You are on page 1of 7

AIP ADVANCES 2, 042140 (2012)

Structure and physical properties of K0.63 RhO2


single crystals
S. H. Yao,1,a B. B. Zhang,1 J. Zhou,1 Y. B. Chen,2 S. T. Zhang,1 Z. B. Gu,1
S. T. Dong,1 and Y. F. Chen1,b
1
National Laboratory of Solid State Microstructures and Department of Material Science and
Engineering, Nanjing University, Nanjing 210093, P. R. China
2
Department of Physics and National Laboratory of Solid State Microstructures, Nanjing
University, Nanjing 210093 P. R. China
(Received 5 May 2012; accepted 31 October 2012; published online 7 November 2012)

K0.63 RhO2 single crystals were successfully grown by the flux method. Rietveld
refinement of power X-ray diffraction patterns suggests that K0.63 RhO2 belongs to
the monoclinic P63mmc space group. Transport measurements on K0.63 RhO2 revealed
metallic behavior. The temperature-dependent resistance is well fitted by a differ-
ent power law in two different temperature ranges. Antiferromagnetic ordering is
observed in the ab-plane of K0.63 RhO2 below 50 K. The most attractive feature of
K0.63 RhO2 is its significant Seebeck coefficient at room temperature (46 μV/cm),
which is much greater than that of normal metals. Considered all together, the metal-
lic conductivity, the significant Seebeck effect, and the non-hygroscopic properties
of K0.63 RhO2 make it a promising candidate material for thermoelectric applications.
Copyright 2012 Author(s). This article is distributed under a Creative Commons
Attribution 3.0 Unported License. [http://dx.doi.org/10.1063/1.4767464]

I. INTRODUCTION
Compounds of the form Ax BO2 (A = alkali metal, B = transition metal), with a bronze-type
crystal structure, have attracted much attention among physicists and materials scientists. Most work
on bronze-type oxides to date has focused on Nax CoO2 because of its various physical properties.1
The physical properties are strongly dependent on the sodium stoichiometry, x. For example, the
giant Seebeck effect (with a Seebeck coefficient ∼100 μV/K at 300 K) is displayed when x = 0.75
(i.e., Na0.75 CoO2 ),1, 2 and superconductivity is observed when x = 0.35 in Na0.35 CoO2 · 1.3H2 O with
a Tc of approximately 5 K.3 These novel physical properties have two main origins. One is the
delicate balance between electron-electron interactions, the crystal field, the degenerate d-orbits, the
spin-orbit interaction, and electron-phonon interactions. The other is the intrinsic two-dimensional
triangle lattice structure, which produces spin frustration and breaks the long-range magnetic order.
It is proposed that the different Co ion spin states and the spin-frustrated state together produce the
giant Seebeck effect in metallic Na0.75 CoO2 .4, 5
Besides Nax CoO2 , a similar thermoelectric material, rhodium oxide, has also been studied
recently. Because rhodium is located below cobalt in the periodic table, it is naturally expected to
have similar properties. For example, Terasaki et al. find that the thermopower of La1-x Srx RhO3 is
approximately 100 μV/K up to 500 ◦ C, comparable to the value in Na0.75 CoO2 .2 Another rhodium
oxide, K0.49 RhO2 , also displays a large Seebeck coefficient (∼40 μV/K at 300 K), which is much
greater than in many normal metals.6 A giant thermoelectric effect was predicted in K7/8 RhO2 by
the first principles calculations.7 However, many important problems remain unresolved with regard
to the Kx RhO2 system. For example, how does the thermopower of Kx RhO2 change as x varies?

a Electronic address: shyao@nju.edu.cn (S.H. Yao)


b Corresponding author: yfchen@nju.edu.cn (Y.F. Chen)

2158-3226/2012/2(4)/042140/7 2, 042140-1 
C Author(s) 2012
042140-2 Yao et al. AIP Advances 2, 042140 (2012)

Can there be superconductivity in the water-intercalated Kx RhO2 ? What is the magnetic structure
of Kx RhO2 ? To address these questions, we have successfully grown K0.63 RhO2 single crystals and
studied their electrical, magnetic, and thermoelectric properties.

II. EXPERIMENTAL
Kx RhO2 single crystals were grown by the self-flux and spontaneous nucleation technique from
high-temperature solutions, with K2 O chosen as the solvent. The mass ratio between solute and
solvent was 1:10. The starting materials for sample synthesis were Rh2 O3 and K2 CO3 with purities
of 99.99%. The weighed materials were completely mixed, heated to 1100 ◦ C, and kept at this
temperature for 24 h. After superheating, the melt was cooled to 880 ◦ C quickly and then slowly
decreased at a rate of 1–2 ◦ C/h from 880 ◦ C to 600 ◦ C. The melt was then allowed to cool to room
temperature naturally. Many shiny and thin hexagonal plate-like crystals were obtained after rinsing
with distilled water.
The composition of Kx RhO2 was determined by energy-dispersive spectroscopy (EDS) in a
scanning electron microscope (SEM) using a standard reference sample. Its crystal structure was
determined by powder X-ray diffraction (XRD, Rugaku). These results were analyzed by Rietveld
refinement, using the general structure analysis software package GSAS-EXPGUI.8 The initial
structure for the refinement was the structure of Nax CoO2 (space group P63mmc ).9 Rietveld refinement
automatically optimizes many parameters, including the background, unit-cell parameters, half-
width, scale factor, atomic coordinates, occupancies, thermal parameters (Uiso ), and preferential
orientation.
The Seebeck coefficient was determined from the slope of the linear relationship between the
thermal electromotive force and the temperature difference (∼10–16 K) between the two sides of
each pellet.10 The temperature-dependent in-plane resistance was measured using a DC four-probe
configuration (PPMS, Quantum Design, Inc.). Both ab-plane and c-plane magnetizations were
measured using a superconductor quantum interference device (SQUID, Quantum Design, MPMS
XL-7) with the temperature varying from 5 to 300 K.
To understand the physical properties of Kx RhO2 , we investigated its electronic structure from
first principles. Calculations were performed using the full-potential linearized augmented plane-
wave method with the generalized gradient approximation, implemented in WIEN2k code.11 The
muffin-tin radii of K, Rh, and O were set to 2.50, 2.09 and 1.85 bohr, respectively. Rmt×Kmax
was set to 8.0, where Rmt is the smallest of the atomic sphere radii and Kmax is the plane-wave
cut-off wave vector. The maximum angular momentum number L = 12 was used for partial waves
inside atomic spheres. We used 1,000 k-points in the self-consistent calculation. The experimental
lattice constant of the Kx RhO2 crystal was used throughout, but the atom positions were optimized.
Spin-orbit coupling, particularly significant in the Rh ions, was included in our calculation. To
simulate the K deficiency, we used the virtual crystal approximation for the K atom. Accordingly,
the effective atomic number of K was set to 18+x in Kx RhO2 .12

III. RESULTS AND DISCUSSIONS


Planar-view and cross-sectional SEM micrographs are shown in Figure 1(a) and 1(b), respec-
tively. Figure 1(a) shows that the K0.63 RhO2 single sheet is composed of hexagonal-shaped tiles. The
macroscopic size of the grown sample is as large as 15 mm × 15 mm, as shown in the optical image
(inset in Fig. 1(a)). The layered structure is clearly seen from the cross-sectional SEM image shown
in Fig. 1(b). The growth direction of layered Kx RhO2 is parallel to the c-axis, as confirmed by the
selected-area electron diffraction (SAED) pattern (inset in Fig. 1(b)). The hexagonal symmetry along
the c-axis of K0.63 RhO2 is also evident from the SAED pattern. The stoichiometries of K, Rh, and
O are 0.63, 1.0, and 2.0, respectively, as confirmed by EDS on the SEM, using a standard reference
sample. The standard deviation in the stoichiometries are of the order of 5.0 %, averaged over thirty
EDS measurements. As-grown K0.63 RhO2 crystal is anti-hygroscopic, intact even exposed to air
over 3 months in our experiment.
042140-3 Yao et al. AIP Advances 2, 042140 (2012)

FIG. 1. (a) Planar-view SEM image of the K0.63 RhO2 single crystal. The inset shows an optical image. (b) Cross-sectional
SEM image of the K0.63 RhO2 single crystal, showing a clear layered structure. The inset shows the selected area electron
diffraction (SAED) pattern, indicating a hexagonal symmetry along the c-axis.

The crystal structure of layered K0.63 RhO2 was determined by powder XRD together with
Rietveld refinement, as shown in Fig. 2. The calculated patterns (solid lines) agree very well with
the experimental data. The unit-cell parameters a, b, and c obtained from the Rietveld refinement
are 3.0645(9) Å, 3.0645(9) Å, and 13.5284(6) Å, respectively. The refined value x in Kx RhO2 is
approximately 0.60, which is close to the measurement 0.63 within the experimental errors. within
the experimental error of the EDS measurement. Close inspection shows some deviations between
the fits and experimental results (two additional weak peaks marked * in the XRD pattern), which
might be correlated with local potassium ordering. Preliminary transmission electron microscopy
results reveal several regions of potassium ordering (typical size 10 nm × 10 nm). The single-
crystalline quality of K0.63 RhO2 is supported by the XRD results, shown in the inset of Fig. 2, which
contain only (00L) reflections. The growth direction of K0.63 RhO2 is therefore along the c-axis, in
agreement with SAED (inset of Fig. 1(b)).
Electrical properties within the ab-plane of single-crystalline K0.63 RhO2 were studied by four-
probe measurements. Figure 3(a) plots the temperature dependence of the in-plane resistivityρ and
shows K0.63 RhO2 demonstrating metallic behavior. The ratio between room temperature and residual
resistance is around 5. This implies the existence of point defects in the K0.63 RhO2 single crystal,
as expected from the non-integer chemical index 0.63. The measured residual resistivity 210 μcm
is slightly greater than the value previously reported for K0.49 RhO2 . This is maybe attributed to the
ordering of K ions in Kx RhO2 (x∼0.5).13 To better understand the carrier-scattering mechanism in
042140-4 Yao et al. AIP Advances 2, 042140 (2012)

FIG. 2. Rietveld refinement of the x-ray diffraction data from polycrystalline K0.63 RhO2 . Plus (+) marks represent the
experimental results and the solid red curve the combined refinement result. The difference between the experimental and
calculated results is shown by the blue curve. The pink tick marks indicate the expected Bragg peak positions for the
K0.63 RhO2 phase. The inset shows the X-ray diffraction intensity distribution for a K0.63 RhO2 single crystal oriented along
the c-axis.

FIG. 3. (a) Temperature dependence of electrical resistivity in the ab-plane (ρ ab ). (b) Comparison of the fit and experimental
results between 50 K and 220 K. (c) Comparison of the fit and experimental results between 220 K and 293 K.

K0.63 RhO2 , we fitted a power-law relation to the ρ-T curves,


ρ = ρ0 + AT α , (1)
where the exponent α depends on the scattering mechanism and ρ0 is the residual resistance.14 The
best fits are shown in Fig. 3(b) and 3(c). The nearly constant resistance (ρ0 ) below 50 K might be due
to extrinsic factors, such as impurity scattering or magnon scattering. A similar phenomenon was
also observed in K0.49 RhO2 .6 The T2 relationship in the resistance (Fig. 3(b)) indicates that electron-
electron interactions are the dominant scattering mechanism between 50 and 220 K. Figure 3(c)
shows that ρ obeys a T 3/2 law over the temperature range 220 K < T < 300 K, suggestive of
locally cooperative bond-length fluctuations. A similar power-law behavior was reported in several
materials such as Na0.5 CoO2 , BaCo0.9 Ni0.1 S1.85 Cl0.15 .14
042140-5 Yao et al. AIP Advances 2, 042140 (2012)

FIG. 4. Temperature dependence of magnetization (M) within the ab- and c-planes in different external magnetic fields.

FIG. 5. Temperature dependence of the K0.63 RhO2 Seebeck coefficient.

Temperature-dependent magnetization within the ab- and c-planes was also characterized, as
shown in Fig. 4. The ab-plane magnetization has a clear peak around 50 K. Considering both
the intrinsic two-dimensional triangular lattice symmetry and the sharp magnetic susceptibility
maximum, we infer an antiferromagnetic-like behavior within the ab-plane. The relationship between
magnetization and temperature within the c-plane, shown in Fig. 4, does not display well-defined
conventional magnetism. The origin of the steep increase in magnetization below 20 K is not currently
understood, but might be related to a spin-glass or to the existence of paramagnetic impurities. Similar
magnetic phenomena were reported in Nax CoO2 (x > 0.5), a Curie-Weiss metal.15, 16 With the limited
size of the single crystals (along c direction) currently available and weak magnetic, it is technical
difficult to determine to magnetic structure by neutron experiment.
The most attractive property of K0.63 RhO2 is its significant Seebeck effect, a phenomenon
observed in metals. The Seebeck coefficient near room temperature, plotted in Fig. 5, is approxi-
mately 46.3 μV/K at 300 K, smaller than the value for Nax CoO2 but still slightly greater than that
for K0.49 RhO2 . The difference in Seebeck coefficient between the rhodate and the cobaltate might
be related to the greater itinerancy of rhodium compared with cobalt. The Seebeck coefficient of
K0.63 RhO2 in the high-temperature limit can be calculated as
κB g 3+ x
S= log Rh , (2)
e gRh4+ 1 − x
042140-6 Yao et al. AIP Advances 2, 042140 (2012)

FIG. 6. Electronic structure of (a) KRhO2 and (b) K0.63 RhO2 .

where gRh3+ and gRh4+ are the degeneracies of the Rh3+ and Rh4+ respectively, and x is the Rh4+
content. The Seebeck coefficient in the high-temperature limit can thus reach 200 μV/K in theory.
Our experimental value is much smaller than the high-temperature classical limit, which might
suggest the possibility to improve the Seebeck coefficient in experiments, and there are not phase
transition point before the melting. So we believe K0.63 RhO2 is a candidate thermoelectric material
at high temperature.
Considering also the resistivity of K0.63 RhO2 at 300 K, the power factor of K0.63 RhO2 can be
estimated as 2.2 W cm−1 K−2 . This value is a slightly smaller than that of K0.49 RhO2 , owing to the
smaller resistance of K0.49 RhO2 . It is generally believed that a better crystalline quality decreases
resistance. Consequently, K0.63 RhO2 samples with a better crystalline quality are being prepared.
One important feature of Kx RhO2 that is different from Nax CoO2 is the fact that K0.63 RhO2 is
not hygroscopic in an ambient environment. This gives Kx RhO2 a potentially great advantage in
real-world thermoelectric applications.
To better understand the metallic behavior of K0.63 RhO2 , we performed a calculation from first
principles. Using the virtual-crystal approximation, we simulated the physical properties for the case
of a non-integer concentration (x = 0.63) of potassium, based on the structure of KRhO2 (where x
= 1). The energy band structures for x = 1.0 and 0.63 are shown in Fig. 6(a) and 6(b), respectively.
KRhO2 is obviously a semiconductor, with a band gap of approximately 1.1 eV between the t2g
and eg branches, due to the large crystal-field splitting. However, metallic behavior appears when x
decreases to 0.63. Similar behavior is also observed in the Nax CoO2 system. In K0.63 RhO2, there are
many t2g -like bands across the Fermi level, supporting the hypothesis that itinerant electrons mainly
come from the d-level of Rh when K is deficient.

IV. CONCLUSIONS
In conclusion, single crystals of K0.63 RhO2 were grown by the self-flux method. Their crystal
structure was determined by the Rietveld refinement of powder XRD patterns. Transport measure-
ments on K0.63 RhO2 demonstrate clear metallic behavior. The most interesting physical properties,
however, are its relatively large Seebeck coefficient (46 μV/K) at 300 K and the fact that K0.63 RhO2
is non-hygroscopic under ambient conditions, which make it a good candidate material for thermo-
electric applications.
042140-7 Yao et al. AIP Advances 2, 042140 (2012)

ACKNOWLEDGMENTS
The authors are grateful to Prof. C. F. Cai for Seeback measurements. This work is supported
by National Basic Research Program of China (No. 51002074, 51032003, 10974083), We also
acknowledge the support the High Performance Computing Center of Nanjing University.

1 G. Peleckis, T. Motohashi, M. Karppinen, and H. Yamauchi, Appl. Phys. Lett. 83, 5416 (2003).
2 Ichiro Terasaki, Soichiro Shibasaki, Shin Yoshida, and Wataru Kobayashi, Materials 3, 786 (2010).
3 Y. Krockenberger, I. Fritsch, G. Christiani, H.-U. Habermeier, Li Yu, C. Bernhard, and B. Keimer, Appl. Phys. Lett. 88,

162501 (2006).
4 N. P. Ong and R. J. Cava, Science 305, 52 (2004).
5 Yayu Wang, Nyrissa S. Rogado, R. J. Cava, and N. P. Ong, Nature 423, 425 (2003).
6 S Shibasaki, T. Nakano, I. Terasaki, K. Yubuta, and T. Kajitani, J. Phys.: Condens. Matter 22, 115603 (2010).
7 Y. Saeed, N. singh, and U. Schwingenschlögl, Advanced Functional materials 22, 2792 (2012).
8 B. H. Toby, J. Appl. Cryst. 34, 210 (2001).
9 Q. Huang, M. L. Foo, R. A. Pascal, Jr., and J. W. Lynn, Phys. Rev. B 70, 184110 (2004).
10 Y. Y. Wang, K. F. Cai, and X. Yao, ACS Appl. Mater. Interfaces 3, 1163 (2011).
11 P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, and J. Luitz, WIEN2k, An Augmented Plane Wave Local Orbitals

Program for Calculating Crystal Properties, edited by Karlheinz Schwarz (Techn. Universität Wien, Austria, 2001). ISBN
3-9501031-1-2.
12 D. Johannes, I. I. Mazin, and D. J. Singh, Phys. Rev. B 71, 214410 (2005).
13 K. Yubuta, Soichiro Shibasakib, Ichiro Terasakib, and Tsuyoshi Kajitani, Philosophical Magazine 89, 2813 (2009).
14 F. Rivadulla, J. S. Zhou, and J. B. Goodenough, Phys. Rev. B 67, 165110 (2003).
15 G. Bergmann and P. Marquardt, Phys. Rev. B 17, 1355 (1978).
16 Maw Lin Foo, Yayu Wang, and Satoshi Watauchi, Phys. Rev. Lett. 92, 247001 (2010).

You might also like