Full text of DOI:10.1557/JMR.2002.0334

Effect of Si N O content on the microstructure, properties,
2
2
and erosion of silicon nitride–Si N O in situ composites
2
2
a)
Dong-Soo Park
Ceramic Materials Group, Korea Institute of Machinery and Materials, 66 Sang-Nam-Dong,
Chang-Won, Kyong-Nam, Korea
Hyun-Ju Choi
Department of Materials Engineering, Korea University, An-Am-Dong, Sung-Buk-Gu, Seoul, Korea
Byung-Dong Han and Hai-Doo Kim
Ceramic Materials Group, Korea Institute of Machinery and Materials, 66 Sang-Nam-Dong,
Chang-Won, Kyong-Nam, Korea
Dae-Soon Lim
Department of Materials Engineering, Korea University, An-Am-Dong, Sung-Buk-Gu, Seoul, Korea
(Received 18 February 2002; accepted 4 June 2002)
Silicon nitride–Si N O in situ composites were prepared by hot pressing powder
2
2
mixtures of ␣–Si N , 6 wt% Y O , 1 wt% Al O , and 0–12 wt% SiO . X-ray
3
4
2
3
2
3
2
diffraction (XRD) analysis indicated that the volume percents of Si N O were 0, 13,
2
2
3
1, and 54 for the composites prepared with 0, 4, 8, and 12 wt% SiO , respectively.
2
XRD results also indicated that both silicon nitride grains and Si N O grains were laid
2
2
down perpendicular to hot pressing direction. As the volume percent of Si N O
2
2
increased, the width and the amount of elongated silicon nitride grains decreased, but
the fracture toughness increased. Young’s modulus of the in situ composites decreased
as the Si N O content was increased. The erosion rate decreased as the Si N O content
2
2
2 2
was increased, in part, due to both the increased fracture toughness and the reduced
grain size. Erosion of the composites occurred primarily due to the grain dislodgment.
The sample without Si N O experienced micro-chipping due to transgranular fracture.
2
2
I. INTRODUCTION
At high temperatures, Si N O is formed by a reaction
2 2
during sintering or during annealing. Ohashi et al. pre-
Erosion is often observed from ceramic parts, as in the
case of sandblast nozzles. The erosion rate has been re-
pared Si N O by hot pressing an equi-molar powder mix-
2
2
−p
ture of silicon nitride and silica added with ceria as the
ported to be proportional to the (hardness) multiplied
5
−q 1–3
sintering additive and obtained >99% Si N O. Emoto
2
2
by (the fracture toughness) . Wada tested dozens of
different silicon nitride ceramics using SiC particles and
found that the values of p and q were 1.2 and 2.9, re-
spectively for the sintered samples while they were 0.27
et al. reported that large, elongated Si N O grains pref-
2
2
erentially grew in the [010] direction in a fine-grained
-silicon nitride matrix when the silicon nitride mixed
␤
6
1
with cordierite was hot pressed and annealed at 2023 K.
and 7.1, respectively for the hot-pressed samples. Liu
According to Emoto et al., Si N O nucleated only at
2
2
et al. reported p and q values obtained from the erosive
wear tests of ␣/␤ SiAlON ceramics as 1.51 and 3.33,
early stages of annealing and grew quickly in its length
direction. The fracture toughness was increased linearly
as the Si N O grains grew.
4
respectively. From the literature, it is understood that the
2
2
erosion rate is more dependent on the fracture toughness
than on the hardness of the material. Zhang et al. re-
ported that the erosion rate was lower for ␣–SiAlON with
the interlocking elongated grains than for ␣–SiAlON
with equiaxed grains, emphasizing the importance of the
In this study, four powder mixtures with differing lev-
els of silica were hot pressed to make silicon nitride-
Si N O in situ composites. The effects of the Si N O
2
2
2 2
volume percent on the microstructures, the properties,
and the erosion rates of the composites were investigated.
A gas-blast apparatus was used for the erosion test in the
current study. It is one of the more popular methods for
2
microstructure.
1
,2,7
evaluating the erosion resistance of the materials.
Wada used an abrasive water jet for evaluating the ero-
sion rates of ceramics.
a)
8
Address all correspondence to this author.
J. Mater. Res., Vol. 17, No. 9, Sep 2002
© 2002 Materials Research Society
2275
http://journals.cambridge.org
Downloaded: 14 Mar 2015
IP address: 139.86.13.152

D-S. Park et al.: Effect of Si N O content on the microstructure, properties, and erosion of silicon nitride–Si N O in situ composites
2
2
2 2
II. EXPERIMENTAL
XRD peaks of Si N O phase from the in situ composite
2 2
with those from the reference powder mixtures gave the
volume percent of Si N O.
Powdered ␣-silicon nitride (SN-E10; Ube Industries
Ltd., Tokyo, Japan) was mixed with 6 wt% yttria (C;
H.C. Starck Co., Berlin, Germany), 1 wt% alumina
2
2
Vickers hardness was measured under 9.8 N load.
To calculate the fracture toughness with Evans’ and
(
AKP30; Sumitomo Chemical Co., Osaka, Japan) and 0,
, 8, and 12 wt% silica (EP; Junsei Chemical Co., Tokyo,
9
Charles’s equation, the crack length generated by Vick-
4
ers indentation under 196 N load was measured. Ten
hardness measurements and ten fracture toughness meas-
urements were recorded for statistics. The microstructure
was observed following plasma etching. Widths of ␤-
silicon nitride grains were measured by image analysis
software (Image-Pro Plus, version 3.0, Media Cybernet-
ics, L. D., Silver Spring, MD). Elastic moduli of the
samples were measured by a pulse-echo method as in
Japan) for 8 h by planetary ball milling. Silicon nitride
balls 5 mm in diameter and ethanol were used for mill-
ing. After milling, the slurry was dried with a rotary
evaporator. The powder mixtures were hot pressed by
using a BN-coated graphite mold and punch at 2123 K
for 1.5 h under 30 MPa in a flowing nitrogen atmos-
phere. Samples with 0, 4, 8, and 12 wt% silica were
labeled SO-0, SO-4, SO-8, and SO-12, respectively.
X-ray diffraction (XRD) analysis was performed on
the hot-pressed samples to determine the Si N O content.
1
0
the previous report.
2
2
To quantify the Si N O content, a silicon oxynitride ref-
2
2
erence sample was prepared. An equi-molar powder mix-
ture of ␣-silicon nitride and silica added with 3 mol%
CeO was hot pressed at 2173 K for 2 h under 30 MPa in
2
a flowing nitrogen atmosphere. XRD analysis verified
that silicon nitride peaks were absent in the presence of
Si N O peaks. The silicon oxynitride sample was ground
2
2
using a disc mill consisting of WC–Co parts. The beta
silicon nitride powder was prepared in an identical pro-
cedure as the silicon oxynitride powder and was then
mixed with the silicon oxynitride powder using an agate
mortar and pestle to make the reference powder mixtures.
The ratios of the beta silicon nitride to the silicon oxy-
nitride in the reference powder mixtures in respective
order were 0:100, 20:80, 40:60, 60:40, 20:80, and 100:0
by weight. To change the above weight ratios to volume
ratios, full densities of the ␤-silicon nitride and the sili-
con oxynitride were calculated from the rule of mixture
of ␤–Si N , Y O and Al O , and Si N O and CeO ,
FIG. 2. XRD patterns of samples after hot pressing at 2123 K for
1
.5 h; peaks with closed circles belong to Si N O and the others to
2 2
␤
–Si N .
3 4
3
4
2
3
2
3
2
2
2
respectively. Comparing the relative intensity of the
FIG. 3. Relative intensity of XRD peaks versus volume percent of
silicon oxynitride. In the reference powder mixtures, the polynomial fit
equation for the data was (vol% ס −0.0078 + 0.8812 × I + 0.9806 ×
R
2
3
FIG. 1. Schematic diagram for the erosion tester.
IR − 0.8605 × IR ).
2
276
J. Mater. Res., Vol. 17, No. 9, Sep 2002
http://journals.cambridge.org
Downloaded: 14 Mar 2015
IP address: 139.86.13.152

D-S. Park et al.: Effect of Si N O content on the microstructure, properties, and erosion of silicon nitride–Si N O in situ composites
2
2
2 2
5
An erosion test was performed at room temperature in
an ambient environment using the gas-blast tester shown
schematically in Fig. 1. The velocity of SiC abrasive par-
ticles with an average size of 131 ␮m was measured at
to have elongated shapes, which were laid down on the
plane perpendicular to the hot-pressing direction like sili-
con nitride grains. Figure 2 suggests that Si N O grains
2
2
had (200) plane perpendicular to the pressing direction.
Since the hot-pressed samples showed highly anisotropic
XRD patterns, samples SO-4, SO-8, and SO-12 were
pulverized to obtain the volume fraction of Si N O.
12 m/s. The impact angles were 45°, 60°, and 90°. The
inner diameter of the nozzle was 3 mm, and the sample–
nozzle stand-off distance was 10 mm. Each test was re-
peated three times, and each run was finished in 15 min
during which 400 g of the SiC abrasive particles were
consumed. Weight loss due to the erosion was measured
with a precision balance that can measure items as light
as 0.01 mg. The erosion rate was obtained by dividing
weight loss of the sample by weight of the erodent (SiC
abrasive particles) consumed. The eroded sample was
observed with a scanning electron microscope equipped
with a field emission gun (FEG-SEM).
2
2
Relative peak intensity (I ) defined by Eq. (1) was ob-
R
tained from the XRD patterns of the pulverized samples.
O
͒
O
͒
I͑
200
111
+ I͑
111
I =
.
(1)
R
O
͒
O
␤
͒
␤
͑
101
͒
I
+ I
+ I
+ I
͑
200
͑
͒
͑
210
O
O
␤
␤
I(200) , I(111) , I(210) , and I(101) in Eq. (1) are intensities
of the (200) and (111) peaks of Si N O and the (210) and
2
2
(
101) peaks of ␤–Si N , respectively. The Si N O con-
3 4 2 2
tent was obtained from comparison of the relative peak
III. RESULTS AND DISCUSSION
intensities (I ) of the samples and the calibration curve
R
Figure 2 illustrates the XRD patterns of the samples
obtained from the surfaces normal to the pressing direc-
tion except the top pattern, which is from the surface
parallel to the pressing direction of sample SO-12. The
anisotropy of hot-pressed silicon nitride has been well
from the reference powder mixtures (Fig. 3). The Si N O
2
2
contents of samples with 4, 8, and 12 wt% SiO were 13,
2
31, and 54 vol%, respectively.
Figures 4(a)–4(d) display FEG-SEM micrographs of
the samples after plasma etching. Sample SO-0 consisted
of elongated silicon nitride grains embedded in a glassy
intergranular phase. Figures 4(b)–4(d) illustrate that the
silicon nitride grains were etched deeply while the glassy
phase and the Si N O grains were etched minimally.
11
reported, and consequently will not be explained here.
The XRD patterns from surfaces perpendicular and par-
allel to the pressing direction had strong (200) and (111)
Si N O peaks, respectively. Si N O grains were reported
2
2
2
2
2 2
FIG. 4. SEM micrographs and EDS results of the surface normal to
hot pressing direction of the samples after plasma etching: (a) SO-0,
(b) SO-4, (c) SO-8, and (d) SO-12.
J. Mater. Res., Vol. 17, No. 9, Sep 2002
2277
http://journals.cambridge.org
Downloaded: 14 Mar 2015
IP address: 139.86.13.152

D-S. Park et al.: Effect of Si N O content on the microstructure, properties, and erosion of silicon nitride–Si N O in situ composites
2
2
2 2
Figure 5(a) shows a SEM micrograph of the sample SO-
2 with a higher magnification than that shown in
the light region shown in Figs. 5(b) and 5(c). The dark-
region consisted of Si N O, while the light region was in
1
2
2
Fig. 4(d). The areas with dark and light contrasts are
indicated by arrows. Energy dispersive spectroscopy
a glassy grain boundary phase. Figure 6 shows the width
distributions of silicon nitride grains. The average grain
width decreased as more silica was added to the sample,
up to 8 wt%. This result is, in part, due to the fact that
more Si N was consumed in the making of Si N O. It is
(
EDS) analyses on the two areas indicated that yttrium
was not detected in the dark region, but it was detected in
3
4
2 2
probable that ␤-silicon nitride grains were trapped by
rapidly growing Si N O grains and were dissolved in
2
2
12
the interfacial liquid as reported by Wang et al. Also, the
growth of Si N O grains might have inhibited the growth
2
2
of silicon nitride grains. It is not clear at this point why
the average width of ␤-silicon nitride grains in sample
SO-12 was similar to that in sample SO-8 despite the
latter sample having a smaller amount of Si N O.
2
2
Figure 7 shows Vickers hardness and the fracture
toughness of the samples. The hardness was initially de-
creased from 15 to 14.1 GPa and remained constant at
about 14.1 GPa when the addition of silica was increased
FIG. 6. Width distribution of silicon nitride grains of the samples; the
average widths of the grains of samples SO-0, SO-4, SO-8, and SO-12
were 0.87, 0.53, 0.47, and 0.47 ␮m, respectively.
FIG. 5. SEM micrograph of sample SO-12 and EDS result: (a) SEM
micrograph, (b) EDS result from the “dark” area in (a), and (c) EDS
result from the “light” area in (a).
FIG. 7. Vickers hardness (Hv1 kgf) and the fracture toughness of the
samples; the hardness was measured with a load of 9.8 N and
the fracture toughness with 196 N.
2
278
J. Mater. Res., Vol. 17, No. 9, Sep 2002
http://journals.cambridge.org
Downloaded: 14 Mar 2015
IP address: 139.86.13.152

D-S. Park et al.: Effect of Si N O content on the microstructure, properties, and erosion of silicon nitride–Si N O in situ composites
2
2
2 2
from 0 to 12 wt%. Meanwhile, the fracture toughness
the grains to cause the intergranular cracks followed by
grain dislodgement, as reported for wear of alumina.
1
/2
14
was increased steadily from 5.01 ± 0.12 MPam for the
1
/2
SO-0 sample to 5.84 ± 0.09 MPam for the SO-12
sample when the addition of silica was increased from 0
to 12 wt%. In other words, the fracture toughness was
increased as the sample contained more Si N O, even
Figures 9(a) and 9(b) illustrate SEM micrographs of the
eroded areas of sample SO-0. When the impact angle was
low, grain pull-out occurred as indicated by the arrow in
Fig. 9(a). Even though it is not clear in Fig. 9(a), inter-
granular cracking was suspected to precede the grain
pull-out. As the impact angle was increased, micro-
chipping due to transgranular fracture also occurred as
indicated by the arrow in Fig. 9(b). Meanwhile, Fig. 10
shows a SEM micrograph of the eroded area of sample
SO-12. Although the impact angle was 90°, scratch
marks were observed, and micro-chipping due to trans-
granular fracture was not found in the eroded area.
Figures 8, 9(b), and 10 summarize the following: at low
fracture toughness, the erosion occurred primarily by
the fracture mechanism, and at high fracture toughness, it
occurred by the scratching mechanism.
2
2
though the size and amount of the elongated ␤–Si N
3
4
grains were decreased. The toughening effect of Si N O
2
2
has been reported for the in situ composite of ␤–Si N
with equiaxial grains. Young’s moduli of samples SO-0,
3
4
6
SO-4, SO-8, and SO-12 were 310, 291, 283, and
1
3
2
68 GPa, respectively. According to Ohashi et al.,
Young’s modulus of Si N O was 239 GPa. Young’s
2
2
moduli of the three samples, i.e., SO-4, SO-8, and SO-12,
were very close to the values predicted by the rule of
mixture of ␤–Si N (sample SO-0) and Si N O.
3
4
2 2
Figure 8 displays the erosion rates of the samples ac-
cording to the impact angle. There are two reported
1
mechanisms for the erosion of ceramics; first, the frac-
ture mechanism and second, the scratching mechanism.
The former mechanism produced a high erosion rate
while the latter mechanism produced a low rate. Usually,
erosion of hard and brittle ceramics occurs by the fracture
mechanism. However, in the case where the impact angle
was low, scratching occurred simultaneously and the ero-
sion rate was decreased. The erosion mechanism was
also affected by the fracture toughness of the ceramics.
When the fracture toughness of the ceramics was high,
1
scratching was reported to replace microfracture. The
fracture mechanism produced a high erosion rate in
the low fracture toughness sample SO-0. As the fracture
toughness of the sample was increased, the erosion rate
was decreased, and scratching as well as grain pull-out
occurred. Scratching represents a plastic deformation
caused by the impingement of the erodents. In some
cases, the plastic deformation can be accumulated within
FIG. 9. SEM micrographs of eroded sample SO-0. Impact angles were
(a) 45° and (b) 90°. The arrows indicate the locations of (a) grain
pull-out and (b) micro-chipping due to transgranular fracture.
FIG. 8. The erosion rates of the samples according to the impact
angle.
J. Mater. Res., Vol. 17, No. 9, Sep 2002
2279
http://journals.cambridge.org
Downloaded: 14 Mar 2015
IP address: 139.86.13.152

D-S. Park et al.: Effect of Si N O content on the microstructure, properties, and erosion of silicon nitride–Si N O in situ composites
2
2
2 2
volume percent of Si N O was increased, the fracture
2
2
toughness of the composite increased and the width of
the silicon nitride grain decreased. Elastic moduli of the
in situ composites were close to the values predicted by
the rule of mixture of the two constituents, ␤–Si N and
3
4
Si N O. The erosion rate decreased as the Si N O con-
2
2
2 2
tent was increased in part due to the increased fracture
toughness and the reduced grain size. Erosion of the
composites occurred primarily by grain dislodgment. At
a high impact angle, micro-chipping due to severe trans-
granular fracture also occurred in the sample without
Si N O.
2
2
ACKNOWLEDGMENT
This work was supported by the National Research
Laboratory program of the Korean Ministry of Science
and Technology.
FIG. 10. SEM micrograph of eroded sample SO-12. The impact angle
was 90°. The arrow indicates the scratch mark.
The erosion rates of the samples shown in Fig. 8 can
also be explained in terms of their grain sizes. From the
target’s perspective, the impacts by the erodents are simi-
lar to the repeated contacts in the contact fatigue. As
REFERENCES
1
. S. Wada, in Erosion of Ceramic Materials, edited by J.E. Ritter
(Key Engineering Materials, Trans Tech Publications, Uetikon-
Zuerich, Switzerland, 1992), Vol. 71, p. 51.
1
5
2. Y. Zhang, Y-B. Cheng, and S. Lathabai, J. Eur. Ceram. 21, 2435
2001).
reported by Lawn, the intergranular crack density in the
contact zone of the flat sample increased as the grain size
of the sample was increased. Similarly, more intergranu-
lar cracks were generated by the impact of the erodents,
and the erosion rate increased as the grain size was in-
creased. Elastic modulus may affect the erosion rate as
(
3. B. Lawn, in Fracture of Brittle Solids, 2nd ed. (Cambridge Uni-
versity Press, London, United Kingdom, 1993), pp. 303–304.
4. D-M. Liu, J-T. Lin, and R.R-R. Lee, Ceram. Int. 24, 217 (1998).
5. M. Ohashi, S. Kanzaki, and H. Tabata, Nippon Seramikkusu
Kyokai Gakujutsu Ronbunshi, 96, 1073 (1988).
1
6. H. Emoto, M. Mitomo, C-M. Wang, H. Hirosturu, and T. Inaba,
appeared in one of the proposed equations. However,
J. Eur. Ceram. Soc. 18, 527 (1998).
the erosion rate was dependent on the ratio of the elastic
modulus to the hardness in that equation, and the differ-
ence among those ratios of the current in situ composites
was not big enough to explain the difference among the
measured erosion rates.
7
. S.M. Wiedhorn and B.J. Hockey, J. Mater. Sci. 18, 766 (1983).
8. S. Wada, J. Ceram. Soc. Jpn. 104, 247 (1996).
9. A.G. Evans and E.A. Charles, J. Am. Ceram. Soc. 59, 371 (1976).
0. D-S. Park, T-W. Roh, B-D. Han, H-D. Kim, and C. Park, J. Eur.
Ceram. Soc. 22, 535 (2002).
1
1
1
1
1
1. F. Lee, M.S. Sandlin, and K.J. Bowman, J. Am. Ceram. Soc. 76,
1
793 (1993).
2. C. Wang, H. Emoto, and M. Mitomo, J. Am. Ceram. Soc. 81,
125 (1998).
3. M. Ohashi, K. Nakamura, K. Hirao, M. Toriyama, and S. Kanzaki,
Ceram. Int. 23, 27 (1997).
IV. CONCLUSION
1
The XRD pattern of hot-pressed silicon nitride–
Si N O in situ composite exhibited strong (200) and
2
2
4. S.J. Cho, B.J. Hockey, B.R. Lawn, and S.J. Bennison, J. Am. Ce-
(
111) Si N O peaks from the surface perpendicular
2 2
ram. Soc. 72, 1249 (1989).
and parallel to the pressing direction, respectively. As the
15. B.R. Lawn, J. Am. Ceram. Soc. 81, 1977 (1988).
2
280
J. Mater. Res., Vol. 17, No. 9, Sep 2002
http://journals.cambridge.org
Downloaded: 14 Mar 2015
IP address: 139.86.13.152