Influence of precursor densification on strength retention of zirconia-coated Nextel 610 fibers

Article abstract of DOI:10.1111/j.1551-2916.2008.02282.x

Strength degradation of Nextel 610 fibers by continuous liquid phase coating was investigated for four different zirconia precursors. The precursors differed regarding their chemical composition (with or without yttrium), phase composition (amo

Full text of DOI:10.1111/j.1551-2916.2008.02282.x

J. Am. Ceram. Soc., 91 [4] 1070–1076 (2008)
DOI: 10.1111/j.1551-2916.2008.02282.x
r 2008 The American Ceramic Society
ournal
J
Influence of Precursor Densification on Strength Retention
of Zirconia-Coated Nextelt 610 Fibers
Matthias J. Bockmeyer and Reinhard Kru
¨
gerw
¨ ¨
Fraunhofer-Institut fur Silicatforschung ISC, 97082 Wurzburg, Germany
Strength degradation of Nextelt 610 fibers by continuous liquid
phase coating was investigated for four different zirconia pre-
cursors. The precursors differed regarding their chemical com-
position (with or without yttrium), phase composition
monazite coatings obtained with particulate precursors. These
rabdophane precursors contained no carbon compounds
and were deliberated from nitrate by a washing procedure.
8
1
2
4
Similar results were obtained for CePO coatings.
(
amorphous or crystalline), and decomposition behavior. Phase
Further, it was observed that oxide coatings with a fugitive car-
bon phase did not affect fiber strength as long as carbon remained
7
transformation and densification of the films were characterized
and found to depend on the kind of precursor. Single fiber
Weibull’s strength was measured for calcination temperatures
between 2501 and 11501C for all precursors. Each precursor had
an individual degradation behavior. For an annealing tempera-
ture of 11501C highly damaged (B1600 MPa) and undamaged
in the coating (calcination in inert gas). Degradation occurred
when carbon was removed by an additional thermal oxidation.
A thermo-mechanical cause for the degradation of fibers was
2
,13
pointed out by Parthasarathy et al. : Stresses in the coating
that derive from a larger CTE in the coating, which is strongly
bonded to the fiber surface, can degrade the fiber.
(43300 MPa) fibers were obtained depending on the kind of
precursor. Fiber degradation could be correlated to mechanical
stresses. Stress concentration due to inhomogeneous film thick-
ness distribution is proposed as the cause of fiber strength
degradation. Full strength could be retained for porous coatings
or coatings where stresses were reduced by phase transformation.
The role of a solid-state interaction between the fiber and ions
from the coating was assessed by several authors. Cinibulk
15
investigated the solid-state reactivity by dip coating and
calcination of Nextelt 610 fibers with the respective nitrate
21
21
salt solutions. He reported that Ca or Sr degraded the fiber
31
14
at 10001C (10 h) while La ions did not. Davis et al. found
that Nextelt 610 fibers were degraded after coating with LaPO
4
I. Introduction
from different salt solutions. This degradation could be sup-
pressed by the addition of alumina powder that presumably
acts as a buffer and even slightly enhanced fiber bundle strength.
NTERPHASE materials between fiber and matrix are utilized to
control the mechanical behavior of ceramic matrix compos-
I
1
0,11
1,2
Boakye and colleagues
disproved a degrading solid-state
ites. An economic process to apply this interphase is dip coat-
ing with a precursor liquid (e.g., a sol) and subsequent
calcination. For such liquid-phase-coated fibers strength degra-
dation has been reported by several authors and for several
reaction of lanthanum or phosphate ions with the fiber: Stoic-
hiometric, lanthanum-rich or phosphate-rich monazite coatings
did not result in a significant difference of the fiber strength.
3
–6
Hence, something other than AlPO
(LaAl O ) formation was charged to degrade the fiber.
11 18
4
or magnetoplumbite
coating-fiber systems. Also nondegrading coatings were pre-
1
1
7
8
sented by Cinibulk et al. and Boakye et al. Most studies were
done with either Nextelt 720 (composed of corundum and
mullite) or Nextelt 610 (composed of corundum) fibers. Sev-
eral correlations were investigated and different reasons were
proposed to explain the degradation phenomenon. These cor-
relations and other important findings shall be briefly itemized.
Finally it should be mentioned that coating thickness and the
number of coatings were found to have no influence.
1
0,12
In general, it must be stated that the degrading mechanisms
are not completely understood and individual coating processes/
systems are different. As far as known to the authors, direct
2
evidence of the degradation mechanism has not yet been found.
In this paper we present a study on the strength of Nextelt 610
fibers coated with stabilized and unstabilized zirconia derived
Extensive studies with different monazite (LaPO ) precursors
4
9
–11
coated on Nextelt 720 fibers
showed a dependence on indi-
vidual precursor chemistry. It was concluded that surface-active
decomposition products are the cause of the following degrada-
tion mechanism: Intergranular surface flaws of the fiber are
attacked by surface-active decomposition products of the coat-
ing. Growth of the flaws is driven by the stress from the thermal
expansion difference between alumina and mullite grains within
the fiber. A nitrogen-containing species (released from the
nitrate-based precursor) was speculated to act as the surface-
active gas species, yet could not be unambiguously identified. A
correlation was seen between fiber strength and weight loss of
16
from different precursors. Zirconia coatings on oxide and SiC
have previously been reported. In nonoxide composites
17,18
fibers
these coatings can weaken the fiber–matrix interface especially
when zirconia is not stabilized in its cubic polymorph. For oxide
fibers and composites, zirconia coatings are of little technical
importance and our interest is to understand fundamental aspects
of the coating process and its effect on the fiber properties. From
the phase diagram of the ternary system Al
ACerS-NIST Phase Equilibria Diagram No. 06825) it can be
seen that an alumina fiber cannot form mixed oxides with neither
ZrO nor 8 mol% yttria-stabilized zirconia (8YSZ).
2 3 2 2 3
O –ZrO –Y O
(
8
the monazite precursor at the particular temperature. A low
coating permeability was considered to promote stress corrosion
by trapping decomposition products at high partial pressure at
the fiber-coating interface whereas high a coating porosity al-
lows decomposition products to escape rather than attack the
2
The aim of the study was to investigate which process vari-
ables affect the strength of zirconia coated fibers. Furthermore,
it was intended to assess if one of the above stated degradation
mechanisms applies to our fiber coatings.
4
,9–11
fiber.
Fiber strength could be retained for more porous
M. Cinibulk—contributing editor
II. Experimental Procedure
(
1) Precursor Synthesis
Two amorphous precursor powders were synthesized for
zirconia as follows: Zirconium iso-propylate was reacted with
Manuscript No. 23149. Received April 29, 2007; approved November 26, 2007.
Author to whom correspondence should be addressed. e-mail: krueger@isc.fhg.de
w
1
070

April 2008
Influence of Precursor Densification on Fiber Strength Retention
1071
acetylacetone and afterwards hydrolyzed with water. Then, all
volatile components were removed by rotational evaporation at
reduced pressure and an amorphous precursor powder was ob-
assumed. This value was the average fiber diameter of the lot
determined before by secondary electron microscopy (SEM;
Model S-800, Hitachi Ltd., Tokyo, Japan) analysis of 20 fibers.
Weibull’s reference strengths were calculated from at least 30
tested fibers by the software ExcelToWeibull (Ceramic Materials
and Components, Department of Production Engineering
University of Bremen, Germany).
For SEM analysis original fracture surfaces of tested fibers
were used. Here for secondary fractures caused by the release of
the fracture energy were suppressed.
In tensile tests of single ceramic fibers usually the release of
fracture energy at the rupture event leads to secondary fractures.
In order to observe where the failure of the tested fiber origi-
nated, the primary fracture surface has to be obtained. For this
purpose secondary fractures were suppressed by the application
of grease as a damping medium onto the fiber. Before SEM
investigation the grease was removed by acetone.
2
tained (designated ‘‘A_ZrO ’’). For the yttria-stabilized
precursor (‘‘A_8YSZ’’) 8 mol% yttrium acetate was added
after hydrolysis. Further synthesis details are given in the
1
9–21
literature.
Coating sols with an oxide yield of 8 wt% respective to ZrO
2
or 8YSZ were prepared by dissolution of the amorphous pre-
cursor powder in water. After stirring over night they were fil-
2
0,22
tered through a 0.45 mm membrane.
The cationic surfactant
hexadecyltrimethylammoniumbromide was added below the
critical micelle concentration (0.1 wt% respect to water content)
in order to improve the fiber wetting.
The two crystalline precursor sols—unstabilized ZrO
C_ZrO ) and 8YSZ (C_8YSZ)—were obtained from the Chair
for Chemical Technology of Materials Synthesis, University of
Wurzburg, Germany in an oxide concentration of 8 mass% and
2
(
2
¨
were used without any further purification. Synthesis of the
crystalline precursors occurred in an aqueous medium under
hydrothermal conditions. Preparation route was identical for
C_ZrO2 and C_8YSZ except for the addition of Yttrium
acetate.
III. Results and Discussion
(
1) Characterization of Precursors
Two of the four investigated zirconia precursors contained
mol% yttria as dopant (8YSZ) and two were without dopant
ZrO ). One of the 8YSZ precursors and one of the ZrO pre-
cursors contained amorphous colloidal particles (designation A)
and the other contained crystalline particles (designation C).
XRD measurements of coatings prepared from the sols con-
taining crystalline particles and dried at room temperature re-
vealed that the phase composition with yttria as a dopant is pure
cubic zirconia and without a dopant it is a mixture of mono-
clinic and amorphous phase (Table I).
8
(
2
2
(2) Precursor Characterization
Small-angle X-ray scattering-measurements (SAXS) were done
for evaluating the particle size in the sols. A Kratky compact
camera (Anton Paar GmbH, A-8054 Graz, Austria) was used to
observe the scattering of X-rays obtained from a 2 kW X-ray
generator (Philips PW1731, Eindhoven, the Netherlands).
Thin films were applied on glass plates by multiple dip coat-
ing (approximately 30 times with 50 cm/min and a drying time
of 10 min after each coating step). Then, the dried coating material
was scraped off by a razor blade. As described by Bockmeyer
The precursor sols with the amorphous particles A_8YSZ
and A_ZrO have an appearance of clear orange–brown, while
2
the precursors with crystalline particles (C_8YSZ and C_ZrO2)
show Tyndall’s effect. This indication of larger particles was
proven through particle size measurements by SAXS (Table I).
The particles found in the two precursors are approximately
four times larger in size than the precursors with the amorphous
particles (Table I).
Thermal analysis of the precursor material scraped off coated
glass plates (without calcination) is presented in Figs. 1 and 2.
All coating materials show their major mass loss at temperatures
o4501C (Fig. 1) which can be directly referred to the endother-
mic evaporation of water (201–2001C) and the exothermic de-
2
3
et al. this scraped-off material behaves more similar to a real
coating on a substrate than the bulky precursor and was used
for the thermal analysis of the precursors. Differential thermal
analysis (DTA) and thermogravimetric analysis (TGA; Setaram
TAG24, Caluire, France) of the scraped-off coating materials
were done at a heating rate of 10 K/min in dry air atmosphere
2 2
(499.999% N /O 5 79/21).
Thin film grazing incidence X-ray diffractometry (XRD;
Siemens D-5005, Bruker AXS GmbH, Karlsruhe, Germany)
was performed at an angle of incidence g of 0.351. Macroscopic
film density was measured by X-ray reflectometry (XRR) using
2
0,23
composition reactions of organic residues (Fig. 2).
materials obtained from amorphous particles (A_8YSZ,
A_ZrO ) additionally show a second mass loss at around
001C, which can be explained by an exothermic decomposi-
Coating
a
Siemens D-5005 diffractometer (Bruker AXS GmbH)
equipped with a reflectometry sample stage. The sample height
was adjusted by the edge slit. Thin film density was evaluated by
fitting the measured XRR data with the software DIFFRAC
2
8
plus
2
1
tion reaction of residual carbon.
Generally, the amorphous coating materials (A_8YSZ and
A_ZrO ) show a larger mass loss (B50 mass%) than the crys-
2
4
REFSIM 2.0 (Bruker AXS GmbH). Fitting of the reflection
curves was done based of on a single layer model on a pure Si
substrate. The film compositions were taken as pure 8YSZ and
2
talline coating materials C_8YSZ and C_ZrO (B12 mass%).
2
2
ZrO , respectively.
This points to the fact that the amorphous particles contain
a larger amount of organic residues, like alkoxide and
acetylacetonate groups. For coating materials obtained from
amorphous particles it has been previously reported that the
temperature region where the major part of the mass loss occurs
(3) Coating Procedures
Planar coatings on Si wafers were prepared at a drawing speed
of 50 cm/min calcinated at different temperatures in a preheated
furnace (dwell time 10 min).
Nextelt 610 fibers were continuously desized, coated, and
calcined at a speed of 50 cm/min. The hot zone of the calcination
furnace is approximately 30 cm long resulting in a dwell time of
2
0,23
is also accompanied by a high increase in skeletal densities.
So, the densification behavior of thin films prepared from sols
0
.6 min at the indicated temperature. The coating procedure and
Table I. Phase Composition and Particle Size of Zirconia
Precursors Used for Fiber Coating Experiments
2
0
¨
apparatus is described in more detail by Kruger et al.
For tensile testing at room temperature single fibers were
extracted from the coated tows and fixed in the rubber clamps of
a test machine (Model Zwicki 1120, Zwick GmbH & Co., Ulm,
Germany) with a cross-head speed of 1.45 mm/min and a gauge
length of 20 mm. Alignment of the fiber with the direction of the
load was carefully regarded. From the force–displacement
curves Young’s modulus and tensile strength were determined.
For calculation of the stress a fiber diameter of 11.7 mm was
Precursor
Phase composition (XRD)
Particle size (nm) (SAXS)
A_ZrO2
Amorphous
2.6
12.8
2.1
C_ZrO2
Monoclinic1amorphous
Amorphous
Cubic
A_8YSZ
C_8YSZ
9.5
XRD, X-ray diffractometry; SAXS, small-angle X-ray scattering.

1
072
Journal of the American Ceramic Society—Bockmeyer and Kru¨ger
Vol. 91, No. 4
on Si wafers under comparable conditions. The influence of the
type of substrate (Si wafer of Al fiber) on the thin film
2 3
O
densification and phase transformation was assessed to be minor
compared with the influence of precursor chemistry and
calcination temperature.
The results of the GIXRD measurements of coatings calcined
at temperatures of 4001–11501C are presented in Figs. 3 and 4.
Coatings prepared from sols containing amorphous particles
show reflections for crystalline phases starting at temperatures
of 4001C. However, as expected, coatings prepared from
2
sols C_ZrO and C_8YSZ were already crystalline at room
temperature.
Coatings that were prepared from yttria-doped precursor sols
A_8YSZ and C_8YSZ) resulted in the cubic phase at all inves-
(
tigated temperatures (Fig. 3). The increase of the full width at
half-maximum of X-ray reflections qualitatively indicates an
increase in the crystallite size as it has been shown in detail
2
0,21
previously.
However, when thin films were prepared with nonstabilized
zirconia-coating sols (A_ZrO and C_ZrO ), a more detailed
distinction has to be made between the two precursors types
(Fig. 4). At a calcination temperature of 4001 and 7001C coat-
ings prepared with the sol consisting of amorphous particles
Fig. 1. Thermogravimetric analysis of dried film powders scraped off
glass substrates without calcination. The four different coating sols con-
tained amorphous particles (designation A) or crystalline particles
2
2
(designation C). Two precursors were stabilized by yttria (8YSZ) and
two were undoped (ZrO ).
2
(
2
A_ZrO ) present only reflections of the tetragonal-zirconia
phase (Fig. 4 (left)). After the 10001C calcination step, a
mixture of the monoclinic and the tetragonal polymorphs is
observed. For the calcination temperature of 11501C there are
only reflections for monoclinic phase.
containing amorphous and crystalline particles is expected to
differ considerably. The synthesis procedure of the variants
2
8YSZ and ZrO are very similar for the amorphous and crys-
talline precursors, respectively. Particle sizes hardly differ from
each other (Table I). Their thermal analysis results are also very
likewise (Fig. 2). Thus, for the amorphous and the crystalline
precursors, decompositional release of gaseous species from the
8YSZ variant can be expected to resemble that of the ZrO2
variant.
Thin sol–gel films that are exposed to low calcination tem-
peratures develop a crystallite size in the nanometer (nm) range.
Owing to the enhanced contribution of surface energy to the
total free energy, phase stability regions of such small-grained
microstructures can be shifted. Instead of the monoclinic phase,
which is the common low-temperature polymorph of undoped
26
zirconia, tetragonal zirconia can become thermodynamically
By raising the temperature, grain growth in the
2
7,28
stable.
(
2) Crystallization and Densification of Zirconia Coatings
zirconia films leads to an increasing formation of the mono-
clinic phase.
A different phase formation was observed in the case of the
Film porosity analysis from transmission electron microscope
images is mostly vague and insufficient for quantitative deter-
mination of the densification behavior. By XRR reliable quan-
tification of film densities is possible, provided the composition
coatings synthesized with the sol C_ZrO , which contains
2
a mixture of mainly monoclinic and only a minor fraction of
amorphous particles (Table I). Over the whole temperature
range the monoclinic modification of zirconia was identified
(Fig. 4, right). However, at relatively low temperatures like
7001C the observed phase composition additionally consists of
a minor portion of tetragonal crystallites.
2
5
is known. Unfortunately, due to the curved surface of coated
fibers, it is neither possible to measure the macroscopic density
by XRR, nor can phase composition be determined in good
quality with grazing incidence X-ray diffraction (GIXRD).
Hence, in order to obtain information about macroscopic den-
sity and phase development of zirconia films, XRR and GIXRD
measurements were carried out with coatings that were prepared
This tetragonal phase presumably evolves from the amor-
phous particles contained in the sol. Analogous to the phase
Fig. 2. Differential thermal analysis of dried film powders scraped off glass substrates without calcination. The four different coating sols contained
amorphous particles (designation A) or crystalline particles (designation C). Two precursors were stabilized by yttria (8YSZ) and two were undoped
(
2
ZrO ).

April 2008
Influence of Precursor Densification on Fiber Strength Retention
1073
Fig. 3. Grazing incidence X-ray diffraction pattern of 8YSZ coatings prepared on Si-wafer using a coating sol either consisting of amorphous particles
A_8YSZ) (left) or crystalline particles (C_8YSZ) (right).
(
evolution of the coating sol A_ZrO , these amorphous particles
2
residues decompose and the material crystallizes which leads to
a substantial densification and shrinkage process in the same
temperature regime. Considering this circumstance the amor-
phous films annealed at 5501C have already gone through con-
siderable densification. Additional heating up to 11501C causes
skeletal densification and sintering, which lead to a further
increase of the macroscopic film density close to the theoretical
tend to form tetragonal zirconia at low temperatures due to their
very small grain size. At higher temperatures, transformation to
monoclinic zirconia is induced by crystallite coarsening.
Figure 5 shows the evolution of the macroscopic density in
dependence of the calcination temperatures of the coatings.
Measurement of macroscopic density by XRR usually requires
an exact knowledge of the elementary composition of the coat-
ings as otherwise the error of measurement is significantly large.
Thus, due to the high organic content, no measurements were
done for coatings prepared with amorphous coating sols for
calcination temperatures below 4001C. The given densities
represent the overall density of the film.
2
1
density of zirconia.
The densification behavior of coatings prepared with crystal-
line particles is in contrast to that of coatings prepared with sols
containing amorphous particles. The macroscopic density of the
former coatings starts at a very low level and stays nearly
unchanged until a first significant rise in density is observed at
calcination temperatures ꢀ10001C. From 10001C up to
11501C, the largest increase of density close to the theoretical
density of zirconia occurs.
When already crystalline particles with a low content of or-
ganic residues and a high skeletal density are used, during the
dip coating and pyrolysis (o7001C) procedure, no major addi-
tional network densification can occur. The film density is most-
ly defined through the packing and aggregation of the particle
during the dip coating procedure. Further increase in thin film
density can then only be achieved by sintering at the corre-
sponding temperatures (47001C).
Densification of this kind of the amorphous precursors leads
to a dense surface layer and closed porosity underneath as soon
2
3,29
as organic is degassed ꢀ3001C.
Elimination of the
closed pores occurs with higher calcination temperatures. To
control XRR densities of films, SEM investigations were done
for fiber coatings calcinated at 7001C (Fig. 6). Coatings showed
an open porous microstructure for the film derived from the
crystalline precursors and dense coating surfaces for films
from amorphous precursors. This is in good correlation with
the densification behavior of coatings on Si wafers observed
by XRR.
As already described in Section III (1), the observed densifi-
cation behavior of coatings prepared from sols containing either
amorphous particles or already crystalline particles is consider-
ably different. The retarded densification of the crystalline pre-
cursors can be explained from the larger size and crystalline
nature of the particles. The evolution of density is only slightly
influenced no matter whether yttria-doped or undoped zirconia
coating sols were used.
(3) Strength of Coated Fibers
The tensile strengths were measured for single fibers extracted
from the bundles after coating with one of the four precursors
and for different calcination temperatures. In Fig. 7 the Weibull
reference strengths are displayed as a function of the calcination
temperature. Table II additionally lists strength data and Wei-
bull’s moduli. Strength of uncoated fibers that passed through
the coating apparatus (without a precursor bath) at 11501C is
The density of a green film prepared from amorphous zir-
conia precursors is about 1.6 g/cm . Upon annealing, organic
3
21
Fig. 4. Grazing incidence X-ray diffraction pattern of ZrO
(A_ZrO ) (left) or crystalline particles (C_ZrO ) (right).
2
coatings prepared on Si-wafer using a coating sol consisting either of amorphous particles
2
2

1
074
Journal of the American Ceramic Society—Bockmeyer and Kru¨ger
Vol. 91, No. 4
MPa) for all higher calcination temperatures. For C_8YSZ,
fiber strength degradation starts at 10001C and becomes more
severe at 11501C.
Following the results on the thermal decomposition and
densification, the two major differences between the two pre-
cursors are obvious: the amounts of organic decomposition
products and the temperature intervals where densification of
the films occur.
In the case of the A_8YSZ precursor the major part of de-
composition reactions and film shrinkage both coincide with the
degradation of the fiber strength.
For the C_8YSZ, decomposition reactions are completed
around 5001C (Fig. 2, left) yet strength degradation starts
at temperatures above 7001C simultaneously with the film
densification. Note that the film densification coincides with
the strength degradation for both precursors.
.
.
.
.
.
.
.
.
2
For unstabilized ZrO precursors, up to 7001C the same
strength behavior as for the 8YSZ precursors is observed, the
amorphous precursor degrades the fiber from 5501C on, while
the crystalline precursor does not affect the fiber strength (Fig. 7
(
right)).
Fig. 5. Macroscopic density of thin films on Si-wafer determined by
X-ray reflectometry as function of annealing temperature. Films were
prepared by using four different coating sols consisting of amorphous
particles (designation A) or crystalline particles (designation C). Two
precursor were containing yttria as dopant (8YSZ) and two were
Above 7001C however, the ZrO precursors show a different
2
behavior compared with the 8YSZ precursors. For the crystal-
line precursor C_ZrO , no fiber strength degradation is observed
up to 11501C. The strength of fibers coated with the amorphous
2
2
undoped (ZrO ).
precursor A_ZrO
2
is unaffected (10001C) or only slightly
reduced (11501C).
also indicated (dashed line). The strength of these fibers is
virtually equal to that of as-received fibers. This proves that
fibers are neither mechanically damaged by the transport system
nor from exposure to the calcination temperature. Hence, fibers
with decreased strength are degraded through the coating.
First the strength of fibers coated with yttria-stabilized pre-
cursors is treated (Fig. 7, left). For the amorphous precursor
Precursor decomposition (Fig. 2) and film densification
(Fig. 5) are analogous to the 8YSZ precursors. Up to 7001C
for the amorphous precursor, decomposition reactions and film
densification coincide with fiber strength degradation. For the
crystalline ZrO precursor, the observed decomposition reac-
2
tions do not result in a loss of fiber strength. Only minor film
densification has occurred up to 7001C.
(
A_8YSZ), fiber strength drops off from the value of the un-
Interestingly for high calcination temperatures ( ꢀ10001C),
where decomposition of organics is already finished, doped and
coated fibers at ꢀ 5501C and remains at a low level (B1600
Fig. 6. Secondary electron microscope images of zirconia films on Nextelt 610 fibers after calcination at 7001C (top view, unsputtered). Films were
prepared by using the four different sols consisting of amorphous particles or crystalline particles. Two precursors were containing yttria as dopant
2
(8YSZ) and two were undoped (ZrO ).

April 2008
Influence of Precursor Densification on Fiber Strength Retention
1075
Fig. 7. Single fiber Weibull’s strengths as function of annealing temperature of Nextelt 610 fiber coated with different zirconia precursors. Strength
depends on the particular precursor and temperature. Amorphous (designation A) and crystalline (designation C) precursors exhibit different degra-
dation behavior for yttria doped (left) and undoped (right) zirconia. The strength of an uncoated fiber that passed at 11501C through the coating
apparatus is given as reference (dashed line).
undoped precursors show different degradation behavior al-
though their densification is nearly equal (depending on the
precursor type). The only difference between the respective pre-
cursors is the observed phase evolution (Figs. 3 and 4). The sta-
bilized precursors exhibit no phase transformation (Fig. 3)
whereas in the undoped precursors the tetragonal crystallites
transform to the monoclinic phase between 7001 and 10001C
For the degradation mechanism two explanations are possible.
First, thermal expansion mismatch between a strongly bonded
coating and the fiber leads to mechanical stresses that cause
2,13
defects in the fiber.
The CTE of ZrO or 8YSZ is higher than
2
the one of Al . When the coated fiber cools down, tensile stress
2 3
O
arises in the coating. This thermal stress is highest for dense
coatings which have a higher Young’s modulus than porous
coatings (as an example it is mentioned that the relative Young’s
Modulus of a silica gel with 40% porosity is only 1/5 of the fully
(
Fig. 4). This phase transformation, which is associated with a
volume expansion, occurs in the same temperature regime where
the doped precursors degrade the fiber strength.
In the following, the observed fiber degradation behavior
shall be discussed with regard to the degradation mechanisms
described in ‘Introduction’.
32
dense material). Furthermore bond strength between coating
and fiber surface is expected to be higher for the dense coatings.
Second, constrained shrinkage of the film can also result in
tensile stress of the coating. Stresses in densifying zirconia coat-
3
0
Comparing the strength for 8YSZ- and ZrO -coated fibers, it
2
may be assumed that Y ions undergo a degrading solid-state
ings can be as high as 800 MPa and abruptly rise during
decomposition of organic compounds and densification. As a
31
31
21
reaction with the fiber as found for Ca
Cinibulk Therefore additional coating experiments with an 8
wt% yttrium acetate solution were performed. Up to calcination
temperatures of 11501C the strength of these fibers was un-
21
or Sr ions by
consequence of the strong dependency of elastic modulus on
porosity, large stresses evolve from film densification only for
quite dense coatings.
1
5
For unstabilized zirconia coatings the densification stresses
can be relaxed and/or the interface bonding between the coating
and the fiber can be weakened when tetragonal to monoclinic
phase transformation occurs. This phase transformation in
31
affected (3405 MPa at 11501C). Thus fiber degradation by Y
ions can be ruled out.
Strength reduction of the fibers by stress corrosion from sur-
face-active species which are released during precursor decom-
3
3
unstabilized zirconia leads to a volume expansion of 8 vol%.
Hence, densified and transformed ZrO coatings (10001 and
11501C for A_ZrO and C_ZrO ) do not degrade the fiber
strength while densified but untransformed ZrO2 coatings
(A_ZrO 7001C) do.
6
,8
position, has been mentioned. For our results no consistent
correlation between decomposition of organics and fiber
strength can be seen. For instance, precursors with nearly iden-
tical composition and decomposition reactions (e.g., A_8YSZ
2
2
2
2
and A_ZrO ) lead to very different fiber strengths at high tem-
2
peratures. Therefore this point was excluded.
Irrespective of its origin, the tensile stress of the coating is
compensated by a compressive stress in the fiber. In case of an
ideal homogenous film thickness the resulting homogenous
stress distribution should not lead to strength degradation.
However, real coatings from liquid precursors always have an
inhomogeneous film thickness and there are coating bridges be-
tween neighboring fibers. This results in an inhomogeneous
stress distribution and stress concentration.
This stress concentration is assumed to be the cause of fiber
strength degradation observed for our coatings. From our re-
sults it cannot be distinguished whether these stresses arise from
From our results we reflect that for all degraded fibers the
coatings were nearly densified. For the undamaged fibers the
coating was either still porous or had undergone a phase
4
transformation. Hay et al. also reported a correlation of
2
0
strength reduction of Nextelt 720 fibers and the densification
and pore coarsening of the coating. He explained the benefit of
permeable porosity by an easier release of gaseous decomposi-
tion products which prevents fiber degradation by surface-active
4
,8
species.
Table II. Weibull Strength and Weibull Modulus m of Zirconia-Coated Nextelt 610 Fibers
Temperature (1C)
C_ZrO
2
strength (MPa)
m
C_8YSZ strength (MPa)
m
A_8YSZ strength (MPa)
m
A_ZrO
2
strength (MPa)
m
2
4
5
7
000
150
50
00
50
00
3146
3137
3262
3121
3427
3339
12.0
9.7
10.2
10.2
13.7
10.7
3330
3353
3318
3053
3060
2022
19.1
10.0
10.1
12.1
9.5
3177
3175
1900
1522
1769
1718
9.8
14.7
17.7
9.4
9.2
7.3
3088
2948
2012
1849
3213
2887
14.5
15.6
13.9
14.1
9.2
1
1
5.7
10.6
Strength of fibers passing through the apparatus at 11501C without coating is 3395 MPa (m 5 7.8).

1
076
Journal of the American Ceramic Society—Bockmeyer and Kru¨ger
Vol. 91, No. 4
4
R. S. Hay, E. E. Boakye, and M. D. Petry, ‘‘Effect of Coating Deposition
Temperature on Monazite Coated Fiber,’’ J. Eur. Ceram. Soc., 20, 589–97 (2000).
5
J. B. Davis, D. B. Marshall, and P. E. D. Morgan, ‘‘Monazite-Containing
Oxide/Oxide Composites,’’ J. Eur. Ceram. Soc., 20, 583–7 (2000).
R. S. Hay and E. E. Boakye, ‘‘Monazite Coatings on Fibers: I, Effect of Tem-
6
perature and Alumina Doping on Coated-Fiber Tensile Strength,’’ J. Am. Ceram.
Soc., 84 [12] 2783–92 (2001).
M. K. Cinibulk, T. A. Parthasarathy, K. A. Keller, and T.-I. Mah, ‘‘Porous
7
Yttrium Aluminum Garnet Fiber Coatings for Oxide Composites,’’ J. Am. Ceram.
Soc., 85 [11] 2703–10 (2002).
E. E. Boakye, R. S. Hay, P. Mogilevsky, and L. M. Douglas, ‘‘Monazite
8
Coatings on Fibers: II, Coating without Strength Degradation,’’ J. Am. Ceram.
Soc., 84 [12] 2793–801 (2001).
E. E. Boakye, R. S. Hay, M. D. Petry, and T. A. Parthasarathy, ‘‘Sol–Gel
9
Synthesis of Zircon–Carbon Precursors and Coatings of Nextel 720t Fiber
Tows,’’ Ceram. Eng. Sci. Proc. A, 20 [3] 165–72 (1999).
E. E. Boakye, R. S. Hay, and M. D. Petry, ‘‘Continuous Coating of Oxide
10
Fiber Tows Using Liquid Precursors: Monazite Coatings on Nextel 720t,’’ J. Am.
Ceram. Soc., 82 [9] 2321–31 (1999).
R. S. Hay and E. E. Boakye, ‘‘Monazite Coatings on Fibers: I, Effect of
11
Temperature and Alumina Doping on Coated-Fiber Tensile Strength,’’ J. Am.
Ceram. Soc., 84 [12] 2783–92 (2001).
E. E. Boakye and P. Mogilevsky, ‘‘Fiber Strength Retention of Lanthanum-
12
and Cerium Monazite-Coated Nextelt 720,’’ J. Am. Ceram. Soc., 87 [2] 314–6
(2004).
T. A. Parthasarathy, C. A. Folsom, and L. P. Zawada, ‘‘Combined Effects of
Exposure to Salt (NaCl) Water and Oxidation on the Strength of Uncoated and
Fig. 8. Secondary electron microscope images of a typical fracture
surface of degraded coated Nextelt 610 fiber.
13
BN-Coated Nicalont Fibers,’’ J. Am. Ceram. Soc., 81 [7] 1812–8 (1998).
J. B. Davis, D. B. Marshall, and P. E. D. Morgan, ‘‘Monazite-Containing
Oxide/Oxide Composites,’’ J. Eur. Ceram. Soc., 20, 583–7 (2000).
M. K. Cinibulk, ‘‘Hexaluminates as a Cleavable Fiber-Matrix Interphase:
Synthesis, Texture Development, and Phase Compatibility,’’ J. Eur. Ceram. Soc.,
a thermal expansion mismatch, the constrained film shrinkage,
or both.
14
15
For fractographic analysis original fracture surfaces of de-
graded coated fibers were studied with SEM. As a typical ex-
ample a fracture surface of a fiber coated with C_8YSZ and
calcination at 11501C is given in Fig. 8. It can be clearly seen
that the fracture starts from the surface at a position where film
thickness is larger. This observation appears for all four coating
precursors. This confirms the assumption that stress concentra-
tion due to inhomogeneous coating thickness is the cause of
fiber degradation.
2
0, 569–82 (2000).
X. Gu, P. A. Trusty, E. G. Butler, and C. B. Ponton, ‘‘Deposition of Zirconia
16
Sols on Woven Fibre Performs Using a Dip-Coating Technique,’’ J. Eur. Ceram.
Soc., 20, 675–84 (2000).
17
N. I. Baklanova, A. T. Titov, A. I. Boronin, and S. V. Kosheev, ‘‘The Yttria-
Stabilized Zirconia and Interfacial Coating on Nicalont Fiber,’’ J. Eur. Cer. Soc.,
6 [9] 1725–36 (2006).
2
18
H. Li, J. Lee, M. R. Libera, W. Y. Lee, A. Kebbede, M. J. Lance, H. Wang,
and G. N. Morscher, ‘‘Morphological Evolution and Weak Interface Develop-
ment within Chemical-Vapor-Deposited Zirconia Coating Deposited on Hi-
Nicalont Fiber,’’ J. Am. Ceram. Soc., 85 [6] 1561–8 (2002).
1
9
P. C. Lo
Prepared from Soluble Powders and their Application,’’ J. Sol–Gel Sci. Technol.,
9, 473–7 (2000).
¨
bmann, R. Jahn, S. Seifert, and D. Sporn, ‘‘Inorganic Thin Films
IV. Conclusion
1
20
Crystalline and amorphous ZrO
2
and 8YSZ precursors showed
¨
¨
R. Kruger, M. J. Bockmeyer, A. Dutschke, and P. C. Lobmann, ‘‘Continuous
Sol–Gel Coating of Ceramic Multifilaments: Evaluation of Fiber Bridging by
Three Point Bending Test,’’ J. Am. Ceram. Soc., 89 [7] 2080–8 (2006).
different effects on the strength coated Nextelt 610 fibers. The
precursors and resulting films differed in their decomposition,
phase transformation and densification behavior.
21
¨
C. Peters, M. Bockmeyer, R. Kruger, A. Weber, and E. Ivers-Tiffee, ‘‘Pro-
´
cessing of Dense Nanocrystalline Zirconia Thin Films by Sol–Gel, in Current and
Future Trends of Functional Oxide Films’’; pp. GG16-01 in Materials Research
Society Symposium Proceedings, Vol. 928E, Edited by D. Kumar, V. Craciun,
M. Alexe, and K. K. Singh. Warrendale, PA, 2006.
It was found that stress corrosion by precursor decomposi-
tion products was not a sufficient explanation for the observed
fiber degradation. Rather, the individual densification behavior
of the coatings was in good correlation with fiber strength.
In densified coatings large tensile stresses are present by ther-
mal expansion mismatch and the shrinking process constrained
by the fiber. Stress concentration arising from inhomogeneous
film thickness was identified as the cause of fiber degradation.
Degradation can be avoided for coatings that scarcely
densify, have weak adhesion to the fiber or can relax stresses
by e.g. phase transformation.
22
¨ ¨
2
P. C. Lobmann and P. Rohlen, ‘‘Industrial Processing of TiO Thin Films
from Soluble Precursor Powders,’’ Glass Sci. Technol., 76 [3] 1–7 (2003).
¨
M. Bockmeyer and P. Lobmann, ‘‘Densification and Microstructural Evolu-
23
tion of TiO2 Films Prepared by Sol–Gel Processing,’’ Chem. Mater., 18 [18]
4478–85 (2006).
24 plus
Bruker advanced X-ray solutions: DIFFRAC REFSIM Version 2.0. Bruker
AXS, Karlsruhe, Germany, 2000.
K. Maex, M. R. Baklanov, D. Shamiryan, F. Iacopi, S. H. Brongersma,
25
and Z. S. Yanovitskaya, ‘‘Low Dielectric Constant Materials for Microelectron-
ics,’’ J. Appl. Phys., 93 [11] 8793–884 (2003).
26
M. H. Bocanegra-Bernal and S. D’iaz de Torre, ‘‘Review: Phase Transitions in
Zirconium Dioxide and Related Materials for High Performance Engineering
The generalization of the presented results to other coating
materials and fiber types will be the aim of future investigations.
Ceramics,’’ J. Mater. Sci., 37, 4947–71 (2002).
A. Suresh, M. J. Mayo, W. D. Porter, and C. J. Rawn, ‘‘Crystallite and Grain-
27
Size-Dependent Phase Transformations in Yttria-Doped Zirconia,’’ J. Am. Ceram.
Soc., 86 [2] 360–2 (2003).
M. W. Pitcher, S. V. Ushakov, A. Navrotskyz, B. F. Woodfield, G. Li,
Acknowledgments
28
The authors would like to thank Dennis Troegel and Johannes Mederer from
¨
Chair for Chemical Technology of Materials Synthesis, University of Wurzburg,
Germany for synthesis and supply of the crystalline zirconia precursors.
J. Boerio-Goates, and M. B. Tissue, ‘‘Energy Crossovers in Nanocrystalline Zir-
conia,’’ J. Am. Ceram. Soc., 88 [1] 160–7 (2005).
29
M. Bockmeyer, ‘‘Structure and Densification of Thin Films Prepared from
Soluble Precursor Powders by Sol-Gel-Processing’’; Ph.D. Thesis, University
Wurzburg, 2007.
¨
3
0
¨
A. Mehner, H. Klumper-Westkamp, F. Hoffmann, and P. Mayr, ‘‘Crystalli-
References
zation and Residual Stress Formation of Sol–Gel-Derived Zirconia Films,’’ Thin
Solid Films, 308–309, 363–8 (1997).
R. Brenier, A. Gagnaire, C. Urlacher, J. Mugnier, and M. Brunel, ‘‘Stress
1
¨
J. B. Davis, J. P. A. Lofvander, A. G. Evans, E. Bischoff, and M. L. Emiliani,
‘Fiber Coating Concepts for Brittle-Matrix Composites,’’ J. Am. Ceram. Soc., 76
5] 1249–57 (1993).
R. J. Kerans, R. S. Hay, T. A. Parthasarathy, and M. K. Cinibulk, ‘‘Interface
31
‘
[
Development in Amorphous Zirconium Oxide Films Prepared by Sol–Gel
Processing,’’ Thin Solid Films, 338, 136–41 (1999).
2
32
Design for Oxidation-Resistant Ceramic Composites,’’ J. Am. Ceram. Soc., 85 [11]
2
D. Ashkin, R. A. Haber, and J. B. Wachtman, ‘‘Elastic Properties of Porous
Silica Derived from Colloidal Gels,’’ J. Am. Ceram. Soc., 73 [11] 3376–81 (1990).
599–632 (2002).
T. Helmer, H. Peterlik, and K. Kromp, ‘‘Coating of Carbon Fibers—The
Strength of the Fibers,’’ J. Am. Ceram. Soc., 78 [1] 133–6 (1995).
3
33
H. Salmang and H. Scholze, Keramik, Teil 2. Springer-Verlag, Heidelberg,
&
1983.