High temperature oxidation of silicon hexaboride ceramics

Article abstract of DOI:10.1016/S0025-5408(01)00560-8

Isothermal oxidation of silicon hexaboride (SiB₆) ceramic powder at high temperature was investigated. The SiB₆ powder samples were heated from room temperature to 1273K for 25 hours in air, and weight changes were measured to estimate the oxidation resistance and nature of the oxide layers on the SiB₆ ceramics. The samples oxidized in the temperature range of 673 to 1273 K for 25 hours exhibited increasing weight gain with increasing oxidation temperature. The oxidation proceeded in accordance with the parabolic law during the initial oxidation stage. The weight gain increased with increasing oxidation temperature. Based on the results of X-ray diffraction analysis, silicon oxide (SiO₂) and boron oxide (B₂O₃) were present on the surface of samples oxidized in the temperature range of 873 to 1273 K.

Full text of DOI:10.1016/S0025-5408(01)00560-8

Materials Research Bulletin 36 (2001) 1083–1089
Pergamon
High temperature oxidation of silicon hexaboride
ceramics
ଁ
J. Matsushita , S. Komarneni*
Materials Research Laboratory, The Pennsylvania State University, University Park, PA 16802, USA
(Refereed)
Received 21 September 2000; accepted 11 October 2000
Abstract
Isothermal oxidation of silicon hexaboride (SiB6) ceramic powder at high temperature was
investigated. The SiB6 powder samples were heated from room temperature to 1273K for 25 hours in
air, and weight changes were measured to estimate the oxidation resistance and nature of the oxide
layers on the SiB6 ceramics. The samples oxidized in the temperature range of 673 to 1273 K for 25
hours exhibited increasing weight gain with increasing oxidation temperature. The oxidation pro-
ceeded in accordance with the parabolic law during the initial oxidation stage. The weight gain
increased with increasing oxidation temperature. Based on the results of X-ray diffraction analysis,
silicon oxide (SiO2) and boron oxide (B2O3) were present on the surface of samples oxidized in the
temperature range of 873 to 1273 K. © 2001 Elsevier Science Ltd. All rights reserved.
Keywords: A. Ceramics; A. Electronic materials
1. Introduction
Several silicon boride phases have been registered in the X-ray cards of the International
Center for Diffraction Data. Among them, silicon tetraboride, SiB4 and silicon hexaboride,
SiB6 have been shown to be potentially useful materials because of their excellent chemical
* Corresponding author. Fax: 814-865-2326.
E-mail address: komarneni@psu.edu (S. Komarneni).
ଁ
Permanent address: Department of Applied Chemistry, Tokai University, 1117 Kitakaname,
Hiratsuka 259-1292, Japan.
0025-5408/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0025-5408(01)00560-8

1084
J. Matsushita, S. Komarneni / Materials Research Bulletin 36 (2001) 1083–1089
Table 1
Impurity of as-received SiB6 powder as starting material
Impurity (wt%)
Al
Ca
Co
Cr
Cu
0.1
Fe
0.1
0.05
0.01
0.001
0.02
Mg
Mn
Ni
0.15
0.02
0.01
0.01
Sr
resistance properties. SiB6 is is more stable at high temperature than SiB4 in the B-Si binary
alloy phase diagram [1].
SiB6 has been attracting a great deal of attention as a material for thermoelectric
applications at high temperatures, because of its favorable properties such as high melting
point, chemical stability and high electrical conductivity, low thermal conductivity, and high
Seebeck coefficient at high temperature. Chen et al. reported that the Si-B composite had
good thermoelectric property at high temperature [2].
Unfortunately, although the monolithic SiB6 is known to be chemically stable and
resistant to high temperatures in air, it is rarely used at high temperatures. To date, there have
only been a few reports regarding the oxidation properties of SiB6. In this study, the
oxidation of SiB6 powder samples is investigated from room temperature to high tempera-
tures in order to determine its suitability for advanced high-temperature applications.
2. Experimental procedure
SiB6 powder used in this study is a commercially available powder. The starting material
was made by Cerac Corp., U.S.A. and one as-received batch was used for all tests. The purity
and median particle size of the sample are 98% and 3 ␮m, respectively. The impurities of the
sample by chemical analysis are shown in Table 1.
Differential thermal analysis (DTA) and thermogravimetric analysis (TG) for SiB6 start-
ing powder was conducted with a thermal analyzer (Model: TG8120, Rigaku Co., Ltd.,
Japan) using a ␣-alumina pan and high-purity ␣-alumina standard powder in flowing air (0.5
l/min) heated at 10 K/min.
Isothermal oxidation in air was carried out in an electric furnace maintained at the desired
temperature. The sample to be evaluated usually consisting of about 500 mg of the powder,
was weighed carefully and placed in a high-purity ␣-alumina boat. This high-purity ␣-alu-
mina boat containing several samples was directly introduced in the hot zone of the furnace
at a predetermined operating temperature onto a low-density stabilized zirconia setter. After
the prescribed period, the sample, together with its boat, was removed from furnace, allowed
to cool and reweighed carefully. Mass data were obtained using an accurate electric balance.
The value of weight gain of the sample includes the cooling period in addition to the total
oxidation time at treatment temperature.
A few typical samples of partly oxidized samples were subjected to X-ray diffraction

J. Matsushita, S. Komarneni / Materials Research Bulletin 36 (2001) 1083–1089
1085
Fig. 1. Results of differential thermal analysis (DTA) and thermogravimetric analysis (TG) for SiB6 powder.
(XRD) analysis for phase evolution using an X-ray diffractometer (Cu K␣/Ni filter, 40 kV,
30 mA) from Philips Ltd., Netherlands (Model: APD-1700). The powder X-ray diffraction
peaks were identified using the computer program, which incorporates all of the data of the
X-ray cards of the International Center for Diffraction Data (ICDD).
Morphological changes of microstructures and the specimen surfaces were examined by
scanning electron microscopy (SEM) using a microscope from Elionics Co., Ltd., Japan
(Model: ESA-2000).
3. Results and discussion
In general, the evaluation of oxidation resistance of silicon components of silicon carbide
(SiC) or silicon nitride (Si3N4) is carried out by investigating the weight gain accompanying
the following reaction [3–5]:
SiC͑s͒ ϩ 3/2O2͑ g͒ ϭ SiO2͑s͒ ϩ CO͑ g͒
(1)

1086
J. Matsushita, S. Komarneni / Materials Research Bulletin 36 (2001) 1083–1089
Fig. 2. Change of weight gain with oxidation as a function of temperature and time.
Si3N4͑s͒ ϩ 3O2͑ g͒ ϭ 3 SiO2͑s͒ ϩ 2N2͑ g͒
(2)
(3)
SiB6, however, is thought to form the following oxides:
SiB6͑s͒ ϩ 11/2O2͑ g͒ ϭ SiO2͑s͒ ϩ 3B2O3͑s͒
Even if the weight is apparently decreased by the vaporization of B2O3 formed as one of
the oxidation products at a high temperature, the weight of the sample should increase with
the passage of oxidation time. Thus, the weight gain due to the oxidation of the sample was
measured in order to evaluate relative oxidation resistance [6, 7].
The result of the differential thermal analysis (DTA) and thermogravimetric analysis (TG)
for SiB6 starting powder are shown in Fig. 1. During the initial heating between room
temperature to about 850 K, absorbed water is removed from the surface of the sample
corresponding to an initial weight loss of about 0.5%. The weight gain increased with
increasing oxidation temperature.

J. Matsushita, S. Komarneni / Materials Research Bulletin 36 (2001) 1083–1089
1087
Fig. 3. X-ray diffraction patterns of the oxidation surfaces at a): room temperature (as-received), b): 873K, c):
1073K and d): 1273K. The oxidation time was 25 hours.
Oxidation weight increased sharply near 850 K, so thermogravimetric analysis (TG) and
thermal analysis (DTA) were conducted to confirm oxidation initiation temperature. An
exothermic peak seen near 850 ϳ 1100 K in DTA curve is assumed to be due to SiB6
oxidation. In the TG curve, weight gain by oxidation was confirmed from 850 K, and about
55% of weight gain occurs due to oxidation up to 1273K. Similar measurement was
conducted using mixed 50% SiB6 and 50% ␣-alumina powder and the result was almost the
same as monolithic SiB6, suggesting that no reaction occurred between SiB6 and ␣-alumina.
The oxidative weight gain of the sample at room temperature to 1273 K versus oxidation
time is shown in Fig. 2. The weight gain data of the different samples is the average of three

1088
J. Matsushita, S. Komarneni / Materials Research Bulletin 36 (2001) 1083–1089
Fig. 4. Scanning electron microscope (SEM) photographs of the surfaces at a): room temperature (as-received),
b): 873K, c): 1073K and d): 1273K. The oxidation time was 25 hours.
measurements. The weight gain of the sample oxidized below a temperature of 673 K for 1
hour was approximately 1 %. However, when the oxidation time was extended, additional
weight change did not occur. Significant oxidation of samples started at 873 K, and the
weight gain increased with increasing oxidation temperature. The samples oxidized at and
above 873 K also exhibited increasing weight gain with increasing oxidation time. The
oxidation proceeded in accordance with the parabolic law during the initial oxidation stage.
This trend implies that the silicon oxide layer produced by oxidation serves as a protective
layer against oxygen, causing the oxidation rates to be determined by the diffusion of oxygen
through the oxidation layer.
Figs. 3 shows the results of the powder X-ray diffraction analysis of samples oxidized at
873 to 1273 K for 25 hours in air. The X-ray diffraction data of the starting powder only
showed peaks of SiB6 (Fig. 3 (a)). A similar result was obtained for samples treated in air
at 473K for 25 hours; the weight gain was below 1%. The result of X-ray diffraction showed
no crystalline phase other than SiB6 in the sample. The samples heated from 873 to 1273 K
for 25 hours in air showed the presence of an additional phase, silicon oxide SiO2 (␣-
cristobalite) and boron oxide B2O3 (Fig. 3 (b) ϳ (d)).
From the above data, we suggest that room temperature to high temperature oxidation
(1273 K) of SiB6 follows the equation:
SiB6͑s͒ ϩ 11/2O2͑ g͒ ϭ SiO2͑s͒ ϩ 3B2O3͑s͒
(4)

J. Matsushita, S. Komarneni / Materials Research Bulletin 36 (2001) 1083–1089
1089
The calculated ideal weight gain change of SiB6 from eq. (4) is approximately 189.5 %.
The weight gain of the sample oxidized at 1273 K for 25 hours was about 95% (Fig. 2). This
value is half the value of theoretical weight gain of SiB6.
SEM photographs of the surface of the sample oxidized at 873 to 1273 K for 25 hours are
shown in Fig. 4. The surface oxidized at 873 K showed a coarsely grained oxide layer (Fig.
4 (b)). Based on the results of X-ray diffraction analysis, silicon oxide and boron oxide were
present on the surface of the sample oxidized at 873 K. On the other hand, the surface of the
sample oxidized at 1073 to 1273 K showed a glassy appearance because of phase transition
to a crystalline phase, SiO2 along with B2O3 as a liquid phase. B2O3 may also react with SiO2
to become amorphous borosilicate glass, and partial vaporization may occur [8–10]. This
borosilicate glass was assumed to be a vitreous substance (Fig. 4 (c), (d)). X-ray diffraction
peaks other than SiO2 and B2O3 could not be confirmed as oxidation products, and a
3-component Si-B-O compound (silicon borate) was not detected.
4. Conclusion
The SiB6 powder samples oxidized from 873 to 1273K exhibited increasing weight gain
with increasing oxidation temperature and time. Based on the results of X-ray diffraction
analysis, silicon oxide (SiO2) and boron oxide (B2O3) were postulated to be present on the
surface of the samples oxidized from 873 to 1273 K; oxidation was assumed to advance as
follows:
SiB6͑s͒ ϩ 11/2O2͑ g͒ ϭ SiO2͑s͒ ϩ 3B2O3͑s͒
References
[1] R.W. Olesinski, G.J. Abbaschian, Bull. Alloy Phase Diagrams 5 (1984) 478–484.
[2] L. Chen, T. Goto, M. Mukaida, M. Niino, T. Hirai, J. Jpn. Soc. Powder and Powder Metallurgy (1994)
1299–1303.
[3] H. Abe, M. Kawai, T. Kanno, K. Suzuki, Engineering Ceramics, Gihoudo, Tokyo, Japan, 1984, p. 161.
[4] T. Sasaki, Y. Fukatsu, T. Iseki, Yogyo-kyokai-Shi 95 (1987) 646–651.
[5] D.S. Fox, J. Am. Ceram. Soc. 81 (1998) 945–950.
[6] J. Matsushita, S. Hayashi, H. Saito, J. Ceram. Soc. Japan 98 (1990) 308–310.
[7] J. Matsushita, N. Nagashima, H. Saito, J. Ceram. Soc. Japan 98 (1990) 1172–1175.
[8] T.J. Rockett, W.R. Foster, J. Am. Ceram. Soc. 48 (1965) 78.
[9] M.A. Janney, Am. Ceram. Soc. Bull. 66 (1987) 322–324.
[10] J.W. Fergus, W.L. Worrell, J. Am. Ceram. Soc. 78 (1995) 1961–1964.