Heinrich J.G., Aldinger F. (Eds.) Ceramic Materials and Components for Engines
Подождите немного. Документ загружается.
Table 3. Grain boundary phases found after performing
dynamic fatigue tests on notched specimens at room
Material
SO2-rich
Y203-rich
temperature
Surface ITemperd ITemperd
1
Surface
Y~S~OS,
(Si04)
30,
Y2Si207,
SiO2
Y2SiOs,
Y2Si207,
Si02
y4.67
I
I
surface
I
below
100
pm
Table
4.
Grain boundary phases found after performing
n notchec
100
pm
below
surface
Y2SiOs,
(Si04)
30,
Y2Si207
Y2SiOs,
Y2Si207
y4.67
below
surface
SO2
The polished and etched cross sections of the tempered
specimens are shown in Figures 9 and 10. The thickness
of the oxidation layer for the SiO2-rich material is ca.
5
pm and for the Y203-rich material, ca. 10 pm.
Figure 9. Polished and etched cross section of the Si02-
rich material after tempering for 100 h in
air
at 1500
"C
Figure 10. Polished and etched cross section of the
Y203-rich material after tempering for
100
h in air at
1500
"C
DISCUSSION
General Differences Between Y203-Rich
and
Si02-Rich Materials
The two materials
had
different grain boundary
compositions and liquidus temperatures (ca. 1820
"C
for
the Y203-rich material and ca. 1940
"C
for the SiO2-rich
material). This is believed to have led to a lower
strength and higher fracture toughness at room
temperature for the SiO2-rich material than for the
Y203-rich material.
With regard to the dynamic fatigue strength, differences
were found between the Y203-rich and the Si02-rich
materials, independent of the flaw type. The Y203-rich
specimens generally failed at higher stresses than the
SiO2-rich specimens,
as
shown
in
Tables
1
and
2.
At
1500
"C,
slow crack growth was observed in the Si02-
rich specimens, whereas typical fast fracture markings
were observed in the Y203-rich specimens. Thus, slow
crack growth during high-temperature testing appears to
be associated with low dynamic fatigue stresses and
with the Si02-rich materials.
The steady-state creep rate for the Y203-rich material
(3.5 x h-')
was
comparable to that of the SO2-rich
material (2.9
x
10" h-' for a material with the same
composition but a lower grain boundary volume fraction
of
0.07;
for a
grain
boundary volume fraction of 0.12,
the steady-state creep rate should be somewhat higher,
but still similar to that of the Y203-rich material); this
indicates that creep probably did not play a significant
role in strengthening the Y203-rich material during
dynamic fatigue testing.
X-ray difhction analysis of notched specimens (see
Tables
3
and
4)
showed that for both test temperatures,
the two materials had different grain boundary phases
on
and 100 pm below the surface, but the same grain
boundary phases on the surface after tempering. 100 pm
below the surface of the tempered materials, however,
different grain boundary phases were found for the two
142
materials. Figures
9
and 10 confm that the Y203-rich
material was oxidized to a greater extent than the Si02-
rich material, but that neither material suffered
significant oxidation damage after being tempered for
100 h in
air
at 1500
"C.
It is assumed that differences in
the condition and grade of crystallinity of the grain
boundaries are mainly responsible for the differences
observed in the two materials.
Dynamic Fatigue Testing
of
Specimens with
Natural Flaws
It can be seen that only at 1500 "C were there
significant differences between the results (see Table 1).
At 1500
"C,
the dynamic fatigue strengths of the Y2O3-
rich specimens were much higher than those of the
SO2-rich specimens. This is probably due to the
differences in the condition and grade of crystallinity of
the grain boundaries. The strengths at 1500
"C
were
lower than those at room temperature, especially for the
SO2-rich material. This is probably related to the
softening of the glassy phase at 1500
"C.
Tempering caused the strength to increase for the Si02-
rich material and to decrease for the Y203-rich material,
especially at 1500 "C. The SO2-rich material was not
oxidized significantly during the tempering experiments
(see Figure
9,
showing an oxidation layer of ca. 5 pm).
The fracture origins for the specimens tested at room
temperature were surface defects; after tempering, the
fracture origins were surface regions rich in Si02 and Y.
At 1500 "C, the fracture origins were surface defects
(see Figure 1); after tempering, the fracture origins were
defects in the interior of the specimens (see Figure 2),
the oxidation layer having effectively healed the surface
defects. Thus, tempering had a positive effect on the
dynamic fatigue strength, especially at 1500
"C.
None
of the typical markings associated with fast fracture
(i.e., fracture mirror, mist, and hackle) could be
observed on the fracture surfaces after testing at 1500
"C.
The Y203-rich material was oxidized to a much greater
extent (see Figure 10, showing an oxidation layer of ca.
10 pm). At room temperature, the fracture origins were
surface defects such
as
large Sic
grains;
after
tempering, the fracture origins were still surface defects,
but were caused by build-up of the oxidation products.
At 1500
"C,
the fracture origins were either large
particles or regions enriched in the grain boundary
phase at or close to the surface (see Figure
3);
after
tempering, the fracture origins were generally regions
consisting
of
Si02
and
large Y2SizO7 crystals close
to
the surface (see Figure
4).
Thus, tempering had a
negative effect on the dynamic fatigue strength,
especially during testing at 1500
"C.
Oxidation of the
grain boundary phase in the top 10
-
30
pm
of the bulk
occurred.
Dynamic Fatigue Testing
of
Specimens with
an Edge Notch
Significant differences between the results could be
found, especially at 1500
"C
(see Table 2). The dynamic
fatigue strengths of the Y203-rich specimens were much
higher than those of the SiO2-rich specimens at 1500 "C,
whereas they did not differ considerably from those of
the SiO2-rich specimens at room temperature. At 1500
"C, flow andor oxidation of the grain boundaries may
occur to hinder crack propagation. Since oxidation is
much more and flow is slightly more pronounced in the
Y203-rich material than in the Si02-material, they may
be responsible for the differences observed at 1500
"C.
For materials with an edge notch, hgher strengths were
obtained at 1500 "C than at room temperature and at
both temperatures after tempering for 100 h at 1500 "C
in air. It is assumed that processes involving blunting of
the notch tip are responsible for the observed behaviour.
X-ray diffraction analysis (see Tables
3
and
4)
showed
that after testing at 1500
"C,
only additional crystobalite
was present on the surface for the SO2-rich materials;
for the Y203-rich materials, crystobalite was present and
Y2O3 was absent. After tempering, the crystalline grain
boundaries found on and 100 pm below the surface after
testing at 1500 "C were identical to those found after
room-temperature testing. Thus, for the tempered
specimens, no significant changes detectable by X-ray
diffraction occurred during dynamic fatigue testing at
1500 "C. The 25% decrease in strength for the SO2-rich
material and
the
50%
increase in strength for the YZO3-
rich material might be related to softening of the glassy
phase during testing at 1500
"C.
The edge notch changes the behaviour in the following
ways:
1.
2.
3.
The Si02-rich material underwent extensive slow
crack growth during testing at 1500
"C;
the
dynamic fatigue strengths for hot-pressed and
tempered specimens were correspondingly low.
Tempering causes the fracture stress to increase for
Y20-rich and for Si02-rich materials, especially at
room temperature. Since the amount of
microstructural damage caused by tempering is the
same for specimens with and without an edge
notch, this increase must be due to blunting of the
crack tip by tempering.
The fracture stress at
1500
"C
is higher than at
room temperature for SiO2-rich (see Figures 5 and
6)
and YzO3-rich hot-pressed materials (see Figures
7
and
8)
and for Y203-rich tempered materials. The
increase may be due to blunting of the crack tip and
lowering of the stress intensity factor by oxidation
or flow of the grain boundaries during high-
temperature testing. However, other mechanisms
may account for
the
increase in dynamic fatigue
strength at 1500 "C. Changes in the amounts of
glassy phase, crystalline phases and microstructural
143
damage in the grain boundaries due to oxidation
may also have an effect on the dynamic fatigue
behaviour.
Conclusions
The dynamic fatigue behaviour of two materials, one
rich in Si02 and the other rich in Y2O3, was examined at
room and at elevated temperatures for unnotched and
notched specimens. At room temperature, no differences
were observed. At 1500 "C, however, significant
differences were observed, regardless of defect type.
The SiO2-rich specimens exhibited less oxidation, more
slow crack growth and lower dynamic fatigue strengths
than the Y203-rich specimens did. Additionally, for
notched specimens, the dynamic fatigue strengths were
higher at 1500 "C than at room temperature and at both
temperatures after tempering. It is assumed that
processes, such
as
oxidation, involving blunting of the
crack tip are responsible for the observed behaviour.
However, further studies must be performed in order to
explain the high-temperature behaviour of these
materials. Not only oxidation, but other mechanisms
must be active at 1500 "C to cause the differences which
were seen especially in the tempered materials between
room temperature and 1500 "C. Dynamic fatigue testing
could be performed in an inert atmosphere to remove
the effects of oxidation. Additionally, the crystallization
behaviour of the two materials could be tested using
techniques such
as
differential thermal analysis, TEM
and quantitative analysis
of
the X-ray diffraction results.
REFERENCES
[I]
Hecht, N.L., Goodrich, S.M., Chuck,
L.,
McCullum,
D.E., Tennery, V.J., Mechanical Properties
Characterization
of
One Sic and Two Si3N4
Commercially Available Ceramics, Ceramic Bulletin 7
1
[4], 1992, 653.
[2] Lee, J.-K., Tan&, H., Kim, H., Movement of liquid
phase and the formation
of
surface reaction layer on the
sintering of p-Sic with an additive of yttrium
aluminium garnet, Journal of Materials Science Letters
15, 1996,409.
[3] Falk, L.K.L., Microstructural Development during
Liquid Phase Sintering
of
Silicon Carbide Ceramics,
Journal of the European Ceramic Society 17,1997,983.
[4] Mulla, M.A., Krstic, V.D., Pressureless sintering
of
p-Sic with A1203 additions, Journal of Materials
Science 29, 1994,934.
[5] Mulla, M.A., Krstic, V.D., Low-Temperature
Pressureless Sintering
of
p-Silicon Carbide with
Aluminum Oxide and Yttrium Oxide Additions,
Ceramic Bulletin 70 [3], 1991,439.
[6]
Lee, J.K., Kang, H.H.,
Kim.
Y.J., Lee, E.G., Kim,
H., Effects of YAG-Phase Amount on the
Microstructure and Phase Transformation during the
Liquid-Phase Sintering of p-Sic, Key Engineering
Materials Vols. 161-163, 1999,263-266.
[7] Nader, M., Aldinger, F., Hoffmann, M.J., Influence
of the alp-Sic phase transformation on microstructural
development and mechanical properties of Iiquid phase
sintered silicon carbide,
J.
Mater. Sci. 34, 1999, 1197-
1204.
[8]
Kim, Y.-W., Kim,
J.-Y.,
Rhee, S.-H., Kim, D.-Y.,
Effect of initial particle size on microstructure of liquid-
phase sintered a-silicon carbide, Journal
of
the
European Ceramic Society 20,2000,945-949.
[9] Winn, E.J., Clegg, W.J., Role of the Powder Bed in
the Densification of Silicon Carbide Sintered with Yttria
and Alumina Additives,
J.
Am. Ceram. SOC. 82 [12],
1999,3466-70.
[lo] Kim, Y.-W., Mitomo, M., Fine-Grained Silicon
Carbide Ceramics with Oxynitride Glass, J.
Am.
Ceram.
SOC. 82
[lo],
1999,2731-36.
[ll] Zhan, G.-D., Mitomo, M., Kim, Y.-W.,
Microstructural Control for Strengthening of Silicon
Carbide Ceramics, J. Am. Ceram. SOC. 82 [lo], 1999,
2924-26.
[12] Kim,
S.,
Kriven, W.M., Preparation,
Microstructure, and Mechanical Properties
of
Silicon
Carbide-Dysprosia Composites, J. Am. Ceram. SOC. 80
[
121, 1997,2997-3008.
[I31 Tanaka, I., Kleebe, H.-J., Cinibulk, M.K., Bmley,
J., Clarke, D.R., and Riihle, M., Ca Concentration
Dependence
of
the Equilibrium Thickness
of
the
Intergranular Film in Si3N4, J. Am. Ceram. SOC. 77 [4],
1994,911.
[14] Quinn, G.D., Fracture Mechanism Maps for
Advanced Structural Ceramics Part 1 Methodology and
Hot-Pressed Silicon Nitride Results,
J.
Mater. Sci. 25,
1990,4361.
[15] Wang, C.-M., Pan, X., Hoffinann, M.J., Cannon,
R.M., Riihle, M., Grain Boundary Films in Rare-Earth-
Glass-Based Silicon Nitride, J.
Am.
Ceram.
Soc.
79
[3],
1996,788-92.
[I61 Wolf, C., Hubner,
H.,
Adler, J., Mechanical
Behaviour of Pressureless Sintered Sic at High
Temperature, pp. 465-470 in Third Euro-Ceramics, Vol.
3. Edited by P. Duran and J.F. Femandez, 1993.
[17] Jou, Z.C., Virkar, A.V., Cutler, R.A., High
temperature creep of Sic densified using a transient
liquid phase, J. Mater. Res. 6 [9], Sep 1991, 1945.
[18] Keppeler, M., Reichert, H.-G., Broadley, J.M.,
Thurn, G., Wiedmann, I., Aldinger, F., High
144
Temperature Mechanical Behaviour of Liquid Phase
Sintered Silicon Carbide, Journal of the European
Ceramic Society 18, 1998, 521-526.
[19] Kinoshita, T., Munekawa,
S.,
Tanaka,
S.-I.,
Effect
of
Grain Boundary Segregation on High-Temperature
Strength
of
Hot-Pressed Silicon Carbide, Acta mater.
Vol. 45, NO. 2, 1997,801-809.
[20] Sajgalik, P., Hnatko, M., Lofaj, F., HvizdoS, P.,
Dusza, J., Warbichler, P., Hofer, F., Riedel,
R.,
Lecomte,
E.,
Hoffmann,
M.J.,
SiC/Si3N4 nano/micro-
composite
-
processing, RT and HT mechanical
properties, Journal of the European Ceramic Society 20,
2000,453-462.
[21] Dusza, J., Sajgalik, P. Steen, M., Fracture
Toughness
of
a Silicon Nitride/Silicon Carbide
Nanocomposite at 1350 "C,
J.
Am. Ceram. SOC. 82 [12],
1999,3613-3615.
[22] Cheong, D.-S., Hwang, K.-T., Kim, C.-S., High-
Temperature Strength and Microstructural Analysis in
Si3N4/20-vol%-SiC Nanocomposites,
J.
Am. Ceram.
SOC. 82
[4],
1999,981-86.
[23] Rouxel, T., Wakai, F., Sakaguchi,
S.,
R-Curve
Behaviour and Stable Crack
Growth
at Elevated
Temperature (1500 "C
-
1650 "C) in a Si3N4/SiC
Nanocomposite,
J.
Am. Ceram. SOC.
77
[12], 1994,
3237-43.
145
This Page Intentionally Left Blank
THERMAL SHOCK PROPERTIES
OF
SIALON CERAMICS
Pernilla Pettersson*, Zhijian Shen, Mats Johnsson, and Mats Nygren
Department
of
Inorganic Chemistry, Stockholm University,
S-106
91
Stockholm, Sweden
ABSTRACT
An indentation-quench method based on Vickers
cracks for measuring thermal shock properties has been
applied to different sialon materials. For sialons based
on the P-phase, Si6-,AI,0,Ng.,
(0
I
z I4.2), the best
thermal shock resistance was found at low
z
values.
The thermal shock properties could be further
improved by adding an intergranular yttrium
containing glass phase. For sialons consisting of a
mixture of the
a-
(YxSil~+,,+n~Al,,,+nOnN16n)
and the
p-
sialon their thermal shock resistance improved with
increasing amount of P-phase and adding an
intergranular glass phase improved the thermal shock
resistance.
INTRODUCTION
a-
and
p-
sialons are
two
solid solution phases with
compositions and crystal structures close to
a-
and
p-
Si3N4 respectively. p-sialon is represented by the
general formula Si6-zAl,0,Ns.Z, with
0
<
z
I
4.2.
a-sialon is represented by MxSi~Z.(m+n~lm+nOnN~~n
where M represents metal ions (most often yttrium or
rare earths) and x, m, and n are parameters that restricts
the formation of a-sialon to small composition areas
[
11. An intergranular glass phase can be added in order
to assist sintering.
Sialon materials based on the
a-
and p-sialon
phases or mixtures thereof are attractive for high
temperature applications. However, at the moment
there is little experimental data available in the
literature concerning their thermal shock behaviour
i.
e.
the effect of thermal cycling on mechanical properties
of sialon based materials.
In this article we present a qualitative investigation
of the thermal shock properties of
a-
and
p-
sialon
materials. An indentation-quench method based on
Vickers cracks for measuring thermal shock properties
[2] has been applied and evaluated.
For P-sialon materials the thermal shock properties
have been correlated with both different z-values
within the solid solution range and the amount of
residual intergranular glass phase present. For
dp-
sialons the thermal shock properties are correlated with
different ratios between the
a-
and the P-phase and
with addition of increasing amount of intergranular
glass phase. For those materials the a-phase
(YxSiiz.(m+n,AIm+nOnNI,n)
composition was fixed at
x
=
0.33, m
=
1.0, and n
=
1.2 while the P-phase was
fixed at z
=
0.6.
We have kept the sample size constant in order to
make it possible to compare the thermal shock
properties for the different sialon materials.
EXPEFUMENTAL
Three different types
of
sialon ceramics were
designed and prepared, see Table I. The first type
consisted of pure p-sialons (Si6.,A1,O,Ns-,) with
z-
values in the range 0.6 to 3.0. In the second ones the
same 0-sialon compositions were prepared but with
an
addition of intergranular glass phase. The added glass
phase had the nominal overall composition
YI.~~S~Z.~Z~A~~.O~~.~NI.Z~
(28 Y; 56 Si; 16 Al;
80
0;
20
N
in equivalent
%
[I]). The third type of sialon material
consists of mixed alp-sialons with various amounts
of
glass added. For those materials the a-sialon
composition was fixed at
Y0.~3Si~.~A1~.20~.2N~4.g
(x
=
0.33, m
=
1.0
and n
=
1.2); the P-sialon composition
was Si5.4A10.600.6N7.4 (z
=
0.6)
and the glass phase
composition was designed to have the same
composition as for the p-sialons.
Table
I
List of prepared compositions. The
different materials are referred to in the text as
samples A to
Y,
GP
=
vol
%
glass phase.
I
I
I
Specimens were prepared from commercial Si3N4
(UBE, SN-ElO), AlN (H.C. Starck-Berlin, grade A),
YzO3
(99.9
%,
Johnson Matthey Chemicals Ltd.),
A1203 (Alcoa, A16SG), and SiOz
(99.9
YO,
<325 mesh,
Johnson Matthey Chemicals Ltd.), and corrections
were made for the amounts
of
oxygen present in the
Si3N4 and AlN raw materials. The starting-material
147
mixtures were milled in water-free propanol for 24
hours in a plastic jar, using sialon-milling media. The
dried powder mixtures were hot pressed in order to
ensure full density, at a temperature of 1750 "C and a
pressure of 30 MPa, in nitrogen atmosphere, for 60
min. However, two monophasic samples (B and
C)
had
to be hot-isostatically pressed at a temperature of
1820 "C and a pressure of 200 MPa for 60 min in order
to become fully dense, due to the complete or nearly
complete lack of intergranular glass phase. Samples A,
0
and P were densified by using spark plasma sintering
technique (SPS) at
1700
"C,
50
MPa for
3
min.
C
haracterisation
The densities of the sintered specimens were
measured according to Archimedes' principle. Before
the mechanical and microstructural studies, the
specimens were carefully polished by standard
diamond polishing techniques. The hardness (HVlo)
and indentation fracture toughness (KI,) at room
temperature were determined by means of a Vickers
diamond indenter with a 98 N (10 kg) load, and the
fracture toughness was evaluated according to the
method of Anstis
et
al.
[3] assuming a value of
300 GPa for Young's modulus for all compositions.
Five indentations were made on each sample for the
hardness measurements. The microstructures of the
samples were investigated in a scanning electron
microscope (SEM, JEOL 880) equipped with an
energy-dispersive spectrometer (EDS, LINK
ISIS)
that
allows detection of boron and heavier elements.
Micrographs of fractured samples were taken in
secondary electron mode (SE). To obtain the best
contrast between different phases on polished surfaces,
the micrographs were recorded in back-scattered
electron mode (BSE) at an acceleration voltage of 20
kV. The amounts of intergranular glass phase were
evaluated with an image-analysing package supplied
with the LINK
ZSIS
system. The contrast difference
between the sialon grains (a and
p)
and the yttrium-
containing glass phase was accounted for. The
estimates
of
the minimum and maximum amount of
glass phase gave an error of +2%.
A focusing X-ray powder diffraction camera of
Guinier-Hagg type with CuLI radiation
(A
=
1.5405981
A)
was used to record the X-ray
powder diffraction
(XRD)
patterns, and powdered
silicon
(a
=
5.430879
A
at 25°C) was added
as
internal
standard. The computer programs SCANPI
[4]
and
PIRUM [5] were used to evaluate the recorded films.
The latter program was also used to determine and
refine the unit cell parameters. Experimental z-values
of the p-sialon phase were obtained from the unit-cell
dimensions of the samples, using the equations given
by Ekstrom
et
al.
[6].
The amounts of
a-
and p-sialon
were estimated by comparison of the intensities of the
two strongest XRD peaks of each phase (1
0
2 and
2 1
0
for a-sialon and 1
0
1 and 2 1
0
for p-sialon).
Thermal
shock
measurements
Cylindrical samples with a diameter of 12
mm
and
a thickness of 3.76
+
0.40 mm with parallel surfaces,
were carefully polished on one side. Well-defined
cracks were initiated with a Vickers indenter (Instron,
model 856 1, High Wycombe, England) allowing
variation of the load and the loading-, holding- and
unloading time. The loading and unloading times were
10
s
each, and the holding time at maximum load was
20
s.
Four indents were made on each sample, and each
indent generated four cracks defined as the distance
from the centre of the indent to the crack tip (see
Figure l),
so
that a total of sixteen cracks were initiated
on each sample. The crack lengths were measured in an
optical microscope (Olympus PMG3, Japan). In order
to compare different samples more easily, the original
crack length was held constant,
e.g.
around 100 pm.
As
a consequence, different samples had to be indented
with different. loads (see Table 11).
A
vertical tubular
furnace was heated to the pre-set starting temperature,
initially 190 "C, which was then normally increased in
steps of 100 "C. The sample was hoisted into the
furnace where it was thermally equilibrated for 20 min
and was then quenched in a 90 "C water bath.
I
Figure
1
Definition of the crack length
used in thermal shock measurements.
The test pieces were heated in air, but in the
experiments presented below the temperature were not
raised above 1000 "C, in order to avoid oxidation. A
90 "C water bath was used as quenching media. This
temperature was kept constant with a thermostat. The
quenching procedure was repeated from stepwise
increasing heating temperatures. The crack growth was
measured at each temperature step and the total
percentage growth was calculated. The critical thermal-
shock temperature (AT,) was defined according to
Anderson and Rowcliffe [2], as the temperature where
25%
of
the cracks have grown more than 10% of their
original length.
For sample M, a comparison was made between the
re-use of the same sample for
a
whole quenching series
and the use of a fresh sample for each new quenching
temperature.
148
Sample
*
The amounts of
a-
and B-sialon were estimated fiom comparison between the intensities of the
two
strongest
XRD-
Thermal shock Mechanical properties
measurements
Load Initial AT,
K1c
HV
Glass Density
z
value
d(a+p)*
(N)
crack
(“C)
(MPa*m”) (GPa) (~01%) (g/cm3)
length
-
lines for the
two
phases respectively (1
0
2 and
2
1
0
for a-sialon and 1
0
1 and
2
1
0
for p-sialon).
**
These samples contain the B-phase, implying that no (or only a small amount
of)
a-sialon phase has formed.
RESULTS
Microstructural evaluation
Measured densities are listed in Table
11.
All
samples were according to SEM studies hlly dense
and the obtained densities for the p-sialons were also
comparable to previously reported ones
[6].
SEM micrographs are given in Figure 2a-d for
some selected materials representing the different
materials investigated: p-sialon, p-sialon
+
glass,
alp-
sialon, and alp-sialon
+
glass. The morphology of the
materials without additional glass phase added is
equiaxed. The materials containing additional glass
phase show elongated grains, in accordance with
previous studies of
a-
and p-materials
[7].
When no extra glass phase is added the p-sialons
with z
5
1.5
fracture intergranularly, while the more
Al-rich p-sialons (z
>
1.5)
fracture transgranularly.
Thus the strength of the P-grains decreases with
increasing Al-content.
Besides the
a-
and P-phases also Y2SiA105N was
identified in some samples of the alp-sialons with an
additional glass content (samples T,
U,
and Y). This
phase is
known
as the B-phase, which most often
crystallises from oxynitride glass in yttrium containing
sialons
[8].
The presence of this phase is due to the
difficulties in designing the composition of the liquid
that is in equilibrium with the
a-
and P-phases at the
sintering temperature. For samples T,
U,
and Y it thus
seems that some of the yttrium designed for the
a-
sialon phase has been used up in the glass formation
during sintering.
The unit cell dimensions of the a-phase were the
same in all samples indicating that the composition
of
the a-sialon is similar in all samples. The z-values
calculated from the unit cell dimensions of the P-phase
are given in Table
11.
Those values show a slight
variation from the aimed values.
Mechanical properties
The hardness values
(HVIo)
and fracture toughness
(K,,)
for the different samples are given in Table
11.
The z-value does not influence on the hardness for the
p-sialons. Addition of extra glass phase increases the
149
hardness and the fracture toughness somewhat for low
z-values, but the hardness for high-z samples do not
improve with addition of glass.
A
general tendency for the
alp
sialons is that both
hardness and fracture toughness increase slightly with
increasing a-sialon content. The hardness decreases
with increasing glass content, but fracture toughness
increases. The measured hardness as well as the
fracture toughness values are consistent with literature
data
[9].
Al-content in the P-phase makes it easy to fracture
upon thermal cycling.
For the alp-sialons the best thermal shock
resistance was found at a low fiaction of
a.
Increasing
the fraction of a-phase decreases the thermal-shock
resistance. Presence of glass slightly improves the
thermal shock resistance and the best resistance is
found at the highest glass content investigated
(20
vol%), see sample
Y
in Figure
4.
High fracture toughness normally correlates with a
good thermal shock resistance.
The critical thermal-shock temperature (AT,) as
defined by Anderson and Rowcliffe
[2]
is
evaluated.
AT, decreases with increasing z-value both for samples
with and without added glass, see Figure 5a. For low
z-
values the addition of glass has a pronounced effect on
AT,, whereas for high
z the glass has no influence. A
high amount of glass seems to have a more pronounced
effect for the p-sialon than for the alp-sialon, see
Figure 5b. For the
alp
sialons the AT, value decreases
with increasing fraction of a-sialon, see Figure 5c.
0-0
Sample
E
h
v-v
Sample
A
4-A
Sample M
3
5
d60
e
M
Figure 2a-d
SEM
micrographs of
(a)
fractured surface of sample A,
z=O.6
(SE mode).
(b)
polished surface
of
sample
M,
z=O.6,
20
vol% glass. The
p-sialon grains are black or dark grey
and the glass phase is white
(BSE
mode).
(c)
Fractured surface of sample
P,
a/a+p=0.3
(SE
mode).
(d)
Polished
ala+p
=
0.3,
10
%
glass. The p-sialon
grains are black, the a-sialon grains are
grey and the glass phase is white
(BSE
mode).
surface of sample
U,
Thermal
shock
properties
For the p-sialon materials it is clear that samples
with low z-values (low Al-content) are more resistant
to thermal shock than those with higher z-values. For
samples with additional glass added, the thermal shock
resistance is improved for all z-values except for the
sample with the highest z-value investigated, z
=
3.0;
see Table I1 and Figure
3.
It is obviously
so
that a high
"
0
200
400 600
800
1000
AT
("C)
Figure
3
Crack growth in percent plotted
versus AT (see text) for some selected p-sialon
samples.
loo,,
I
f
0-
-0
Sample
R
v-v
Sample
X
0-0
Sample
P
h
40
I
AT
("C)
Figure
4
Crack growth in percent plotted
versus AT for some selected a/p-sialon
samples.
150
For sample
M
(z
=
0.6,
20
%
glass) that show the
best thermal shock resistance a comparison was made
between the re-use of the same sample for a whole
quenching series and the use of fresh samples for each
new quenching temperature. The
two
series gave very
similar results (see Figure
6).
This concludes that the
indentation method used is reliable for qualitative
comparison of thermal shock properties of sialon
materials.
800
1
200
1
0
I
0.5
1
.o
1.5
2.0 2.5
3.0
z
value
1000
(b)
/
800
,
z
=
3.0
-----a
0
5
10
15
20 25
30
Glass
(voIYo)
20
40
60 80
100
d(a+P)
Figure
5a-c
The critical thermal-shock temperature
difference
(AT,)
measured
as
a
function
of
(a)
the
nominal z-value
for
p-sialons
with
and
without
additional glass phase added
(b)
the
measured
amount of intergranular glass
phase
for
p-
and
alp-
sialons
(c)
the
nominal
amount
of
a-sialon
for
alp-
sialons.
15,
I
A
0'
-
0
0
200
400
600
800
1000
AT
(OC)
Figure
6
Crack growth in percent plotted
versus AT for sample
M
(z
=
0.6,
20
%
glass);
I
The same compact has been used at each
quenching temperature.
I1
A new compact has
been used at each quenching temperature.
CONCLUSION
An indentation-quench method based on Vickers
cracks for measuring thermal-shock properties has been
applied to sialon materials. The method allows the use
of the same sample during a series of increasing
temperatures. The percentage crack growth is measured
at each temperature-quenching step, and the statistics
are improved by making several Vickers indents on the
same sample.
For p-sialons (Si6.zAl,0,N8.z) the z- value has been
varied in the range
0.6
I
z
I
3.0
and it is found that the
thermal shock resistance is best at low z-values. The
thermal shock resistance improves even further with
addition of yttrium containing glass phase. However,
for high z-values the thermal shock properties do not
improve by addition of glass.
For the dp-sialons the z-value of the P-phase were
kept at
0.6
and the a-phase
(Y,Si~~~m+n~lm+nOnN~~n,)
had a composition of x
=
0.33,
m
=
1.0,
and n
=
1.2.
The thermal shock resistance improves with increasing
fraction of P-phase. Also for these materials the
thermal shock resistance improves with addition of
glass phase.
High fracture toughness normally correlates with a
good thermal shock resistance.
The best thermal-shock properties were found for a
p-sialon with a z-value of
0.6,
containing
20
vol%
glass (ATc
=
900
"C). The poorest resistance
to
thermal-shock was found for a p-sialon with
z
=
3.0
(ATc
=
100
"C),
and the thermal-shock properties of
p-
sialons with high z-values were not improved by
addition of extra glass phase.
151