Materials Transactions, Vol. 44, No. 3 (2003) pp. 401 to 406
#2003 The Japan Institute of Metals
Oxidation of Boron Carbide–Silicon Carbide Composite at 1073 to 1773 K
Takayuki Narushima1 , Takashi Goto2 , Makoto Maruyama*1 , Haruo Arashi3; *2 and Yasutaka Iguchi1
1
Department of Metallurgy, Tohoku University, Sendai 980-8579, Japan
Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan
3
Department of Machine Intelligence and System Engineering, Tohoku University, Sendai 980-8579, Japan
2
The oxidation behavior of B4 C–(25–60 mol%)SiC composites prepared by arc-melting was investigated in the temperature range of 1073
to 1773 K using a thermogravimetric technique. Liquid borosilicate, solid SiO2 and carbon were identified as oxidation products by X-ray
diffraction and Raman spectroscopy. Mass gain was observed during oxidation at 1073 K, while mass loss due to the vaporization of boron oxide
in liquid borosilicate was observed at temperatures of 1273 K and higher. In situ Raman spectra of the surface of B4 C–SiC composites indicated
that the silica concentration in the liquid borosilicate increased with increasing SiC content in the composite. Micro-Raman spectroscopy
showed that carbon was enriched in the borosilicate layer close to the oxide/composite interface. The parabolic rate constants for B4 C–
50 mol%SiC composites at 1073 K were proportional to ambient oxygen partial pressures ranging between 30 and 100 kPa. The diffusion of
oxygen molecules through the liquid borosilicate layer could be the rate-controlling process. The increase of SiC content in the B4 C–SiC
composites improved the oxidation resistance in both the mass gain and mass loss regions.
(Received September 24, 2002; Accepted January 24, 2003)
Keywords: oxidation, Raman spectroscopy, silicon carbide, boron carbide, silica, boron oxide, kinetics, carbon
1.
Introduction
Boron carbide B4 C has been used for engineering
materials such as wear parts and abrasives because of its
high hardness and excellent wear resistance. Recently, the
boron carbide has been expected to be a high temperature
thermoelectric material1,2) and a first-wall material for a
fusion reactor3) due to its high figure of merit and low atomic
number. Oxidation resistance of the boron carbide is an
important issue for its practical applications at high temperatures. The oxidation behavior has been reported by
several research groups.4–10) Liquid B2 O3 was detected as the
oxidation product. The liquid B2 O3 layer can not effectively
protect boron carbide because of its low viscosity and high
vapor pressure at high temperatures.11) It is well known that
SiO2 formed on silicon-based materials by thermal oxidation
is an excellent protective layer for oxidation.12) Telle
reported that B4 C–SiC composites synthesized by hotpressing exhibited better oxidation resistance than single
phase B4 C.13) Liquid B2 O3 14) and borosilicate13,15–18) have
been reported as oxidation products of composites in the
B4 C–SiC–C system. The oxide layers on the composites were
analyzed by X-ray diffraction,13,14,17) X-ray photoelectron
spectroscopy13,17) and IR absorption spectroscopy.18) However, the relationship between the oxidation rate and oxide
layer has not been clearly understood in the B4 C–SiC system.
In the present work, the oxidation behavior of B4 C–SiC
composites was investigated by thermogravimetry and
Raman spectroscopy, and the oxidation mechanism of the
B4 C–SiC composites was discussed.
*1Graduate
Student, Tohoku University. Present address: Furukawa
Electric Co., Ltd., Fukui 913-8588, Japan.
*2Deceased.
2.
Experimental Procedure
B4 C and SiC powders were mixed at specific compositions, and then pressed into 15 mm diameter disks with 5 mm
thickness. The disks were arc-melted in an argon atmosphere
with a tungsten electrode. The B4 C–SiC composites after arcmelting were cut into plates (10 mm 10 mm 1 mm). The
plates were polished with diamond paste to 1 mm. SiC content
in the composites ranged between 25 and 60 mol%. The
quasi-binary B4 C–SiC system contains a eutectic composition at SiC content of 45–50 mol%,19) and the solid
solubilities of both B4 C in SiC and SiC in B4 C are less than
2 mass%.20)
The mass change of the composite due to oxidation was
measured continuously using an electrobalance. The specimen temperature was increased to a specific value in the
flowing argon gas, and then oxygen gas or oxygen–argon gas
mixture was introduced to initiate oxidation. The gases used
in the present experiments were dried by passing them
through CaCl2 and P2 O5 . The total gas pressure was 100 kPa.
The oxygen partial pressures ranged between 30 and 100 kPa
in oxygen–argon atmospheres. The oxidation temperatures
varied from 1073 to 1773 K. The longest oxidation time was
86.4 ks. After the oxidation, the specimen was cooled in
flowing argon gas.
In situ Raman spectra of the surface of B4 C–SiC
composites during the oxidation were measured by a backscattering method (T64000, Jovin-Yvon).21,22) An argon ion
laser was used as an incident beam at 514.5 nm. The
measurement of scattered light was carried out with a triple
monochromator and a multi-channel detector. The measuring
time for each in situ Raman spectrum ranged from 100 to
300 s.
The oxide films on the specimens after the oxidation
experiments were analyzed at room temperature by microRaman spectroscopy (T64000, Jovin-Yvon) and X-ray
diffraction (CuK, Ni-filtered, RAD-C, Rigaku).
402
T. Narushima, T. Goto, M. Maruyama, H. Arashi and Y. Iguchi
Fig. 1 Mass change of B4 C–50 mol%SiC composite due to oxidation in an
oxygen atmosphere.
3.
Results and Discussion
3.1 Mass change during oxidation
Figure 1 shows the mass change due to the oxidation of the
B4 C–50 mol%SiC composite in an oxygen atmosphere at
different temperatures. Mass gain was observed at 1073 K,
while mass loss occurred at temperatures of 1273 K and
higher. The mass loss rates decreased with time and with
decreasing temperature. Figure 2 demonstrates the XRD
patterns of the surface of B4 C–SiC composites after oxidation in an oxygen atmosphere for 86.4 ks. SiO2 was detected
on the B4 C–SiC composites oxidized at 1473 and 1773 K.
Liquid B2 O3 and solid SiO2 could be formed by the oxidation
of B4 C and SiC in the composite, respectively. SiO2 might be
soluble in liquid B2 O3 up to around 60 mol% at 1073 K and
would form liquid borosilicate.23) Ogawa et al. reported
liquid borosilicate would be rapidly formed by the dissolution of SiO2 in liquid B2 O3 .18) Other research groups detected
the borosilicate on the B4 C–SiC composites13) and B4 C–
SiC–C composites15,17,18) by X-ray photoelectron spectroscopy and IR absorption spectroscopy. As described later, the
liquid borosilicate was investigated by Raman spectroscopy
in the present work. The mass loss observed in the oxidation
of B4 C–SiC composites at 1273 K and higher might be
caused by vaporization of B2 O3 in the liquid borosilicate.
Figure 3 shows the surface of the B4 C–50 mol%SiC composite oxidized at 1773 K for 86.4 ks. Pores were observed on
the surface, which were formed by the vaporization of B2 O3 .
The silica concentration in the liquid borosilicate increased
by the vaporization of B2 O3 , and then a solid SiO2 phase
formed in the liquid borosilicate when the SiO2 concentration
exceeded the solubility value.23)
Mass losses due to the oxidation of B4 C8) and the B4 C–SiC
composites in the present work are compared in Fig. 4. The
addition of SiC to B4 C was effective to decrease mass loss by
oxidation. The silica-rich borosilicate and solid SiO2 phases
in the oxide layer could suppress the formation and the
vaporization of B2 O3 .
Fig. 2 XRD patterns of B4 C–SiC composites before and after oxidation in
an oxygen atmosphere for 86.4 ks.
Fig. 3 Scanning electron micrograph of the surface of B4 C–50 mol%SiC
oxidized at 1773 K for 86.4 ks.
3.2 Observation of oxide layer by Raman spectroscopy
Figure 5 shows the in situ Raman spectra of the surface of
the B4 C–50 mol%SiC composite during oxidation in an
oxygen atmosphere at 1073 K. The Raman spectra during
heating to 1073 K and cooling to room temperature in an
argon atmosphere are included in Fig. 5. The Raman peaks B
and S due to liquid borosilicate24) were detected. The Raman
peak B around 800 cm1 was assigned to a breathing
Oxidation of Boron Carbide–Silicon Carbide Composite at 1073 to 1773 K
Fig. 4
Mass loss of B4 C and B4 C–SiC composites due to oxidation.
403
thermodynamically stable at low temperatures. A broad
Raman peak S around 450 cm1 could be assigned to a
bending or rocking of B–O–B, B–O–Si and Si–O–Si bridging
bonds.24) In situ Raman spectra of the surfaces of the B4 C–
60 mol%SiC and B4 C–25 mol%SiC composites during oxidation in an oxygen atmosphere at 1073 K are shown in
Fig. 6. Furukawa and White24) suggested that the Raman
peak S corresponded to silica in liquid borosilicate because
the intensity of the Raman peak S increased with increasing
silica concentration in the liquid borosilicate. Figure 7
demonstrates the change of the intensity ratio of the Raman
peak S to the Raman peak B (IS =IB ) with oxidation time at
1073 K. The value of IS =IB increased with increasing SiC
contents in the composites. At constant SiC contents, the
value of IS =IB increased with oxidation time, particularly up
to 10 ks. This result suggests that B4 C in the composites
might be preferentially oxidized in the initial period.
In addition to the Raman peaks of the liquid borosilicate,
the Raman peaks of carbon were observed at 1360 and
1580 cm1 (see Figs. 5 and 6). Figure 8 shows the microRaman spectra in the borosilicate on the B4 C–50 mol%SiC
composite after oxidation in an oxygen atmosphere at 1073 K
for 86.4 ks. The intensity of carbon peaks near the oxide/
composite interface was stronger than that of the gas/oxide
interface. The carbon was enriched in the borosilicate layer
near the oxide/composite interface because of lower oxygen
partial pressure at the oxide/composite interface than that at
the oxide/gas interface. Shimada et al. reported the formation
of carbon in the initial stage of oxidation of HfC in the
Fig. 5 In situ Raman spectra of B4 C–50 mol%SiC composite during
heating, oxidation and cooling.
vibration of boroxy rings.24–28) The boroxy rings are
equilibrated with isolated BO3 units in the liquid borosilicate.24) The increase of Raman peak B during the cooling
process might be caused by the formation of boroxy rings
from isolated BO3 units. The boroxy rings would be
Fig. 6 In situ Raman spectra of B4 C–60 mol%SiC and B4 C–25 mol%SiC
composites during oxidation in an oxygen atmosphere at 1073 K.
404
T. Narushima, T. Goto, M. Maruyama, H. Arashi and Y. Iguchi
Fig. 9 Equilibrium oxygen partial pressures at the oxide/composite interface.
Fig. 7 Intensity ratio of Raman peak S to the Raman peak B (IS =IB ) during
oxidation of B4 C–SiC composites in an oxygen atmosphere at 1073 K.
Si(s in SiC) þ O2 (g) ¼ SiO2 (s)
2B(s in B4 C) þ 3/2O2 (g) ¼ B2 O3 (l)
ð2Þ
ð3Þ
The equilibrium oxygen partial pressures of eqs. (1), (2) and
(3), which were calculated using the thermodynamic data11)
are shown in Fig. 9. The activities of C, SiO2 and B2 O3 were
assumed to be unity. The activity values of Si and B decided
by eqs. (4) and (5), respectively, were used in the calculation,
e.g. 6:7 104 for Si and 0.19 for B at 1073 K.
Fig. 8 Micro-Raman spectra in the borosilicate on B4 C–50 mol%SiC
composite after oxidation in an oxygen atmosphere at 1073 K for 86.4 ks.
temperature range of 753 to 873 K.29) They assumed that
carbon could be present beneath the HfO2 layer or contained
in the HfO2 layer. Recently, the presence of carbon in silica
layer after oxidation of silicon carbide at 973 K in ozonecontaining atmospheres has been shown.30) These previous
reports and the present result suggest that the oxygen partial
pressure at the oxide/composite interface was lower than the
equilibrium oxygen partial pressure of eq. (1).
C(s) þ 1/2O2 (g) ¼ CO(g)
ð1Þ
The oxidation reactions of Si and B were represented by eqs.
(2) and (3), respectively.
Si(s in SiC) þ C(s) ¼ SiC(s)
ð4Þ
4B(s in B4 C) þ C(s) ¼ B4 C(s)
ð5Þ
The activities of SiC and B4 C were assumed to be unity
because of low solid solubilities of B4 C in SiC and SiC in
B4 C. The formation of carbon at the oxide/composite
interface during the oxidation of B4 C–SiC composites seems
to be thermodynamically possible under the conditions, CO
partial pressure >101 Pa, at 1073 K.
It was reported that the intensity ratio between Raman
peaks at 1360 cm1 and 1580 cm1 (I1360 =I1580 ) was 1.2 for
amorphous carbon,31) and the inverse of the crystallite size
was proportional to I1360 =I1580 .32) The value of I1360 =I1580
evaluated from the Raman spectra in Fig. 8 was above 1,
meaning that the carbon could be amorphous or crystalline
with a small crystallite size (3 nm).
3.3 Oxidation mechanism
Figure 10 shows the relationships between mass gain (M)
and oxidation time (t) for the B4 C–50 mol%SiC composite at
1073 K in oxygen–argon atmospheres. The gradient of each
curve changed around 1 ks. A high oxidation rate at t < 1 ks
may be caused by the preferential oxidation of B4 C in the
composites because liquid B2 O3 -rich borosilicate has high
oxygen permeability and may not be protective for oxidation
above 1050 K.16) In the present work, the gradient at t > 1 ks
in Fig. 10 is around 0.5, and the oxidation rate seemed to be
described by the parabolic law shown in eq. (6).
Oxidation of Boron Carbide–Silicon Carbide Composite at 1073 to 1773 K
Fig. 10 Mass change of B4 C–50 mol%SiC composite due to oxidation in
oxygen–argon atmospheres at 1073 K.
M 2 ¼ kp t
ð6Þ
where kp is the parabolic rate constant. Figure 11 shows the
effects of ambient oxygen partial pressure on the parabolic
rate constant for the B4 C–50 mol%SiC composite at 1073 K.
The parabolic rate constant was proportional to the ambient
oxygen partial pressure. This dependence implies that the
rate-controlling process of the oxidation may be the diffusion
of oxygen molecules through the liquid borosilicate layer and
the oxidation reaction proceeded at the oxide/composite
interface. Other research groups13,16–18) also reported that the
oxygen diffusion in the oxide layer could be the ratecontrolling process of composites in the B4 C–SiC–C system.
Fig. 11 Effect of ambient oxygen partial pressure on the parabolic rate
constant for oxidation of B4 C–50 mol%SiC composite at 1073 K.
405
Fig. 12 Comparison of the present parabolic rate constants of B4 C–
50 mol%SiC composite with those of B4 C, SiC and B4 C–SiC composite
reported by other research groups.
Figure 12 shows the parabolic rate constants obtained in
the present work comparing with those of B4 C, SiC and B4 C–
SiC composites reported in literatures.4,14,33) The reported14)
and present kp values of B4 C–SiC composites had values
between those of B4 C4) and SiC.33) The borosilicate would
provide more effective protection for oxidation than liquid
B2 O3 because of its high viscosity and low oxygen permeability.16)
4.
Conclusions
The oxidation behavior of B4 C–(25–60 mol%)SiC composites were investigated in oxygen–argon atmospheres at
1073 to 1773 K. The following results were obtained:
(1) Liquid borosilicate, solid SiO2 and carbon were
detected as the oxidation products by X-ray diffraction
and in situ Raman spectroscopy. The silica concentration in the liquid borosilicate layer increased with
increasing SiC content in the composite. Carbon
enriched in the borosilicate layer near the oxide/
composite interface.
(2) Mass gain due to the formation of liquid borosilicate
was observed at 1073 K, while mass loss caused by the
vaporization of boron oxide in liquid borosilicate was
observed at temperatures of 1273 K and higher. The
addition of SiC to B4 C was effective to improve the
oxidation resistance both in mass gain and mass loss
regions. This was caused by the increase of silica
concentration in the liquid borosilicate layer and
formation of solid SiO2 .
(3) The parabolic rate constants for oxidation of the
composites were proportional to the ambient oxygen
partial pressure at 1073 K, and the rate-controlling
process of oxidation could be the diffusion of oxygen
molecules through the liquid borosilicate layer.
406
T. Narushima, T. Goto, M. Maruyama, H. Arashi and Y. Iguchi
Acknowledgments
This research was supported by Grant-in-Aid for Scientific
Research under Contract Numbers 11650730 and 14550696
from the Ministry of Education, Culture, Sports, Science and
Technology, Japan. The authors thank Dr. H. Naito from
National Aerospace Laboratory of Japan for his helpful
suggestions on Raman spectroscopic analyses. Thanks to Mr.
T. Uehara, former undergraduate student of Tohoku University, for his assistance in the TG experiments.
REFERENCES
1) C. Wood: Mat. Res. Soc. Proc. Vol. 97, ed. by D. Emin, T. L. Aselage
and C. Wood (MRS, 1987) pp. 335–346.
2) T. Goto, J. Li and T. Hirai: J. Jpn. Soc. Pow. and Pow. Metall. 43 (1996)
306–310.
3) O. I. Buzhinskij and Y. M. Semenets: Fusion Eng. Design 45 (1999)
343–360.
4) V. A. Lavrenko, A. P. Pomytkin, P. S. Kislyj and B. L. Grabchuk: Oxid.
Met. 10 (1976) 85–95.
5) Y. L. Krut-skii, G. V. Galevskii and A. A. Kornilov: Sov. Powder
Metall. Met. Ceram. 22 (1983) 116–118.
6) G. A. Gogotsi, Ya. L. Groushevsky. O. B. Dashevskaya, Yu. G. Gogotsi
and V. L. Lavrenko: J. Less-Common Met. 117 (1986) 225–230.
7) L. N. Efimenko, E. V. Lifshifts, I. T. Ostapenko, I. A. Snezhko and É. P.
Shevyakova: Sov. Powder Metall. Met. Ceram. 26 (1987) 318–321.
8) V. A. Lavrenko and Yu. G. Gogotsi: Oxid. Met. 29 (1987) 193–202.
9) Yu. G. Gogotsi and V. V. Kovylyaev: Sov. Powder Metall. Met. Ceram.
28 (1989) 306–309.
10) J. P. Viricelle, P. Goursat and D. Bahlout-Hourlier: J. Thermal Analysis
and Calorimetry 63 (2001) 507–515.
11) Thermochemicl Properties of Inorganic Substances, ed. by O. Knache,
O. Kubaschewski and K. Hasselmann (Springer-Verlag, Berlin, 1991).
12) T. Narushima, T. Goto, T. Hirai and Y. Iguchi: Mater. Trans., JIM 38
(1997) 821–835.
13) R. Telle: AIP Conference Proceedings 231, Boron Rich Solids, ed. by
D. Emin, T. L. Aselage, A. C. Switendick, B. Morosin and C. L.
Beckel, (AIP, NY, 1991) pp. 553–560.
14) D.-H. Riu, R. Choi, H.-E. Kim and E.-S. Kang: J. Mater. Sci. 30 (1995)
3897–3902.
15) W. G. Zhang, H. M. Cheng, H. Sano, Y. Uchiyama, K. Kobayashi, L. J.
Zhou, Z. H. Shen and B. L. Zhou: Carbon 36 (1998) 1591–1595.
16) Q. Guo, J. Song, L. Liu and B. Zhang: Carbon 37 (1999) 33–40.
17) K. Kobayashi, K. Maeda, H. Sano and Y. Uchiyama: Carbon 33 (1995)
397–403.
18) I. Ogawa, K. Kobayashi and S. Nishikawa: J. Mater. Sci. 23 (1988)
1363–1367.
19) T. Goto, E. Ito, M. Mukaida and T. Hirai: J. Jpn. Soc. Pow. and Pow.
Metall. 43 (1996) 311–315.
20) D. R. Secrist: J. Am. Ceram. Soc. 47 (1964) 127–130.
21) T. Narushima, N. Kikuchi, M. Maruyama, H. Arashi, Y. Nishina and Y.
Iguchi: Tetsu-to-Hagané 81 (1995) 607–612.
22) Y. Iguchi, T. Narushima, M. Maruyama, T. Shiozawa and H. Arashi:
Design Fundamentals of High Temperature Composites, Intermetallics
and Metal-Ceramic Systems, ed. by R. Y. Lin, Y. A. Chang, R. G.
Reddy and C. T. Liu, (TMS, 1996) pp. 337–343.
23) T. J. Rockett and W. R. Foster: J. Am. Ceram. Soc. 48 (1965) 75–80.
24) T. Furukawa and W. B. White: J. Am. Ceram. Soc. 64 (1981) 443–447.
25) J. Krogh-Moe: J. Non-Cryst. Solids 1 (1969) 269–284.
26) F. L. Galeener and A. E. Geissberger: J. De Phys. C9 Suppl. 43 (1982)
343–346.
27) G. E. Walrafen, S. R. Samanta and P. N. Krisnan: J. Chem. Phys. 72
(1980) 113–120.
28) J. P. Bronswijk and E. Strijks: J. Non-Cryst. Solids 24 (1977) 145–147.
29) S. Shimada, M. Inagaki and K. Matsui: J. Am. Ceram. Soc. 75 (1992)
2671–2678.
30) T. Narushima, M. Kato, S. Murase, C. Ouchi and Y. Iguchi: J. Am.
Ceram. Soc. 85 (2002) 2049–2055.
31) F. Tuinstra and J. L. Koenig: J. Chem. Phys. 53 (1970) 1126–1130.
32) J. Heremans, I. Rahim and M. S. Dresselhaus: Phys. Rev. B 32 (1985)
6742–6747.
33) E. C. Ramberg, G. Cruciani, K. E. Spear and R. E. Tressler: J. Am.
Ceram. Soc. 79 (1996) 2897–2911.