Full text of DOI:10.1557/JMR.2000.0025

Journal of
MATERIALS RESEARCH
Welcome
Comments
Help
Nanoindentation and adhesion of sol-gel-derived hard
coatings on polyester
C.M. Chan, G.Z. Cao,a) H. Fong, and M. Sarikaya
Department of Materials Science and Engineering, University of Washington,
Seattle, Washington 98195
T. Robinson and L. Nelson
Korry Electronics, Co., Seattle, Washington 98109
(Received 2 August 1999; accepted 29 October 1999)
We investigated sol-gel-derived silica-based hard coatings on modified polyester
substrates. The silica network was modified by incorporating an organic component
and adding transition metal oxides. These modifications resulted in tailored thermal,
optical, and mechanical properties of the coatings. Various low-temperature
densification techniques were studied including sol-preparation procedure, enhanced
solvent evaporation, ultraviolet irradiation, and low-temperature heating (below
150 °C). Oxygen plasma etching was applied to improve the adhesion of the sol-gel
coatings on the plastic surface. Nanoindentation analysis revealed that the coatings
have a surface hardness up to 2.5 ± 0.27 GPa and an elastic modulus up to 13.6 ±
0.4 GPa, approximately an order of magnitude higher than that of the plastic surface.
I. INTRODUCTION
sition conditions can be established that favor heteroge-
neous growth of the thin film on the substrate without
bulk particle precipitation (homogeneous growth). The
major drawback of this approach is the need of function-
alized surface. Langmuir films and self-assembled
monolayers (SAMs) are often applied to introduce func-
tionality to the polymer surfaces.
The many desired features of plastics such as light
weight, formability, high impact strength, and ductility
have allowed them to be widely used in many optical
applications. However, plastics have poor abrasion resis-
tance, which means they are readily scratched, leading to
decreased optical transparency. One approach is to apply
abrasion-resistant coatings so as to improve their sur-
face mechanical properties such as abrasion resistance.
Vapor-phase techniques, such as evaporation, sputtering,
and chemical vapor deposition (CVD) (e.g., Refs. 1–3),
produce high-quality films with high density and high
mechanical properties, but they are expensive, require
sophisticated equipment, and often require substrates to
be maintained at high temperatures. In addition, uniform
coatings of complex-shaped substrates are difficult with
such line-of-sight vacuum techniques.
Solution deposition is an alternative approach for ap-
plying oxide films to temperature-sensitive substrates.
For example, the biomimetic synthesis (e.g., Refs 4–8)
has two important characteristics: (i) control of solution
conditions, including ion concentration (supersaturation
levels), pH, and temperature; and (ii) the use of function-
alized interfaces to promote mineralization at the sub-
strate surface. By controlling the surface energy of the
substrate along with the solution supersaturation, depo-
Oxide or oxide-based coatings have been deposited on
various substrate materials using the sol-gel tech-
niques.9–19 Protective-coating films derived by the sol-
gel method can be potentially employed commercially
for several reasons, including the method’s low-cost
coating equipment and precursors, room-temperature and
standard pressure processing, complex-shape deposition,
product homogeneity, and fast throughput. Moreover, be-
cause the sol-gel method is a liquid-phase process, large-
area films can be made easily. Sol-gel processing also
offers easy manipulation of chemical composition, pre-
cursor materials are recyclable, and the substrate with-
drawal rate can be controlled to obtain desired film
thickness.19 However, postdeposition heat treatments at
elevated temperatures are commonly required to convert
the porous structure to a dense film. Organically modi-
fied oxide (also known as ormocer or ormosil) coatings
have been studied extensively9–13 and have been success-
fully commercialized as coatings for optical glasses (for
example, Dow Corning ARC™ abrasion-resistant coat-
ings). Organic components are introduced through hy-
drolysis and copolymerization of organic and inorganic
precursors together during sol preparation. The unhydro-
a)Address all correspondence to this author.
e-mail: gzcao@u.washington.edu
148
J. Mater. Res., Vol. 15, No. 1, Jan 2000
© 2000 Materials Research Society
http://journals.cambridge.org
Downloaded: 16 Jul 2014
IP address: 137.207.120.173

C.M. Chan et al.: Nanoindentation and adhesion of sol-gel-derived hard coatings on polyester
lyzable organic ligands are incorporated into the oxide
sols doped with titania and zirconia were prepared with
TEOS, titanium-isopropoxide, Ti(OC3H7)4 (99%, Al-
drich), and zirconium-isopropoxide, Zr(OC3H7)4 (99%,
Aldrich), as precursors using the following procedure.
TEOS was first partially hydrolyzed with a deficient
amount of water with a molar ratio of TEOS:H2O ס 1:1
at 60 °C for 90 min. After completely depleting the wa-
ter, the titanium or zirconium precursor was added to the
partially hydrolyzed TEOS solution. After titanium or
zirconium alkoxide reacted with partially hydrolyzed
TEOS, more water was added into the solution so that
hydrolysis and condensation reactions could proceed fur-
ther at 60 °C for another 30 min. The final sol had a
nominal molar ratio of TEOS:Ti(OC3 H7 )4 :
C2H5OH:H2O:HNO3 of 0.83:0.17:3.8:5:4.8 × 10−3 (re-
ferred to as SiO2–17TiO2). The zirconia-doped silica sol
had the same nominal molar ratio as SiO2–17ZrO2. Sols
were diluted with ethanol with a volume ratio of sol:etha-
nol of 1:2, prior to dip-coating. Undiluted sols were
stored in a freezer at −20 °C or sealed in plastic bottles
for gelation at 60 °C. The resultant xerogels were later
used for x-ray diffraction (XRD) and nitrogen sorption
analysis at 77 K (Micromeritics ASAP 2000M).
network through copolymerization. The oxide in the hy-
brid coatings provides a backbone for improved me-
chanical properties, and organic components make the
coatings flexible and reduce the thermal expansion
coefficient mismatch between polymer substrates and
oxide coatings. A high coating density is necessary to
ensure the mechanical properties required for protective
coatings. However, highly dense inorganic coatings gen-
erally require a heat treatment at elevated temperatures
(>500 °C) that readily damage key properties of the
temperature-sensitive plastic substrates, such as optical
properties and bioproperties.
This paper reports our recent study of sol-gel-derived
hard coatings on modified polyester plastic substrates.
Organic components were introduced into the film as a
means to enhance film flexibility as well as to promote
the collapse of the sol-gel network during drying so as to
obtain a highly dense sol-gel coating. Transition metal
oxides were also incorporated into the silica network to
modify thermal, optical, and mechanical properties of the
sol-gel coatings. Ultraviolet (UV) irradiation and low-
temperature (below 200 °C) firing were applied for fur-
ther densification of sol-gel-derived coatings. The
coatings were characterized by means of ellipsometry
and nanoindentation. The relationships among the chemi-
cal compositions, sol-gel processing conditions, low-
temperature densification, and mechanical properties of
the coatings are discussed. Also discussed are the adhe-
sion of the sol-gel-derived coatings on modified polyes-
ter substrates and the effect of oxygen plasma treatments
on the polyester substrates.
B. Substrate preparation and deposition
Optical-grade modified polyester substrates, with a di-
mension of 1 × 4 cm, were cleaned with ethanol and sub-
jected to oxygen plasma etching for 60 s with an oxygen
pressure of 100 mtorr prior to sol-gel coating. Silicon
wafers were also used for comparison studies and treated
in a similar manner as that for polyester substrates. Sol-
gel films were deposited by dipping the treated substrates
into sols with ultrasonication for 1 min to remove pos-
sible gas bubbles at the substrate surface. They were then
withdrawn at a speed of 3.5 cm/min in the presence of a
dry air flow along the coating surfaces so as to enhance
solvent evaporation. Some of the coatings were UV ir-
radiated using a high-pressure Hg lamp (254 nm) at
2.4 mW/cm2, immediately after withdrawal from the
sols. The samples were kept approximately 3 cm from
the lamp. Low-temperature heating was done on some of
the coatings by exposing them to 150 °C for 3 h in air
under ambient pressure.
II. EXPERIMENTAL
A. Sol preparation
Five types of silica sols were studied, including
silica, organically modified silica, and transition metal-
oxide-doped silica. The silica sol was prepared by an
acid-catalyzed hydrolysis and condensation of tetra-
ethylorthosilicate, Si(OC2H5)4 (TEOS, 98%, Aldrich),
dissolved in a mixture of ethanol (C2H5OH), water
(H2O), and nitric acid (HNO3). The solution was
under vigorous stirring at 60 °C for 90 min to yield a
stable sol with a molar ratio of TEOS:C2H5OH:
H2O:HNO3 of 1:3.8:5:4.8 × 10−3. In addition to the SiO2
sol, two different organic/inorganic hybrid sols were
prepared using the same procedure described above.
The hybrid sols were prepared with the addition of
an organosilane precursor, 3-methacryloxy-propyl-
trimethoxysilane, H2CסC(CH3)CO2(CH2)3Si(OCH3)3
(MPS, 98%, Aldrich), resulting in nominal molar ratios
of TEOS:MPS:C2H5OH:H2O:HNO3 of 0.95:0.05:3.8:5:
4.8 × 10−3 and 0.67:0.33:3.8:5:4.8 × 10−3 and referred to
as SiO2–5MPS and SiO2–33MPS, respectively. Silica
C. Characterization
Nanoindentation on the sol-gel coatings were per-
formed using a Hysitron™ Nano-Mechanical System at-
tached to a Park CP™ Scanning Probe Microscope (SPM).
A Berkovich diamond tip was used as both an indenter
and imaging probe. Hardness and elastic modulus were
determined from the unloading part of the force–depth
(F–d) curve. The tip-area function is calibrated with
fused silica according to the procedure reported in the
literature.20 Hardness was calculated as the maximum
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 16 Jul 2014
149
http://journals.cambridge.org
IP address: 137.207.120.173

C.M. Chan et al.: Nanoindentation and adhesion of sol-gel-derived hard coatings on polyester
applied load over the area of contact, which was cali-
silica coatings. These results are in good agreement with
the literature that the incorporation of organic compo-
nents into the silica network makes the resultant coatings
flexible and soft.21,22 However, the addition of 5 mol%
MPS resulted in an increase in both hardness and elastic
modulus. Such an increase in mechanical properties
might be attributable to a denser microstructure brought
about by the addition of MPS.23 The incorporation of
17 mol% titania and zirconia also led to a reduction in
hardness. It was anticipated that the incorporation of ti-
tanium and zirconium ions into a silica network would
increase the connectivity due to the higher coordination
numbers of Ti4+ and Zr4+ (larger than 6 as opposed to 4
for Si4+). Thus, a higher hardness could be expected,
similar to the addition of Ta5+ into a silica network.24
The reduced hardness of both Ti- and Zr-doped silica
coatings could be attributed to a relatively more porous
structure of the coatings. It is well known that the addi-
tion of titanium or zirconium alkoxide into the silica sol
greatly reduces the gelation time, because titanium or
zirconium alkoxide has a catalytic effect on the conden-
sation reaction;25 a rapid condensation reaction would
lead to a more porous structure. Nitrogen sorption analy-
ses on xerogels confirmed this hypothesis.
Table II summarizes the gelation times, Brunauer–
Emmett–Teller analysis (BET) surface areas, pore vol-
umes, and average pore size for the five types of xerogels
studied in the present work. This table shows that the
incorporation of organic ligands into silica network re-
sulted in a decrease in condensation reaction rate, which
is reflected by a significant increase in gelation time.
Xerogels for nitrogen sorption analyses were made by
pouring the sols to petri dishes and drying at ambient
conditions. After all the solvent was removed, xerogels
were ground to fine powders and thoroughly degassed
at 150 °C to remove any residual volatile components.
Nitrogen sorption isotherms showed that the xerogels
of silica and silica incorporated with MPS exhibited no
detectable nitrogen sorption at 77 K (with BET sur-
face areas less than 1 cm2/g and pore volume less than
1 × 10−3 cm3/g). Thus, these xerogels are considered to
be dense.26,27 However, nitrogen sorption revealed that
brated as a function of contact depth up to 100 nm. The
hardness measurements on the sol-gel coatings were per-
formed with a load ranging from 10 to 300 ␮N, corre-
sponding to a contact depth from 15 to 90 nm.
Thickness and refractive indices of the films on both
the polyester and the silicon substrates were measured
with a He–Ne laser ellipsometer (␭ ס 632.8 nm). The
coatings were subjected to various thermal-cycling/
adhesion tests. Dry thermal-shock tests were conducted
by cycling between 150 °C on a hot plate for 55 min and
−20 °C for 5 min for a total of 5 cycles. Wet thermal-
shock tests were performed by (i) cycling between
100 °C for 55 min to 0 °C for 5 min in water for a total
of 5 cycles and (ii) submerging in water at 100 °C for 7 h.
The tested films were then analyzed by optical and scan-
ning electron microscopy (SEM).
III. RESULTS AND DISCUSSION
XRD was used to analyze the xerogels; it was found
that no crystalline phase was detectable in any of the
five samples. Furthermore, SEM observations revealed
that both the coatings and the xerogels had a homoge-
neous morphology with no phase segregation detectable.
From these results we can infer the following: (i) the
unhydrolyzable organic component is homogeneously
dispersed within the silica network, and (ii) the titanium
and zirconium oxides are incorporated into an amor-
phous silica network.
Table I shows the hardness and elastic modulus values
of the different sol-gel coatings using a load of 50 ␮N
except for SiO2–17ZrO2 coatings and the depth of pen-
etration relative to coating thickness. The coatings were
not subjected to any treatment; i.e., neither UV irradia-
tion, nor low-temperature heating. The incorporation of
organic components into silica coatings (the unhydrolyz-
able organic ligands in MPS precursors) resulted in a
significant change in both hardness and modulus. With a
33 mol% concentration of MPS, the resultant coating
exhibited much lower hardness and elastic modulus than
TABLE I. A composition and properties of sol-gel-derived silica-
based coatings on modified polyester substrates. Neither UV irradia-
tion nor low-temperature heating was applied.
TABLE II. Composition, gelation time, and microporous structure of
silica-based xerogels.
Gelation
time
Pore
volume
BET surface
area
Average
pore size
Elastic
modulusa
Contact
deptha
Composition
a
Composition
Hardnessa
Thickness
SiO2
36 h
88 h
>500 h
20 min
60 min
<10−3 cm3/g
<10−3 cm3/g
<1 m2/g
<1 m2/g
<1 m2/g
70 m2/g
11 m2/g
N/Ab
N/A
N/A
2.0 nm
2.5 nm
a
SiO2
1.45 ± 0.08 GPa
1.77 ± 0.08 GPa
0.20 ± 0.01 GPa
0.95 ± 0.05 GPab
1.30 ± 0.08 GPa
10.7 ± 0.3 GPa
12.5 ± 0.4 GPa
2.0 ± 0.2 GPa
13.6 ± 0.4 GPab
11.5 ± 0.2 GPa
20 ± 1 nm
20 ± 1 nm
40 ± 2 nm
43 ± 2 nmb
15 ± 1 nm
128 nm
188 nm
250 nm
190 nm
120 nm
0.5 MPS–SiO2
SiO2–5MPS
SiO2–33MPS
SiO2–17ZrO2
SiO2–17TiO2
0.33 MPS–SiO2
0.17 ZrO2–SiO2
0.17 TiO2–SiO2
<10−3 cm3/g
37 × 10−3 cm3/g
7 × 10−3 cm3/g
aNitrogen sorption produces typical type-VI isotherms, which is typical for dense
materials.
bN/A ס not available.
aBoth hardness and modulus were determined using a load of 50 ␮N.
bThese two data were measured using a load of 100 ␮N.
150
J. Mater. Res., Vol. 15, No. 1, Jan 2000
http://journals.cambridge.org
Downloaded: 16 Jul 2014
IP address: 137.207.120.173

C.M. Chan et al.: Nanoindentation and adhesion of sol-gel-derived hard coatings on polyester
both SiO2–17TiO2 and SiO2–17ZrO2 xerogels were po-
crease with increasing contact depth. The reduction in
coating hardness and elastic modulus could be due to the
density gradient of the sol-gel coatings. The sol-gel-
derived coatings may have a density gradient with den-
sity decreasing from the surface toward the interface
between the coating and the substrate.14 The density gra-
dient could be due to the more restricted shrinkage and a
slower drying rate at the interface compared with the
relative free shrinkage and rapid solvent evaporation at
the surface. However, our ellipsometry analysis did not
produce convincing data in supporting or opposing the
hypothesis of whether a density gradient exists in the
coatings.
There is, however, another possible mechanism. The
reduction of both hardness and elastic modulus of the
sol-gel coatings could be attributable to substrate effect,
because the plastic substrates are much softer (with a
hardness of approximately 0.29 ± 0.04 GPa under a load
of 200 ␮N) than the coating materials. Such an influence
of soft substrates is a well-known phenomenon in the
literature.29 Figure 2 compares the depth dependence of
both hardness and elastic modulus of sol-gel-derived
rous with pore volumes of 7 and 37 × 10−3 cm3/g, re-
spectively (data shown on Table II). It is noticed that
both the pore volume and BET surface area of SiO2–
17ZrO2 xerogel are significantly higher than that of
SiO2–17TiO2 xerogel, which is in good agreement with
the gelation time. Bulk xerogels generally exhibit a dif-
ferent microstructure than that of thin films or coatings
due to the significantly different drying process, with thin
films or coatings having a much denser structure. The
general trend obtained in the xerogels would provide a
good indication for the thin films or coatings. Signifi-
cantly porous structure of both SiO2–17TiO2 and SiO2–
17ZrO2 xerogels may indicate a more porous structure in
both SiO2–17TiO2 and SiO2–17ZrO2 coatings in com-
parison with the microstructures of silica and silica–MPS
coatings and, thus, explain the relatively low hardness
and elastic modulus of the former.
Figure 1 depicts the dependence of the hardness and
elastic modulus of sol-gel coatings on contact depth,
which is directly related to the indentation load; i.e., an
increased load results in a greater contact depth. For all
coatings, both the hardness and elastic modulus de-
FIG. 2. (a) Hardness and (b) elastic modulus of sol-gel-derived silica
coatings on modified polyester and silicon substrates. The coatings
were made from the same batch of sol, coated with the same proce-
dure, and ambient dried prior to nanoindentation tests.
FIG. 1. (a) Hardness and (b) elastic modulus of various ambient dried
sol-gel coatings deposited on modified polyester substrates as a func-
tion of the contact depth.
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 16 Jul 2014
151
http://journals.cambridge.org
IP address: 137.207.120.173

C.M. Chan et al.: Nanoindentation and adhesion of sol-gel-derived hard coatings on polyester
silica coatings on polyester and silicon substrates. All of
doped silica films, low-temperature heat treatment or
UV irradiation resulted in a much more significant
improvement.
the chemical compositions and the processing conditions
were kept the same. The silicon wafer has a significantly
higher hardness (approximately 12.2 ± 0.04 GPa) and
elastic modulus in comparison to that of sol-gel derived
silica coating. Nanoindentation measurements indeed re-
vealed that both hardness and elastic modulus of sol-gel
coatings on silicon wafer increased as contact depth in-
creased. Opposite results were obtained for the same
coating on polyester substrates. Thus, the above results
support that both hardness and elastic modulus are
indeed dependent on the mechanical properties of the
substrates.
Low-temperature heating and UV irradiation promote
surface condensation and structural relaxation, and thus
lead to denser sol-gel coatings. Table III compares the
layer thickness, refractive index, and hardness and elastic
modulus of sol-gel silica coatings with various treat-
ments. The thickness of the coatings was reduced notice-
ably with either UV irradiation or heat treatment. The
coating hardness and elastic modulus increased accord-
ingly. The reduction in thickness following UV irradia-
tion and thermal treatment suggests increased film
density or reduced film porosity. One might expect that a
density increase would result in a higher refractive index.
However, the refractive indices of the coatings measured
by ellipsometry showed a different trend. The coating
without any treatment had the highest refractive index,
whereas both UV-irradiated and heat-treated sol-gel
coatings had relatively lower refractive indices. One
might anticipate that both UV irradiation and low-
temperature heat treatment promoted evaporation of re-
sidual water or organic solvent from sol-gel coatings,
whereas a relatively large amount of residual water could
be present in the coating without any treatment. It is
noticeable, however, that both UV irradiation and low-
temperature heat treatment did not improve the hardness
of silica coatings as significantly as reported in the open
literature.23,29 A possible explanation is that the sol-gel
silica coating without any treatment was already rela-
tively dense. For more porous coatings such as zirconia-
Figure 3 compares the hardness and elastic modulus of
sol-gel SiO2–17ZrO2 coatings before and after a low-
temperature heat treatment at 150 °C as a function of
contact depth. Figure 3(a) shows that the hardness of the
coatings increased significantly, although the elastic
modulus did not have an appreciable increase. For ex-
ample, the hardness (measured with a load of 100 ␮N)
of sol-gel SiO2–17ZrO2 coatings increased from
0.96 ± 0.05 GPa to 1.80 ± 0.04 GPa with a heat treat-
ment at 150 °C, whereas the elastic modulus remained
almost the same. After the low-temperature heat treat-
ment, the same contact depth dependence of hardness
and elastic modulus was found. It is also worth noting
that there was no appreciable difference in hardness and
modulus between using UV irradiation and low-
temperature heating. Preliminary abrasion-resistance
tests showed that both SiO2 and SiO2–5MPS coatings
were not damaged by the 2.5-lb eraser rub test described
in MIL-C-675.
TABLE III. The variation of layer thickness, refractive indices, hard-
ness, and elastic modulus of silica coatings on modified polyester
substrates with different treatments.a
Layer
thickness
Refractive
index
Elastic
modulus
Treatment
Hardness
No treatment
UV irradiation 114.5 nm
Heating at
128.0 nm
1.437
1.432
1.11 ± 0.11 GPa 5.6 ± 0.1 GPa
1.23 ± 0.15 GPa 6.7 ± 0.4 GPa
150 °C
108.8 nm
1.436
1.37 ± 0.11 GPa 9.0 ± 0.2 GPa
aBoth hardness and elastic modulus were determined by a load of 30 ␮N,
which corresponds to a penetration depth (or contact depth) of approxi-
mately 15 nm.
FIG. 3. The change of (a) hardness and (b) elastic modulus of sol-
gel-derived zirconia-doped silica coatings with a low-temperature fir-
ing at 150 °C for 2 h in air.
152
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 16 Jul 2014
http://journals.cambridge.org
IP address: 137.207.120.173

C.M. Chan et al.: Nanoindentation and adhesion of sol-gel-derived hard coatings on polyester
Silica-based sols can be readily applied to coat-
modified polyester substrates by either dip- or spin-
coating; a thin ( 100 nm in thickness) uniform film is
formed. However, peel testing using Scotch™ tape re-
vealed that the sol-gel-derived coatings can readily be
peeled off completely from the modified polyester
substrates, indicating that the adhesion of sol-gel coat-
ings on modified polyester substrates is very weak. In the
current study, oxygen plasma etching was thus applied to
modify the surface chemistry of the substrates, so as to
improve the adhesion of sol-gel coatings on modified
polyester.
After oxygen plasma etching, the advancing DI–H2O
contact angle of the plastic substrates changed from >90°
to <2° and tape testing showed that the sol-gel coatings
had very good adhesion on the plastic substrates. Oxygen
plasma etching introduced various oxygen-containing
species into the plastic surface, including carboxylates
and hydroxyl groups30,31 possibly leading to the forma-
tion of Si–O–C bonds at the interface between the coat-
ing and the plastic substrate. As a result, very good film
adhesion was achieved. The sol-gel coatings that were
subjected to the dry thermal-cycling tests remained very
stable; neither crack nor delamination was observed.
However, coatings that were thermally cycled with water
or continuously submerged in boiling water developed
cracks and delamination; no appreciable difference was
observed between continuous submersion in boiling wa-
ter and thermal cycling with water. Figure 4 shows op-
tical micrographs of the silica sol-gel coatings on plastic
substrates after 5 cycles of 55 min in boiling water
(100 °C) and 5 min in ice water (0 °C). The coatings
were severely cracked and delaminated. It is interesting
to note that large delaminated fragments have been re-
deposited onto the substrate. Therefore, the adhesion be-
tween the plastic and the coating is reversible and subject
to degradation when exposed to boiling water or ice wa-
ter cycling, while the film cohesion remains quite strong.
Although adhesion mechanisms can be complex and dif-
ficult to infer without careful surface analysis, one pos-
sible failure mechanism can be explained by the
hydrolytic instability of the Si–O–C bonds between the
coating and the plastic substrate. Figure 5 schematically
shows a suggested “chemistry” evolution of the surface
of modified polyester substrates and/or the interface be-
tween substrates and coatings. Oxygen plasma etching
resulted in the formation of hydroxyl groups on the poly-
mer surface, albeit a small fraction.30 These hydroxyl
groups would react with partially hydrolyzed TEOS, re-
sulting in the formation of Si–O–C bonds. Such chemical
bonds link sol-gel-derived coatings with the polymer
substrates and result in a strong adhesion. However, the
Si–O–C bonds formed at the interface between the sol-
gel coating and plastic substrate may undergo hydrolysis
reaction with water at elevated temperatures catalyzed by
FIG. 4. (a) Formation of cracks and (b) delamination of sol-gel-
derived silica coatings on modified polyester substrates after thermal
cycling (water-boiling test) at 100 °C for 55 min and 0 °C for 5 min in
water for 5 h in total.
the interfacial stress, induced from the mismatch of the
thermal expansion coefficients of sol-gel-derived coat-
ings and modified polyester substrates, resulting in dis-
sociation of sol-gel coatings from plastic substrates. The
above discussion is highly speculative and will be a sub-
ject of future investigation.
IV. SUMMARY
Five types of sol-gel coatings have been studied. Rela-
tively dense sol-gel-derived coatings were obtained with
low-temperature densification methods including UV ir-
radiation and low-temperature heating. The incorporation
of organic component resulted in an increased density.
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 16 Jul 2014
153
http://journals.cambridge.org
IP address: 137.207.120.173

C.M. Chan et al.: Nanoindentation and adhesion of sol-gel-derived hard coatings on polyester
REFERENCES
1. J.J. Senkevich and S.B. Desu, Chem. Vap. Deposition 4, 92
(1998).
2. I.F. Hu, P.J. O’Connor, J.C. Tou, J.H. Sedon, S.E. Bales, and
D.J. Perettie, U.S. Patent No. 5 718 967 (17 February 1998).
3. F.A. Sliemers, U.S. Nandi, P.C. Behrer, and G.P. Nance, U.S.
Patent No. 4,778,721 (18 October 1988).
4. H. Shin, R.J. Collins, M.R. DeGuire, A.H. Heuer, and C.N. Sukenik,
J. Mater. Res. 10, 692 (1995).
5. H. Shin, R.J. Collins, M.R. DeGuire, A.H. Heuer, and C.N.
Sukenik, J. Mater. Res. 10, 699 (1995).
6. P. Calvert and P. Rieke, Chem. Mater. 8, 1715 (1996).
7. B.J. Tarasevich, P.C. Rieke, and J. Liu, Chem. Mater. 8, 292
(1996).
8. B.C. Bunker, P.C. Rieke, B.J. Tarasevich, S.B. Bentjen, G.E.
Fryxell, and A.A. Campbell, Science 264, 48 (1994).
9. B.E. Yoldas and C.C. Lin, U.S. Patent No. 4 753 827 (28 June
1988).
10. J. Wen and G.L. Wilkes, J. Inorganic and Organometallic Poly-
mers 5, 343 (1995).
11. J. Wen V.J. Vasudevan, and G.L. Wilkes, J. Sol-Gel Sci. Technol.
5, 115 (1995).
12. N. Gupta, T.J.M. Sinha, and I.K. Varma, Indian J. Chem. Technol.
4, 130 (1997).
13. H. Schmidt and H. Walter, J. Non-Cryst. Solids 121, 428 (1990).
14. L.F. Francis, Mater. Manufacturing Proc. 12, 963 (1997).
15. L.L. Hench and J.K. West, Chemical Processing of Advanced Ma-
terials (John Wiley & Sons, New York, 1992).
16. B.E. Yoldas and C.C. Lin, U.S. Patent No. 4 754 012 (28 June
1988).
17. C.S. Ashley and S.T. Reed, U.S. Patent No. 4 929 278 (29 May
1990).
FIG. 5. A suggested schematic diagram showing the possible chem-
istry evolution of the polymer surface and/or the interface between the
sol-gel coating and polymer substrate under various treatments.
18. V.D. McGrinniss, Prog. Organic Coatings, 27, 153 (1996).
19. C.J. Brinker and A.J. Hurd, J. Phys. III (France) 4, 1231 (1994).
20. W.C. Oliver and G.M. Pharr, J. Mater. Res. 4, 1564 (1992).
21. P. Innocenz, M.O. Abdirashid, and M. Guglielmi, J. Sol-Gel Sci.
Technol. 3, 47 (1994).
We have found that the hardness and elastic modulus of
a coating may be enhanced with a small amount of or-
ganic component or reduced with a large amount of or-
ganic component. The introduction of transition metal
oxides led to a coating with reduced hardness. Presum-
ably, the reduced hardness resulted from a porous struc-
ture brought about by rapid condensation reactions due
to the catalytic effect of transition metal oxides. UV
irradiation and low-temperature heating effectively en-
hanced the density and hardness of transition metal-
oxide-doped silica coatings. Oxygen plasma etching was
proven effective in improving the adhesion of sol-gel
coatings on plastic substrates. Nevertheless, the adhesion
between the plastic substrate and the coating degrades
severely when exposed to boiling water.
22. A. Matsuda, Y. Matsuno, M. Tatsumisago, and T. Minami, J. Am.
Ceram. Soc. 81, 2849 (1998).
23. G.Z. Cao, Y.F. Lu, L. Delattre, C.J. Brinker, and G.P. Lopez,
Adv. Mater. 8, 588 (1996).
24. P. Belleville, P. Prene, S. Petit, and R. Pieri, SID 97 Digest, edited
by J. Morreale (Society for International Display, Santa Ana, CA,
1997), p. 1065–1068.
25. C.J. Brinker and G.W. Scherer, Sol-Gel Science: The Physics and
Chemistry of Sol-Gel Processing (Academic Press, San Diego,
CA, 1990).
26. S.J. Gregg and K.S.W. Sing, Adsorption, Surface Area and Poro-
simetry, 2nd ed. (Academic Press, New York, 1982).
27. D. Ruthven, Principles of Adsorption and Desorption Process
(John Wiley & Sons, New York, 1982).
28. J. Mencik, D. Munz, E. Quandt, and E.R. Weppelmann, J. Mater.
Res. 12, 2475 (1997).
29. H. Imai, M. Yasumori, H. Hirashima, K. Awazu, and H. Onuki, J.
Appl. Phys. 79, 8304 (1996).
30. F.D. Egitto and L.J. Matienzo, IBM J. Res. Develop. 38, 423
(1994).
31. L.J. Gerenser, J. Adhesion Sci. Technol. 7, 1019 (1993).
ACKNOWLEDGEMENTS
The authors are grateful for the partial financial sup-
port from the Washington Technology Center and Korry
Electronics Co. Professor Oscar Vilches is acknowledged
for porosity measurement, Dr. Sho Fuji for the oxygen
plasma etching, and Dave Rice and Scott Bulkley for
general technical support.
154
J. Mater. Res., Vol. 15, No. 1, Jan 2000
Downloaded: 16 Jul 2014
http://journals.cambridge.org
IP address: 137.207.120.173