1: \documentstyle[twocolumn,aps,prl,psfig,floats]{revtex}
2:
3: \begin{document}
4:
5: \draft
6:
7: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname @twocolumnfalse\endcsname
8:
9: \title{Long-range antiferromagnetic order in the S=1 chain compound LiVGe$_2$O$_6$ }
10:
11: % repeat the \author\address pair as needed
12: \author{M.D. Lumsden$^1$, G.E. Granroth$^1$, D. Mandrus$^{1,2}$,
13: S.E. Nagler$^1$,
14: J.R. Thompson$^{1,2}$, J.P. Castellan$^3$ and B.D. Gaulin$^3$}
15: \address{$^1$ Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge,
16: TN, 37831 U.S.A.}
17: \address{$^2$ Department of Physics, University of Tennessee, Knoxville, TN,
18: 37996-1200 U.S.A.}
19: \address{$^3$ Department of Physics and Astronomy, McMaster University,\\
20: Hamilton, Ontario, L8S 4M1, Canada}
21: %\date{\today}
22: \maketitle
23:
24: \begin{abstract}
25: % insert abstract here
26: The phase transition in the compound LiVGe$_2$O$_6$ has been proposed as
27: a unique example of a spin-Peierls transition in an S=1 antiferromagnetic
28: chain. We report neutron and x-ray diffraction measurements of LiVGe$_2$O$_6$
29: above and below the phase transition at T=24 K. No evidence is seen for any
30: structural distortion associated with the transition. The neutron results
31: indicate that the low temperature state is antiferromagnetic, driven by
32: ferromagnetic interchain couplings.
33:
34: \end{abstract}
35:
36: % insert suggested PACS numbers in braces on next line
37: \pacs{75.25.+z,75.50.Ee,75.40.Cx}
38:
39: ]
40:
41: \narrowtext
42: %\twocolumn
43: % body of paper here
44:
45: The physics of one-dimensional (1D)
46: spin systems has been a field of considerable
47: theoretical and experimental interest in recent years. The enhanced
48: importance of quantum fluctuations in these systems manifests itself in
49: many exotic phenomena, two of which are the Haldane gap
50: in integer spin chains and the spin-Peierls transition in 1D half-integer spin
51: systems. Haldane predicted in 1983 \cite{Haldane},
52: based on field-theoretical models, that Heisenberg, antiferromagnetic (AF),
53: integer spin chains possess a spectrum of magnetic excitations in which
54: lowest lying excited states are separated by a finite energy gap
55: from the ground state. This remarkable result contrasts the
56: half-integer case where the excitation spectrum is gapless. However,
57: in the presence of magnetoelastic coupling, the energy of 1D, AF,
58: half-integer spin systems can be lowered by a dimerization of the uniform
59: chains. The resulting alternating spin chains possess a gapped
60: magnetic excitation spectrum separating a singlet ground state from the
61: triplet of excited states. This structural phase transition, driven
62: by gains in magnetic energy, from the uniform to dimerized spin chain is known
63: as the spin-Peierls transition \cite{spinpeierls}.
64:
65: Experimentally, the Haldane gap has been observed in numerous S=1 chain systems
66: including CsNiCl$_3$ \cite{Buyers}, AgVP$_2$S$_6$ \cite{Mutka},
67: YBaNiO$_5$ \cite{Xu}, Ni(C$_2$H$_8$N$_2$)$_2$NO$_2$ClO$_4$ \cite{Renard}
68: and Ni(C$_3$H$_{10}$N$_2$)$_2$N$_3$ClO$_4$ \cite{Zheludev}.
69: The spin-Peierls transition has
70: fewer experimental realizations most of which are organic, such as
71: Cu(Au)BDT \cite{spinpeierls} and MEM(TCNQ)$_2$ \cite{spinpeierls}
72: with the only known inorganic example being CuGeO$_3$ \cite{Hase}.
73: Recently, results were published on a new S=1, chain
74: material LiVGe$_2$O$_6$ \cite{Millet}
75: which were interpreted as being a remarkable union of
76: these two, seemingly unrelated, phenomena. Powder susceptibility
77: measurements showed a rapid decrease on passing through a transition
78: temperature of about 22 K reminiscent of the isotropic decrease in
79: magnetization, associated with the opening of a singlet-triplet gap in the
80: excitation spectrum, which accompanies a spin-Peierls
81: transition \cite{spinpeierls}. However, the
82: pre-existing gapped excitation spectrum normally precludes the
83: possibility of a spin-Peierls phase transition in integer spin chain systems.
84: It was suggested \cite{Millet} that the Haldane gap
85: was closed in LiVGe$_2$O$_6$ due to a novel, anomalously large
86: biquadratic exchange
87: interaction and this now gapless integer spin chain system undergoes
88: a spin-Peierls phase transition at T=22 K \cite{Millet}.
89:
90: However, very recent $^7$Li nuclear magnetic resonance (NMR) \cite{Gavilano}
91: experiments produced results
92: which seemed to contradict this interpretation. These measurements indicated
93: that the observed phase transition was
94: not a spin-Peierls transition and that the low temperature
95: phase was antiferromagnetic in nature. Their results are also interpreted
96: as being consistent with a discontinuous phase transition.
97: In addition, these measurements suggest unusual
98: behavior in the ordered phase including a gapped excitation spectrum with
99: a gap $\Delta$/$k_B$=83 K \cite{Gavilano}, roughly twice the
100: observed coupling constant J/$k_B$$\approx$45 K \cite{Millet},
101: despite an anisotropy D/$k_B$ of less than 20 K \cite{Millet}.
102: The authors of this \cite{Gavilano} and a recent theory
103: paper \cite{Mila} discuss possible mechanisms for the low-temperature
104: properties of LiVGe$_2$O$_6$ resulting from
105: splitting of the three $t_{2g}$ orbitals which accomodate the two V$^{3+}$
106: electrons. This splitting results from a small trigonal
107: distortion \cite{Millet} and separates the $d_{xz}$ orbital from the
108: low-lying d$_{xy}$, d$_{yz}$ doublet.
109: Two scenarios are suggested, the first of which
110: leads to a reduced intrachain coupling and subsequent enhancement
111: of biquadratic and next near-neighbor interactions \cite{Gavilano,Mila}.
112: The second scenario leads to increased importance of orbital degrees of
113: freedom \cite{Gavilano} and possible orbital ordering. However, until
114: now, the precise nature of the ordered state was unclear.
115:
116: To elucidate the nature of the low-temperature ordered state,
117: we have performed temperature dependent neutron diffraction
118: studies, and complimentary x-ray diffraction studies, above and below
119: the transition temperature. These measurements indicate a commensurate,
120: antiferromagnetic, long-range ordered low-temperature phase and provide
121: no evidence for a structural distortion associated with the transition.
122:
123: LiVGe$_2$O$_6$ crystallizes in a monoclinic unit cell with space group
124: P2$_1$/c and room temperature lattice constants a=9.863(4) \AA, b=8.763(2) \AA,
125: c=5.409 \AA, and $\beta$=108.21(1)$^\circ$ \cite{Millet}.
126: The structure is composed of chains of edge sharing
127: VO$_6$ octahedra which are connected by GeO$_4$ tetrahedra. Previous
128: measurements \cite{Millet,Gavilano} have indicated that the V-ions
129: exist in a 3+ oxidation state with an effective S=1 magnetic moment.
130:
131: \begin{figure}[bt]
132: %\centering
133: %\includegraphics[angle=270,origin=c,width=6in]{fig1.eps}
134: \centerline{
135: \psfig{file=fig1.eps,width=\columnwidth,angle=270,clip=}}
136: \vspace{2ex plus 1ex minus 0.5ex}
137: \caption{
138: DC molar susceptibility (a) and specific heat (b) as a function of
139: temperature for temperatures in the vicinity of the transition temperature.
140: The DC magnetization was measured in an applied field of 2.5 kOe while the
141: specific heat was a zero-field measurement. Both measurements clearly
142: show the presence of a phase transition at a temperature of about 24 K
143: indicated by the arrows.
144: }
145: \label{fig1}
146: \end{figure}
147:
148: Polycrystalline samples of LiVGe$_2$O$_6$ were prepared
149: using a slightly different
150: method than reported by Millet $et$ $al$.\cite{Millet}.
151: Stoichiometric quantities
152: of GeO$_2$, Li$_2$O, V$_2$O$_5$, and V metal were sealed in evacuated
153: Pt tubes and heated for 2 days at 900 $^\circ$C. The samples were single
154: phase according to powder x-ray diffraction.
155: To reproduce the magnetization results reported previously, a small amount
156: of powder was loaded in a Quantum Design SQUID magnetometer and the DC
157: magnetization was measured as a function of temperature in an applied
158: magnetic field of 2.5 kOe; the magnetization varied linearly with H
159: for fields up to 65 kOe . The results are shown in Fig. 1(a) for temperatures
160: up to 50 K and clearly demonstrate the same rapid decrease in
161: magnetization, below a transition temperature of about 24 K, reported
162: previously \cite{Millet,Gavilano}. The Curie tail below 10 K corresponds
163: to less than 40 ppm of 3d impurity content.
164: In addition, specific heat measurements
165: were performed on a 13.7 mg cold-pressed pellet of LiVGe$_2$O$_6$ using a
166: commercial heat-pulse, Quantum design calorimeter, the results of
167: which are shown in Fig. 1(b). These
168: results show a clear phase transition at a temperature of about 24 K,
169: consistent with the magnetization results.
170:
171: For the neutron diffraction measurements, approximately 10g of
172: LiVGe$_2$O$_6$ powder was loaded in an aluminum sample
173: can in the presence of He exchange gas and attached to the cold finger
174: of a closed-cycle He refrigerator. The neutron diffraction measurements
175: were performed on the HB1 triple-axis spectrometer at the High Flux Isotope
176: Reactor, Oak Ridge National Laboratory. Neutrons
177: of wavelength 2.44 \AA~ were selected using a pyrolitic graphite (PG) (002)
178: monochromator crystal and elastically scattered neutrons were measured
179: using a second PG (002) analyzer crystal.
180: To prevent higher-order contamination, a graphite
181: filter was placed in the incident beam. The intensity of diffracted
182: neutrons was measured for scattering angles between 5$^\circ$ and 100$^\circ$
183: for temperatures of 50 K and 11 K, well above and below the transition
184: temperature, respectively.
185:
186: The 50 K data was refined using Rietveld analysis and the observed pattern
187: was found to be consistent with the previously determined
188: crystal structure \cite{Millet}
189: with low temperature lattice constants of a=9.800 \AA, b=8.709 \AA,
190: c=5.364 \AA, and $\beta$=108.21$^\circ$.
191:
192: \begin{figure}[bt]
193: %\centering
194: %\includegraphics[angle=270,origin=c,width=6in]{fig2.eps}
195: \centerline{
196: \psfig{file=fig2.eps,width=\columnwidth,angle=270,clip=}}
197: \vspace{2ex plus 1ex minus 0.5ex}
198: \caption{
199: The upper panels show powder neutron diffraction data for the (110) and (001)
200: reciprocal lattice positions at temperatures above and below T$_N$.
201: The lower panels show identical scans performed using x-ray
202: diffraction techniques. The neutron diffraction data indicates the
203: presence of additional scattering below the transition temperature. The
204: absence of excess scattering seen in the x-ray diffraction results confirms
205: the magnetic nature of the scattering.
206: }
207: \label{fig2}
208: \end{figure}
209:
210: A portion of the resulting diffraction pattern is shown in the two upper panels
211: of Fig. 2 for temperatures of 11 K and 50 K.
212: We see clear, excess scattering at positions which can be
213: described by Miller indices (110) and (001).
214: These two reflections were the most pronounced but
215: other, weaker reflections were seen at (111) and (220) positions
216: with possible reflections at (-1-11) and (020). No
217: additional scattering was observed for reflections with scattering angles
218: in excess of about 50$^\circ$. This enhacement of the scattering at low angles
219: is consistent with it being magnetic in origin.
220:
221: To further confirm the magnetic nature of
222: the scattering, the measurements were repeated using x-ray diffraction
223: techniques. Approximately 2g of LiVGe$_2$O$_6$ powder was pressed
224: into a pellet and loaded
225: in a Be can in the presence of He exchange gas. This can was then connected
226: to the cold finger of a closed-cycle He refrigerator and mounted
227: in a Huber four-circle diffractometer. The incident radiation was
228: Cu-K$_\alpha$ x-rays from an 18 kW rotating anode x-ray generator which were
229: further monochromatized by Bragg reflection from a PG monochromator
230: crystal. To allow comparison with the neutron diffraction results,
231: the intensity of diffracted x-rays was measured for scattering angles
232: between 5$^\circ$ and 70$^\circ$ at
233: temperatures of 33 K and 11 K, again above and below the
234: transition temperature. The resulting data in the vicinity of the (110) and
235: (001) reflections is shown in the lower two panels of Fig. 2. Clearly,
236: no additional scattering is seen on passing through the transition temperature
237: and, in fact, no excess scattering was observed at any angle. This further
238: confirms the magnetic nature of the excess scattering observed in the
239: neutron diffraction experiment. This coupled with the resolution limited
240: nature of the excess neutron intensity, indicates that the low
241: temperature phase is, in fact, a long-range ordered, commensurate magnetic
242: state.
243:
244: The temperature dependence of the (110) and (001) magnetic Bragg reflections
245: was measured from 11 K to about 40 K and the resulting data are shown in the
246: upper panel of Fig. 3. For reference, we have plotted the temperature
247: derivative of the molar susceptibility and replotted the specific heat data
248: in the lower panel of Fig. 3. The excess neutron intensity clearly disappears
249: on passing through the transition temperature indicating its
250: association with the phase transiton. The temperature
251: dependence can be well described by a power law in the reduced temperature
252: suggestive of a continuous phase transition. In addition, no hysteresis
253: has been observed in either the magnetization or the specific heat and, thus,
254: in contrast to the NMR results \cite{Gavilano},
255: we conclude that the transition is continuous in nature.
256: While we can rule out the possibility of the transition being strongly
257: first-order in nature, it is impossible to
258: rule out a weakly first order phase transition from the existing data.
259:
260: \begin{figure}[bt]
261: %\centering
262: %\includegraphics[angle=270,origin=c,width=6in]{fig3.eps}
263: \centerline{
264: \psfig{file=fig3.eps,width=\columnwidth,angle=270,clip=}}
265: \vspace{2ex plus 1ex minus 0.5ex}
266: \caption{
267: The upper panel (a) shows the temperature dependence of the diffracted
268: neutron intensity for the (001) and (110)
269: magnetic Bragg reflections. The solid lines represent fits to a power
270: law in the reduced temperature, $t$=$|T-T_N|/T_N$.
271: For reference, the lower panel (b) shows
272: the temperature derivative of the susceptibility and the specific heat.
273: The dashed line represents the transition temperature (about 24 K). This
274: clearly shows that the extra magnetic neutron scattering disappears upon
275: passing through the transition temperature.
276: }
277: \label{fig3}
278: \end{figure}
279:
280: Despite the small number of observed reflections, we have sufficient
281: information to determine
282: the spin arrangement in the ordered state. As all reflections can be
283: indexed with integer indices, the magnetic unit cell is
284: identical in size to the nuclear unit cell and, hence, contains four
285: magnetic V$^{3+}$ sites. The observed magnetic peak intensities, coupled with
286: the absence of (100) and (010) reflections, are most consistent with the
287: arrangment of spins shown in Fig. 4 where we have superimposed the spins
288: on the known crystal structure. In comparing various magnetic structures,
289: an analytical approximation for the V$^{3+}$ magnetic form factor was employed
290: \cite{Brown}.
291: The observed spin structure has spins aligned
292: antiferromagnetically along the chain direction with ferromagnetic alignment
293: between neighboring chains suggestive of ferromagnetic interchain coupling.
294: Unfortunately, the limited amount of data makes it impossible to uniquely
295: determine the spin direction but the strength of the (001) reflection indicates
296: a small spin component along that direction. By
297: considering several plausible spin directions an
298: ordered moment ($\mu$=g$\mu_B$$<$S$>$) of 1.14(8) $\mu_B$ was extracted.
299: This is reduced
300: from the expected value of 1.79 $\mu_B$ (for g=1.79 \cite{Millet}
301: and $<$S$>$=1) and this difference is likely due to the importance of quantum
302: fluctuations in the low-dimensional system. In fact, the observed ordered
303: moment is similar to that seen in other Haldane chain systems which
304: order magnetically: CsNiCl$_3$ is found to possess an ordered moment of
305: 1.05 $\mu_B$ \cite{Minkiewicz}
306: and the set of systems RBaNiO$_5$, which if R $\not=$ Y order magnetically,
307: display an ordered moment of about 1.1 $\mu_B$ \cite{Zheludev2}.
308:
309: The observed ordering in LiVGe$_2$O$_6$ is very simple and, as mentioned
310: above, exhibits properties consistent with other Haldane chain systems which
311: order magnetically due to significant interchain
312: couplings. These interchain interactions are the likely cause of the
313: phase transition in LiVGe$_2$O$_6$ and the parallel alignment of
314: neighboring spin chains suggests that they are
315: ferromagnetic in nature. The strength of the interchain coupling is
316: estimated, using the Ginzberg-Landau theory of Scalapino, Imry, and
317: Pinkus \cite{Scalapino}, to be $J_{\perp}$ $\approx$ 1.4 K (assuming
318: $J_{\parallel}$=45 K, four neighboring chains, and $T_N$=24 K). These
319: interchain interactions suggest that LiVGe$_2$O$_6$ is a quasi-1D system
320: with $J_{\perp}/J_{\parallel}$ $\approx$ 0.03 (it is important
321: to note that this this mean-field
322: calculation results in an underestimated value of $J_{\perp}$).
323: This ratio of interchain to intrachain coupling may be slightly larger than
324: expected given the 1D nature of the crystal structure, but is
325: very similar to that observed in the quasi-1D Haldane system
326: CsNiCl$_3$ \cite{Buyers} and certainly seems feasibile.
327: Clearly, both quantum chemistry calculations
328: and further experiments are needed to investigate the actual magnitude
329: of the interchain couplings in LiVGe$_2$O$_6$.
330:
331: \begin{figure}[bt]
332: %\centering
333: %\includegraphics[width=6in]{fig4.eps}
334: \centerline{
335: \psfig{file=fig4.eps,width=\columnwidth,angle=0,clip=}}
336: \vspace{2ex plus 1ex minus 0.5ex}
337: \caption{
338: The structure of LiVGe$_2$O$_6$ is shown for two neighboring chains
339: of edge-sharing VO$_6$ octahedra together with the connecting GeO$_4$
340: tetrahedra. Superimposed on the VO$_6$ octahedra is the arrangement
341: of spins determined from the powder neutron diffraction data showing
342: antiferromagnetic alignment along the chains and ferromagnetic alignment
343: between chains. The spin direction could not be determined unambiguously
344: due to insufficient data and the indicated directions are for demonstration
345: purposes only.
346: }
347: \label{fig4}
348: \end{figure}
349:
350: We cannot rule out more
351: exotic mechanisms for the ordering in this system and if the transition is
352: discontinuous, as suggested by NMR studies, a simple ordering mediated
353: by interchain coupling may be unlikely.
354: However, of the theoretical ideas postulated to date, the notion of
355: weakened intrachain
356: coupling resulting from crystal-field splitting is particularly
357: intriguing \cite{Gavilano,Mila}.
358: Perhaps, in addition to enhancing novel interactions such
359: as next near-neighbor coupling or biquadratic exchange \cite{Gavilano,Mila},
360: this reduction of intrachain coupling also increases the relative
361: strength of the interchain interactions leading
362: to a rather conventional ordering phase transition. In other words,
363: the weakened intrachain interaction makes the system less one-dimensional.
364: The mechanism
365: involved in this phase transition clearly deserves further study and
366: several properties, such as the large magnetic gap observed
367: in the low temperature phase \cite{Gavilano} remain unexplained.
368:
369: In summary, we have performed neutron diffraction, and complimentary x-ray
370: diffraction, measurements on LiVGe$_2$O$_6$ and see clear evidence for
371: simple, commensurate long-range antiferromagnetic order below a
372: temperature of 24 K. In addition, we see no evidence for structural changes
373: at this temperature. These results contradict the complicated picture
374: of a S=1 spin-Peierls system originally proposed \cite{Millet} and seems to
375: indicate surprisingly large interchain interactions.
376:
377: We acknowledge fruitful discussions with D.A. Tennant, I. Affleck, and
378: E. S{\o}renson.
379: Oak Ridge National Laboratory is managed by
380: UT-Battelle, LLC, for the U.S. Dept. of Energy under
381: contract DE-AC05-00OR22725.
382:
383:
384: % now the references. delete or change fake bibitem. delete next three
385: % lines and directly read in your .bbl file if you use bibtex.
386: \begin{references}
387:
388: \bibitem{Haldane}
389: F.D.M. Haldane, Phys. Lett. {\bf 93A}, 464 (1983); F.D.M. Haldane, Phys. Rev.
390: Lett. {\bf 50}, 1153 (1983).
391:
392: \bibitem{spinpeierls}
393: See J.W. Bray $et$ $al$., in
394: $Extended$ $Linear$ $Chain$ $Compounds$, edited by J.S. Miller (Plenum
395: Press, New York, 1983), Vol. 3, p. 353 and references therein.
396:
397: \bibitem{Buyers}
398: W.J.L. Buyers $et$ $al$., Phys. Rev. Lett. {\bf 56}, 371 (1986).
399:
400: \bibitem{Mutka}
401: H. Mutka $et$ $al$., Phys. Rev. B {\bf 39} 4820, 1989.
402:
403: \bibitem{Xu}
404: Guangyong Xu $et$ $al$., Phys. Rev. B {\bf 54} R6827 (1996).
405:
406: \bibitem{Renard}
407: J.P. Renard $et$ $al$., Europhys. Lett. {\bf 3}, 945 (1987).
408:
409: \bibitem{Zheludev}
410: A. Zheludev $et$ $al$., Phys. Rev. B {\bf 53} 15004 (1996).
411:
412: \bibitem{Hase}
413: M. Hase, I. Terasaki, and K. Uchinokura, Phys. Rev. Lett. {\bf 70},
414: 3651 (1993).
415:
416: \bibitem{Millet}
417: P. Millet $et$ $al$., Phys. Rev. Lett. {\bf 83} 4176 (1999).
418:
419: \bibitem{Gavilano}
420: J.L. Gavilano $et$ $al$., cond-mat/0005534.
421:
422: \bibitem{Mila}
423: F. Mila and F.C. Zhang, cond-mat/0006068.
424:
425: \bibitem{Brown}
426: P.J. Brown, in {\it International Tables for Crystallography}, edited
427: by A.J.C. Wilson (Kluwer Academic Publishers, London, 1995), Vol. C,
428: Chap. 4.4.5.
429:
430: \bibitem{Minkiewicz}
431: V.J. Minkiewicz, D.E. Cox, and G. Shirane, Solid State Commun. {\bf 8},
432: 1001 (1970).
433:
434: \bibitem{Zheludev2}
435: See A. Zheludev, Physica B {\bf 241-243} 495 (1998) and references therein.
436:
437: \bibitem{Scalapino}
438: D.J. Scalapino, Y. Imry, and P. Pinkus, Phys. Rev. B {\bf 11}, 2042 (1975).
439:
440: %\bibitem{Garrett}
441: %A.W. Garrett $et$ $al$., Phys. Rev. Lett. {\bf 79}, 745 (1997).
442:
443: \end{references}
444:
445: % figures follow here
446: %
447: % Here is an example of the general form of a figure:
448: % Fill in the caption in the braces of the \caption{} command. Put the label
449: % that you will use with \ref{} command in the braces of the \label{} command.
450: %
451:
452: % Uncomment the following to put the figures into the latex file.
453:
454: \end{document}
455:
456: