Thin film growth of boron nitride on α-Al2O3 (0 0 1) substrates by reactive sputtering

Article abstract of DOI:10.1016/j.materresbull.2011.09.006

Boron nitride thin films were grown on α-Al₂O₃ (0 0 1) substrates by reactive magnetron sputtering. Infrared attenuated total reflection (ATR) spectra of the films gave an intense signal associated with in-plane B-N stretching TO mode of short range ordered structure of BN hexagonal sheets. X-ray diffraction for the film prepared at a low working pressure (ca. 1 × 10^-3 Torr) gave a diffraction peak at slightly lower angle than that corresponding to crystal plane h-BN (0 0 2). It is notable that crystal thickness calculated from X-ray peak linewidth (45 nm) was close to film thickness (53 nm), revealing well developed sheet stacking along the direction perpendicular to the substrate surface. When the substrates of MgO (0 0 1) and Si (0 0 1) were used, the short-range ordered structure of h-BN sheet was formed but the films gave no X-ray diffraction. The film showed optical band gap of 5.9 eV, being close to that for bulk crystalline h-BN.

Full text of DOI:10.1016/j.materresbull.2011.09.006

Materials Research Bulletin 46 (2011) 2230–2234
Contents lists available at SciVerse ScienceDirect
Materials Research Bulletin
journal homepage: www.elsevier.com/locate/matresbu
Thin film growth of boron nitride on
sputtering
a-Al2O3 (0 0 1) substrates by reactive
Atsushi Anzai a, Fumitaka Nishiyama b, Shoji Yamanaka a, Kei Inumaru a,
*
a Department of Applied Chemistry, Graduate School of Engineering, Hiroshima University, Higashi-Hiroshima 739-8527, Japan
b Technical Center, Hiroshima University, Higashi-Hiroshima 739-8527, Japan
A R T I C L E I N F O
A B S T R A C T
Article history:
Boron nitride thin films were grown on
a-Al2O3 (0 0 1) substrates by reactive magnetron sputtering.
Received 10 August 2011
Infrared attenuated total reflection (ATR) spectra of the films gave an intense signal associated with in-
plane B-N stretching TO mode of short range ordered structure of BN hexagonal sheets. X-ray diffraction
for the film prepared at a low working pressure (ca. 1 Â 10À3 Torr) gave a diffraction peak at slightly
lower angle than that corresponding to crystal plane h-BN (0 0 2). It is notable that crystal thickness
calculated from X-ray peak linewidth (45 nm) was close to film thickness (53 nm), revealing well
developed sheet stacking along the direction perpendicular to the substrate surface. When the substrates
of MgO (0 0 1) and Si (0 0 1) were used, the short-range ordered structure of h-BN sheet was formed but
the films gave no X-ray diffraction. The film showed optical band gap of 5.9 eV, being close to that for
bulk crystalline h-BN.
Received in revised form 23 August 2011
Accepted 9 September 2011
Available online 16 September 2011
Keywords:
A. Nitride
B. Sputtering
C. X-ray diffraction
D. Crystal structure
ß 2011 Elsevier Ltd. All rights reserved.
1. Introduction
consists of BN layers which resemble the sheets of graphite and the
sheets are stacking to form a sequence of B-N-B-N Á Á Á along the c
An increasing interest in the chemical and physical properties of
nitrides has recently accelerated studies of the binary and ternary
nitrides [1–15]. Nitrides are categorized into several groups such as
‘‘interstitial nitrides’’, ‘‘ionic nitrides’’ and ‘‘covalent nitrides’’ [16].
Many transition metal binary nitrides are often categorized as
interstitial nitrides [16], in which it is considered that the nitrogen
atoms occupy interstitial sites between the atoms of the parent
metals. These nitrides show a wide variety of physical properties:
for example 3d transition metal nitrides have been shown to
exhibit antiferromagnetism (MnN [7,8], CrN [9–11], FeN [12]),
Pauli paramagnetism [13] (CoN), ferromagnetism [14] (Cr0.5Ti0.5N)
and superconductivity [15] (TiN, VN).
Covalent nitrides include silicon nitride, carbon nitride, and
boron nitride (BN). They are usually electronically insulating or
semiconducting. BN adopts crystal structures similar to those of
elemental carbon, because BN is isoelectronic to Carbon. Cubic
boron nitride (c-BN) adopts zinc blende structure, which corre-
sponds to diamond structure of carbon. Since c-BN is a very hard
material, many studies are devoted to synthesize c-BN coatings for
axis, as shown in Fig. 1. h-BN is attracting much attention as an
optical material. Recently synthesis of single crystal of h-BN was
reported. The crystal was demonstrated to be promising as ultra-
violet illuminator [17].
BN thin films have been prepared by various methods. Chemical
Vapor Deposition (CVD) is a useful method to synthesize highly
covalent compounds, because chemical reactions to form covalent
bonds can be controlled by selecting appropriate source volatile
materials and deposition conditions [18,19]. Physical deposition
techniques have been also reported for preparation of BN films:
They include methods such as sputtering [20], ion beam-assisted
deposition [21], and pulsed laser deposition [22]. Many research-
ers report formation of c-BN-like structure prepared by sputtering
[20–26]. Formation of h-BN-like structure has been also reported
by many authors but, there are only a few reports of the evidence
for crystalline h-BN formation (i.e., X-ray diffraction). Noreika et al.
[27] and Tanaka et al. [28] reported the formation of microcrystal-
line h-BN by reactive sputtering. X-ray diffraction of the samples in
these studies were very broad due to the microcrystallinity and the
cell parameters of h-BN were determined by electron diffraction.
Here we report formation of BN films having short range
ordered structure of h-BN by reactive magnetron sputtering under
nitrogen gas using elemental boron target. Contrary to the reports
on the formation of microcrystalline h-BN by reactive sputtering
mentioned above, in the present study, the crystal thickness
calculated from X-ray peak linewidth (45 nm) was close to film
cutting tools. c-BN is also promising as
a wide band gap
semiconductor and power electronic devices [16]. Another
polymorph of BN is a hexagonal structure, h-BN. This structure
* Corresponding author. Tel.: +81 82 424 7741; fax: +81 82 424 5494.
E-mail address: inumaru@hiroshima-u.ac.jp (K. Inumaru).
0025-5408/$ – see front matter ß 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2011.09.006

A. Anzai et al. / Materials Research Bulletin 46 (2011) 2230–2234
2231
operated in tapping mode. The values for the films prepared under
the conditions were ca. 50 nm. In order to determine the accurate
film thickness of Film A, we analyzed boron in the film by using
11B(p,
a)
8Be nuclear reaction [29]. The acceleration energy was
650 keV per proton, and the accumulation was carried out up to
50
m
C of the incident beam charge. The result was 1.45 Â 1017
B
atom cmÀ2, which corresponds to a thickness of 53 nm.
2.3. X-ray diffraction analysis of the crystalline nitride phases
The crystal structure of the nitride phase formed on the
substrates was characterized by X-ray diffraction with a D8
Advance diffractometer (Bruker AXS) using CuK
a radiation.
Another trial was carried out with an Xpert MRD diffractometer
(Panalytical) to detect diffractions form crystal planes not parallel
to the substrate surface.
Fig. 1. Crystal structure of h-BN.
thickness (53 nm), revealing well developed stacking of the BN
sheets along the direction perpendicular to the substrate surface.
IR ATR spectroscopy, X-ray diffraction, X-ray photoelectron
spectroscopy, and UV–Vis absorption spectroscopy were applied
for the characterization of the films.
2.4. Other characterizations
Infrared spectra were measured with an FT/IR-4200 spectrom-
eter (JASCO, Tokyo). This spectrometer can record transmittance
and attenuated total reflection (ATR, once-reflection, with a
diamond prism) spectra. The UV–Vis transmittance spectra were
recorded with Lambda-900 (PerkinElmer). The compositions of the
thin films were determined by X-ray photoelectron spectroscopy
(ESCA-3400, Kratos Analytical): B 1s1/2 and N 1s1/2 spectra were
measured for the crystalline h-BN bulk sample, and the data was
used to calibrate the relative sensitivity factors of the spectrometer
for the B and N signals. The compositions of the thin films were
calculated using the calibrated sensitivity factors.
2. Experimental
2.1. Materials
Elemental B powder (purity 99%) was purchased from High
Purity Chemical Co. (Tokyo) and was pressed into a disk to use as
the target of sputtering. –Al2O3 (0 0 1), MgO (0 0 1) and Si (0 0 1)
were used as substrates. The substrates were ultrasonically
washed with anhydrous methanol in prior to use. N2 gas
(99.9999%) and Ar gas (99.9999%) were obtained from Nippon
Sanso Co. Bulk h-BN (purity 99.5%) as a reference sample was
obtained from Denki Kagaku Kogyo (Tokyo).
3. Results
Table 1 lists the films prepared in this study. The table contains
compositions (atomic ratio of N/B) of the samples analyzed by XPS.
All films had N/B atomic ratio close to 0.8 or more, showing that the
elemental boron reacts with the nitrogen introduced as N2 gas
during the reactive sputtering process to give highly nitrided
compositions. As shown in Table 1, working pressure and gas
2.2. Synthesis of hexagonal boron nitride films on
substrates
a-Al2O3 and other
The vacuum system (base pressure 5 Â 10À8 Torr) was
equipped with a magnetron sputtering gun (Model A315-ua, AJA
international, MA, USA). The gun was operated in RF mode
(13.56 MHz, 50 W). A mixed gas of Ar and N2 was fed to the
sputtering gun and 50 W of RF power was applied. The gas feeding
rates, working pressure, and other conditions for samples prepared
in this study were listed in Table 1. The typical synthetic conditions
were as follows: Working pressure ꢀ1 Â 10À3 Torr, gas feeding
rate, Ar 2 cm3 minÀ1, N2 15 cm3 minÀ1, deposition time ca. 9 h. The
substrates were heated to 823 K during the deposition of all films.
The film thicknesses were measured using an atomic force
microscope (AFM, Nanoscope D-3100, Digital Instruments, USA)
feeding ratios were changed for a-Al2O3 (0 0 1) substrate. Working
pressure was not influential to the compositions (compare films A
and B). Film C had slightly higher N/B ratio compared to Film B. The
nitrogen gas feeding rate seems to affect the composition. BN films
were grown also on MgO (0 0 1) and Si (0 0 1) substrates (films D
and E). These films had N/B ratios similar to those grown on
Al2O3 (0 0 1).
a-
Infrared spectra were taken by ATR method for the boron
nitride thin films. The spectra were shown in Fig. 2. The ATR
spectrum of bulk polycrystalline h-BN as a reference gave a broad
peak centered at 1320 cmÀ1. This signal was assigned to in-plane
TO mode vibration in many reports [30–32]. Films A–D had intense
Table 1
Deposition condition and properties of thin films synthesized in this study.
W.P.a (10À3 Torr)
Gas flow rateb
Ar
td (h)
N/Bd
B.G.e (eV)
c
Film
Substrate
N2
A
B
C
D
E
a
a
a
-Al2O3 (0 0 1)
-Al2O3 (0 0 1)
-Al2O3 (0 0 1)
ꢀ1
5.2
7.4
ꢀ1
ꢀ1
2
2
5
2
2
15
15
20
15
15
9.3
8.5
3
0.83
0.79
0.88
0.83
0.85
5.9
5.9
5.9
6.0
–
MgO (0 0 1)
Si (0 0 1)
9
9
a
Working pressure.
In cm3 (STP) minÀ1
b
c
.
Deposition time.
d
e
Atomic ratio determined by XPS.
Optical band gap.

2232
A. Anzai et al. / Materials Research Bulletin 46 (2011) 2230–2234
transmission
E
D
E
C
B
A
ATR
E
bulk hBN
Si
2000 1800 1600 1400 1200 1000
wave number / cm-1
800
600
2000
1800
1600
1400
1200
1000
800
600
wave number / cm-1
Fig. 2. Infrared ATR spectra of thin films and bulk h-BN sample.
Fig. 3. Infrared spectra of BN/Si (Film E) and Si substrate.
signal at 1374 cmÀ1, which had a frequency slightly higher thanthe
TO mode of bulk h-BN but coincided with the peak positions of BN
thin films reported in the literature [18,21], and the signal was also
assigned to the in-plane TO mode. The TO mode signal demonstrates
that BN in the films have short-range order of h-BN structure, i.e.,
hexagonal BN sheets shown in Fig. 1. Besides the TO mode
absorption, all films prepared in this study gave a signal at a higher
frequency around 1540 cmÀ1. Fig. 3 compares transmission and ATR
spectra of the BN film grown on a Si substrate (Film E). In the ATR
spectrum (the bottom panel) an intense structure appeared at
1537 cmÀ1 besides signals due to silicon substrates. On the other
hand, in the transmission spectra the structure was not observed.
This indicates that the structure in the ATR spectrum is assignable to
LO mode, which is detectable at 1510 cmÀ1 in reflection spectra of
bulk h-BN [18,30] but is inactive in transmission absorption spectra.
It was reported that ATR spectrum is sensible to LO mode signals
[33]. Thus, we assigned the signal around 1540 cmÀ1 appeared in
ATR spectra to LO mode of h-BN.
In order to investigate crystallinity of the films, X-ray diffraction
measurements were carried out. Fig. 5 focuses on h-BN 002 region
of X-ray diffraction patterns of the films and bulk h-BN. The bulk h-
BN gave an intense 0 0 2 peak. In the films prepared in this study,
Film A showed a diffraction peak and its position was 2
(d = 0.337 nm), of which d value is close to but slightly larger than
that of bulk h-BN (2 = 26.788, d = 0.333 nm). This indicates that
u = 26.648
u
crystalline BN is formed in Film A. The size of the crystal was
estimated to be 48 nm from the FWHM of the diffraction peak. That
is close to the film thickness (53 nm). This point will be discussed
later. We could not detect other diffraction peaks from crystal
planes not parallel to the substrate surface. No diffraction peak of
crystalline BN was observed for the other films. It should be noted
here that Film A was prepared at the lowest working pressure
among the films prepared in this study.
Fig. 6 shows Tauc plots of Film A. The ordinate of the left panel is
0.5, and that of the right panel expresses ( )2. The linearity
(ahn) ahn
of the plots was higher for the right panel than for the left panel,
and this suggest that the BN film had a direct band gap rather than
an indirect one. The band gap was estimated to be 5.9 eV from the
plots in the right panel. The energy gaps determined from Tauc
plots were also listed in Table 1. Single crystal of h-BN has been
reported to have direct band gap of 5.8 eV [34]. The optical band
gap of the films prepared in this study was close to the value
reported for single crystal h-BN.
Fig. 4 shows XPS spectra of thin films prepared in this study and
bulk h-BN. Table 2 summarizes peak positions and peak widths of
the XPS spectra. The peak positions of B 1s1/2 and N 1s1/2 for the
thin films were identical to that of bulk h-BN, and obvious
oxidation of boron cannot be measured by XPS. Peak widths of the
signals are comparable or slightly wider, probably due to lower
crystallinity of h-BN in the film compared to that of bulk h-BN.
N 1s
B 1s
E
D
E
D
C
B
C
B
A
A
bulk hBN
bulk hBN
200
195
190
185
180 410
405
400
395
390
Binding energy / eV
Fig. 4. X-ray photoelectron spectra for thin films and bulk h-BN sample.

A. Anzai et al. / Materials Research Bulletin 46 (2011) 2230–2234
2233
Table 2
E
Data derived from XPS spectra.
Film or sample
Binding energy (eV)
FWHM (eV)
B 1s1/2
D
C
B 1s1/2
N 1s1/2
N 1s1/2
A
190.4
190.5
190.3
190.7
190.2
190.5
193.0
398.2
398.3
398.0
398.4
397.9
398.1
2.4
2.6
2.4
2.4
2.4
2.1
2.0
2.1
2.0
2.0
2.0
2.0
B
B
C
D
E
Bulk h-BN
A
a
B2O3
a
bulk hBN x 0.03
Ref. [36].
25
26
27
28
2Theta / degree
deposited at the lowest working pressure among the films. A
possibility is that deposition at a high working pressure decreases
the film crystallinity by the impacts of argon ions onto the surface
of the films during the deposition. On the other hand, TO mode
absorption observed in IR spectra revealed that all films in this
study had short range ordered structure of h-BN. Being consistent
with this, the films had an optical band gaps close to that of bulk h-
BN. It was easily demonstrated by theoretical calculations that the
energy gaps are governed by the structure in the h-BN sheets and
are little influenced by the stacking way of the sheets. By using a
first principle band calculation program (DMol3 solid, Accelrys) we
calculated and compared energy gaps of h-BN and of a hypothetical
structure in which inter-layer distance of h-BN was expanded by a
factor of 10. The energy gaps of the two structures were very close
(not shown). These are consistent with the fact that similar optical
band gap was observed for all films, which include crystallized and
non-crystallized ones, since all of them had short range ordered
structure of h-BN indicated by the in-plane TO signals.
Fig. 5. X-ray diffraction patterns of thin films and bulk h-BN sample.
4. Discussion
X-ray diffraction revealed that BN in Film A is crystalline and a
peak corresponding to 0 0 2 diffraction of h-BN was observed. The
crystal thickness from X-ray diffraction peak width was close to
the film thickness. This indicates the stacking of the BN sheets
continues through almost the whole thickness of the film. In the
crystal structure of h-BN, the layers stacks so that the atoms align
linearly perpendicular to the BN sheets (i.e., parallel to the c axis)
like BÁ Á ÁNÁ Á ÁBÁ Á ÁN and the layers interact with adjacent ones
through the BÁ Á ÁN bonds (Fig. 1). The inter-plane distance for Film
A was slightly larger than that of the (0 0 2) planes of bulk h-BN.
Disorders in the stacking of the BN sheets like turbostratic
structure may weaken the interaction between them and bring
about the slight increase in the inter-layer distance. Formation of
turbostratic BN has been reported at the interfaces between
substrate surface and cubic BN. In this case, the layer of
turbostratic BN grew perpendicular to the substrate surface
[35]. In the present study, it is notable that the BN sheets were
parallel to the substrate surface and stacked in a long distance
along the direction perpendicular to the substrate surface. A
5. Conclusions
It was demonstrated that reactive RF magnetron sputtering
using an elemental boron target could grow BN films on
a-Al2O3
(0 0 1). X-ray diffraction, IR, and XPS of the film showed the
stacking structure of the h-BN sheets in a long distance, the short
range ordered structure of the h-BN sheets, and the nitrogen-well
incorporated composition, respectively. The control of the working
pressure was crucial to obtain crystallized films. The results of the
present study highlight that reactive sputtering is versatile also for
possibility is that the hexagonal symmetry of
a-Al2O3 (0 0 1)
substrate surface may enhanced the growth of the first BN sheet
along the substrate surface. The result that only Film A crystallized
may be associated with the working pressure: Film A was
ν
ν
α
α
3
4
5
6
7
3
4
5
6
7
Photon energy / eV
Fig. 6. Tauc plot of BN/a-Al2O3 (0 0 1) (Film A).

2234
A. Anzai et al. / Materials Research Bulletin 46 (2011) 2230–2234
[14] K. Inumaru, K. Koyama, Y. Miyaki, K. Tanaka, S. Yamanaka, Appl. Phys. Lett. 91
(2007) 152501-1–152501-3.
[15] H.F. George, H.K. John, Phys. Rev. 93 (1954) 1004.
forming crystalline films of covalent nitride such as hexagonal
boron nitrides.
[16] O.N. Pierson, Handbook of Refractory Carbides and Nitrides. Properties, Char-
acteristics, Processing and Applications, Noyes Publications, New Jersey,
1996 .
Acknowledgments
[17] T. Taguchi, K. Watanabe, J. Cryst. Growth 303 (2007) 525–529.
[18] M.C. Polo, M. Ben el Mekki, J.L. Andu´ jar, N. Mestres, J. Pscual, Diamond Relat.
Mater. 6 (1997) 1550–1554.
[19] M. Ben el Mekki, N. Mestres, J. Pascual, M.C. Polo, J.L. Andu´ jar, Diamond Relat.
Mater. 8 (1999) 398–401.
[20] K. Kojima, K. Nose, M. Kambara, T. Yoshida, J. Phys. D: Appl. Phys. 42 (2009)
055304-1–055304-5.
[21] M. Ben el Mekki, M.A. Djouadi, E. Guiot, V. Mortet, J. Pascallon, V. Stambouli,
D. Bouchier, N. Mestres, G. Nouet, Surface Coat. Technol. 116–119 (1999)
93–99.
This work was partially supported by a Grant-in-Aid on Priority
Areas (Nos. 436 and 472) from the Japan Ministry of Education,
Culture, Sports, Science and Technology (MEXT), and by a Grant-in-
Aid for Scientific Research (S) (No. 19105006) and (B) (No.
20350095), and a Funding Program for World-Leading Innovative
R&D on Science and Technology (FIRST) Program from the Japan
Society for the Promotion of Science (JSPS).
[22] B. Mitu, P. Bilkova, V. Marotta, S. Orlando, A. Santagata, Appl. Surf. Sci. 247 (2005)
123–127.
References
[23] D. Schild, S. Ulrich, J. Ye, M. Stu¨ ber, Solid State Sci. 12 (2010) 1903–1906.
[24] S. Ulrich, J. Ye, M. Stu¨ ber, Surf. Coat. Technol. 205 (2010) S96–S98.
[25] T. Ohori, H. Asami, J. Shirahata, T. Suzuki, T. Nakayama, H. Suematsu, K. Niihara, J.
Ceram. Soc. Jpn. 118 (2009) 164–166.
[26] T. Ohori, H. Asami, J. Shirahata, T. Suzuki, T. Nakayama, H. Suematsu, K. Niihara, J.
Ceram. Process. 10 (2009) S14–S18.
[27] A.J. Noreika, M.H. Francombe, J. Vac. Sci. Technol. 6 (1969) 722–726.
[28] K. Tanaka, Y. Uemura, M. Iwata, Oyobutsuri 46 (1977) 120–127.
[29] J.R. Bird, J.S. Williams (Eds.), Ion Beam for Materials Analysis, Academic Press,
New South Wales, Australia, 1989, pp. 149–207.
[30] R. Geick, H. Perry, G. Pupprecht, Phys. Rev. 146 (1966) 543–547.
[31] E. Borowiak-Palen, T. Pichler, G.G. Fuentes, B. Bendjemil, X. Liu, A. Graff, G. Behr,
R.J. Kalenczuk, M. Knupfer, J. Fink, Chem. Commun. (2003) 82–83.
[32] D.M. Hoffman, G.L. Doll, P.C. Eklund, Phys. Rev. B 30 (1984) 6051–6056.
[33] N. Kuroda, Y. Iida, T. Shigeta, Hasanudin, J. Watanabe, Jpn. J. Appl. Phys. 42 (2003)
L1241–L1243.
[1] A. Simon, Angew. Chem. Int. Ed. Eng. 36 (1997) 1789.
[2] R. Niewa, F.J. DiSalvo, Chem. Mater. 10 (1998) 733–2752.
[3] D.H. Gregory, Coord. Chem. Rev. 215 (2001) 301–345.
[4] E. Horvath-Bordon, R. Riedel, R. Zerr, P.F. McMillan, G. Auffermann, Y. Prots, W.
Bronger, R. Kniep, P. Kroll, Chem. Soc. Rev. 35 (2006) 987–1014.
[5] R. Juza, Adv. Inorg. Chem. Radiochem. 9 (1996) 81–131.
[6] J.L. Hunting, M.M. Szymanski, P.E. Johnson, C.B. Kellar, F.J. DiSalvo, J. Solid State
Chem. 180 (2007) 31–40.
[7] K. Suzuki, T. Suzuki, Y. Fujiwara, T. Kaneko, H. Yoshida, Y. Obi, S. Tomiyoshi, J.
Alloys Compd. 360 (2003) 34.
[8] H.Q. Yang, H. Al-Brithen, A.R. Smith, J.A. Borchers, R.L. Cappelletti, M.D. Vaudin,
Appl. Phys. Lett. 78 (2001) 3860.
[9] L.M. Corliss, N. Elliot, J.M. Hastings, Phys. Rev. 117 (1960) 929.
[10] X.Y. Zhang, J.S. Chawla, B.M. Hawe, D. Gall, Phys. Rev. B 83 (2011), 165205-1–
165205-10.
[34] K. Watanabe, T. Taguchi, H. Kanda, Nat. Mater. 3 (2004) 404–409.
[35] M.P. Johansson, L. Hultman, S. Daaud, K. Bewilogua, H. Luthje, A. Schutze, S.
Kouptsidis, G.S.A.M. Theunissen, Thin Solid Films 287 (1996) 193–201.
[36] D. Briggs, M.P. Seah (Eds.), 2nd ed., Practical Surface Analysis, vol. 1, Wiley, 1990,
p. 599.
[11] K. Inumaru, K. Koyama, N. Imo-oka, S. Yamanaka, Phys. Rev. B 75 (2007), 054416-
1–054416-5.
[12] T. Hinomura, S. Nasu, Physica B 237–238 (1997) 557.
[13] K. Suzuki, T. Kaneko, H. Yoshida, H. Morita, F. Fujimori, J. Alloys Compd. 224
(1995) 232.