Full text of DOI:10.1007/s003390000605

Appl. Phys. A 71, 469–472 (2000) / Digital Object Identifier (DOI) 10.1007/s003390000605
Applied Physics A
Materials
Science & Processing
Rapid communication
The effects of dc bias voltage on the crystal size and crystal quality
of cBN films
W.J. Zhang, S. Matsumoto
National Institute for Research in Inorganic Materials (NIRIM), 1-1 Namiki, Tsukuba City, 305-0044, Japan
(Fax: +81-298/52-7449, E-mail: wjzhang@nirim.go.jp, matsumot@nirim.go.jp)
Received: 13 June 2000/Accepted: 21 June 2000/Published online: 23 August 2000 – Springer-Verlag 2000
Abstract. For the deposition of cubic boron nitride thin films
in Ar−N2−BF3−H2 system by dc jet plasma chemical vapor
deposition, the role of dc substrate bias ranging from −70 V
to −150 V was investigated. A critical bias voltage was ob-
served for the formation of cBN phase. The cBN content in
the film increased with bias voltage and reached a maximum
at the bias voltage of −85 V. Increasing the bias voltage fur-
ther caused a decrease in cBN content and peeling of the films
from the substrate. By combining the results of infrared spec-
troscopy, Raman spectroscopy and X-ray diffraction, the bias
voltage was also found to strongly affect the crystal size, crys-
tal quality and residual stress of the deposited films. A bias
voltage a little higher than the critical value was demonstrated
to be favorable for the deposition of a high-quality cBN film
with large crystal size and low residual stress.
sputtering, which suggested an improvement in film crystal
quality as well as a decrease of stress with decreasing ion en-
ergy. A “thick” film (∼ 2500 Å) with high cBN content was
obtained by this method. It is, however, difficult to evaluate
the crystal quality and stress of the thick cBN films (near or
thicker than 1 µm) by using only the IR technique. There-
fore a combined study by means of IR spectroscopy, Raman
spectroscopy and X-ray diffraction (XRD) were performed in
this work to investigate the dependence of crystal size, crystal
quality and stress of the cBN films on the negative substrate
biases applied during the depositions.
The cBN films studied in this paper were deposited by dc
jet plasma CVD which is well known as a successful method
for the high-rate deposition of CVD diamond. In Ar, N2 and
BCl3 gas systems, Bern et al. reported the deposition of cBN
films in arc jet flows in low pressure (0.1−1 Torr) [18, 19].
In this report, the deposition of cBN films was performed
in an Ar−N2−BF3−H2 gas mixture where argon was used
as a plasma gas, and nitrogen, boron fluoride and hydrogen
were used as reactant gases. The experimental setup has been
shown elsewhere [20]. (001) Silicon wafers of size 14×14×
0.5 mm3 were used as substrates. Silicon substrate were pre-
scratched with 5–10 µm diamond powders and then chemi-
cally cleaned with acetone, deionized water and ethanol, in
succession, in an ultrasonic bath before deposition. The sub-
strate was put on a water-cooled Mo/Cu substrate holder
that can be dc biased. During deposition, an hBN cover was
used to expose only the silicon substrate to the plasma and
thus to concentrate the ion-bombardment on the substrates
and to avoid the contamination of the samples from the sub-
strate holder. The details of the experimental conditions are
listed in Table 1. The substrate temperature was measured
from the backside with a sapphire rod sensor and a glass fiber
transmitting system (Optonico OT-100). After deposition, the
phase composition, structure and crystal quality of the films
were investigated by Fourier-transform infrared spectroscopy,
glancing-angle XRD, scanning electron microscopy (SEM)
and macro-Raman spectroscopy.
PACS: 61.10.Eq; 78.30.-j; 81.15.Gh
Due to its unique properties, i.e., its hardness being second
only to diamond, high thermal conductivity, chemical inert-
ness, high refractive index and wide band gap in connection
with p- and n-type dopability, cubic boron nitride (cBN) is an
extremely promising material for applications in optical, elec-
tronic and mechanical techniques. During recent years, much
effort has been put into improving the preparation of cBN
films; different kinds of deposition methods including both
physical vapor deposition (PVD) [1–7] and chemical vapor
deposition [8–12] were developed. For all these methods, the
production of energetic ions and the ion-bombardment of the
substrate was found to play an important role in the formation
of the cubic phase of boron nitride. By involving low-energy
ion bombardment during film growth, the deposited cBN
films generally exhibited high compressive stress, which was
believed necessary in order to obtain the cubic phase [13–15].
This stress results in a restricted maximum film thickness
(several hundred nanometers) while leading to the destruction
of the cBN samples when the film was too thick. Litvinov
et al. [16, 17] observed narrowing and downshifting of the
characteristic infrared (IR) absorption of the cBN phase by re-
ducing the bias voltage in a two-step process of ion-assisted
Figure 1 shows the IR spectra of the samples deposited
under different substrate biases ranging from −70 V to
−150 V. For the substrate bias voltage of −70 V (curve (a)),

470
Table 1. Experimental parameters for the deposition of cBN films by dc jet
plasma CVD
Figure 2 shows glancing-angle XRD patterns of the sam-
ple deposited under substrate biases ranging from −80 to
−150 V. The incident X-ray angle α was set at 0.05◦ and
the diffraction pattern was taken by 2θ continuous mode with
a step width of 0.02◦. For the sample deposited under −80 V,
diffraction appeared for both cBN and hBN. When the the
voltage was increased to −85 V, the diffraction for the cBN
phase intensively increased. In agreement with the IR obser-
vations, the cBN diffraction became weak and broad at higher
bias voltage of −100 and −150 V. The crystal size can be ob-
tained from the diffraction patterns by using the well-known
Sherrer’s equation:
Ar (slm)
20
N2 (slm)
1.5
BF3 (sccm)
3
H2 (sccm)
5
980
50
Substrate tempreture (◦C)
Pressure (Torr)
Bias voltage (V)
Duration (min)
−70−−150
10
Kλ
Lhkl = ꢀ
(1)
(B2 −b2) cos θ
where K = 1, λ = 1.5418 Å, B and b are the half-maximum
width of the diffraction peak and instrument, respectively.
Furthermore, the strain εψ was calculated from lattice dis-
tortion (d0 −d)/d0, as listed in Table 2. Assuming a biaxial
stress and neglecting shear components in the film, the stress
can be roughly estimated by using Hooke’s law:
σ = εψ E/(1−γ),
(2)
where E = 833 GPa and γ = 0.112 [15]. To ensure the pre-
cision, we used the strongest 111 peak for the calculation of
crystal size, and both 111 and the second strongest peak 220
for the calculation of stress. The calculated crystal sizes and
stresses of the samples deposited under different substrate bi-
ases are summarized in Table 3. The tendency of the increase
of stress and the decrease of crystal size with increasing bias
voltage can be clearly seen. The small crystal size and high
residual stress at higher bias voltages may be due to the exces-
sive damage caused by ion bombardment. It should be noticed
that the compressive stress of the sample deposited under
−85 V is only 1.4−2.1 GPa, which is remarkably lower than
that reported before. The effective decrease of stress allowed
us to increase the film thickness. Figure 3 shows the SEM
cross-sectional morphology of a cBN film deposited for 3 h
under the substrate bias of −85 V. A columnar growth of
Fig. 1. IR spectra of the films deposited under the substrate bias voltage of
(a) −70, (b) −80, (c) −85, (d) −100 and (e) −150 V, respectively
only two intensive absorption bands at about 1380 and
780 cm−1 were observed, which can be assigned to in-plane
and out-of-plane TO modes of hBN, respectively. Increasing
the substrate bias voltage to −80 V (curve (b)), an absorption
band at about 1080 cm−1 appeared, which can be assigned to
the presence of cubic phase of boron nitride. The deviation of
this absorption to the higher wave number from that of HPHT
single-crystal cBN (∼ 1065 cm−1) is believed to be due to
the compressive stress caused by ion bombardment. The cBN
absorption dramatically increased with a small increase of
the bias voltage and reached a maximum at Vb = −85 V.
The corresponding IR spectrum (curve (c)) shows a top-cut
absorption band from about 1000 to 1200 cm−1. The film
thickness was observed to be 3.4 µm with SEM. In this case,
it is not feasible to evaluate accurately the cBN content in
the film by using the normal definition (I1065/(I1065 + I1370),
where I1065 and I1370 are IR absorption-integrated intensities
of cBN and hBN TO modes at 1065 and 1370 cm−1, respec-
tively). Meanwhile, it is also difficult to evaluate the residual
stress and crystal quality from the peak position and linewidth
of such top-cut absorption band. Increasing the bias voltage
further leads to the peeling of the films from the substrates
just after the deposition. Meanwhile, the cBN content de-
creased and the cBN characteristic absorption became broad
from the IR investigations (curve (d) and (e) for bias voltage
of −100 and −150 V, respectively).
Fig. 2. XRD patterns of the films deposited under the substrate bias of
(a) −80, (b) −85, (c) −100 and (d) −150 V, respectively

471
Table 2. The measured lattice spacing (d) of each cBN reflection of the sample deposited under different substrate bias compared with JCPDS-35-1365. The
strain (εψ ) calculated from the lattice distortion for each reflection is also presented
JCPDS 35-1365
Sample A (Vb = −80 V)
Sample B (Vb = −85 V) Sample C (Vb = −100 V) Sample D (Vb = −150 V)
εψ εψ εψ
hkl
d0 (Å)
I/I0
d (Å)
I/I0
d (Å)
d (Å)
d (Å)
111
200
220
311
2.0872
1.8081
1.2786
1.0900
100
5
24
8
2.0924
1.8111
1.2790
1.0911
−0.0025
−0.0017
−0.0003
−0.0010
2.0920
1.8146
1.2805
1.0914
−0.0023
−0.0036
0.0015
−0.0012
2.0978
1.8149
1.2832
1.0929
−0.0051
−0.0038
−0.0036
−0.00265
2.0983
1.8153
1.2862
1.0893
−0.0053
−0.0040
−0.0060
0.0007
Table 3. Calculated stresses and crystal sizes for the samples deposited
under different bias voltage. The crystal size was obtained from the
strongest reflection 111, and the stress was obtained from 111 and the
second strongest reflection 220
Stress
(GPa)
Crystal size L111
(Å)
Sample A (Vb = −80 V)
Sample B (Vb = −85 V)
Sample C (Vb = −100 V)
Sample D (Vb = −150 V)
0.3∼2.3
1.4∼2.1
3.4∼5.0
5.0∼5.6
129
158
94
71
Fig. 4. Raman spectra of the films deposited under the substrate bias of
(a) −80, (b) −85, (c) −100 and (d) −150 V, respectively
about 4 and 2 cm−1 with respect to that of single-crystal
cBN by HPHT (1056 and 1304 cm−1) [21]. The shift of the
TO and LO modes to the lower wave number has been re-
ported to be due to the small grain size [21] The full-width
at half maximum (FWHW) of TO and LO phonon modes
is measured to be about 28.1 and 25.4 cm−1, respectively.
At a bias voltage of −100 V, the intensity of the hBN peak
increased, and both cBN TO and LO modes became broad
and moved more towards the lower wave number. Finally,
when the bias voltage increased to −150 V, only weak and
broad LO mode could be observed. It is known that, as
the crystal quality is reduced or defects increase, the selec-
tion rules that allow only the near-zone-center phonons to
be observed break down, leading to broadening and shift-
ing of the peaks. Therefore, it can be concluded that high-
quality cBN films were deposited under a bias voltage of
−85 V. As the bias voltage increased, the crystal size de-
creased and the defects in the films increased, demonstrat-
ing a decrease in the film quality. The Raman results show
a good agreement with those obtained from IR and XRD
measurements.
Fig. 3. SEM cross-sectional morphology of a film deposited for 3 h under
the substrate bias of −85 V. XRD, IR and Raman measurements were used
to identify it as a cBN film with a high cubic content
the film can be clearly seen. The film thickness was meas-
ured to be about 25 µm. The film shows good stability and
no peeling was observed even after deposition for over nine
months.
In this work, the effects of substrate bias voltage on the
film quality were also investigated by macro-Raman spec-
troscopy. The laser spot size is about 200 µm so that an
overall contribution of phase composition and crystal qual-
ity could be obtained. Figure 4 shows Raman spectra of
the samples deposited under different bias voltage. For the
bias voltage of −80 V, a silicon vibration band from 900
to 1000 cm−1 and a peak at about 1367 cm−1 were ob-
served. The latter is due to the hBN phonon mode (2E2g),
as marked in the figure. With an increase in the bias volt-
age to −85 V, two clear peaks located at about 1051.9 and
1302.4 cm−1 appeared, which can be assigned to scattering
by the TO and LO phonon modes of cBN. The TO and LO
phonon modes of cBN shifted to a lower wave number by
In conclusion, by combining XRD, IR and Raman spec-
troscopy, the effects of substrate bias voltage on the quality
of the deposited cBN films, i.e., phase composition, residual

472
stress, crystal size and quality, were systematically studied. It
was found that the quality of the boron nitride films depended
strongly on the bias voltage. A critical value for the formation
of cubic phase was observed, which was found to be about
−80 V in our experimental conditions. A bias voltage a little
higher than this critical value is favorable for the deposition
of cBN films with high cubic content, large crystal size and
low stress. At this optimized bias voltage, the deposition of
cBN films with a thickness over 25 µm was demonstrated to
be possible. Too high bias voltage led to the peeling of the
films from the substrates and had negative effects on the film
quality.
6. T.A. Friedmann, P.R. Mirkarimi, D.L. Meddlin, K.F. McCarty,
E.J. Klaus, D.R. Boehme, H.A. Johnsen, M.J. Mills, D.K. Ottesen,
J.C. Barbour: J. Appl. Phys. 76, 3088 (1994)
7. M. Sueda, T. Kobayashi, T. Rokkaku, M. Ogawa, T. Watanabe, S. Mo-
rimoto: J. Vac. Sci. Technol., A 16, 3287 (1998)
8. M. Okamoto, Y. Utsumi, Y. Osaka: Plasma Sources Sci. Technol. 2, 1
(1993)
9. A. Weber, U. Bringmann, R. Nikulski, C.P. Klages: Surf. Coat. Tech-
nol. 60, 493 (1993)
10. W. Dworschak, K. Jung, H. Ehrhardt: Diamond Relat. Mater. 3, 337
(1994)
11. T. Ichiki, T. Yoshida: Appl. Phys. Lett. 64, 851 (1994)
12. M. Kuhr, S. Reinke, W. Kulisch: Surf. Sci. Technol. 74–75, 806 (1995)
13. D.R. McKenzie, W.D. McFall, W.G. Sainty, C.A. Davis, R.E. Collins:
Diamond Relat. Mater. 2, 970 (1993)
14. D.R. McKenzie: J. Vac. Sci. Technol. B 11, 1928 (1993)
15. G.F. Cardinale, D.G. Howitt, K.F. McKarty, D.L. Medlin, P.B. Mirka-
rimi, N.R. Moody: Diamond Relat. Mater. 5, 1295 (1996)
16. D. Litvinov, R. Clarke: Appl. Phys. Lett. 71, 1969 (1997)
17. D. Litvinov, C.A. Taylor II, R. Clarke: Diamond Relat. Mater. 7, 360
(1998)
References
1. K. Inagawa, K. Watanabe, H. Ohsone, K. Saitoh, A. Itoh: J. Vac. Sci.
Technol., A 5, 2696 (1987)
2. M. Mieno, T. Yoshida: Jpn. J. Appl. Phys. 29, L1175 (1990)
3. S. Kinder, C.A. Taylor II, R. Clarke: Appl. Phys. Lett. 64, 1859 (1994)
4. D.J. Kester, R. Messier: J. Appl. Phys. 72, 504 (1992)
5. H. Hofsäss, C. Ronning, U. Griesmeier, M. Gross, S. Reinke, M. Kuhr:
Appl. Phys. Lett. 67, 46 (1995)
18. D.H. Berns, M.A. Cappelli: Appl. Phys. Lett. 68, 2711 (1996)
19. D.H. Berns, M.A. Cappelli: J. Mater. Res. 12, 2014 (1996)
20. S. Matsumoto, W.J. Zhang: Jpn. J. Appl. Phys. 39, L442 (2000)
21. T. Werninghause, J. Hahn, F. Richter, D.R.T. Zahn: Appl. Phys. Lett.
70, 958 (1997)