Yttrium silicate coatings on chemical vapor deposition-SiC-precoated C/C-SiC: Thermodynamic assessment and high-temperature investigation

Article abstract of DOI:10.1111/j.1551-2916.2005.00077.x

Yttrium silicates are promising materials for improved oxidation and erosion protection for carbon fiber-reinforced composites. A two-layer coating system of low-pressure plasma-sprayed yttrium silicate on chemical vapor deposition-SiC-precoated C/C-SiC was tested under atmospheric re-entry conditions simulated within a plasma wind tunnel test facility. The thermal expansion behavior of Y₂SiO₅ and Y₂Si ₂O₇ was investigated. The chemical compatibility with and without increasing oxygen partial pressure at the interface of the two-layer system was calculated by the CALPHAD method. The calculations were compared with experimental results. Furthermore, a thermodynamic explanation is presented to understand and predict the observed coating failure mechanism, identified as blister formation.

Full text of DOI:10.1111/j.1551-2916.2005.00077.x

J. Am. Ceram. Soc., 88 [2] 424–430 (2005)
DOI: 10.1111/j.1551-2916.2005.00077.x
ournal
J
Yttrium Silicate Coatings on Chemical Vapor Deposition-SiC-Precoated
C/C–SiC: Thermodynamic Assessment and
High-Temperature Investigation
,w
Hans J. Seifert*
Department of Materials Science and Engineering, University of Florida, Gainesville, FL 32611-6400
*
Sigrid Wagner, Olga Fabrichnaya, Hans L. Lukas, and Fritz Aldinger
¨ ¨
Max-Planck-Institut fur Metallforschung and Institut fur Nichtmetallische Anorganische Materialien,
¨
Universitat Stuttgart, Pulvermetallurgisches Laboratorium, 70569 Stuttgart, Germany
Thomas Ullmann
German Aerospace Center, Institute of Structures and Design, 70569 Stuttgart, Germany
*
¨
Martin Schmucker and Hartmut Schneider
German Aerospace Center, Institute for Materials Research, 51147 Cologne, Germany
Yttrium silicates are promising materials for improved oxidation
and erosion protection for carbon fiber-reinforced composites.
A two-layer coating system of low-pressure plasma-sprayed
yttrium silicate on chemical vapor deposition-SiC-precoated
C/C–SiC was tested under atmospheric re-entry conditions sim-
ulated within a plasma wind tunnel test facility. The thermal
expansion behavior of Y SiO and Y Si O was investigated.
The chemical compatibility with and without increasing oxygen
partial pressure at the interface of the two-layer system was
calculated by the CALPHAD method. The calculations were
compared with experimental results. Furthermore, a thermo-
dynamic explanation is presented to understand and predict
the observed coating failure mechanism, identified as blister
formation.
as possible. It is generally accepted that to fulfill all these fun-
damental requirements in a high-temperature oxidizing environ-
ment, a multilayer coating is needed rather than a single-layer
1
–3
coating.
Silicon-based ceramic coatings, like SiC or Si N , are attrac-
3
4
tive materials for oxidation protection at elevated temperatures
for extended time periods because of their thermophysical prop-
erties, as well as the formation of a thin amorphous SiO scale
2
5
2
2
7
2
4
with simultaneous low oxidation rate. The occurrence of mi-
crocracks within the coating, which is a result of thermo-
mechanically induced stress, is problematic and to avoid this
effect it is proposed to introduce single layers or inlays of a glass-
2
,3
forming phase to provide self-healing capabilities. Many mul-
tilayer systems have been developed, containing one or more
glass sealant layers covered with inner and outer ceramic layers.
3
5
In view of applicability as a coating material, Ogura et al.
investigated the thermal properties of monosilicates from lan-
I. Introduction
thanides (Ln
2 5 2 5
SiO ) and yttrium (Y SiO ). It has been demon-
ECAUSE of their mechanical strength at high temperatures
and their thermal shock resistance in combination with low
strated that yttrium monosilicate (Y SiO , MS) has the potential
2
5
B
to be used as an erosion-resistant layer. Therefore, a two-layer
system for C/C composites has been proposed, consisting of an
inner SiC bonding layer and an outer oxidation/erosion protec-
tion layer of yttrium monosilicate. The aim is to design a coating
system that is inherently crack resistant without using expensive
sealing layers, as for example B C applied by chemical vapor
specific mass, carbon fiber-reinforced composites like C/C–SiC
are promising materials for high-temperature structural appli-
cations in gas turbines or heat shields of aerospace re-entry ve-
hicles. In oxidative environments, protective coatings are
required because of the degradation of carbon at temperatures
above 673 K. The coating has to be oxidation resistant and has
to prevent oxygen diffusion inwards as well as carbon diffusion
outwards in order to avoid oxidation or carbothermic reduction.
During the atmospheric re-entry maneuver of a space vehicle,
the low-pressure environment and the surrounding high-velocity
plasma flow have to be considered. Therefore, the coating ma-
terial must possess low volatility to prevent excessive surface
erosion. To provide adequate adhesion between the coating and
the substrate, and mechanical stability of the system, the coef-
ficient of thermal expansion (CTE) of both have to be as similar
6
4
deposition (CVD). Because of the low Young’s modulus of
Y SiO and the high compatibility of the CTEs between SiC and
2
2
5
5
Y
SiO
, the mechanical-induced stress at the Y
SiO –SiC inter-
2 5
5
face is decreasing significantly. On the other hand, the forma-
tion of gaseous oxidation products between the oxide and the
3
,7–9
carbide layers becomes serious with increasing temperature.
To produce dense yttrium silicate coatings, it has been found
that the presence of yttrium disilicate (Y Si , DS) in the outer
In this work, a low-pressure
2
2 7
O
9
,10
yttrium silicate layer is useful.
plasma spray (LPPS) technique is used to produce yttrium sil-
icate coatings on CVD-SiC-precoated C/C–SiC specimens
D. Butt—contributing editor
(
58.4% carbon fiber content), manufactured via liquid silicon
1
1
infiltration. The oxidation and erosion protection efficiency
under simulated atmospheric re-entry conditions are compared
with specimens protected only with a CVD-SiC layer and with
industrial-manufactured multilayer coatings, based on SiC.
During the tests in a plasma wind tunnel facility, at tempera-
Manuscript No. 187370. Received October 26, 2001; approved September 3, 2004.
Supported by the German Research Foundation (SFB259, High Temperature Problems
of Reusable Space Transportation Systems).
*
Member, American Ceramic Society.
w
10
Author to whom correspondence should be addressed. e-mail: hseif@mse.ufl.edu
tures above 1923 K, a sudden coating failure has been
4
24

February 2005
Yttrium Silicate Coatings on CVD-SiC-Precoated C/C–SiC
425
observed, identified as the formation of blisters. Spallation of a
yttrium silicate coating was also observed by Webster et al.
during continuous isothermal mass change measurements in air
at 1873 K. Discussed as a mismatch of CTE, the aim of this
work is to give an explanation for the failure mechanism. The
2800
2600
9
Y O (h)
Liquid
+
Liquid
1
0
MS +
Liquid
2400
2 5 2 2 7
thermal expansion behavior of Y SiO and Y Si O is investi-
gated and compared with published results. Apparently, the
mismatch in CTE cannot be the only reason for the observed
coating failure. Thermodynamic calculations are carried out to
understand possible interface reactions between yttrium silicate
and SiC with increasing oxygen partial pressure.
DS +
Liquid
Y O (r) + Liquid
2200
Liquid 1
Liquid 2
+
2
1
1
000
800
600
MS +
δ-DS
Liquid + Cr
δ-DS + Cr
γ-DS + Cr
MS +
γ-DS
II. Thermodynamic Database
γ-DS + Tr
1
2
The CALPHAD (CALculation of PHAse Diagrams ) method
is used to determine phase equilibria and phase reactions in
multicomponent systems. A consistent set of thermodynamic
data for the Y–Si–C–O was developed and the calculations were
MS +
β-DS
β-DS + Tr
α-DS + Tr
1400
200
MS +
α-DS
1
3
carried out using ThermoCalc software. The unary data for
the pure elements and the description for the gaseous species
1
0
0.2
0.4
0.6
0.8
1.0
1
4
were taken from the SGTE substance database. The thermo-
dynamic description for the binary systems Y–Si, Y–C, Si–C,
Y2O3
Mole Fraction SiO2
SiO2
1
5,16
and Y–O were derived by thermodynamic optimization
and
20
Fig. 1. The calculated Y
O
2 3
–SiO system.
2
1
7
the system Si–O was taken from the literature. Based on phase
diagram data and measurements of enthalpy of formation and
heat capacity for the solid phases, a thermodynamic assess-
ment of the system Y O –SiO was performed. For the de-
1
8,19
To examine possible reactions between SiC and Y SiO , the
2
Y SiO was milled and mixed with b-SiC (Cerac Inc., 99.9%,
o1 mm, Milwaukee, WI). Bulk samples were prepared as de-
scribed above and annealed for 8 and 128 h, respectively, at 1923
K in air. Additionally, thermogravimetric analysis (TGA) was
¨
carried out on these samples in air up to 1923 K (Bahr STA501)
using an Al crucible, which was covered inside with platinum
5
2
0
2
3
2
2 5
scription of the liquid phase, the two-sublattice liquid model was
used. The descriptions were combined in one single database
to calculate the phase equilibria in the quaternary Y–Si–C–O
system.
2
O
3
foil. The phase compositions were analyzed by XRD (Siemens
Diffractometer D5000/Kristalloflex, CuKa1 radiation, Bruker
AXS, Karlsruhe, Germany). The evaluation was performed by
using Bruker AXS DIFFRAC AT software, comparing with the
Powder Diffraction File data of the International Centre for
Diffraction Data.
III. The Pseudobinary System Y O –SiO
2
2
3
In the system Y O –SiO , two yttrium silicate compounds are
2
3
2
2
1,22
known: Y SiO5 and Y Si O .
2
The monosilicate (Y SiO )
2
2
2
7
5
occurs in two different modifications (designated as X
and the disilicate (Y Si
1
2
and X )
2
O
2 7
) occurs in four different modifica-
Yttrium silicate coatings were deposited on CVD-SiC-pre-
tions (referred to as a, b, g, and d; in order of increasing tem-
perature). The phase transitions are of reconstructive character
10
coated C/C–SiC specimens by an LPPS technique. This proc-
ess is based on the generation of a plasma jet consisting of an
argon gas flow with admixtures of hydrogen and helium. The
gas flow is ionized while passing a high current arc discharge
within the plasma torch. The plasma jet leaves the torch by
passing a Laval-nozzle with a plasma jet velocity of up to 3000
m/s. When powder is injected into the plasma jet, the particles
will be accelerated (4800 m/s) and molten before being pro-
2
1
and the transition temperatures are given by Felsche. Those
between the Y Si modifications are very sluggish. The mod-
ification change of Y SiO is martensitic; both Y SiO struc-
2
2 7
O
2
3
2
5
2
5
tures are of monoclinic symmetry and crystallize within point
group 2/c. The difference lies in the unit cell: While the high-
temperature modification (X ) is base centered, X
2
1
crystallizes in
2
a primitive cell. The calculated pseudobinary system is shown
2
2
4
2
0
jected onto the substrate. Two yttrium silicate powders of
different composition were sprayed, as given in Table I. The
coating thickness of the yttrium silicate layers varied between
in Fig. 1. The calculations do not take into account the poly-
morphic transformation of Y SiO . According to the calcula-
tions, Y SiO melts congruently at 2232 K, and d-Y Si O melts
2
5
2
5
2
2
7
2
5
1
00 and 200 mm.
The C/C–SiC specimens with coatings obtained by spray
powder 1 (Table I) were tested within a plasma wind tunnel
incongruently at 2060 K.
2
6
IV. Experimental Procedure
Y SiO and Y Si O were prepared by the mixed-oxide tech-
test facility to assess the oxidation and erosion resistance un-
der simulated atmospheric re-entry conditions. The vacuum
chamber of the test facility is a 5 m long steel tank with a di-
ameter of 2 m, divided into three cylindrical segments, which all
2
5
2
2
7
2 3
nique. Starting materials were powders of high purity of Y O
(
Alfa, 99.99%, o1 mm, Alfa Aesar, Karlsruhe, Germany) and
10
have a double-wall cooling segment. The sample geometry is
shown in Fig. 2. During these tests, the sample was attached to a
SiO (Chempur, 99.99%, o45 mm, Karlsruhe, Germany). Prior
2
to use, the oxides were dried at 1273 and 1073 K, respectively,
and subsequently stored in a desiccator. Appropriate amounts
of oxides were mixed for 24 h in a 3-D powder mixer. The pow-
der mixtures were isostatically cold pressed at 800 kPa and an-
nealed for 120 h at 1873 K in air, with intermediate milling. By
this treatment, pure yttrium silicates were obtained as confirmed
by XRD and energy-dispersive X-ray spectroscopy (EDS) meas-
urements.
Table I. Chemical Composition and Particle Size Fraction of
the Plasma-Sprayed Powders
Spray powder 1
Spray powder 2
Composition Y
O
:SiO
50:50 wt%
21:79 at.%
75.5:24.5 wt%
45:55 at.%
5–30
2
3
2
The CTEs of the pure yttrium silicates (MS and DS)
were measured on cylindrical samples (0.5 cmꢀ 0.3 cm, 98.7%
theoretical density) with a differential dilatometer (Model
¨
802, Bahr, Hullhorst, Germany) and heating up to 1673 K in
¨
air with a gradient of 5 K/min.
Particle size distribution 8–25
mm)
Median size (mm)
Phase content
(
19.9
16.7
Y Si O 1SiO Y SiO 1Y Si O
2
2
7
2
2
5
2
2
7

4
26
Journal of the American Ceramic Society—Seifert et al.
Vol. 88, No. 2
LPPS-yttrium silicate
porous
dense
CVD-SiC
C/C fiber bundle
SiC matrix
C/C-SiC substrate
Fig. 4. 1B0lister formation as observed at test temperatures above
923 K.
1
sintering effects. However, investigations by SEM did not show
any apparent degradation within the C/C–SiC substrate mate-
rial. As shown in Fig. 3, at 1623 K, the weight change of C/C–
SiC specimens with yttrium silicate protection is less than 10%
that of samples with a CVD-SiC coating. Also, the C/SiC ma-
terial with coating systems from industrial suppliers show a
Fig. 2. Sample geometry for the plasma wind tunnel tests.
2
7
weight loss approximately six times (MAN Technology) and
1
2
0 times (Astrium) higher than that of C/C–SiC material with
7
water-cooled support system, which is mounted on a movable
platform of the positioning system. The sample itself was
mounted inside a sample support cap, which is made of sinte-
red SiC and faced toward the plasma source. As indicated by the
arrow in Fig. 2 (plasma flow direction), a circular-shaped area of
the sample’s surface (diameter: 26.5 mm) was exposed to a plas-
ma flow of approximately 800 m/s.
an LPPS-yttrium silicate coating. Higher weight loss indicates
higher erosion and/or active oxidation of the materials.
1
0
At temperatures above 1923 K, a sudden coating failure
was observed by visual inspection of the samples during the
running tests. The failure always started with the formation of
some local hot spots on the sample’s surface. Overheating effects
like this may be set off by additional heat fluxes resulting from
catalytic interaction between the coating’s surface and high re-
active components of the plasma flow, as for example the re-
combination reaction of atomic oxygen. By the following SEM
investigation, these hot-spot areas have been identified as areas
of blister formation. As an example, Fig. 4 shows the beginning
spallation of yttrium silicate coating that was obtained with
spray powder of type 2 (Table I). Besides the cracks with parallel
The chosen test parameters were the same as those from pre-
with CVD-
1
0,27
vious successful test campaigns of other projects,
SiC-coated C/C–SiC and multilayer-coated C/SiC samples sup-
plied by MAN Technology (Munich, Germany) and by Astrium
(
of 1.6 g/s N and 0.4 g/s O , respectively, the stagnation pressure
Friedrichshafen, Germany), respectively. With mass flow rates
2
2
was approximately 4.1 mbar. For the possibility of direct com-
parison of specimens with different coating systems, the first
tests with the two-layer LPPS/CVD coating system were per-
formed at 1623 K for 23 min at maximum temperature. The
following tests at 1773, 1873, and 1923 K, respectively, were
performed by increasing the current of the plasma generator.
The specific mass loss was determined by the sample’s change in
mass versus testing time.
2 5 2 2 7
orientation to the Y SiO /Y Si O –SiC interface, numerous
small cavities or bubble-like structures are visible within the
loosened silicate layer, which indicates increasing porosity and
gaseous reaction products. These areas are in contrast to the
almost undisturbed coating in between the blisters (Fig. 4, mid-
dle). Here the original low-pressure plasma-sprayed coating with
its dense microstructure (porosity r1%) as obtained with spray
powder 2 is still visible.
In view of coating failure mechanisms, the deficiencies of me-
chanical or chemical compatibility between the coating materi-
als and the substrate are important aspects to discuss.
V. Results and Discussion
(
1) Mass Loss During Plasma Wind Tunnel Tests
After 23 min at 1623 K in the plasma wind tunnel, the visual
appearance of the test specimens with yttrium silicate coatings
obtained from powder 1 changed. The typical gray color of the
coating turned into white—just as the color of the original spray
powder—and the coating surface appeared smoother because of
(2) Mechanical Compatibility
Because of a mismatch in the CTE of phases within a multilayer
coating and the substrate, the cooling process after the coating
deposition will cause multiple microcracks and thus increasing
ingress of oxygen.
In Fig. 5, the results of thermal expansion measurements of
both yttrium silicates are shown. For comparison, literature
0
.00
0.00
5
,28
data of SiC are also shown.
2 2 7
The mean CTE of Y Si O
ꢁ
6
is determined to 4.84 ꢀ 10 /K (300–1673 K) and therefore,
−
−
−
−
0.20
0.40
0.60
−0.20
−0.40
−0.60
−0.80
it is presumed to be close to that of the SiC substrates
ꢁ
6
(
and from 811 to 2477 K, respectively). This was also found
4.3–5.4 ꢀ 10 /K determined in the ranges from 297 to 1144 K
2
8
ꢁ6
by Aparicio and Duran (4.6ꢀ 10 /K for 300–1273 K). Thus
C/C-SiC (LPPS-yttrium silicate coating)
C/C-SiC (CVD-SiC coating)
C/SiC (coated by MAN-Technologie)
C/SiC (coated by Astrium)
Y Si O and SiC can be considered as mechanically compatible.
2
2
7
2 2 7 2 5
Therefore, with increasing amount of Y Si O /Y SiO ratio, the
plasma-sprayed coating is expected to show higher mechanical
compatibility to the SiC sublayer and the C/C–SiC substrate.
Ogura et al. reported that the mean CTE of polycrystalline
0.80
1
5
550
1650
1750
1850
1950
ꢁ6
Y
2
SiO
5
is 4.8 ꢀ 10 /K (300–1623 K), which is also close to that
surface temperature [K]
of SiC. According to Ogura’s XRD analysis, the bulk samples
contained Y O as a minor constituent. Aparicio and Duran
determined for the yttrium monosilicate a higher value of
6.9 ꢀ 10 /K in the range from 300 to 1273 K, measured on
28
Fig. 3. Relative mass loss with coated C/SiC and C/C–SiC specimens
after plasma wind tunnel testing (23 min at maximum temperature with
2
3
ꢁ
6
N
2
: O
2
5 80: 20 wt% and p 5 4.1 mbar).

February 2005
Yttrium Silicate Coatings on CVD-SiC-Precoated C/C–SiC
427
SiC (Ogura et al., 1995)
SiC (Aparicio et al., 2000)
Y Si O (Aparicio et al., 2000)
Y Si O (this work)
Y SiO (this work)
Y SiO (Ogura et al., 1995)
Y SiO (Aparicio et al., 2000)
0
0
0
0
0
0
0
0
0
.8
.7
.6
.5
.4
.3
.2
.1
.0
400
600
800
1000 1200 1400 1600
Temperature [K]
Fig. 5. Thermal expansion of yttrium silicates and SiC.
2 5 2 2 7
Fig. 6. Chemical compatibility of a (Y SiO /Y Si O ) layer on a SiC
layer.
bulk sample prepared by mixing oxides and heat treating at
833 K for 3 h. The preparation in this way also leads to a small
amount of residual Y O , as shown by XRD patterns in
2 2 7 2 5 2
tions at the interface of the LPPS-Y Si O /Y SiO –SiO layer
and the CVD-SiC coating have not been observed even after
thermal treatment up to 1873 K.
Figure 8(a) shows the calculated phase distribution caused by
chemical reactions at the interface of yttrium silicates and SiC at
1
2
3
2
9
ꢁ6
28
Aparicio et al. The mean CTE of Y
which means that the presence of some Y
2
O
3
is 9.5ꢀ 10 /K,
2
O
3
should increase the
thermal expansion somewhat. As depicted in Fig. 5, both au-
thors are in disagreement with the data of this work. However,
the general shape of the mean thermal expansion curve coincides
1
923 K with increasing oxygen partial pressure (pO ). To sim-
2
ulate such conditions, argon (Ar) was formally introduced as an
inert gas. This simulation for a closed system is based on the
assumption of a silicate composition of 0.33 mole Y Si O and
5
with Ogura et al. At temperatures above approximately
1273 K, the curve of the MS shows obviously an increasing
gradient. This can be explained by the mentioned phase trans-
2
2
7
0.67 mole Y SiO to react with 1 mole of SiC. The existence of a
2 5
small amount of gaseous phase, resulting for example from the
ingress through microcracks at ambient pressure of 1 bar, was
also taken into account. Because of an accurate determination
of the partial pressure of oxygen (pO
some predicted chemical reaction mechanisms can be verified.
2
3
formation of Y
2 5
SiO . According to Nowok et al., the phase
transformation takes place at about 1123 K. This is in good
3
0
agreement with our result, observing a transformation tem-
perature of approximately 1173 K. The mean CTE was deter-
2
) within the gaseous phase,
ꢁ
6
ꢁ6
mined to 3.05 ꢀ 10 /K for 300–1173 K and 4.09 ꢀ 10 /K in
the temperature range between 1173 and 1673 K. Thus the phase
transformation of Y SiO probably causes microcracks during
According to Fig. 8(a), Y
2
SiO , Y
5
2
Si
2
O
7
, SiC, and gas phase
ꢁ16
stay in equilibrium as long as the pO does not exceed 3ꢀ 10
2
2
5
bar. Above this value, Y SiO and SiC react to form Y Si O
7
2
5
2
2
high-temperature application of coatings containing the mono-
silicate Y SiO
and gas phase (mainly consisting of CO), as predicted by Web-
ster et al. (Eqs. (1) and (2)). The calculations show that re-
maining SiC may react with oxygen according to the reaction
5
.
6
2
(
3) Chemical Compatibility
According to thermodynamic calculations by Webster et al.,
carried out for a dense two-layer coating consisting of pure
SiO and SiC, chemical reactions are not expected between
the two compositions at temperatures up to 2023 K. Increasing
the oxygen content at the Y SiO –SiC interface leads to the
formation of Y Si and gas (CO/CO ) at 1873 K and a mass
6
1
2
SiC þ 2 2 O2 ! SiO2 þ SiO þ 2CO
(3)
Y
2
5
15
exceeds 5 ꢀ 10ꢁ bar, solid
As the oxygen partial pressure pO
SiO becomes a stable phase. According to the calculations for
2
2
2
5
higher oxygen partial pressures (including conditions in air),
2
O
2 7
2
2 2 7 2
Y Si O and SiO remain as the only stable solid phases with
gain of approximately 6 mass%. Possible reactions between the
silicate and the carbide material in the presence of oxygen are
given by Webster et al. :
a total sample mass gain of 8.6%. It is important to note that
the experimentally observed transition from active to passive
oxidation of SiC at 1923 K occurs at remarkably higher
6
Y2SiO5ðsÞ þ SiCðsÞ þ 3=2 O2ðgÞ¼ Y2Si2O7ðsÞ
þ COðgÞ
(1)
SiO2
Y2SiO5ðsÞ þ SiCðsÞ þ 2O2ðgÞ¼ Y2Si2O7ðsÞ þ CO2ðgÞ (2)
yttrium silicate
According to our thermodynamic calculations, in the absence of
any oxidizing gas, the phases Y
equilibrium up to the incongruent melting of d-Y
K (Fig. 6). Therefore, the chemical compatibility is not influ-
enced by the presence of Y Si . The SEM image of such a
coating system is shown in Fig. 7. In the cross-section, the light
gray LPPS–Y Si /Y SiO layer, as obtained with spray pow-
der of type 1 (Table I), and the dark gray CVD-SiC bonding
layer underneath are clearly visible. Besides the silicate phases,
the plasma-sprayed coating is interspersed by angular—some-
times even rounded—particles of amorphous silica as confirmed
by EDS analysis. Obviously, they are residuals resulting from
the spray powder that also contains free silica. Chemical reac-
2
SiO
5
, Y
2
Si
2
O
7
, and SiC are in
2
Si at 2060
2 7
O
LPPS-coating
CVD-SiC
2
2 7
O
2
O
2 7
2
5
C/C-SiC substrate
200 µm
Fig. 7. Cross-section of a two-layer coating system (SEM image).

4
28
Journal of the American Ceramic Society—Seifert et al.
Vol. 88, No. 2
4
3
3
2
2
1
1
00
50
00
50
00
50
00
∆
Y Si O
2 2 7
Y Si O
+ SiO2
ο Y SiO
# β-SiC
1
28h,1923K, air
2
5
Y SiO
8
air
h, 1923K,
Y Si O
SiO
5
0
SiC
gas
gas
−15.0
starting mixture
0
−
17.0
−16.0
−14.0
Log10 (po )/1atm
2
(
a)
10
20
30
40
50
60
2
θ [°]
1
.0
.9
.8
.7
.6
.5
.4
.3
.2
.1
2 5
Fig. 9. XRD patterns of the investigated Y SiO /SiC mixture before
and after heat treatment at 1923 K in air.
0
0
0
0
0
0
0
0
0
Ar
CO
identified by XRD, in agreement with the calculations. The
TGA (Fig. 10) indicates a mass gain of 8.5 mass%, which is in
good agreement with the calculated value of 8.6 mass%.
The calculations show that the gas phase consists mainly of
CO and SiO species interacting with the solid phases of the sys-
tem (Fig. 8(b)). Figure 11 shows three calculated isothermal po-
tential phase diagrams (1853, 1903, and 2000 K) illustrating the
stabilities of the solid phases as a function of the CO and SiO
partial pressures. A triangle-shaped three-phase field can be
Ar
CO
2 5
seen, in which SiC and Y SiO are in equilibrium with the gas
SiO
0
−
phase, existing in between four-phase equilibria lines with, e.g.
Y Si O or graphite additionally in equilibrium. Since CVD-SiC
coatings often contain very small and more or less dispersed in-
clusions of residual free carbon, the intercept point of the
lower and upper limiting line of the triangle-shaped phase field
is of special interest. At this point, all five phases (gas, SiC,
17
−16
−15
−14
−13
2
2
7
(
b)
Log10 (po )/1atm
2
3
5
Fig. 8. Thermodynamic simulation of the interface reactions of
Si (0.33 mole) and Y SiO (0.67 mole) with SiC (1 mole) at 1923
K with increasing oxygen partial pressure. (a) Calculated phase equili-
bria. (b) Related formation of gas species.
Y
2
2
O
7
2
5
Y
2
SiO
At 1903 K the five-phase equilibria point (gas, SiC, Y
Si , and graphite) reaches a pCO of 1 bar. At 2000 K, the
5 2 2 7
, Y Si O , and graphite) are in equilibrium.
2 5
SiO ,
Y
2
2 7
O
intersection is outside the chosen axis limit in Fig. 11. For all
temperatures, the composition of the gas phase is clearly dom-
inated by pCO (approximately 10 times higher than pSiO). Figure
12 illustrates the evolution of the vapor pressure in equilibrium
with the four solid phases, graphite, SiC, Y SiO , and Y Si O ,
as a function of temperature. As the total pressure of the gen-
erated gaseous phase increases drastically with temperature, an
equilibrium vapor pressure of approximately 10 bar was calcu-
lated for example at 2175 K. These calculations show that
ꢁ
3
31–33
pO (B10 bar)
ꢀ 10
as compared with the calculated value of
bar for the SiO formation. To understand this effect,
2
ꢁ
15
5
2
the mass balance conditions in the treated thermodynamic sys-
tem have to be taken into account as discussed in detail by
Heuer and Lou. However, the mechanism of active–passive
oxidation is different from the one treated in this work for re-
actions proceeding at the interface of yttrium silicate and SiC.
2
5
2
2
7
3
4
2 5 2 2 7
The coating of 100–200 mm Y SiO /Y Si O protects the SiC
layer until the reaction on the interface creates a gas phase of
sufficiently high pressure to delaminate the coating as discussed
below.
Figure 8(b) shows the evolution of the relative gas phase
composition related to the oxygen partial pressure at 1923 K.
The dominating gas species are CO and SiO. At low-oxygen
partial pressures, the ratio of both gaseous components is equiv-
alent because both of them are generated from oxidation of SiC
only. With higher partial pressures of oxygen, the formation of
10
2
000
1800
8
6
4
2
0
1
1
1
600
400
200
2 2 7
Y Si O also increases the production of CO significantly while
SiO is not involved.
To confirm the results of the simulation, a corresponding
1000
800
powder mixture of 0.67 mole Y
statically cold pressed at 800 kPa and annealed at 1923 K in air
for 8 or 128 h, respectively. Y Si was not added, in order to
2 5
SiO and 1.0 mole SiC was iso-
6
00
400
500
2
2 7
O
observe its expected formation. The XRD patterns for this ex-
perimental investigation are shown in Fig. 9. After annealing the
0
100
200
time [min]
300
400
2 5
sample for 8 h, Y SiO is still present, but the formation of
Y Si O has started and SiO is clearly detectable. After anneal-
7
Fig. 10. Thermogravimetric analysis of the Y SiO /SiC mixture in air
2
2
2
2
5
ing for 128 h, only Y Si O and SiO (cristobalite) could be
2
between RT and 1923 K.
2
7
2

February 2005
Yttrium Silicate Coatings on CVD-SiC-Precoated C/C–SiC
429
0.05
0.04
0.03
0.02
0.01
0
anisms a deficiency of mechanical or chemical compatibility
between the coating materials and the substrate has to be
discussed.
2
1
1
000 K
903 K
853 K
Because of the difference in the thermal expansion coefficients
of Y Si O and the two modifications of Y SiO , the resulting
2
2
7
2
5
mechanical strain may also be responsible for microscopic sur-
face cracks and failure during high-temperature application.
Thermodynamic calculations indicate good chemical compat-
ibility between both layer materials, yttrium silicate and SiC.
However, increasing partial pressures of oxygen interface
reactions at high temperatures are expected from thermody-
namic calculations and verified experimentally. As evidenced by
calculated isothermal potential phase diagrams, the partial pres-
sure of a formed gaseous phase, mainly consisting of CO and
SiO, is of great influence on the interface stability. The forma-
tion of gas bubbles and blisters within the silicate layer is ob-
viously resulting from the significantly increasing equili-
brium vapor pressure of the system. For atmospheric pressure,
a critical surface temperature of 1903 K was predicted by
calculations.
Gas+SiC+Y SiO
2
5
+
Y Si O
2
2 7
Gas+SiC+
Y2SiO5
Gas+SiC+Y SiO +Graphite
2
5
0
0.2
0.4
0.6
0.8
1.0
PCO [bar]
Acknowledgments
Fig. 11. Potential phase diagrams at 1853, 1903, and 2000 K, respec-
tively.
The authors thank H. Kummer for support in thermal analyses. Special thanks
are due to the members of the Institute of Space Systems, University of Stuttgart,
for performing plasma wind tunnel tests.
1
1
1
1
03
02
01
00
References
1
J. R. Strife and J. E. Sheehan, ‘‘Ceramic Coatings for Carbon-Carbon Com-
posites,’’ Am. Ceram. Soc. Bull., 67 [2] 369–74 (1988).
E. Fitzer and L. M. Manocha, Carbon Reinforcements and Carbon/Carbon
Composites. Springer, Berlin, 1998.
M. E. Westwood, J. D. Webster, R. J. Day, F. H. Hayes, and R. Taylor, ‘‘Re-
view: Oxidation Protection for Carbon Fibre Composites,’’ J. Mater. Sci., 31,
389–97 (1996).
N. S. Jacobson, ‘‘Corrosion of Silicon-Based Ceramics in Combustion Envi-
ronments,’’ J. Am. Ceram. Soc., 76 [1] 3–28 (1993).
5
2 5
Y. Ogura, M. Kondo, and T. Morimoto, ‘‘Y SiO as Oxidation Resistant
Coating for C/C Composites;’’ pp. 767–74, Proceedings of the 10th International
Conference on Composite Materials (ICCM-10), IV, Whistler, BC, Canada, 1995.
J. D. Webster, M. E. Westwood, F. H. Hayes, R. Taylor, A. Duran,
2
3
1
4
−
1
2
3
4
1
0
−
1
1
1
0
6
−
M. Aparicio, K. Rebstock, and W. D. Vogel, ‘‘Oxidation Protection Coatings
0
for C/SiC Based on Y
2
SiO
Y. Ogura, M. Kondo, T. Morimoto, A. Notomi, and T. Sekigawa, ‘‘Oxidation
Behavior of Y SiO /SiC Coatings in Low Pressures of Oxygen,’’ J. Jpn. Inst. Met-
5
,’’ Key Eng. Mat., 132–136, 1641–4 (1997).
7
−
2
5
0
als, 65 [1] 12–3 (2001).
1
000
1500
2000
2500
8
2 5
M. Kato, T. Fukasawa, and Y. Goto, ‘‘Reactions in Y SiO –SiC and
Y
3
Al
5
O12–SiC in Yttrium-Silicate/Silicon Carbide Layered Composites,’’ J. Ceram.
T [K]
Soc. Jpn., 108 [9] 861–4 (2000).
9
Fig. 12. Vapor pressure in equilibrium with SiC, Y
carbon.
2 5 2 2 7
SiO , Y Si O , and
J. D. Webster, M. E. Westwood, F. H. Hayes, R. J. Day, R. Taylor, A. Duran,
M. Aparicio, K. Rebstock, and W. D. Vogel, ‘‘Oxidation Protection Coatings for
C/SiC Based on Yttrium Silicate,’’ J. Eur. Ceram. Soc., 18, 2345–50 (1998).
10
T. Laux, T. Ullmann, M. Auweter-Kurtz, H. Hald, and A. Kurz, ‘‘Investiga-
tions of Thermal Protection Materials Along an X-38 Re-entry Trajectory by
Plasma Wind Tunnel Simulations;’’ Proceedings of the 2nd International Sympo-
sium, Atmospheric Reentry Vehicles and Systems, 26–29 March 2001, Arcachon,
France, on CD (2001).
during the plasma wind tunnel testing, the generation of CO gas
and the significant pressure increase can result in the formation
of blisters (see Fig. 4) near the Y SiO /Y Si O –SiC interface.
Certainly, the critical temperature for expected coating failure
depends on the ambient pressure and exposition time, but also
on the existence of dispersed carbon inclusions within the un-
derlying CVD-SiC bonding layer.
2
5
2
2
7
11
J. Fabig and W. Krenkel, ‘‘Tailoring of Microstructure in C/C–SiC Compos-
ites;’’ pp. 759–65, Proceedings of the 10th International Conference on Composite
Materials (ICCM-10), Vol. IV, Whistler, BC, Canada, 1995.
N. Saunders and A. P. Miodownik, ‘‘CALPHAD (Calculation of Phase Di-
12
agrams): A Comprehensive Guide;’’ in Materials Series, Vol. 1, Edited by R. W.
Cahn. Pergamon, Oxford, 1998.
13
B. Sundman, B. Jansson, and J.-O. Andersson, ‘‘The Thermo-Calc Databank
System,’’ Calphad, 9, 153–90 (1985).
14
Scientific Group Thermodata Europe, Grenoble Campus, Saint Martin
D’Heres, France, http://www.sgte.org.
¨
J. Grobner, ‘‘Konstitutionsberechnungen im System Y–Al–Si–C–O (Phase
VI. Conclusions
15
Plasma wind tunnel tests between 1623 and 1923 K have shown
that a two-layer coating system of LPPS yttrium silicates de-
posited on a CVD-SiC bond coating provides good oxidation
protection and erosion resistance for C/C–SiC composites. The
relative mass loss was determined to be less than 0.6% and no
apparent degradation of the surface material was observed up to
Diagram Calculations in the System Y–Al–Si–C–O);’’ Ph.D. Thesis, University of
Stuttgart, 1994.
V. Swamy, H. J. Seifert, and F. Aldinger, ‘‘Thermodynamic Properties of
Y O Phases and the Yttrium–Oxygen Phase Diagram,’’ J. Alloys Comp., 269,
2 3
16
2
01–7 (1998).
B. Hallstedt, ‘‘Thermodynamic Assessment of the Silicon-Oxygen System,’’
Calphad, 1, 53–61 (1992).
17
18
1923 K. At higher temperatures, a formation of local hot spots
2 3
U. Kolitsch, ‘‘Hochtemperaturkalorimetrie und Phasenanalytik in SE O –
Al
2
O
3
–SiO
–Al
2
Systemen (High Temperature Calorimetry and Phase Analysis in
–SiO Systems);’’ Ph.D. Thesis, University of Stuttgart, 1995.
was visually detected at the sample’s surface followed by coating
spallation. Overheating effects like these may indicate catalytical
interaction between the sample’s surface and the plasma flow.
Additionally, in view of numerous possible coating failure mech-
RE
2
O
3
2
O
3
2
19
J.-J. Liang, A. Navrotsky, T. Ludwig, H. J. Seifert, and F. Aldinger, ‘‘En-
thalpy of Formation of Rare-Earth Silicates Y SiO and Yb SiO and N-contain-
ing Silicate Y10(SiO ,’’ J. Mater. Res., 14 [4] 1181–5 (1999).
2
5
2
5
4 6 2
) N

4
30
Journal of the American Ceramic Society—Seifert et al.
Vol. 88, No. 2
2
0
28
O. Fabrichnaya, H. J. Seifert, R. Weiland, T. Ludwig, F. Aldinger, and
M. Aparicio and A. Duran, ‘‘Yttrium Silicate Coatings for Oxidation Protec-
tion of Carbon–Silicon Carbide Composites,’’ J. Am. Ceram. Soc., 83 [6] 1351–5
(2000).
A. Navrotsky, ‘‘Phase Equilibria and Thermodynamics in the Y
System,’’ Z. Metallkd., 92, 1083–97 (2001).
2 3
O
–Al
2
O
3
–SiO
2
21
29
J. Felsche, ‘‘Crystal Data on the Polymorphic Disilicate
Naturwissenschaften, 57, 127–8 (1970).
Y
2
Si ,’’
2
O
7
M. Aparicio, R. Moreno, and A. Duran, ‘‘Colloidal Stability and Sintering of
Yttria–Silica and Yttria–Silica–Alumina Aqueous Suspensions,’’ J. Eur. Ceram.
Soc., 19, 1717–24 (1999).
S. Wagner, H. J. Seifert, and F. Aldinger, ‘‘High Temperature Reactions of
22
K. Liddell and D. P. Thompson, ‘‘X-ray Diffraction Data for Yttrium Sili-
cates,’’ Br. Ceram. Trans. J., 85, 17–22 (1986).
30
23
J. W. Nowok, J. P. Kay, and R. J. Kulas, ‘‘Thermal Expansion and High-
Temperature Phase Transformation of the Yttrium Silicate Y SiO ,’’ J. Mater.
Res., 16 [8] 2251–5 (2001).
C/C–SiC Composites with Ceramic Coatings;’’ pp. 71–8 in Proceedings of the In-
ternational Conference on Advances in Materials and Materials Processing,
ICAMMP-2002, 1–3 February 2002, Kharagpur, India. Edited by N. Chakr-
aborti, and U. K. Chatterjee. Tata McGraw-Hill, New Delhi, 2002.
2
5
24
R. Henne, ‘‘Direct Current Plasma Torches for Thermal Spraying—Present
Status and New Developments,’’ Weld. Cutting, 2, E31–5 (1993).
31
W. L. Vaughn and H. G. Maahs, ‘‘Active-to-Passive Transition in the Oxi-
dation of Silicon Carbide and Silicon Nitride in Air,’’ J. Am. Ceram. Soc., 73 [6]
1540–3 (1990).
25
T. Ullmann, M. Schmu
¨
cker, H. Hald, R. Henne, and H. Schneider, ‘‘Oxida-
tion Protection of C/SiC Composites Based on Plasma-Sprayed Y-Silicates;’’ Pro-
ceedings of the 8th International Symposium on Materials in Space and Environment
and 5th International Conference on Protection Materials and Structures from LEO
Space Environment, 5–9 June 2000, Arcachon, France, on CD, 2000.
32
T. Goto, T. Narushima, Y. Iguchi, and T. Hirai, ‘‘Active to Passive Transition
in the High-Temperature Oxidation of CVD SiC and Si ;’’ pp. 165–76 in Cor-
rosion of Advanced Ceramics, Edited by K. G. Nickel. Kluwer Academic Publish-
3 4
N
26
M. Auweter-Kurtz, H. Habiger, H. Kurtz, S. Laure, W. Ro
nos, ‘‘Die Plasmawinkanale PWK-IRS-Werkzeuge zur Untersuchung von Hit-
zeschutzmaterialien fur Ruckkehrfahrzeuge,’’ Z. Flugwiss. Weltraumforsch., 17,
–15 (1993).
¨
ck, and N. Tuba-
ers, The Netherlands, 1994.
33
¨
T. Goto, H. Homma, and T. Hirai, ‘‘Effect of Oxygen Partial Pressure on the
High-Temperature Oxidation of CVD SiC,’’ Corrosion Sci., 44, 359–70 (2002).
A. H. Heuer and V. L. K. Lou, ‘‘Volatility Diagrams for Silica, Silicon Nitride,
¨
¨
34
1
27
T. Ullmann, H. Hald, O. Vollrath, H. H. Grotheer, O. Haidn, and M. Au-
weter-Kurtz, ‘‘Behaviour of C/C–SiC Material in Reentry and H /O Combustion
and Silicon Carbide and Their Application to High-Temperature Decomposition
and Oxidation,’’ J. Am. Ceram. Soc., 73 [10] 2789–803 (1990).
2
2
35
Environment;’’ pp. 291–301, Proceedings of the 3rd European Workshop on Ther-
mal Protection Systems, 25–27 March 1998, ESA-WPP-141, 1998.
K. Brennfleck, G. Porsch, H. Reich, and H. J. Scheiffarth, ‘‘CVD of SiC on an
Industrial Scale,’’ Proc. Int. Conf. CVD, CVD-XI, 395–403 (1990).
&