1: \documentclass[letter]{jpsj2}
2:
3: \def\runtitle{Antiferro-quadrupole Ordering of CeB$_6$ Studied
4: by Resonant X-ray Scattering}
5: \def\runauthor{Hironori {\sc NAKAO}}
6:
7: \title {Antiferro-quadrupole Ordering of CeB$_6$ Studied
8: by Resonant X-ray Scattering}
9: \author{Hironori {\sc Nakao},
10: \footnote{E-mail: hironori.nakao@kek.jp}
11: Ko-ichi {\sc Magishi},$^1$ Yusuke {\sc Wakabayashi},$^{2}$
12: Youichi {\sc Murakami}, Kuniyuki {\sc Koyama},$^1$
13: Kazuma {\sc Hirota},$^3$ Yasuo {\sc Endoh},$^4$ and Satoru {\sc Kunii}$^3$}
14:
15: \inst{
16: Photon Factory, Institute of Materials Structure Science, High Energy
17: Accelerator Research Organization (KEK), Tsukuba 305-0801, Japan\\
18: $^1$ Faculty of Integrated Arts and Sciences, The University of Tokushima\\
19: $^2$ Department of Physics, Keio University, Yokohama 223-8522, Japan\\
20: $^3$ Department of Physics, Tohoku University, Sendai 980-8578, Japan\\
21: $^4$ Institute for Materials Research, Tohoku University,
22: Sendai 980-8577, Japan\\ }
23: \recdate{\today}
24:
25: \abst
26: {
27: Under zero magnetic field, a quadrupolar order parameter
28: at $\mib{q}_Q=(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
29: in a typical antiferro-quadrupole (AFQ) ordering compound CeB$_6$
30: has been observed for the first time
31: by means of a resonant X-ray scattering (RXS) technique.
32: The RXS is observed at the $2p\rightarrow 5d$
33: dipole transition energy of the Ce $L_3$-edge.
34: Using this RXS technique to observe the pure order parameter of the AFQ state,
35: the magnetic phase diagram of Phase II is first determined.
36: }
37: \kword
38: { CeB$_6$, antiferro-quadrupole ordering,
39: orbital ordering, resonant X-ray scattering, magnetic phase diagram}
40:
41:
42: \begin{document}
43:
44: \sloppy
45: \maketitle
46:
47: For both $d$- and $f$-electron systems, various phase transitions
48: involving the orbital degree of freedom are attracting much interest.
49: The orbital degree of freedom is usually coupled with atomic displacements
50: by the Jahn-Teller effect.
51: For example, $3d$ electrons are spread in space so that the orbital
52: ordering is definitely accompanied with lattice distortion.
53: In the $f$-electron system, however,
54: orbital ordering without atomic displacement
55: owing to electron-electron interaction exists,
56: which is called antiferro-quadrupole (AFQ) ordering.
57:
58: CeB$_6$ has been considered a typical compound for AFQ ordering.%
59: ~\cite{ceb_st,gamma8,specific_heat,neutron1,%
60: nmr,theory_octa,ultrasonic,fluctuation-th,sample,neutron2}
61: CeB$_6$ has a cubic CsCl-type structure with Ce and B$_6$ octahedra
62: as shown in Fig.~\ref{fig1}.
63: The space group is $Pm\bar{3}m$ and
64: the lattice constant is $a=4.14$~\AA .~\cite{ceb_st}
65: The ground state of the Ce$^{3+}$ multiplet is a $\Gamma_8$ quartet
66: of the spin doublet (Kramers doublet) and the orbital doublet.~\cite{gamma8}
67: The specific heat measurement indicates the anomaly of a sharp peak
68: at $T_N\sim2.3$~K and a broad satellite peak at $T_Q\sim3.2$~K.%
69: ~\cite{specific_heat}
70: When a magnetic field is applied,
71: $T_Q$ shifts to a higher temperature with a rise in the peak height
72: and $T_N$ decreases rapidly.
73: This unusual behavior of $T_Q$ has attracted much attention.
74: \begin{figure}[bht]
75: \noindent
76: \includegraphics[width=7.5cm,angle=0]{fig1.eps}
77: \caption[Structure]{
78: (left) Crystal structure of CeB$_6$ at $T>T_Q$.
79: (right) AFQ structure at $T<T_Q$.
80: This is an NaCl-type ordering of Ce A, B sites;
81: the structure expects the order parameter
82: at $\mib{q}_Q=(\frac{1}{2},\frac{1}{2},\frac{1}{2})$.
83: B$_6$ octahedron is omitted.
84: }
85: ~\label{fig1}
86: \end{figure}
87: %
88: The magnetic scattering of neutron experiments proved
89: that the phase below $T_N$ (Phase III) is an antiferro-magnetic state.
90: In the phase at $T_N<T<T_Q$ (Phase II), on the other hand,
91: there is no magnetic ordering
92: although $T_Q$ is strongly affected by the magnetic field.%
93: ~\cite{neutron1}
94: Because the entropy calculated from the specific heat
95: reaches $R\ln2$ at $T_N$,~\cite{specific_heat}
96: the Kramers degeneracy was expected to lift below $T_N$.
97: Based on these experimental results,
98: the AFQ transition is expected at $T_Q$,
99: although the AFQ order parameter has never been observed
100: under zero magnetic field.
101: This is a famous hidden orbital ordering of the $f$-electron.
102:
103: A magnetic field was applied
104: to confirm the orbital ordering in Phase II.
105: The field-induced moment was observed
106: by neutron scattering~\cite{neutron1} and NMR.~\cite{nmr}
107: The neutron experiment observed the order parameter
108: at $\mib{q}_Q=(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
109: indicating a NaCl-type AFQ ordering as shown in Fig.~\ref{fig1},
110: while the results of the NMR measurements indicated a complicated
111: triple-$\mib{q}$ structure.
112: Recently, however, theories including the influence of octapoles
113: have clarified that the antiferromagnetic
114: structure of $\mib{q}_Q=(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
115: can account for the NMR data.~\cite{theory_octa}
116: In ultrasonic measurements, the softening of elastic constants was observed
117: toward $T_Q$,
118: which implied the AFQ ordering of $\Gamma_5$ symmetry.~\cite{ultrasonic}
119: The mysterious behavior of Phase II under the magnetic field was
120: theoretically explained by
121: the effect of the octupole-octupole interaction~\cite{theory_octa}
122: and the fluctuation of the quadrupole moment.~\cite{fluctuation-th}
123: All the results suggest the existence of the O$_{xy}$-type AFQ
124: orbital ordering in Phase II, but the AFQ order parameter
125: $\mib{q}_Q$ under zero magnetic field has not been observed up to now.
126:
127: Recently, the resonant X-ray scattering (RXS) technique
128: to observe orbital ordered states
129: has been developed for the $3d$-electron system.~\cite{murakami1,murakami2}
130: When the orbital of $3d$-electrons is ordered,
131: the energy level of the $4p$-orbital splits
132: owing to some interactions.
133: This $4p$-level splitting is the origin of the RXS.~\cite{murakami2}
134: Two mechanisms for the $4p$-level splitting are
135: theoretically proposed, however,
136: the mechanisms are a controversial problem.~\cite{RXS}
137: In this $3d$-electron system, the orbital ordering can also be discussed
138: on the basis of the state of the lattice distortion
139: studied by structural analysis.~\cite{kaji}
140: %
141: In the $4f$-electron system, on the other hand,
142: the RXS is the only technique for measuring the order parameter
143: of AFQ without magnetic ordering under zero magnetic field.
144: For the $L_3$-edge of the Ce ion, the energy level splitting of
145: the $5d$-orbital is expected to induce the RXS.
146: The $5d$-level splitting is caused by
147: the Coulomb interaction between $4f$- and $5d$-orbitals
148: rather than by the lattice distortion,
149: because the localized $4f$-electron has the weaker coupling with
150: the lattice.
151: Actually, no superlattice peak due to lattice distortion~\cite{kasuya}
152: has been observed below $T_Q$ in CeB$_6$ up to now.
153: In DyB$_2$C$_2$ of the $4f$-electron system, Hirota {\it et al.}
154: successfully measured an RXS signal of the AFQ ordering
155: for the first time.~\cite{dybc1}
156: In order to develop the RXS technique to
157: observe the orbital ordering in $f$-electron systems,
158: a study on AFQ ordering in CeB$_6$ is very important.
159: In this letter, we present the AFQ order parameter in CeB$_6$
160: which has been observed for the first time by using the RXS technique
161: under zero magnetic field and a magnetic field up to 2~T.
162:
163:
164: A single crystal of CeB$_6$ was fabricated by the
165: floating zone method under pressurized high-purity Ar gas.~\cite{sample}
166: The $(1\ 1\ 1)$ surface, $\sim$3$\times$1.5 mm$^2$, was cut
167: and polished with fine emery paper.
168: X-ray scattering experiments were carried out at the beam lines (BL)
169: 4C and 16A2 of the Photon Factory in KEK.
170: At BL-16A2, the incident beam was monochromatized
171: by Si(111) double crystals;
172: the second crystal was used for horizontal sagittal focusing.
173: The beam was vertically focused using a mirror.
174: The incident X-ray energy
175: is about 5.72~keV of a Ce $L_{3}$-edge
176: and the energy resolution is about 1~eV.
177: A four-circle diffractometer equipped with a He-flow cryostat was used.
178: The flow of He gas in the sample space prevented the sample surface
179: from heating up due to X-ray irradiation.
180: The temperature can be decreased to 2.5~K by He pumping,
181: and the magnetic field can be applied up to $H=2$~T
182: using a superconducting magnet.
183: The polarization vector $\sigma$ of the incident X-ray
184: was parallel to $[11\bar{2}]$ of the crystal structure and
185: the magnetic field was also applied along the $[11\bar{2}]$
186: in this experiment.
187:
188: \begin{figure}[tbh]
189: \noindent
190: \includegraphics[width=7.5cm,angle=0]{fig2.eps}
191: \caption[RXS-Edep]{
192: Energy dependence of scattering intensities
193: at $(\frac{1}{2},\frac{1}{2},\frac{1}{2})$ (filled circles)
194: and at $(0.49,0.49,0.49)$ (open circles) with $H=0$~T and $T=2.7$~K.
195: The difference of intensities between
196: $(\frac{1}{2},\frac{1}{2},\frac{1}{2})$ and $(0.49,0.49,0.49)$
197: (filled triangles).\\
198: Inset: Temperature dependence of $(hhh)$ scans
199: at 5.722~keV and at $H=0$~T.
200: }
201: ~\label{fig2}
202: % \vspace*{-1cm}
203: \end{figure}
204:
205: The RXS of the order parameter
206: at $\mib{q}_Q=(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
207: due to AFQ ordering has been searched under zero magnetic field.
208: The energy dependence of scattering intensities
209: at $(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
210: is measured when $H=0$~T and $T=2.7$~K, as shown in Fig.~2.
211: The intensity at $(0.49,0.49,0.49)$
212: is mainly attributed to the fluorescence of the Ce ion so that
213: we regard the intensity as the background.
214: The difference between the two intensities
215: is also shown in Fig. 2.
216: The weak signal of RXS is observed
217: at the $2p\rightarrow 5d$ dipole transition energy of 5.722~keV
218: which is denoted by an arrow.
219: The $(hhh)$ scans of the reciprocal lattice space at 5.722~keV
220: are shown in the inset of Fig.~\ref{fig2}.
221: The half-width at half-maximum of the peak at $T=3.02$~K ($<T_Q$)
222: is 0.00053\AA$^{-1}$ which is
223: resolution-limited in this experimental configuration;
224: namely, this phase is a long-range ordered state.
225: The dependence of the intensity on the energy and ${\mib q}$-space
226: is strong evidence that this peak intensity is attributable to
227: the AFQ ordering in CeB$_6$.
228: The signal of RXS disappears at $T=3.20$~K ($>T_Q$),
229: as shown in the inset.
230: Consequently, the temperature dependence of the AFQ order parameter
231: is clearly observed.
232:
233: \begin{figure}[tbh]
234: \noindent
235: \includegraphics[width=7.5cm,angle=0]{fig3.eps}
236: \caption[1.9T-th2th-tdep]{
237: Temperature dependence of $(hhh)$ scans
238: at $E=5.722$~keV and $H=1.9$~T.
239: }
240: ~\label{fig3}
241: \end{figure}
242:
243: Next, we investigate the AFQ ordering in a magnetic field
244: using this RXS technique.
245: Figure \ref{fig3} shows the temperature dependence
246: of the RXS of $(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
247: at $E=5.722$~keV and $H=1.9$~T.
248: The RXS intensity is much larger than that of the RXS
249: under zero magnetic field.
250: The peak profiles can be fitted by Gaussian curves
251: and the peak widths are also resolution-limited.
252: The dependence of the RXS integrated intensities
253: on the temperature and field is summarized in Fig.~\ref{fig4}.
254: As the magnetic field is increased, the RXS intensity
255: becomes larger and the transition temperature $T_Q$ shifts to higher ones.
256: The intensities of RXS at low temperatures grow
257: with increasing field owing to the development of the order parameter
258: of the AFQ state.
259: These temperature dependences of the intensities are fitted well by
260: $I_{(\frac{1}{2},\frac{1}{2},\frac{1}{2})} \propto
261: \bigl(\frac{T_Q-T}{T_Q}\bigr)^{2\beta}$
262: with the critical exponent $\beta=0.37$
263: which was determined by the neutron scattering experiment.~\cite{neutron2}
264: The solid lines in Fig.~\ref{fig4} indicate this function.
265: All the data can be fitted by this function with the same $\beta$
266: within experimental error.
267: Based on this fitting, the $H-T$ phase diagram is determined,
268: as shown in the inset of Fig.~\ref{fig4}.
269: This phase diagram is in good agreement
270: with previous result of neutron diffraction.~\cite{neutron1}
271: The increase in $T_Q$, which is an unusual behavior of CeB$_6$, is clearly
272: shown in the phase diagram.
273: \begin{figure}[t]
274: \noindent
275: \includegraphics[width=7.5cm,angle=0]{fig4.eps}
276: \caption[Phase diagram]{
277: Temperature dependence of RXS intensities of CeB$_6$
278: with various magnetic fields, $H\parallel [11\bar{2}]$.
279: The solid lines are results of fitting with the critical index $\beta=0.37$.
280: The transition temperatures $T_Q$ are indicated by thick vertical lines.
281: The baselines are shifted for clarity\\
282: Inset: $H-T$ phase diagram for the present result
283: and the result of the neutron experiment
284: with the magnetic field $H\parallel [111]$ (Ref.[4]).
285: }
286: ~\label{fig4}
287: \end{figure}
288: %
289: All the peak widths below $T_Q$ are resolution limited
290: and have no temperature dependence.
291: The short-range ordered state could not be observed
292: at $T>T_Q$ owing to a weak signal although
293: the short-range ordering is expected theoretically.~\cite{fluctuation-th}
294: %
295: Thus, by using the RXS technique at the Ce $L_3$-edge
296: $2p\rightarrow 5d$ dipole transition energy,
297: the order parameter of AFQ has been measured
298: and the phase diagram has also been determined.
299: %
300: It is noted that the RXS technique is the only experimental technique
301: for observing the order parameter of AFQ under zero magnetic field.
302:
303: When a magnetic field is applied, X-ray scattering,
304: as well as neutron scattering, can be used to detect
305: magnetic scattering of the field-induced moment.
306: Therefore we should pay attention to the origin of the observed intensity.
307: Under zero magnetic field, on the other hand,
308: the intensity of RXS corresponds not to magnetic scattering,
309: but to the pure AFQ order parameter.
310: The scattering intensity does not change much when the field is applied.
311: Therefore, it is expected that the RXS under magnetic field also
312: corresponds to the pure order parameter of AFQ ordering and
313: has no component of magnetic scattering.
314:
315: On the basis of the azimuthal angle dependence of RXS,
316: the wave function of the ordered orbital can be determined quantitatively,
317: which has been demonstrated in the case of YTiO$_3$.~\cite{ytio_nakao}
318: We attempted to measure the azimuthal angle dependence of RXS
319: under zero magnetic field.
320: The intensity was so weak in CeB$_6$ that
321: a quantitative discussion could not be made.
322: However, it is possible to determine
323: the wave function of the ordered orbital in the AFQ state
324: based on the azimuthal angle dependence of RXS.
325: Moreover, there is a possibility of determining
326: the various types of the induced quadrupole ordered state
327: depending on the applied field directions,
328: which have been predicted by some theories.~\cite{theory_octa}
329: %
330: This RXS at the dipole transition energy is represented by a tensor of
331: the atomic scattering factor, which only reveals the symmetry
332: of the quadrupole ordered state.
333: On the basis of RXS at the dipole transition, therefore,
334: we unfortunately can not observe octapole.
335:
336: The RXS observed here is considered to reflect the energy level split
337: of the Ce $5d$-orbital due to
338: the Coulomb interaction between $4f$- and $5d$-orbitals,
339: however, the scattering mechanism is still a controversial problem.
340: In fact, we also must consider an anisotropic potential
341: of the core $2p$-hole.
342: Development of a theory concerned with the RXS mechanism
343: in the $f$-electron system is strongly desired.
344:
345:
346: We have succeeded in measuring the order parameter
347: $\mib{q}_Q=(\frac{1}{2},\frac{1}{2},\frac{1}{2})$
348: under zero magnetic field
349: in a typical AFQ-ordered system, CeB$_6$, for the first time.
350: The RXS is observed at the $2p\rightarrow 5d$ dipole transition energy
351: of the Ce $L_3$-edge owing to
352: the energy level split of the Ce $5d$-orbital.
353: Only the RXS technique can be used to observe the order parameter
354: including information on Q of AFQ ordering under zero magnetic field,
355: as opposed to neutron scattering observing the field-induced moment.
356: Using this observation of the pure order parameter of AFQ,
357: the magnetic phase diagram is elucidated.
358: This success with the typical AFQ compound, CeB$_6$,
359: will be crucial in the study of AFQ materials.
360:
361:
362: The authors would like to thank Professor Hishizume for the use
363: of the He-flow-type cryostat with a superconducting magnet.
364: This work was supported by Core Research for Evolutional Science
365: and Technology Corporation (CREST).
366: %
367: This study has been performed under the approval of the Photon
368: Factory Program Advisory Committee (Proposal No. 99G211).
369:
370: %References
371: \begin{thebibliography}{99}
372: \bibitem{ceb_st}
373: P. P. Blum and E. F. Bertaut: Acta Cryst. {\bf 7} (1954) 81.
374: \bibitem{gamma8}
375: E. Zirngieble, B. Hillebrands, S. Blumenr\"oder, G. G\"untherodt,
376: M. Loewenhaupt, J. M. Carpenter, K. Winzer and Z. Fishk:
377: Phys. Rev. B {\bf 30} (1984) 4052;
378: B. L\"uthi, S. Blumenr\"oder, B. Hillebrands, E. Zirngiebl,
379: G. G\"untherodt and K. Winzer: Z. Phys. B {\bf 58} (1984) 31;
380: N. Sato, S. Kunii, I. Oguro, T. Komatsubara and T. Kasuya:
381: J. Phys. Soc. Jpn. {\bf 53} (1984) 3967.
382: \bibitem{specific_heat}
383: T. Fujita, M. Suzuki, T. Komatsubara, S. Kunii,
384: T. Kasuya and T. Ohtsuka: Solid State Commun. {\bf 35} (1980) 569.
385: \bibitem{neutron1}
386: J. M. Effantin, J. Rossat-Mignod, P. Burlet, H. Bartholin,
387: S. Kunii and T. Kasuya: J. Magn. \& Magn. Matter. {\bf 47\&48} (1985) 145.
388: \bibitem{nmr}
389: M. Takigawa, H. Yasuoka, T. Tanaka and Y. Ishizawa:
390: J. Phys. Soc. Jpn. {\bf 52} (1983) 728.
391: \bibitem{theory_octa}
392: R. Shiina, H. Shiba and P. Tahlmeier:
393: J. Phys. Soc. Jpn. {\bf 66} (1997) 1741;
394: O. Sakai, R. Shiina, H. Shiba and P. Tahlmeier:
395: {\it ibid.} {\bf 66} (1997) 3005;
396: R. Shiina, O. Sakai, H. Shiba and P. Tahlmeier:
397: {\it ibid.} {\bf 67} (1998) 941.
398: \bibitem{ultrasonic}
399: S. Nakamura, T. Goto, S. Kunii, K. Iwashita and A. Tamaki:
400: J. Phys. Soc. Jpn. {\bf 63} (1994) 623.
401: \bibitem{fluctuation-th}
402: N. Fukushima and Y. Kuramoto:
403: J. Phys. Soc. Jpn. {\bf 67} (1998) 2460.
404: \bibitem{sample}
405: M. Kawakami, S. Kunii, T. Komatsubara and T. Kasuya:
406: Solid State Commun. {\bf 36} (1980) 435.
407: \bibitem{neutron2}
408: J. Rossat-Mignod: Methods of Experimental Physics, ed. K. Sk\"old
409: and D. L. Price (Academic Press, New York, 1978) Vol. 23C, p.69.
410:
411: \bibitem{murakami1}
412: Y. Murakami, H. Kawada, H. Kawata, M. Tanaka, T. Arima,
413: T. Morimoto and Y. Tokura: Phys. Rev. Lett. {\bf 80} (1998) 1932.
414: \bibitem{murakami2}
415: Y. Murakami, I. Koyama, M. Tanaka, H. Kawata, J. P. Hill, D. Gibbs,
416: M. Blume, T. Arima, Y. Tokura, K. Hirta and Y. Endo:
417: Phys. Rev. Lett. {\bf 81} (1998) 582.
418: \bibitem{RXS}
419: S. Ishihara and S. Maekawa: Phys. Rev. Lett. {\bf 80} (1998) 3799;
420: I. S. Elfimov, V. I. Anisimov and G. A. Sawatzky:
421: {\it ibid.} {\bf 82} (1999) 4264;
422: M. Benfatto, Y. Joly and C. R. Natoli:
423: {\it ibid.} {\bf 83} (1999) 636;
424: M. Takahashi, J. Igarashi and P. Fulide:
425: J. Phys. Soc. Jpn. {\bf 68} (1999) 2530.
426: \bibitem{kaji}
427: For example: R. Kajimoto {\it et al.}:
428: Phys. Rev. B {\bf 60} (1999) 9506;
429: M. Kubota {\it et al.}: J. Phys. Soc. Jpn. {\bf 69} (2000) 1606;
430: C. D. Ling {\it et al.}:
431: Phys. Rev. B {\bf 62} (2000) 15096.
432: \bibitem{kasuya}
433: Though Kasuya predicted a charge dipolar ordering at ${\mib q}_Q$
434: owing to a pair distorted Jahn-Teller effect,
435: the superlattice peak due to the distortion has never been observed;
436: T. Kasuya: J. Phys. Soc. Jpn. {\bf 66} (1997) 2950.
437: \bibitem{dybc1}
438: K. Hirota, N. Ohmi, T. Matsumura, H. Nakao, Y. Wakabayashi,
439: Y. Murakami and Y. Endo: Phys. Rev. Lett. {\bf 84} (2000) 2706.
440: \bibitem{ytio_nakao}
441: H. Nakao {\it et al.}, to be submitted.
442: \end{thebibliography}
443: \end{document}
444: