Surface and Coatings Technology 169 – 170 (2003) 281–286
Synthesis and characterization of BON thin films using low frequency RF
plasma enhanced MOCVD: effect of deposition parameters on film
hardness
G.C. Chena, M.C. Kima,b, J.G. Hanb, S.-B. Leea, J.-H. Booa,*
a
Department of Chemistry and Institute of Basic Science, Sungkyunkwan University, Suwon 440-746, South Korea
b
Center for Advanced Plasma Surface Technology, Sungkyunkwan University, Suwon 440-746, South Korea
Abstract
With the expectation of getting hard material, we have firstly grown the BON thin film by radio frequency plasma enhanced
metal-organic chemical vapor deposition with 100 kHz frequency and trimethyl borate precursor. The plasma source gases used
in this study were Ar and H2, and two kinds of nitrogen source gases, N2 and NH3, were also employed. The as-grown films
were characterized with XPS, IR, SEM and Knoop microhardness tester. The film growth rate was influenced both by substrate
temperature and by nitrogen source gas. It decreased with increasing the substrate temperature, and was higher by using NH3
rather than by N2. The hardness of the film was dependent on several factors such as nitrogen source gas, substrate temperature
and film thickness due to the variation of the composition and the structure of the film. Both nitrogen and carbon-content could
raise the film hardness, on which nitrogen content had stronger effect than carbon. The smooth morphology and continuous
structure yielded high hardness. The maximum hardness of BON film was approximately 10 GPa.
䊚 2003 Elsevier Science B.V. All rights reserved.
Keywords: BON film; Plasma enhanced MOCVD; Low frequency RF; Hardness
1. Introduction
Superhard materials were widely applied in the cutting, polishing and wear protection engineering w1x. The
successful synthesis of BN and fabrication of BCN w2x
as well as BCO w3x indicated that the materials, composed of boron (B) with other light elements, such as
nitrogen (N), carbon (C) and oxygen (O), were attractive candidate of superhard material. Recently, the possibility of existence of BON material has been proposed
w4x. So far, however, there was no report on the growth
and hardness of this material in detail. Radio frequency
(RF) plasma enhanced metal-organic chemical vapor
deposition (PEMOCVD) has been successfully applied
to the fabrication of oxide w5x, nitride w6x and boroncontaining w7x materials. The frequency was usually
13.56 MHz in these cases. The high ratio of gas-phase
molecule dissociation w8x was expected by use of such
a high frequency that might cause the multi-deposit in
fabrication of multi-element compounds. To avoid this
*Corresponding author. Tel.: q82-31-290-7072; fax: q82-31-2907075.
E-mail address: jhboo@chem.skku.ac.kr (J.-H.-H. Boo).
disadvantage, a deposition process with low frequency
is highly desirable.
In this paper, therefore, low frequency (100 kHz) RF
plasma enhanced MOCVD was used to grow this new
material, BON, with trimethyl borate precursor acting
as boron and oxygen source. The effects of deposition
parameters, such as nitrogen source gas, substrate temperature, and film thickness, on the hardness of film
were also investigated.
2. Experimental
The procedure of BON growth was done in a set of
parallel plate electrode discharge deposition system. The
plasma was derived by low frequency RF with 100 kHz
and 500 W power. In the case of high frequency, 13.56
MHz, a high ratio of gas-phase molecule dissociation
was expected and resulted in the multi-deposit fabrication of multi-element compounds. To avoid these disadvantages, we used low frequency RF rather than that
of high frequency. The plasma source gases were Ary
H2, in which the flux was 200 and 20 sccm, respectively.
N2 or NH3 was also used as nitrogen source gas with
0257-8972/03/$ - see front matter 䊚 2003 Elsevier Science B.V. All rights reserved.
PII: S 0 2 5 7 - 8 9 7 2 Ž 0 3 . 0 0 0 5 6 - 2
282
G.C. Chen et al. / Surface and Coatings Technology 169 – 170 (2003) 281–286
Fig. 1. OES results of the plasma obtained in situ by different nitrogen
source gases: (a) N2 and (b) NH3.
20 sccm flux. The precursor applied as boron and
oxygen source was trimethylborate, B(OCH3)3. It was
introduced into the reactive zone of deposition chamber
by a gas distribution ring with pine-holes. The flux was
2.2 sccm without the carrier gas. The working pressure
was 266 Pa in the deposition chamber.
The substrates employed in this study were one-side
polished single crystalline silicon wafer with (0 0 1)
crystal surface. Before they were placed on the substrate
holder in the deposition chamber, they underwent
degreasing and drying in the vacuum. The substrate
temperature measured by a thermocouple and controlled
by a commercial temperature controlling system was
changed in the range from 280 to 500 8C at different
deposition procedures. The deposition time was from 2
to 5 h. The as-grown films were characterized by XPS
(model: ESCALAB MKII, Al Ka, 15 kV, 20 mA),
SEM (model: JSM-840A), and FT-IR (model: Nicolet
205).
3. Results and discussion
The plasma condition for BON growth was investigated by in situ OES measurement. Fig. 1 is the result
of the plasma gained by introducing different nitrogen
source gases, N2 and NH3, respectively. Both Fig. 1a
and b show that the main plasma optical emission occurs
in the range of 300–500 and 600–800 nm. These two
ranges are related to N radical (Nq
2 and N) and Ar
radical. So, the radicals in both plasmas are almost the
same. The only difference is that the intensities of
nitrogen related species in the NH3 plasma is much
higher than those in the N2 plasma.
Fig. 2 shows the high resolution XPS results obtained
from a BON film with 4 mm thickness (measured by
a-step surface profiler) that was grown by using N2 at
500 8C. B, C, N, O are determined as the main species
in the film. The binding energies of each element are
191.5 eV (B1s), 531.2 eV (O1s), 399.4 eV (N1s) and
284.4 eV (C1s). Among them, the binding energies of
B, O, N are similar to the reported ones in Ref. w4x
where B(ON) was affirmed. Moreover, the X-ray photoelectron survey spectra (not present here) showed that
the carbon-content decreased from 23 to 1% after the
Ar ion sputtering for 3 min. This means that neither
precursor itself nor residual gas in the reaction chamber
attribute to the carbon-content of the film surface at this
substrate temperature. The reason for arising the carboncontent on the surface is mainly due to the air contamination after deposition. So, the main composition of
the film should be BON. The composition of the asgrown BON film at 500 8C with the N2 plasma measured
by XPS is B1.0O1.4N0.9.
The effects of substrate temperature and nitrogen
source gas on the growth rate and hardness of the film
were studied, and the results were summarized in Fig.
3. The substrate temperatures were changed from 280
to 500 8C under the N2 and NH3 plasmas. When
measuring the growth rate, we fixed the growth time to
be 2 h in each case. From Fig. 3, one can see that the
films grow fast in the NH3 plasma rather than N2
plasma, especially at low substrate temperature. In both
N2 and NH3 plasma, the film growth rates decrease with
increasing the substrate temperature. The phenomenon
of growth rate decrease with the increase of substrate
temperature has been suggested due to the effect of
precursor mediation andyor deposition reaction on the
growth surface in the deposition of CNx by N2 plasma
sputtering w9x. Also, it was specified by the Refs. w10,11x
that the deposition rate decreases with increasing substrate temperature because of an increased desorption
process such as organic C–C fragments which will not
contribute any more to the film growth. In CVD procedure, the film growth is controlled, at least, by three
mechanisms: surface reaction kinetic, mass transfer and
gas-phase reaction. The feature in surface reaction kinetic controlling growth is that the growth rate increases
with increasing the substrate temperature. So, this mechanism is not suitable for our case as well as the reported
cases w9–11x. It is well known that the boundary layer
G.C. Chen et al. / Surface and Coatings Technology 169 – 170 (2003) 281–286
between gas-phase and substrate surface becomes thick
with increasing the substrate temperature. The gas-phase
particles need to diffuse long distance to arrive at the
growth surface on the substrate. The growth rate would
decrease with increasing the substrate temperature. BON
material and the materials reported in Refs. w9–11x are
formed under plasma environment. There is the possibility that they are formed directly from the gas-phase
reaction. The formation of the film is the cause of these
gas-phase particles arriving at the growth surface. It is
consequently influenced by the boundary thickness. It
can be understood that the growth rate decreases with
increasing the substrate temperature in both N2 and
NH3 plasma as well as in reported cases. However, our
results further find that the decreasing tendency of the
growth rate is dependent on the nitrogen source gas. It
decreased linearly in N2 plasma. Otherwise, the decreasing tendency in NH3 plasma is close to that in N2 over
350 8C, but it is far from that in N2 plasma below 350
8C. As known, the ionized energy of NH3 is approximately 4 eV, and N2 is approximately 9.4 eV. At the
same RF condition, NH3 is easier to be activated than
N2. The results in Fig. 1 also show that the plasma
emission is stronger in NH3 plasma than that in N2
plasma. Thus, the effective radicals are more in NH3
plasma than those in N2 plasma. The growth rate is
higher, therefore, in NH3 plasma. The different decreasing tendency between N2 plasma and NH3 plasma below
350 8C indicates that the gas-phase reaction overwhelm
283
Fig. 3. Dependence of growth rate and hardness as a function of substrate temperature.
the mass transport. Over 350 8C, mass transport is the
main factor to control the film growth. Thus, both mass
transport and gas-phase reaction influence the BON film
growth. The gas-phase reaction is relatively more important than mass transport under our PEMOCVD
condition.
Fig. 3 also presents the dependence of the hardness
on the substrate temperature. In this research, the thickness was over 3 mm for each tested film in order to
avoid the substrate effect w12x. It can be seen that under
Fig. 2. High resolution X-ray photoelectron spectra of a BON film grown on Si(1 0 0) at 500 8C with N2 plasma.
G.C. Chen et al. / Surface and Coatings Technology 169 – 170 (2003) 281–286
284
Table 1
The variation of composition in the films gained at different substrate temperatures and nitrogen source gases
Temperature
(8C)
Nitrogen source gas
NH3
N2
Content (%)
280
350
430
450
500
B
O
N
C
NqC
B
O
35.0
34.9
28.8
56.0
58.1
53.3
2.9
3.5
16.7
6.1
3.5
1.2
9.0
7.0
17.9
41.4
39.6
50.8
52.4
32.0
42.1
25.8
0.1
25.9
38.1
37.9
50.1
49.8
our experimental temperature regions below 550 8C, the
hardness increases linearly with increasing of substrate
temperature in both N2 plasma and NH3 plasma. The
highest value, approximately 10 GPa in N2 plasma and
9 GPa in NH3 plasma, occurs at 500 8C. The higher
hardness can be easily obtained by using N2 rather than
by NH3, but it is not always so. Generally speaking,
high hardness easily occurs in closed film structure,
rather than loose one w13x. The growth rate is directly
related to the film structure. The closed film structure is
expected to gain at low growth rate, rather than high
one w14x. It is reasonable that the higher hardness occurs
at lower growth rate case, since low deposition rate is
related to low amount of carbon-content fragments on
the films w10,11x. As shown in the growth rate measurement Fig. 3, the growth rate decreases with increasing
the substrate temperature in both N2 and NH3 case. So,
it can be understood that the hardness increases with
increasing the substrate temperature in both cases. It is
the same reason that the most films gained by using
N2 are harder than that by NH3. However, the higher
hardness at 350 8C occurs not in the N2 plasma, but in
the NH3 plasma. This means that the tendency of
hardness change on deposition parameters such as temperature and growth rate as well as different plasmas is
not simply explained with a simple model. Table 1 is
the XPS results of the composition changes in the
hardness-tested films studied in Fig. 3. It can be seen
that the main contents of the films are boron and oxygen,
like the ‘matrix’ of the deposit. The N- and C-contents
are then ‘bonded’ to this B–O matrix. The content of
nitrogen and carbon varies with the deposition temperature. It increases with increasing the temperature for
N-content, and decreases for C-content. In generally, the
sum of N- and C-contents in films gained by either N2
or NH3 increase with the substrate temperature, except
the data at 350 8C by N2. Comparing the composition
in the films gained by N2 with those by NH3, the sum
of N- and C-contents is higher in former cases. Meanwhile, the hardness in these cases is also higher than
the later ones. For a film gained at 350 8C by N2
plasma, for example, the sum of N- and C-contents
N
C
NqC
2.0
2.9
5.6
5.1
7.6
8.0
8.9
10.9
2.9
1.4
11.8
12.3
(7%) is lower than that (8%) by NH3 at the same
temperature, and even lower than that (9%) by N2 at
280 8C. The hardness of the film is also lower than
either the film gained at the same temperature with
NH3 plasma or the one grown at 280 8C with N2 plasma.
Thus, the dependence of the hardness and the substrate
temperature can be attributed to the sum of N- and Ccontent in the film. This result indicates that both
nitrogen and carbon-contents can raise the hardness of
B–O matrix. Especially, nitrogen has stronger effect on
raising the hardness of B–O matrix than carbon.
The hardness of BON is not high in Fig. 3. So, it can
be expected that the film thickness will strongly influence the hardness due to the substrate effect. This
relationship between thickness and hardness was also
studied by controlling the deposition time to obtain
different thickness. The samples were gained at 500 8C
by using N2. The results are shown in Fig. 4, in which
the curve is divided into three zones according to the
varied tendency of hardness. As expected, the hardness
changes obviously with the thickness. It decreases sharply with the thickness in zone I. Then, it increases with
the increase of thickness in zone II. Finally, arrives at a
thickness-independent value, approximately 10 GPa, in
zone III. This value is comparable with the one obtained
from a BNCO film with carbon-content less than 10%
reported previously w15x. The typical morphology of the
film in different zones is also appeared in the inset of
Fig. 4. It is smooth and continuous in zones I and III,
but becomes rough and even discontinuous in zone II.
The change of morphology indicates the change of film
structure. Thus, the variation of hardness with the
thickness is due to not only the substrate effect, but also
the change of film structure. To confirm this point, FTIR was employed, and the typical results of the films in
each zone are shown in Fig. 5. Usually, the peaks at
1200–1600 cmy1 are regarded as B–O w16x, graphite
carbon structure w17x and B–N w18x. The late theoretical
research on B–O–N shows that angular and linear
B–O–N structure has the feature vibration at 1469 and
1455 cmy1 w19x. The result in zone III accords well
with the reported value. In addition, as the XPS results
G.C. Chen et al. / Surface and Coatings Technology 169 – 170 (2003) 281–286
285
show the compositions are B, O, N in the films, we
regard the peak at this range is B–O–N feature peak. In
zone II, the spectra have two small peaks that are,
respectively, related to BON and BN in the feature
wavenumber range. This means that there are two phases
in the film, which reveals the change of film structure.
In the zone I, there is no obvious peak in the feature
wavenumber range. This means that the films are in the
initial stage of growth. Neither BON nor BN structure
possesses strong IR intensity. The disappearance of BN
in zone III indicates that BN structure is unstable with
the existence of oxygen. The other peaks in Fig. 5 are
due to Si–O (gs1106 cmy1) and B–O–Si (gs920
cmy1, ds612 cmy1) w20x. The IR results also confirm
that the hardness test value in zone III should present
the hardness of BON.
4. Conclusions
BON film was able to be grown by low frequency
RF plasma enhanced MOCVD with trimethyl borate
precursor. The film growth rate was influenced both by
substrate temperature and by nitrogen source gas. It
decreased with increasing the substrate temperature, and
it was more highly increased by using NH3 than that by
N2. Several factors such as nitrogen source gas, substrate
Fig. 5. FT-IR results obtained from the BON films with different film
thickness corresponding to the same zones as Fig. 4.
temperature and film thickness could influence the
hardness of BON films. The reason was regarded due
to the change of composition and the structure of BON
film. The more the sum of N- and C-contents in the
film, the higher the hardness was obtained. Both nitrogen
Fig. 4. Dependence of hardness as a function of film thickness. The insets of the figure are the changes of film morphology for each thickness
zones.
286
G.C. Chen et al. / Surface and Coatings Technology 169 – 170 (2003) 281–286
and carbon could raise the hardness of B–O matrix, on
which the nitrogen had stronger effect than carbon.
Moreover, the film with smooth morphology and continuous structure will have relatively high hardness. The
maximum hardness of BON film obtained in this study
is approximately 10 GPa.
Acknowledgments
One of the authors, G.C. Chen, would like to thank
the BK21 project of the Ministry of Education, Korea.
Support of this research by the Korea Research Foundation though 2000BSRI project (KRF-2000-015BPO0195), and the Center for Advanced Plasma Surface
Technology of SungKyunKwan University through the
ERC project of Korea Science and Engineering Foundation is gratefully acknowledged.
References
w1x S. Veprek, J. Vac. Sci. Technol. A 17 (5) (1999) 2401.
w2x R.B. Kaner, J. Kouvetakis, C.E. Warble, M.L. Sattler, N.
Bartlett, Mater. Res. Bull. 22 (1987) 399.
w3x H. Hubert, L.A.J. Garvie, D.R. Buseck, W.T. Petuskey, P.F.
McMillan, J. Solid State Chem. 133 (1997) 556.
w4x S. Sahu, S. Kavecky, L. Illesova, J. Madejova, I. Bertoti, J.
Szepvolgyi, J. Eur. Cer. Soc. 18 (1998) 1037.
w5x K.H.A. Bogart, N.F. Dalleska, G.R. Bogart, E.R. Fisher, J. Vac.
Sci. Technol. A 13 (1995) 476.
w6x M.R. Hilton, L.R. Narasimhan, S. Nakamura, M. Salmeron,
G.A. Somorjal, Thin Solid Films 139 (1986) 247.
w7x H. Miyamoto, M. Hirose, Y. Osaka, Jpn. J. Appl. Phys. 22
(1983) L216.
w8x C.R. Abernathy, J.D. Mackenzie, S.M. Donovan, J. Cryst.
Growth 178 (1997) 74.
w9x W.T. Zheng, N. Hellgren, H. Sjostrom, J.-E. Sundgren, Surf.
Coat. Technol. 100–101 (1998) 287.
w10x H. Holzschuh, Plasma-aktivierte Abscheidung duenner OxidFilme des Yttriums, Bariums, Kupfers, und Zirconums, Ph.D.
Dissertation, University of Tuebingen, 1990.
w11x A. Gebauer, Plasmagestuetztes CVD mit metallorganischen
und analogen Verbindungen Fortschrittberichte VDI, Nr. 344,
Duesseldorf, 1994.
w12x G.M. Pharr, W.C. Oliver, MRS Bull. 17 (1992) 28.
w13x T.A. Rawdanowicz, V. Godbole, J. Narayan, J. Sankar, A.
Sharma, Composites Part B: Eng. 30 (1999) 657.
w14x H. Zhang, X. Xue, D. Wang, Y. He, S. Peng, Mater. Chem.
Phys. 58 (1999) 1.
w15x C.W. Ong, K.F. Chan, X.A. Zhao, C.L. Choy, Surf. Coat.
Technol. 115 (1999) 145.
w16x W. Zimmerman, A.M. Murphy, C. Feldman, Appl. Phys. Lett.
10 (1967) 71.
w17x X.A. Zhao, C.W. Ong, Y.C. Tsang, Y. Wong, P.W. Chan, C.L.
Choy, Appl. Phys. Lett. 66 (1995) 2652.
w18x R.F. Davis, M.J. Paisley, Z. Sitar, et al., J. Cryst. Growth 178
(1997) 87.
w19x S. Su, J. Mol. Struct. 430 (1998) 137.
w20x G.D. Soraru, N. Dallabona, C. Gervais, F. Babonneau, Chem.
Mater. 11 (1999) 910.