^\text{Cr}$ = 0\,eV (LSDA, top panel), as well as
212: $U_\text{eff}^{Co}$ = 4\,eV and $U_\text{eff}^\text{Cr}$ = 2\,eV (lower panel)
213: and a collinear N{\'e}el-type ordering.} \end{figure}
214: 
215: We also extracted the Heisenberg exchange coupling constants by calculating the
216: total energy differences for different collinear magnetic configurations. It has
217: been shown by Kaplan \cite{Kaplan} that using a simple Heisenberg model for the
218: cubic spinel structure with only nearest neighbor couplings $J_{AB}$ and
219: $J_{BB}$, the magnetic ground state structure is determined by the parameter $u
220: = (4 J_{BB} S_B)/(3 J_{AB} S_A)$, where $S_A$ and $S_B$ denote the spins of the
221: $A$ and $B$-site cations respectively. The collinear ferrimagnetic N{\'e}el
222: state is the stable ground state for $u < u_0 = 8/9$, whereas for larger values
223: of $u$ non-collinear spiral ordering is expected. The Heisenberg exchange
224: constants extracted from our calculations, using the reasonable values
225: $U_\text{eff}^{Co}$ = 4\,eV and $U_\text{eff}^\text{Cr}$ = 2\,eV, are: $S_A
226: J_{AB} S_B$ = 5.0\,meV, $S_B J_{BB} S_B$ = 2.4\,meV, and $u$ = 0.65. Increasing
227: $U_\text{eff}^\text{Co}$ and decreasing $U_\text{eff}^\text{Cr}$ leads to a
228: larger value of $u$, eventually exceeding the critical value $u_0 = 8/9$. In
229: addition, we also extracted the coupling constant $J_{AA}$ which is neglected in
230: the treatment of Kaplan. For the same values of $U$ as above
231: ($U_\text{eff}^{Co}$ = 4\,eV and $U_\text{eff}^\text{Cr}$ = 2\,eV) we obtain
232: $S_A J_{AA} S_A$ = 0.7\,meV. This value is not necessarily negligible and
233: indicates that the coupling between the A-site cations could play an important
234: role in the magnetic properties of cubic spinels and in fact lead to
235: non-collinear magnetic order even for values of $u$ smaller than $u_0=8/9$. The
236: possible importance of A-site coupling has also been pointed out in a recent
237: experimental study of the spinel systems $M$Al$_2$O$_4$ ($M$=Co, Fe, Mn)
238: \cite{Loidl}. A detailed analysis of the $U$ dependence of the electronic
239: structure and magnetic couplings in CoCr$_2$O$_4$ is in progress. From the
240: preliminary results presented in this work it can be concluded that the
241: calculated values of the exchange coupling constants are consistent with the
242: appearance of non-collinear magnetic order and that $A$ site coupling might be
243: important to understand the magnetic properties of this system.
244: 
245: In an attempt to establish whether the spiral magnetic transition at 
246: $T_S$ = 27\,K  is long range\cite{menyuk} or short range\cite{tomiyasu}, we measured
247: the  specific heat of our polycrystalline \cco\, sample.  The thermodynamic
248: data  presented in FIG.\,5, including both the lattice and magnetic
249: contributions  to heat capacity, strongly suggests that there are two
250: transitions to magnetic states with long range order, one occurring with \Tc =
251: 94\,K and the second at  $T_S$ = 27\,K.  Specifically, the sharp peak in $C/T$ at
252: $T_S$ = 27\,K is  indicative of long range collective ordering.  We can calculate
253: the entropy  under the peaks in the upper panel of FIG.\,5 to estimate the
254: fraction of  spins participating in the long range order at each transition. 
255: The entropy lost by \cco\, at \Tc\, is approximately 0.1\,J/K per mole, 
256:  while the entropy change at $T_S$ = 27\,K is approximately  0.15\,J/K
257: per mole, both at $H$ = 0.  These changes in entropy are  small
258: compared to the expected change in entropy for fully spin  ordered \cco\, of
259: 34.6 J/K per mole (using the spin-only values for Co$^{2+}$ and
260: Cr$^{3+}$).  This is consistent with the relatively large specific heat
261: exhibited by \cco\, over the entire temperature range shown, which suggests that
262: significant short range magnetic order is likely developing over a wide range of
263: temperatures.  We do not see any features in specific heat which could be
264: associated with a thermodynamic transition occurring at $T$ = 13\,K; this low
265: temperature anomaly in the magnetization does not involve any measurable change
266: in entropy of the system.
267: 
268: \begin{figure} \epsfig{file=fig5.eps, width=8cm} \caption{Upper Panel:  Specific
269: heat measurements on \cco\, plotted as $C/T$ vs $T$.  Lower Panel:  Specific
270: heat of \cco\, at the  $T_S$ = 27\,K transition, measured at $H$ = 0\,T and $H$ =
271: 5\,T.} \end{figure}
272: 
273: The lower panel of FIG.\,5 plots the dependence of the $T_S$ = 27\,K transition at
274: two different magnetic fields. The $H$ = 0 peak appears to be  somewhat larger
275: than that plotted in the upper panel. This is because the  higher density of
276: data points more accurately captures the proper shape of the  peak.  The amount
277: of entropy under the peak is practically unchanged. The  transition temperature
278: is almost completely independent of applied magnetic  field, but the magnitude
279: of the peak is suppressed by over a factor of two at  $H$ = 5\,T as compared to
280: the zero field data.  This suggests that fewer  spins undergo long-range spiral
281: order at $T_S$ = 27\,K at larger magnetic  fields, but it is impossible to tell
282: from the data if the extra entropy is  removed at higher or lower temperatures. 
283: 
284: One of the primary motivations for  these experiments was to investigate what
285: effect the ferrimagnetic and spiral  magnetic transitions would have on the
286: dielectric constant of \cco.  These data  are plotted in the upper and middle
287: panels FIG.\,6, which shows the  dependence of the dielectric constant on
288: temperature for two different  polycrystalline \cco\, samples measured on
289: warming.  In both these samples,  the dielectric constant depends only weakly on
290: temperature, with sample B  showing a variation of approximately 1\% over the
291: entire temperature range,  and sample A showing an even smaller change.
292: Differences in the magnitude of  the dielectric constant can be attributed to
293: errors in determining the  geometrical capacitance factor. Differences in the
294: temperature dependence  may possibly arise from small differences in
295: conductivity between the two  samples.  Preliminary measurements suggest that
296: sample B is slightly more  conducting than sample A, although this needs to be
297: investigated in more  detail. Both samples were structurally identical (verified
298: by powder x-ray  diffraction) but sample B was better sintered. 
299: 
300: \begin{figure} \epsfig{file=fig6.eps, width=6cm} \caption{Top Panel:  Dielectric
301: constant of \cco\, sample A measured at H=0 as a function of temperature
302: measured at $\omega/2\pi$=300 kHz.  Middle Panel:  Dielectric constant of \cco\,
303: sample B measured at H=0 as a function of temperature measured at
304: $\omega/2\pi$=30 kHz.  The arrows indicate the location of the observed
305: dielectric anomalies.  Bottom Panel:  Relative dielectric constant of \cco\,
306: sample A at the spiral ordering transition at $H$=0 and at $H$=5 T.}
307: \end{figure}
308: 
309: The general features in the dielectric constant are very similar for both these
310: samples.  Most notably, there is a well-defined drop in dielectric constant
311: coincident with the onset of long range spiral magnetic order at $T_S$ = 27\,K
312: (indicated with an arrow).  This sharp decrease is a signature of the coupling
313: between the dielectric properties and magnetic structure.  The change in
314: dielectric constant is fairly small, only 0.2\% between $T_S$ = 27\,K and the 
315: base temperature. This should be compared with EuTiO$_3$\cite{katsufuji} and
316: SeCuO$_3$\cite{lawes} which show larger changes on ordering (3.5\% and 2\%
317: respectively).  The small magnitude of the shift in the dielectric constant
318: \cco\, at the spiral ordering transition may result from the the fact that only
319: a small fraction of spins order at this temperature, as evidenced by the
320: thermodynamic data discussed previously.   This dielectric anomaly is almost
321: completely independent of magnetic field as well (lower panel of FIG.\,6).  An
322: applied magnetic field of $H$ = 5\,T introduces almost no shift in either  the
323: temperature or magnitude of the anomaly. A magnetic field does however 
324: introduce a small shift in the overall dielectric constant, on the order of 
325: 0.01\% in a field of $H$ = 1\,T (not shown). 
326: 
327: There is no clear dielectric feature at the ferrimagnetic ordering temperature
328: \Tc. There may be a subtle change of slope observable in Sample A, but this is
329: not seen in Sample B, which shows much larger intrinsic dependence of dielectric
330: constant on temperature.  The lack of a sharp anomaly at \Tc\, suggests that
331: there is only weak coupling between the dielectric constant of \cco\, and the
332: ferrimagnetic component of the magnetic structure.  Furthermore, there is no
333: dielectric feature associated with the $T$ = 13\,K transition. Since this 
334: magnetic feature is believed to be related to the spiral  magnetization
335: component\cite{tomiyasu}, one  would expect to see a clear anomaly at 13\,K. 
336: However, specific heat measurements show that only a small number of spins are
337: likely to be involved in this transition, which perhaps explains the absence of
338: any dielectric feature.
339: 
340: The dielectric constant of \cco\, couples only to the spiral spin structure, and
341: not to the ferrimagnetic structure.  The shift in dielectric constant observed
342: in many magnetodielectric materials is believed to arise from strong spin-phonon
343: coupling\cite{katsufuji,lawes}.  Within this framework, the phonon mode giving
344: rise to this magnetodielectric behavior must not couple to magnetic order along
345: the (001) direction, but should be sensitive to spiral spin-ordering with a
346: propagation vector along the $Q$=($\delta$, $\delta$,0) direction in the (001)
347: plane\cite{tomiyasu}.  This observation provides an important symmetry
348: restriction on allowed forms for the spin-lattice coupling in \cco.  Further
349: investigations on the microscopic mechanisms for spin-phonon coupling in \cco\,
350: will rely on a detailed understanding of the phonon modes in this system.  
351: 
352: We find evidence for one further temperature dependent dielectric anomaly in
353: \cco. The dielectric constant shows a distinct peak at approximately  $T$ =
354: 50\,K for the two samples (indicated by arrows in FIG. \,6). More precisely, 
355: measurements on Sample B seem to indicate the presence of a series of peaks, 
356: while Sample A rather exhibits one broader peak. The magnitude of the increase 
357: in dielectric constant at this temperature is comparable to the decrease in 
358: dielectric constant below $T_S$ = 27\,K.  However, the origins of this feature 
359: are less clear.  Neither the magnetization data in FIG.\,2 nor the specific 
360: heat data in FIG.\,6 show any evidence for a phase transition at this 
361: temperature.  Additionally, this anomaly has a rather broad onset, which  varies
362: somewhat with sample history.  Since \cco\, does not have any  structural phase
363: transitions at low temperatures, this anomaly must have a  magnetic origin. 
364: Specifically, we believe that this dielectric anomaly arises  from the onset of
365: short range spiral magnetic order, which has also been  shown to occur at
366: approximately $T$ = 50\,K\cite{tomiyasu}.
367: 
368: As a further probe of this anomaly, we measured the dielectric constant on
369: warming and cooling at different rates.  FIG.\,7 plots the dielectric constant
370: of \cco\, measuring on warming and cooling at 1\,K/min. (upper panel) and 
371: 5\,K/min. (lower panel). In both cases, there is {\it no} anomaly when the 
372: dielectric constant is measured on cooling, but there is a clear feature in  the
373: dielectric constant when the sample is warmed. Furthermore, the peak 
374: temperature for this anomaly occurs at a higher temperature ($T$ = 50\,K)  when
375: the sample is warmed at 5\,K/min. as compared to when it is warmed at  1\,K/min.
376: ($T$ = 48\,K).  This thermal hysteresis may arise from a  non-equilibrium
377: distribution of regions of short range spiral magnetic order.  Below $T_S$ =
378: 27\,K, the thermodynamic and dielectric data suggest that \cco\,  develops long
379: range spiral order.  On warming, the disappearance of a  non-equilibrium
380: distribution of short range spiral magnetic clusters may be  responsible for the
381: dielectric signal at $T$ = 50\,K.  
382: The shift of this anomaly to lower  temperatures as the
383: sample is warmed more slowly supports this idea.  Conversely, when the sample is
384: cooled from higher temperatures, there would be no spiral magnetic clusters
385: present, and therefore no shift in the dielectric constant. 
386: 
387: In summary, we have presented extensive magnetic, thermodynamic, neutron, and
388: dielectric data on \cco.  These results are consistent with a phase transition
389: to a ferrimagnetically ordered state at $T$=94\,K and we have conclusive
390: evidence for a second transition to a phase with long-range spiral magnetic
391: order below $T_S$ = 27\,K in our polycrystalline sample.  Magnetocapacitive
392: measurements show that the dielectric constant of \cco\, couples strongly to the
393: spiral magnetic order parameter, but is insensitive to the ferrimagnetic spin
394: component. This limits the allowed spin-phonon coupling symmetry in this system,
395: and places restrictions on the possible microscopic mechanism for the observed
396: magnetodielectric shifts. The dielectric constant of \cco\, is also found to be
397: affected by the development of {\emph short-range} spiral magnetic order.  This
398: suggests that in materials with strong spin-phonon coupling, it may be possible
399: to use capacitive measurements to probe short-range magnetic correlations.
400: 
401: \begin{figure} \epsfig{file=fig7.eps, width=6cm} \caption{Upper Panel: 
402: Dielectric constant of \cco\, near $T$=50 K measured on cooling and warming at 1
403: K/min.  Lower Panel:  Dielectric constant of \cco\, near $T$=50 K measured on
404: cooling and warming at 5 K/min.  In both plots, the arrows indicate the
405: temperature change.} \end{figure}
406: 
407: \section*{Acknowledgements}
408: This work in UCSB was supported by National Science Foundation through the MRL 
409: program (DMR00-80034), through a Chemical Bonding Center (CHE04-34567).
410: Measurements at the Lujan Center at Los Alamos Neutron Science Center were
411: supported by the Department of Energy Office of Basic Energy Sciences and
412: Los Alamos National Laboratory funded by Department of Energy under contract
413: W-7405-ENG-36. MAH thanks the Royal Society for funding.
414: 
415: \begin{thebibliography}{99} 
416: 
417: \bibitem{CMR} S. Jin, T.H. Tiefel, M. McCormack, R.A. Fastnacht, R. Ramesh, L.H.
418: Chen, {\it Science} {\bf 264}, 413 (1994).
419: 
420: \bibitem{DMS} H. Ohno, {\it Science} {\bf 281}, 951 (1998).
421: 
422: \bibitem{GotoPRL}
423: T. Goto, T. Kimura, G. Lawes, A.P. Ramirez, and Y. Tokura, 
424: \textit{Phys. Rev. Lett.\/} \textbf{92}, 257201 (2004).
425: 
426: \bibitem{KimuraNature}
427: T. Kimura \textit{et al.\/}, \textit{Nature\/} \textbf{425}, 55 (2003).
428: 
429: \bibitem{tomiyasu} K. Tomiyasu, J. Fukunaga, and H. Suzuki, {\it Phys. Rev. B}
430: {\bf 70}, 214434 (2004).
431: 
432: \bibitem{Proffen} Th. Proffen, T. Egami, S.J.L. Billinge, A.K. Cheetham, D.
433: Louca, and J.B. Parise, \textit{Appl. Phys. A} \textbf{74}, 163 (2002).
434: 
435: \bibitem{paw} P. E. Bl\"ochl, \textit{Phys. Rev. B} \textbf{50}, 17953 (1994).
436: 
437: \bibitem{vasp} Vienna Ab-Initio Simulation Package,
438: http://cms.mpi.univie.ac.at/vasp/vasp/vasp.html
439: 
440: \bibitem{dudarev} S. L. Dudarev, G. A. Botton, S. A. Savrasov, Z. Szotek, W. M.
441: Temmerman and  A. P. Sutton, \textit{Phys. Status Solidi A\/} \textbf{166}, 429
442: 1998.
443: 
444: \bibitem{Kaplan} T.~A.~Kaplan, {\it Phys Rev.} {\bf 119}, 1460 (1960).
445: 
446: \bibitem{Loidl} N.~Tristan, J.~Hemberger, A.~Krimmel, H.-A.~Krug von Nida, 
447:   V.~Tsurkan, and A.~Loidl, {\it Phys. Rev. B} {\bf 72}, 174404 (2005).
448: 
449: \bibitem{menyuk} N. Menyuk, K. Dwight, and A. Wold, {\it J. Phys.} (Paris)  {\bf
450: 25}, 528 (1964). 
451: 
452: \bibitem{katsufuji} T. Katsufuji and H. Takagi, {\it Phys. Rev. B} {\bf 64},
453: 054415 (2001). 
454: 
455: \bibitem{lawes} G. Lawes, A.P. Ramirez, C.M. Varma, and M.A. Subramanian, {\it
456: Phys. Rev. Lett.} {\bf 91}, 257208 (2003). 
457: 
458: \end{thebibliography}
459: 
460: \clearpage
461: 
462: \end{document}
463: