Journal of Crystal Growth 154 (1995) 322-328
Growth and characterization of
Tong-rong Zhao *, Masashi Hasegawa, Humihiko Takei Institute for Solid State Physics, The University of Tokyo, 7-22-1, Roppongi, Minato-ku, Tokyo 106, Japan
Received 13 February 1995; manuscript received in final form 16 March 1995
In this paper, the first successful growth of CuFeO 2 single crystals by applying the traveling solvent floating-zone technique is reported. Typical crystals of 5-8 mm in diameter and 10-30 mm in length have been obtained from the starting materials Cu20 and Fe20 3 with the molar ratio 1:1. The as-grown crystal is homogeneous and of the trigonal system with cell parameters a h 3.04 ,~ and c h = 17.17 .~ in the hexagonal description. In M6ssbauer effect (ME) measurements, the isomer shift for 57Fe in CuFeO 2 at room temperature is 0.40 m m / s with respect to metallic iron, which offers experimental confirmation that the iron is present in the trivalent state. Magnetic susceptibility measurements indicate that CuFeO 2 has two successive magnetic transition temperatures at T N = 13.5 K and TN = 9.5 K, respectively. Thermal analysis reveals that CuFeO 2 partially decomposes into Fe20 3 and Cu20 at about 1180°C. =
1. I n t r o d u c t i o n
C u F e O 2 is one of the stable compounds in the C u - F e - O ternary system which is historically the first known compound exhibiting the so-called delafossite structure [1,2]. The chemical structure belongs to the space group R3m with a h 3.03 and c h = 17.09 ~ in the hexagonal description . Such a structure consists of hexagonal layers of Cu, O and Fe with Cu at ( 0 , 0 , 0 ) , Fe at ( 1 / 2 , 1/2, 1 / 2 ) and two O at _+(1/9, 1 / 9 , 1/9), which stack in a sequence of A - B - C along the c-axis to form a layered triangular lattice antiferromagnet (T N = 13 K) where the triangular lattices of the magnetic Fe 3+ are separated by the =
nonmagnetic ion layers of Cu ÷ and 0 2- [3,4]. It could be anticipated from the structure that C u F e O 2 shows stronger interactions within the layers than along the c-axis. Also, the M6ssbauer effect (ME) investigation of C u F e O 2 by Muir and Wiederisch  showed that the iron is present in the trivalent and the copper in the monovalent state (Cu + Fe 3 + 022- ). As the Cu-delafossite compounds melt incongruently , and because of the strong influence of oxygen partial pressure on the C u 2 0 / C u O equilibrium , and liquid copper(I) oxide attacks all known crucible materials , it is very difficult to grow bulk C u F e O 2 single crystals. So far, mmsized CuFeO2 single crystals could only be grown by flux or hydrothermal methods, as described in the literature [2,5,7]. The purpose of the present work is to grow large C u F e O 2 single crystals with high quality by
0022-0248/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0022-0248(95)00172-7
T.-R. Zhao et al. /Journal of Crystal Growth 154 (1995) 322-328
means of a floating-zone method. Studies on the structural, chemical and thermal properties of the crystal have been carried out on the basis of X-ray diffraction (XRD), electron microanalysis, ME investigation and T G - D T A measurement. Studies of the magnetic properties of these crystals has also been carried out.
2. Experimental procedure 2.1. Crystal growth
As starting materials, powders of Fe203 (reagent grade, Kanto Chemical Co., Inc.) and C u 2 0 (99.7%, K~jundo Chemicals) were well mixed in the molar ratio 1:1 and filled into an alumina boat which was introduced into a silica reaction furnace (Ohkura Riken Co., Ltd.). The solid state reaction between C u 2 0 and FezO 3 was carried out at 850-900°C for 20-30 h in a flow of argon gas (20 ml/min): C u 2 0 + Fe203 ~ 2CuFeO 2.
The X R D technique was used to confirm that the synthesized powder was of CuFeO 2 and to check for the presence of extraneous phases. The cylindrical feed rod of about 8-10 mm in diameter and 60-70 mm in length was formed from the synthesized CuFeO 2 powder under the hydrostatic pressure of 3.9 × 107 Pa and then sintered again in argon at 850-900°C for 5 h to increase the density of the rod. Crystal growth was carried
out in an infrared radiation furnace, which has been described elsewhere . 2.2. Characterization
As-grown crystals are black in appearance with a typical size of 5 - 8 mm in diameter and 10-30 mm in length, as shown in Fig. 1. In the case of a size below 1 m m / h , necking was not performed because of the low growth velocities. 2.2.1. Structure analysis The structure and unit cell parameters of the as-grown crystals were examined and determined by means of the X-ray powder diffraction method using Cu K a radiation. The crystal perfection and the space group of the crystals at room temperature were determined by means of the precession method using a Buerger precession camera.
2.2.Z Chemical analysis To characterize the as-grown crystals, they were cut perpendicularly to and along the elongation axis. The specimens were then buried in resin, lapped with sandpaper of decreasing coarseness and finally polished with diamond slurry. The polished planes were examined by means of a stereo type microscope (Type SZH10, Olympus Optical Co., Ltd.) and a metallurgical microscope (Type BH-2, Olympus Optical Co., Ltd.). The compositional analysis was carried out
Fig. 1. CuFeO2 crystalsgrown by the floating-zonemethod. (a) Sample a: at the growth speed of 1 mm/h in Ar gas. (b) Sample b: at the growth speed of 2 mm/h in Ar gas.
T.-R. Zhao et al. /Journal of Crystal Growth 154 (1995) 322-328
by using a scanning electron microscope (SEM) together with electron probe microanalysis (EPMA).
2.2.3. M6ssbauer effect The M6ssbauer effect measurement was also made on the as-grown single crystal to investigate the metallic iron oxidation state in CuFeO2. For M/Sssbauer effect investigation, two pieces of sample crystal were cut perpendicular to the elongation axis, ground and polished to a thickness of about 50/zm and were then cemented to a lead aperture corresponding to the lateral dimensions of the crystal. The ME measurement was carried out by means of a conventional constant acceleration spectrometer in transmission arrangement, y-rays were detected in the direction parallel to the c-axis, which is perpendicular to the cleavage plane of the crystal. The velocity was calibrated with Fe metal and Fe20 3 absorbers at room temperature. The center shift is referred to that of Fe metal. The calibration of the hyperfine field was done by using a value of 517 kOe for Fe 20 3 at room temperature. The temperature was kept stable within 0.05 K throughout the experiment, and the temperature gradient across the absorber was estimated to be less than 0.05 K.
2.2.4. Magnetic susceptibility For magnetic property investigation, a sample of 2.0 X 2.7 x 3.5 mm from crystal b was attached to a holder. Laue back-reflection patterns were used to establish the orientation of the sample crystal. The magnetic susceptibility was measured in the magnetic field of 1 kOe by using a SQUID magnetometer (Quantum Design MPMS) from 300 to 5 K. The magnetic field was applied perpendicular ( H _1_c) or parallel ( H IIc) to the caxis.
2.2.5. Thermal analysis For TG-DTA measurement, powders from synthesized polycrystalline CuFeO 2 and single crystals were packed in a platinum pan, respectively, heated from room temperature to 1200°C at a rate of 10°C/min in flowing argon gas (500 ml/min) and then cooled to room temperature.
The thermal reactivities of the specimens were measured by means of a TG-DTA 2000 system (MAC Science Co., Ltd.). Moreover, a cylindrical rod of about 5 mm in diameter was formed from the synthesized CuFeO 2 powder, sintered in argon at 900°C for 5 h. The sintered rod was then suspended from the upper shaft in the infrared radiation furnace with a flow of argon gas (400 ml/min). After melting the tip of the rod at about 1180°C, the temperature was first lowered to 1170°C at a slow speed of 1°C/h, then to 1155°C at 5°C/h, and finally to room temperature at 80°C/h. The solidified rod was then cut along the elongation axis, buried in the resin and ground and polished with diamond slurry. Metallurgical microscope and SEM observations were employed on the section of the sample, and composition and structure were determined by means of EPMA and XRD.
3. Results and discussion
3.1. Crystal structure Fig. 2 shows the XRD pattern of the as-grown crystal. The main spacings agreed with those from the literature . Moreover, all diffraction peaks are of the delafossite structure which confirms that the structure of the as-grown CuFeO 2 is of the trigonal system. The hexagonal structure lattice constants of the crystals are calculated from the XRD pattern and the results are listed in Table 1. The lattice constants of the as-grown
20 Fig. 2. X-raypowderdiffractionpattern for sample a.
T.-R. Zhao et al. /Journal of Crystal Growth 154 (1995) 322-328
Table 1 Lattice constants of CuFeO 2 Sample
Crystal (0.5 r a m / h ) Crystal a (1 r a m / h ) Crystal b (2 r a m / h ) Sintered powder JCPDS
Lattice constant a h (A)
CuFeO 2 singleocrystals are a h = 3.036(1) ,~ and c h = 17.169(3) A, respectively. The results are in good agreement with those from the sintered polycrystalline CuFeO 2 sample (a h = 3.036(1) .~, c h = 17.171(9) ,~) and from the Joint Committee on Powder Diffraction Standards (JCPDS) data card ( a h = 3.035(0) ,~, c h = 17.161(1) A) . 3.2. Chemical analysis The compositional homogeneities of metallic iron and copper along the elongation axis of the as-grown CuFeOz single crystals are detected by SEM-EPMA. The metallic iron and copper distributions are uniform (Cu: 24.98 _+ 0.06 mol% and Fe: 25.01 _+ 0.04 mol%) and the molar ratio of copper to iron in the sample is found to be about 1 s (0.999+ 0.004), which indicates that the metallic compositional distributions in the asgrown single crystals are homogeneous and the obtained single crystals are considered to be of stoichiometric CuFeO 2. 3.3. MOssbauer effect The transmission spectra of the CuFeO 2 single crystal at room temperature is shown in Fig. 3, from which a quadrupole split absorption doublet is observed. The isomer shift arises from the electrostatic interactions of the nuclear charge and the electron charge, which is affected by the degree of occupation not only of the s, but also of the d electron orbitals, by means of their shielding ef-
Velocity (mm/see) Fig. 3. M6ssbauer spectra for the CuFeO 2 single crystal sample at room temperature.
fect on the s electrons. From our experiment, the room temperature spectrum, which covers a wide velocity range ( - 1 0 . 0 to 10.0 m m / s ) , indicates that only a pure CuFeO a phase can be observed. The isomer shift between observer and source is 0.40 m m / s , which is typical for iron in the trivalent state . The interaction of the nuclear magnetic dipole moment with a magnetic field at the site of the nucleus, splits the nuclear state with spin I ( I > O) into ( 2 1 + 1) sublevels. 57Fe has I = 1 / 2 for the ground state and I = 3 / 2 for the 14.4 keV first excited state. The electronic quadrupole interaction of the first excited state of the 57Fe nucleus ( I = 3 / 2 ) with an axially symmetric electric field gradient splits the first nuclear excited state into two sublevels, I z = + 3 / 2 and + 1/2. If the z-axis (the direction of the principal axis of the electric field gradient) is chosen as the c-axis of CuFeO 2, the quadrupole interaction energy, e becomes e = e2qQ/4
and the observed peak separation, A E, will be 2e* 1. In our e x p e r i m e n t , the observed quadrupole interaction at room temperature is
I For a detailed discussion, see Ref. .
T.-R. Zhao et aL / Journal of Crystal Growth 154 (1995) 322-328
Table 2 M6ssbauer parameters for CuFeO 2 Temperature (K)
Relative line intensity
Sh St S 
0.395 0.395 0.389
0.313 0.312 0.311
2.05 : 1 2.19 : 1 2.2:1
a Sh represents the specimen from the head part of the crystal, S t represents the one from the tail part of the crystal.
e = 0.31 m m / s , which coincides with the result from Muir and Weidersich . The angular dependences of the radiation pattern produced at the 57Fe nucleus by an electric field gradient with axial symmetry are 1 + cos20 for the + 3 / 2 ~ + 1 / 2 transition and 2 / 3 + sin20 for the + 1 / 2 ~ + 1 / 2 transition, respectively. O represents the angle between the direction of the principal axis of the electric field gradient and the propagation direction of the y-ray. For the single crystal absorber, the relative line intensities of the two quadrupole split lines are 3:1 for 0 = 0 ° and 3 : 5 for 0 = 90 °, respectively. The observed intensity ratio from our experiment is listed in Table 2.
3.4. Magnetic susceptibility Fig. 4 shows the thermal variation of the magnetic susceptibilities and their reciprocals parallel and perpendicular to the c-axis for the CuFeO2 single crystal. As the temperature decreases, CuFeO 2 undergoes a transition from a paramagetic to an antiferromagnetic phase and it will be noticed that Curie-Weiss-like behavior is found above about 170 K. The Weiss temperature O, obtained from the intercept of the reciprocal curve with the temperature axis, was found to be about - 1 0 5 K which is close to the result of Doumerc et al. . The magnetic susceptibility from the inset shows a dull maximum value around TN1 = 13.5 K, which corresponds exactly to the N6el point ( T N = 14 K) as revealed from the M6ssbauer resonance study by Muir et al. , and an abrupt decrease a[ about TN2 = 9.5 K. Little anisotropy was o b s e ~ e d aboue TN1 and at lower tempera-
tures ( T < TNI), X lL is observed to be less than X ± and tends to vanish at 0 K while X ± remains almost constant. Consequently, the magnetic susceptibility is found to be much more anisotropic in the low temperature phase (T < TN2) than in the intermediate temperature phase (TN2 < T < TNI). Mitsuda et al. reported that CuFeO 2 is considered to have two collinear magnetic ordered phases with different unit cells from each other below TN1, the monoclinic magnetic unit cell for the intermediate temperature phase (TN2 < T < TN1) and an orthorhombic magnetic unit cell for the low temperature phase (T < TN2). 
3.5. Thermal analysis The results of differential thermal analysis (DTA) for the CuFeO 2 single crystal and the synthesized CuFeO 2 powder are shown in Fig. 5. Both specimens show an endothermal peak at about 1180°C as temperature increases, which is thought to be the melting point of CuFeO 2, and an exothermic one at about 1170°C during the cooling process, the solidification temperature of CuFeO 2. From this experiment, CuFeO 2 acts like a kind of compound with a congruent melting point which is inconsistent with that reported before . To investigate the thermal properties
' , / ' ~,.,
,.' ] ~.~'
Temperature (K) Fig. 4. Temperature dependence of the magnetic susceptibilities XII and X± as well as the reciprocals X~ ~ and XS. 1 of the CuFeO 2 single crystal along and perpendicular to the c-axis, respectively. The inset shows the enlarged pattern of X, and X J_ below 25 K.
T.-R. Zhao et al. /Journal of Crystal Growth 154 (1995) 322-328
system. Fe304 is thought to be produced from the reduction of Fe203 in the argon atmosphere following the partial decomposition of CuFeO 2 at high temperature,
/\ 168,3 "c 20
2CuFeO 2 ~ Cu20 + Fe203,
-20 V 1181.4 ~2 i
t 50 ~-
. . . .
(b) sintered powder Fig. 5. DTA pattern of CuFeO2.
of CuFeO 2 in more detail, an experiment involving slow cooling of a molten rod sample was carried out in an infrared radiation furnace. The solidified rod was then examined using a metallurgical microscope, SEM-EPMA and XRD. The results are indicated in Fig. 6. During the cooling process, the molten zone of the sample is divided into three parts: the lowest part is CuzO-rich, the middle part is the CuFeO 2 single phase and the upper part is Fe304-rich. The co,.position of the CuEO-rich part, which was determined to be 56 mol% Cu20 and 44 mol% CuFeO 2 by EPMA, is thought to be that of the eutectic phase in this
Considering the narrow temperature difference between the endothermal peak corresponding to the decomposition of CuFeO2 and the exothermic peak corresponding to the eutectic phase of CuFeO 2 + Cu20 and the limitations of the experimental conditions, if the experimental temperature in TG-DTA measurement is raised highly, another endothermal peak corresponding to the melting of Fe203 or Fe304 should appear above 1550°C. As a result, CuFeO 2 has been demonstrated to be a compound with a incongruent melting point, and the growth of CuFeO2 single crystals occurs from the traveling solvent zone. A partial phase diagram for the CuzO-Fe203 binary system can be roughly sketched on the basis of DTA and the zone-cooling data, as shown in Fig. 7. If CuFeO 2 is heated from a temperature below 1180°C it will decompose entirely upon reaching 1180°C. In the decomposition process, a
CuFeO 2 polycrystal
I II. ' Fe O 3 4 + CuFe(
Cu~O + CuFe
6Fe203 --~ 4Fe304 + 02.
(a) single crystal
l CuFeO2 single phase
Fig. 6. S c h e m a t i c r e p r e s e n t a t i o n of the vertical d i s t r i b u t i o n of the c o m p o s i t i o n a l o n g the h a n g e d rod.
Mol % Fig. 7. P a r t i a l p h a s e d i a g r a m in the C u 2 0 - F e 2 0 ) system.
T.-R. Zhao et aL /Journal of Crystal Growth 154 (1995) 322-328
liquid of composition L 2 and a solid of composition S 2 will form. At the t e m p e r a t u r e 1180°C, three condensed phases, i.e., solid F e 2 0 3 or Fe304, solid CuFeO2, and liquid, coexist: C u F e O 2 ~ F e 2 O a ( F e 3 0 4 ) + liquid.
When the composition in question is cooled, first crystallization of F e 2 0 3 or F e 3 0 4 occurs at L~, with the liquid composition upon further cooling following the liquidus segment L 1 - L 2. When the t e m p e r a t u r e reaches 1180°C (M), C u F e O 2 begins to crystallize out. The liquid phase at L 2 is considered unstable and tends to follow the liquidus segment L 2 - E . Further cooling causes the solidification of C u F e O 2 and C u 2 0 . The composition of E should be the same as that of the molten zone in the present floating zone growth.
tals display a clear anistropic behavior with the successive magnetic phase transitions at TN~ = 13.5 K at TN2 = 9.5 K, respectively. (6) Thermal analysis demonstrates that CuFe0 2 has an incongruent melting point of about 1180°C.
Acknowledgements T h e authors gratefully acknowledge the M/Sssbauer effect m e a s u r e m e n t by Professor A. Ito and Dr. S. Morimoto at Ochanomizu W o m e n ' s University, and the magnetic m e a s u r e m e n t by Dr. M. Tamura. We also want to thank Mr. M. Koike for skilled experimental help, especially for the X-ray precession experiment.
4. Conclusions References (1) Bulk C u F e O 2 single crystals of 5 - 8 m m in diameter and 10-30 m m in length have been successfully grown by using the floating-zone method under the traveling solvent condition. (2) Structural analysis reveals that the as-grown C u F e O 2 single crystal is of the trigonal system with unit cell .parameters a h = 3.036(1) A and c h = 17.169(3) A, respectively. (3) Chemical analysis by means of E P M A confirms that the as-grown crystal is almost homogeneous in its concentration of copper and iron along the elongation axis. (4) The M6ssbauer effect measured at room t e m p e r a t u r e indicates that C u F e O 2 crystals exhibit an absorption spectra characterized by a quadrupole interaction energy of 0.31 m m / s and an isomer shift of 0.40 m m / s relative to metallic iron, the latter providing experimental confirmation of the ionic character of C u + F e 3 + 0 2 - . (5) Magnetic susceptibility m e a s u r e m e n t s show that with decreasing temperature, C u F e O 2 crys-
 E.M. Levin, C.R. Robbins and H.F. McMuride, Phase Diagrams for Ceramists (American Ceramic Society, Columbus, Ohio, 1964) p. 53.  P. Dordor, J.P. Chaminade, A, Wichainchai, E. Marquestaut, J.P. Doumerc, M. Pouchard and P. Hagenmuller, J. Solid State Chem. 75 (1988) 105.  A. Pabst, Am. Mineral. 31 (1946) 539.  R.D. Schmidt-Whitley, M. Martinez-Clemente and A. Revcolevschi, J. Crystal Growth 23 (1974) 113.  A.H. Muir, Jr. and H. Wiedersich, J. Phys. Chem. Solids 28 (1967) 65.  R.S. Zucker, J. Eiectrochem. Soc. 112 (1965) 417.  Klockow and W. Eysel, Mineral. Petrograph. Institut der Universit~it Heidelberg, Germany, JCPDS Grant-in-Aid Report (1988).  T.R. Zhao, M. Hasegawa, M. Koike and H. Takei, J. Crystal Growth 148 (1995) 189.  J.P. Doumerc, A. Wichainchai, A. Ammar, M. Pouchard and P. Hagenmuller, Mater. Res. Bull. 21 (1986) 745.  A.H. Muir, Jr., R.W. Grant and H. Wiedersich, Proc. Conf. on the Application of M6ssbauer Effect (Tihany, 1969).  S. Mitsuda, H. Yoshizawa, N. Yaguchi and M. Mekata, J. Phys. Soc. Jap. 60 (1991) 1885.