Effects of oxygen partial pressure on the nucleation behavior and morphology of chemically-vapor-deposited zirconia on Hi-Nicalon fiber and Si

Article abstract of DOI:10.1111/j.1151-2916.2003.tb03604.x

A ZrO₂ coating was prepared on Hi-Nicalon fiber and single-crystal Si by chemical vapor deposition (CVD) using ZrCl₄, CO₂, and H₂ as precursors at 1050°C. The effects of oxygen partial pressure on the nucleation behavior of the CVD-ZrO₂ coating were systematically studied by intentionally varying the controlled amount of O₂ into the CVD chamber. Characterization results suggested that the number density of tetragonal ZrO₂ nuclei apparently decreased with increasing the oxygen partial pressure from 4 × 10^-3 to 1.6 Pa. Also, the coating layer became more columnar and contained larger monoclinic ZrO₂ grains. The observed relationships between the oxygen partial pressure and the nucleation and morphologic characteristics of the ZrO₂ coating were attributed to the grain size and oxygen deficiency effects, which have been previously reported to cause the stabilization of the tetragonal ZrO₂ phase in bulk ZrO₂ specimens.

Full text of DOI:10.1111/j.1151-2916.2003.tb03604.x

J. Am. Ceram. Soc., 86 [12] 2031–36 (2003)
journal
Effects of Oxygen Partial Pressure on the Nucleation Behavior and
Morphology of Chemically-Vapor-Deposited Zirconia on
Hi-Nicalon Fiber and Si
,
†
Jinil Lee, Hao Li,* and Woo Y. Lee*
Department of Chemical, Biochemical and Materials Engineering Stevens Institute of Technology,
Hoboken, New Jersey 07030
Michael J. Lance*
Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6068
A ZrO coating was prepared on Hi-Nicalon fiber and single-
grains and (2) an outer coating layer where m-ZrO grains
predominantly grew in a columnar manner, but with the presence
2
2
2
crystal Si by chemical vapor deposition (CVD) using ZrCl4,
CO , and H as precursors at 1050°C. The effects of oxygen
2
2
of small t-ZrO crystals of ϳ50–100 nm on the coating surface.
2
partial pressure on the nucleation behavior of the CVD-ZrO2
coating were systematically studied by intentionally varying
The outer coating layer was detached from the inner layer. Also,
these observations suggested that t-ZrO nuclei were continuously
2
the controlled amount of O into the CVD chamber. Charac-
formed throughout the deposition process (i.e., continuous nucle-
2
2
,3
terization results suggested that the number density of tetrago-
nal ZrO2 nuclei apparently decreased with increasing the
ation of t-ZrO2).
2
We have previously proposed that the delamination within the
؊3
oxygen partial pressure from 4 ؋ 10 to 1.6 Pa. Also, the
ZrO layer occurred as a result of the following sequences: (1)
2
coating layer became more columnar and contained larger
formation of t-ZrO and m-ZrO nuclei on the fiber surface in the
2
2
monoclinic ZrO grains. The observed relationships between
early stage of the deposition, (2) martensitic transformation of the
2
the oxygen partial pressure and the nucleation and morpho-
t-ZrO nuclei to m-ZrO on reaching a critical grain size, and (3)
2
2
logic characteristics of the ZrO coating were attributed to the
development of significant compressive hoop stresses due to the
volume dilation associated with the transformation. Note that
m-ZrO is a thermodynamically stable phase up to 1205°C, t-ZrO
2
grain size and oxygen deficiency effects, which have been
previously reported to cause the stabilization of the tetragonal
ZrO phase in bulk ZrO specimens.
2
2
2
2
from 1205° to 2377°C, and cubic ZrO (c-ZrO ) from 2377°C to
2 2
the melting point at 2710°C. The martensitic transformation from
t-ZrO to m-ZrO is known to be diffusionless, and accompanies
2
2
4
9
° shear and 4.5% volume dilation. When this delamination
I. Introduction
behavior was observed, the ZrO coating did not degrade the fiber
2
3
strength.
N SiC/SiC ceramic matrix composites, toughness is obtained by
adding a fiber coating which provides a weak interface for crack
Interestingly, the nucleation of t-ZrO was not evident when the
air leak rate was high (e.g., ϳ2 Pa/min), and the coating layer
mainly consisted of columnar m-ZrO2 grains. In the apparent
absence of t-ZrO2 nucleation, the ZrO2 coating was bonded
strongly to the Hi-Nicalon fiber surface without any evidence for
delamination at the fiber/coating interface region. The strong
bonding significantly decreased the tensile failure load of the
coated fiber tows. We suspected that the additional oxygen partial
pressure introduced by the higher air leak rate into the CVD
chamber might have mitigated the continuous nucleation of
I
2
deflection and debonding between the SiC fiber and the brittle SiC
matrix. However, the most commonly used fiber coatings, carbon
and boron nitride, are unstable in oxidative environments. The
oxidation problem is considered to be one of the major problems
remaining to be solved before SiC/SiC composites can be reliably
introduced into next-generation aircraft engines, power generation
3
3
1
turbines, space vehicles, and other industrial applications.
2
,3
In our earlier studies, we examined the feasibility of using a
chemically-vapor-deposited zirconia (CVD-ZrO ) fiber coating,
2
t-ZrO . If this speculation is true, the oxygen partial pressure (P )
prepared from ZrCl , CO , and H precursors, as an oxidation-
2
O
2
4
2
2
would be a critical parameter for controlling the ZrO2 coating
process and for reproducibly producing the desired weak interface
behavior. The purpose of this investigation was to study the effects
of PO2 on the nucleation behavior and morphology of the CVD-
resistant interphase for SiC/SiC composites. When the air leak rate
was low (e.g., ϳ0.3 Pa/min as measured by the procedure
described in Section II), the ZrO coating typically consisted of
2
two distinct regions: (1) an inner layer (ϳ50–100 nm) consisting
of both tetragonal ZrO (t-ZrO ) and monoclinic ZrO (m-ZrO )
ZrO coating by intentionally adding the controlled amounts of O
2
2
2
2
2
2
into the reactor chamber.
II. Experimental Procedure
R. Naslain—contributing editor
As shown in Fig. 1, a hot-wall CVD reactor was used to prepare
ZrO coating specimens from ZrCl , CO , and H precursors. The
2
4
2
2
reactor chamber made of a fused SiO2 tube (3.4 cm outside
Manuscript No. 186523. Received December 3, 2002; approved August 5, 2003.
This work was supported by the National Science Foundation (DMR-GOALI
971623) with cofunding from the Air Force Research Laboratory (AFRL). M. Lance
diameter and 34 cm in length, Q-glass, Towaco, NJ) was heated
TM
using a resistance furnace (L. H. Martial
Model 1123, Therm-
9
craft, Inc. Winston-Salem, NC). The effective heating zone of the
furnace was ϳ30 cm in axial length. A K-type thermocouple was
used to measure the temperature at the center region of the reactor.
The temperature was controlled using a temperature controller
was sponsored by the U.S. Department of Energy, Office of Basic Energy Sciences,
Division of Materials Sciences and Engineering, under Contract No. DE-AC05-
0
0OR22725 with UT-Battelle, LLC.
Member, American Ceramic Society.
*
†
To whom all correspondence should be addressed.
2
031

2
032
Journal of the American Ceramic Society—Lee et al.
Vol. 86, No. 12
Fig. 1. Schematic diagram of the CVD system used for CVD-ZrO experiments.
2
(Type 818, Eurotherm, Leesburg, VA). The pressure of the reactor
CO and H were used as an oxygen source for the CVD-ZrO
2
2
2
was controlled using a mechanical vacuum pump and an exhaust
valve (Type 253, MKS, Andover, MA) interfaced with a pressure
controller (Type 252A, MKS) and a capacitance manometer
process. As summarized in Table I, experiments Z2, Z3, Z4, and
Z5 were designed to introduce small amounts of O in addition to
2
the CO /H mixture. Two Ar bottles containing 100 ppm O and
2
2
2
(
Baratron, Type 222BA, MKS).
Hi-Nicalon fiber tows (Nippon Carbon Co., Ltd., Japan) and
1000 ppm O were specially purchased and used for this work
2
TM
(99.99% Ar, Matheson Tri-Gas, East Rutherford, NJ). A separate
gas line was added, as shown in Fig. 1, to introduce the Ar/O2
stream into the reactor chamber.
Si (100) (Silicon Quest International, Inc. Santa Clara, CA) were
used as substrates. A long Hi-Nicalon fiber tow was wrapped
around a fused SiO holder to make about 10 fiber tow specimens
Deposition experiments were performed using the following
general procedure. The reactor chamber was evacuated slowly
after the substrate holder was loaded into the reactor to ϳ5 Pa. The
reactor chamber was heated to 1050°C in flowing Ar (100
2
(ϳ25 cm long) for each deposition experiment. Si substrates were
placed on the bottom wall of the SiO2 reactor chamber at the
middle location of the Hi-Nicalon fiber tows (i.e., 12.5 cm
measured from the inlet end of the fiber holder). Si substrates are
denoted as “SZ”, as tabulated in Table I. The Si substrates were
used as a model substrate material for studying the phase contents
of the ZrO2 coating by X-ray diffraction (XRD) and Raman
spectroscopy.
3
cm /min). After the desired temperature was reached, the Ar flow
3
(100 cm /min) into the reactor chamber was stopped and then the
H , CO , and/or Ar/O streams were introduced into the reactor
2
2
2
chamber for ϳ1 min. Subsequently the ZrCl /Ar stream was added
4
to the H , CO , and Ar/O streams to start the deposition process.
2
2
2
A stainless steel vaporizer was used to contain ZrCl powder
After the desired deposition time was elapsed, the ZrCl , H , CO ,
4
4 2 2
(99.5% purity, Strem Chemicals, Inc., Newburyport, MA). The
and Ar/O streams were stopped, and the reactor chamber was
2
vaporizer was heated to ϳ190°C to generate a sufficient vapor
evacuated. The reactor chamber was cooled to room temperature
while flowing Ar at 100 cm /min, and then the substrate holder
3
pressure of ZrCl (0.13 kPa). The ZrCl vapor was transferred into
4
4
the CVD reactor chamber using Ar as a carrier gas. A stainless
was retrieved.
steel/Pyrex/fused SiO transition was used to heat the inlet part of
the CVD chamber to ϳ240°C. This design feature was important
Before each experiment, the leak rate of air into the CVD
system was measured without any gas flow using a thermocou-
ple gauge controller (Model 801, Varian, Lexington, MA) and
2
for avoiding the condensation of ZrCl at this inlet part of the
4
®
reactor chamber. The line from the outlet valve of the vaporizer to
the reactor inlet was gradually heated from ϳ200° to 230°C. At the
outlet end of the reactor chamber, an elastomeric O-ring was used to
an isolation valve (HPS Series 170, MKS) located between the
CVD chamber and the vacuum pump. The average leak rate
measured for the experiments was ϳ0.3 Pa/min on average, and
the leak rate variations for the experiments were less than Ϯ0.1
seal this end of the SiO reactor chamber with a stainless steel flange.
2
Table I. Processing Conditions Used for CVD ZrO Experiments
2
†
3
3
3
‡
3
§
Sample
T (°C)
P (kPa)
Time (min)
CO
2
(cm /min)
H
2
(cm /min)
Ar (cm /min)
Ar (cm /min)
Leak rate (Pa/min)
O
2
P (Pa)
Z1
Z2
Z3
Z4
Z5
Z6
Z7
SZ1
SZ2
SZ3
SZ4
SZ5
—
1050
1050
1050
1050
1050
1050
1050
4
120
120
120
120
120
120
180
120
120
120
120
120
120
120
120
120
120
120
120
120
120
‡
40
40
40
40
40
40
40
—
0.29
0.34
0.24
0.39
0.29
0.23
0.32
¶
0.004
0.04
0.4
¶
4
30
†
†
†
†
†
†
4
30
120
187
—
4
1.2
4
1.6
0.003
0.004
1.3
1.3
—
—
†
§
4 2
Z denotes coating specimens deposited on Hi-Nicalon, and SZ denotes those deposited on Si. Pure Ar used to carry the ZrCl vapor. Ar containing O . Ar containing 100
†
†
ppm of O
2
.
2
Ar containing 1000 ppm of O .

December 2003
Chemically-Vapor-Deposited Zirconia on Hi-Nicalon Fiber and Silicon
2033
Pa/min (Table I). The flow rate of air into the CVD chamber at
STP and PO2 could be estimated by the following equations:
⌬
pV
f ϭ
P0
0
.21f
PO2 ϭ Ptotalͩ ͪ
f
total
where f is the equivalent volumetric flow rate of air leaked into the
CVD chamber at STP, ⌬p is the experimentally measured air leak
rate, V is the reactor chamber volume which includes the volume
of the gas lines up to mass flowmeters and the outlet valve of the
ZrCl vaporizer, P is the atmospheric pressure, f is the total
4
0
total
gas flow of all gas streams introduced into the reactor chamber
during the CVD experiment at STP, and Ptotal is the reactor
pressure during the CVD experiment. For example, for experiment
Z1, the following values were used to calculate P to be ϳ4 ϫ
O
2
Ϫ3
Ϫ3
1
0
Pa (i.e., base line P due to the air leaks): f ϭ 1.7 ϫ 10
O
2
3
3
3
cm /min, f
ϭ 280 cm /min, P
ϭ 4 kPa, and V ϭ 580 cm .
total
total
By adjusting the flow rate of the Ar/O2 stream, PO2 for
experiments Z2, Z3, Z4, and Z5 was controlled to be 0.04, 0.4, 1.2,
and 1.6 Pa, respectively. For example, using the following equa-
tion and values, PO2 was estimated to be 0.04 Pa for experiment
Z2:
1
00 ppm ϫ fAr/O2
PO2 ϭ Ptotal
ͩ
ͪ
ftotal
3
where fAr/O2 is the flow rate of the Ar/O stream (30 cm /min), f
ϭ 310 cm /min, and P
2
total
Fig. 2. SEM images of Z1 at the 12.5 cm location: (a) cross-section and
3
ϭ 4 kPa. For this estimation, we
total
(b) fiber surface.
assumed that the f value would not change at the operating reactor
pressure (e.g., 4 kPa) although it was actually measured when the
chamber was evacuated to ϳ5 Pa without any controlled gas flow.
In these experiments, the oxygen partial pressure was dominantly
during the CVD step. Throughout this paper, we will refer to these
small crystals as “t-ZrO nuclei.”
controlled by the O in Ar and air leaks, but not by the equilibrium
2
2
Sample Z3 was produced by increasing PO2 to 0.4 Pa by adding
the Ar/O stream with 1000 ppm O . As shown in Fig. 3, the inner
oxygen partial pressure of the CO ϩ H mixtures. The equilib-
2
2
Ϫ5
rium oxygen pressure was calculated to be in the range of 10 and
Pa, whereas the oxygen partial pressure introduced by the
O /Ar and air leaks was on the order 10 to 10 .
2
2
Ϫ6
layer remained the fiber surface composed of some t-ZrO nuclei,
1
0
2
Ϫ3
0
but the t-ZrO nuclei did not fully cover the fiber surface. Also, the
2
2
A field-emission scanning electron microscope (FEG-SEM,
LEO DSM 982, LEO Electron Microscopy, Inc. Thornwood, NY)
was used for coating morphology characterization. A Dilor XY
8
00 triple stage Raman microprobe (JY, Inc., Edison, NJ) and an
Innova 308C argon ion laser (Coherent, Inc., Santa Clara, CA)
operating at 514.5 nm with a 100 mW output power were used to
examine the phase contents of the CVD-ZrO coating as described
2
5
in detail elsewhere. An X-ray diffractometer (Rigaku, Rint2000,
Woodlands, TX) was used to identify the crystalline phases of the
ZrO coating specimens deposited on the Si substrates.
2
III. Results and Discussion
Figure 2(a) shows a cross-section SEM image from sample Z1
as obtained from the 12.5 cm location measured axially from the
inlet end of the substrate holder (i.e., “12.5 cm axial location”).
This sample was produced without any intentional O introduction.
2
Therefore, PO2 for this sample was assumed to be the base line PO2
value produced by the air leak into the reactor chamber. The
coating thickness of Z1 was in the range of 1.5 to 3.5 ␮m. The
coating contained two distinct layers. An inner layer of the coating
remained on the fiber surface as delamination occurred within the
ZrO2 layer. This inner layer composed mostly of many small
crystals (ϳ50 to 150 nm) as can be seen from Fig. 2(b). The outer
layer also consisted of many small crystals (as indicated by arrows
in Fig. 2), but also with the development of larger grains toward
the surface of the coating layer.
As mentioned in the Introduction, we interpret that some of
these small crystals are t-ZrO2 nuclei which include m-ZrO2
crystals that used to be t-ZrO2 nuclei and subsequently trans-
formed on growth. Furthermore, we interpret the presence of these
small crystals as evidence of the continuous nucleation of t-ZrO2
Fig. 3. SEM images of Z3 at the 12.5 cm location: (a) cross-section and
(b) fiber surface.

2
034
Journal of the American Ceramic Society—Lee et al.
Vol. 86, No. 12
Fig. 5. XRD patterns of SZ1 through SZ5.
Fig. 4. SEM images of Z5 at the 12.5 cm location: (a) cross-section and
b) fiber surface.
(
thermodynamically stable above 1205°C. Several explanations
have been proposed for the stabilization of the high-temperature
t-ZrO (or c-ZrO ) phase observed at lower temperatures. Garvie
number density of t-ZrO2 nuclei observed for this sample was
considerably reduced from that observed for Z1. Figure 4(a) shows
a cross-section SEM image from sample Z5 produced with a
further increase of PO2 of 1.6 Pa. For this sample, only a few
2
2
8
et al. proposed that the lower surface energy of the t-ZrO phase
2
was the reason for the stabilization of t-ZrO when its size is below
2
a critical grain size (e.g., Ͻ10 nm at room temperature). This
phenomenon is often referred as the “size effect” in the literature.
There is also some evidence in the literature to support the
t-ZrO nuclei were observed in the coating layer as indicated by
2
the arrows in Fig. 4(a). Larger and columnar ZrO grains were
2
mostly observed in the coating layer. Also, it was interesting to
possible role of PO2 in controlling the nucleation behavior of ZrO .
observe that no t-ZrO nuclei were observed on the fiber surface as
2
2
9
Tomaszewski et al. observed that c-ZrO2 and/or t-ZrO2 were
shown in Fig. 4(b).
favorable to form when PO2 was below a certain limit in case of
The XRD patterns of the ZrO coating specimens deposited on
2
sintering undoped ZrO in an Al O matrix. c-ZrO was the only
Si substrates (SZ1, SZ2, SZ3, SZ4, and SZ5) are shown in Fig. 5.
2
2
3
2
Ϫ2
phase detected at PO2 Ͻ 2.80 ϫ 10 Pa, t-ZrO was the only
The coating specimens consisted of both t-ZrO and m-ZrO . The
2
2
2
phase detected at PO2 ϭ 0.18 Pa, and a mixture of t-ZrO and
(
111) peak at ϳ30.4° was detected for all of the specimens. With
2
t
m-ZrO was detected at P Ͼ 1.07 Pa. It appeared that these
increasing P , the relative amount of the t-ZrO phase continu-
2
O
2
O
2
2
observations were in qualitative agreement with our results,
ously decreased as shown in Fig. 6. The t-ZrO2 content was
៮
estimated from the peak intensity ratio of (111) /[(111) ϩ (111)
t
m
m
6
,7
ϩ (111)t]. Raman spectroscopy also confirmed the presence of
t-ZrO and m-ZrO phases in the coating specimens (not shown).
2
2
However, it was difficult to detect significant variations in the
t-ZrO phase content by the Raman analysis, as the t-ZrO peak
2
2
intensities were much weaker than those of the m-ZrO phase.
2
From our experimental results, two major trends in the coating
structure were observed with increasing P : (1) the degree of
O
2
continuous t-ZrO nucleation was decreased and (2) the coating
2
layer became more columnar and contained larger m-ZrO grains.
2
We note that it is difficult to directly relate the t-ZrO content with
2
the degree of continuous t-ZrO nucleation since (1) the t-ZrO
2
2
content was measured after the deposition and (2) a significant
fraction of t-ZrO nuclei might have been converted to m-ZrO
2
2
during the deposition. Nevertheless, we could expect that the
higher degree of t-ZrO continuous nucleation would result in the
2
presence of more t-ZrO content in the coating layer. When the
2
continuous nucleation of t-ZrO nuclei was suppressed, uninter-
2
rupted vertical growth of m-ZrO2 would be favorable at the
solid–vapor interface, resulting in the development of columnar
m-ZrO grains.
2
We consistently observe that t-ZrO could nucleate during the
Fig. 6. t-ZrO2 content as a function of PO2. The t-ZrO content is
2
2
៮
CVD process performed at 1050°C, whereas t-ZrO is known to be
estimated from the XRD peak ratio of (111) /[(111) ϩ (111)m ϩ (111)t].
2
t
m

December 2003
Chemically-Vapor-Deposited Zirconia on Hi-Nicalon Fiber and Silicon
2035
Fig. 8. Coating thickness measured by SEM in the axial direction and
within the fiber tow.
We compared the Raman peak intensity ratios between (1)
Ϫ1
Ϫ1
carbon and m-ZrO [i.e., I (1350 cm )/I (473 cm )] and (2)
2
c
m
Ϫ1
Ϫ1
carbon and t-ZrO [I (1350 cm )/I (260 cm ] as a function of
2
c
t
P . These ratios did not correlate with P , suggesting that the
O
2
O
2
carbon amount and the stabilization of the t-ZrO2 phase were
probably not related. At this time, we also do not understand why
the carbon phase was codeposited with the ZrO phases. The
2
1
3
thermodynamic calculations by Minet et al. indicated that high
temperatures, lower pressures, and high H /CO ratios would
2
2
result in the codeposition of carbon and ZrO . However, the
2
codeposition was not predicted at the experimental conditions that
Minet et al. and we used. This discrepancy could not be explained
with our current results. More work is needed to address these
carbon issues.
Fig. 7. Raman spectra showing carbon peaks of samples SZ1 through
SZ5.
Figure 8 shows the relationship between the reactor pressure
and coating uniformity for samples Z1, Z6, and Z7. Coating
thickness was measured from three axial locations, i.e., 7.5, 12.5,
and 17.5 cm measured at the inlet end of the substrate holder. Note
that temperature variations along the axial distance were measured
to be Ϯ40°C with the maximum temperature at the axial center.
The error bars in Fig. 8 represented thickness variations of the
coating within the fiber tow in the radial direction. The fiber tow
coated in experiment Z1 deposited at 4 kPa showed significant
thickness variations within the fiber tow as well as along the gas
flow direction. As Fig. 8 shows, we were able to improve coating
uniformity for samples Z6 and Z7 by reducing the reactor pressure
to 1.3 kPa. The reduced pressure clearly improved the uniformity
in the axial and radial directions because of (1) less depletion of
regents due to slow growth rate, (2) increased diffusivity of
although the situations were quite different, as we observed the
continuous decrease in the t-ZrO content with increasing P from
2
O
2
Ϫ3
4
ϫ 10 to 1.6 Pa.
10
Collins reported that oxygen vacancies played an important
role in the nucleation behavior of ZrO . The oxygen deficiency
2
effect is similar to that observed for the stabilization of t- and
c-ZrO by alloying them with another oxide such as Y O or CaO.
2
2 3
In the m-ZrO form, Zr atoms are 7-coordinated, whereas they are
2
8
-coordinated in the t- and c-ZrO2 structures. A high oxygen
vacancy concentration can provide a mechanism for the Zr atoms
in the t- or c-ZrO phase to have, on average, less than 8 oxygen
2
neighbors. The deficiency of oxygen could be, therefore, another
14,15
gaseous species, and (3) higher velocity of the gas mixture.
reason for promoting the continuous nucleation of t-ZrO at low
2
Figure 9 shows the surface morphology of Z6 along the flow
direction, i.e., 7.5, 12.5, and 17.5 cm axial locations. It is evident
that t-ZrO2 nuclei were continuously formed along the flow
direction as evident from the presence of small crystals on the
coating surface despite the local axial temperature variations. The
PO2 conditions in addition to the size effect.
Another important issue in understanding our results was the
1
1
presence of free carbon in the ZrO2 coating. Minet et al.
observed the presence of free carbon in the ZrO coating prepared
2
with the same precursors as we have used. They speculated that the
carbon amount in the ZrO2 coating could be influenced by
deposition temperature and reactor pressure. Furthermore, they
proposed that carbon might be a factor that can cause t-ZrO2
stabilization during the CVD process.
We also detected the presence of a small amount of carbon from
our coating samples by Raman spectroscopy as shown in Fig. 7.
The carbon phase in the coating samples appeared to be nanoc-
results from this study were used to prepare a set of ZrO fiber
2
coating specimens (including Z6 and Z7) for the SiC matrix
infiltration by chemical vapor infiltration (CVI) to prepare SiC/SiC
minicomposite specimens containing the ZrO fiber coating for
2
room-temperature mechanical and stress-rupture evaluation stud-
16
ies, as described elsewhere.
rystalline graphite, as indicated by the Raman peaks at 1350 and
IV. Conclusions
Ϫ1
1
600 cm . It has been reported that single-crystal graphite shows
Ϫ1
a Raman line at 1575 cm , whereas polycrystalline graphite
A set of ZrO coating specimens was prepared at 1050°C on
2
Ϫ1 12
TM
exhibits Raman bands at 1355 and 1575 cm
.
However, we
Hi-Nicalon
fiber and single-crystal Si by CVD using ZrCl4,
were not able to detect any crystalline or turbostratic carbon phase
from our XRD studies. Also, we did not detect any evidence of the
carbon phase throughout our SEM and TEM studies.
CO , and H as precursors while intentionally and systematically
2 2
varying the controlled amount of O into the CVD chamber. We
2
2
found that PO2 in the reactor chamber was a critical process

2
036
Journal of the American Ceramic Society—Lee et al.
Vol. 86, No. 12
0
.004 to 1.6 Pa. Also, the coating layer became more columnar and
contained larger m-ZrO grains. The observed relationships be-
2
tween PO2 and the nucleation and morphologic characteristics of
the ZrO2 coating were mainly attributed to the grain size and
oxygen deficiency effects which have been previously reported to
cause the stabilization of the t-ZrO phase in bulk ZrO specimens.
2
2
We detected the presence of a small amount of carbon in our
coating samples by Raman spectroscopy. However, it appeared
that the t-ZrO stabilization behavior was not strongly affected by
2
the presence of carbon in the coating layer since there was no clear
correlation among the carbon, t-ZrO , and m-ZrO amounts by the
2
2
Raman analysis. We reduced the reactor pressure to improve
coating uniformity in the axial and radial directions and prepared
a set of uniform fiber coating specimens for Hi-Nicalon/ZrO /SiC
2
minicomposite fabrication and evaluation.
Acknowledgments
We thank Dr. Hongyu Wang at General Electric Power Systems, Dr. Anteneh
Kebbede at General Electric Global Research Center, and Dr. Gregory N. Morscher
at NASA Glenn Research Center for insightful discussions during our investigation.
References
1
L. U. J. T. Ogbuji, “Pervasive Mode of Oxidative Degradation in a SiC–SiC
Composite,” J. Am. Ceram. Soc., 81 [11] 2777–84 (1998).
2
H. Li, J. I. Lee, M. Libera, W. Y. Lee, M. J. Lance, A. Kebbede, H. Wang, and
G. Morscher, “Morphological Evolution and Weak Interface Development within
CVD-Zirconia Coating Deposited on Hi-Nicalon Fiber,” J. Am. Ceram. Soc., 85 [6]
1
561–68 (2002).
3
H. Li, J. Lee and W. Y. Lee, “Effects of Air Leaks on the Phase Content,
Microstructure, and Interfacial behavior of CVD Zirconia on SiC Fiber,” Ceram. Eng.
Sci. Proc., 23 [3] 261–68 (2002).
4
E. H. Kisi and C. J. Howard, “Crystal Structure of Zirconia Phases and Their
Inter-Relation,” Key Eng. Mater., 153–154, 1–36 (1998).
5
M. J. Lance, J. A. Haynes, M. K. Ferber, and W. R. Cannon, “Monoclinic Zirconia
Distributions in Plasma-Sprayed Thermal Barrier Coatings,” J. Therm. Spray Tech-
nol., 9 [1] 68–72 (2000).
6
R. C. Garvie and R. S. Nicholson, “Phase Analysis in Zirconia Systems,” J. Am.
Ceram. Soc., 55 [6] 303–305 (1972).
7
D. J. Kim, “Influence of Aging Environment on Low-Temperature Degradation of
Tetragonal Zirconia Alloys,” J. Eur. Ceram. Soc., 17 [7] 897–903 (1997).
8
R. C. Garvie, “The Occurrence of Metastable Tetragonal Zirconia as a Crystallite
Size Effect,” J. Phys. Chem. Solids, 69, 214–20 (1991).
9
H. Tomaszewski and K. Godwod, “Influence of Oxygen Partial Pressure on the
Metastability of Undoped Zirconia Dispersed in Alumina Matrix,” J. Eur. Ceram.
Soc., 15, 17–23 (1995).
1
0
D. E. Collins, K. A. Rogers, and K. J. Bowman, “Crystallization of Metastable
Tetragonal Zirconia from the Decomposition of a Zirconium Alkoxide Derivative,”
J. Eur. Ceram. Soc., 15, 1119–24 (1995).
1
1
J. Minet, F. Langlais, and R. Naslain, “On the Chemical Vapour Deposition of
Zirconia from ZrCl –H –CO –Ar Gas Mixtures: II, An Experimental Approach,”
J. Less-Common Met., 132, 273–89 (1987).
4
2
2
1
2
F. Tuinstra and J. L. Koenig, “Raman Spectrum of Graphite,” J. Am. Ceram. Soc.,
3 [3] 1126–30 (1972).
5
1
3
J. Minet, F. Langlais, and R. Naslain, “On the Chemical Vapour Deposition of
Zirconia from ZrCl –H –CO –Ar Gas Mixtures: I, A Thermodynamic Approach,”
J. Less-Common Met., 119, 219–35 (1986).
4
2
2
1
4
L. F. Cheng, Y. Xu, L. Zhang, and X. Yin, “Preparation of an Oxidation
Protection Coating for C/C Composites by Low Pressure Chemical Vapor Deposi-
tion,” Carbon, 38 [10] 1493–98 (2000).
1
5
Fig. 9. SEM surface images of sample Z6 along the flow direction.
W. Y. Lee and G. Y. Kim, “Kinetic Considerations for Processing a Diffusion
NiAl Coating Uniformity Doped with a Reactive Element by Chemical Vapor
Deposition”; p. 149 in Elevated Temperature Coatings: Science and Technology III.
Edited by J. M. Hampikian and N. B. Dahotre. TMS, Warrendale, PA, 1999.
H. Li, G. Morscher, J. Lee, and W. Y. Lee, “Tensile and Stress-Rupture Behavior
of SiC/SiC Minicomposite Containing Chemically Vapor Deposited Zirconia Inter-
parameter for controlling the nucleation and growth characteristics
of the ZrO2 coating. It appeared that the degree of continuous
t-ZrO nucleation with m-ZrO decreased with increasing P from
16
phase,” J. Am. Ceram. Soc., submitted.
Ⅺ
2
2
O
2