This article was downloaded by: [Dalhousie University]

On: 26 September 2012, At: 21:15

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T
3JH, UK

Journal of Adhesion Science

and Technology
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/tast20

The influence of surface

modifications of glass on glass
fiber/polyester interphase
properties
V. Cech, R. Prikryl, R. Balkova, J. Vanek & A.
Grycova

Version of record first published: 02 Apr 2012.

To cite this article: V. Cech, R. Prikryl, R. Balkova, J. Vanek & A. Grycova (2003):
The influence of surface modifications of glass on glass fiber/polyester interphase
properties , Journal of Adhesion Science and Technology, 17:10, 1299-1320

To link to this article: http://dx.doi.org/10.1163/156856103769172751

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-

and-conditions

This article may be used for research, teaching, and private study
purposes. Any substantial or systematic reproduction, redistribution,
reselling, loan, sub-licensing, systematic supply, or distribution in any form
to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any
representation that the contents will be complete or accurate or up to
date. The accuracy of any instructions, formulae, and drug doses should
be independently verified with primary sources. The publisher shall not
 be liable for any loss, actions, claims, proceedings, demand, or costs or
damages whatsoever or howsoever caused arising directly or indirectly in
connection with or arising out of the use of this material.
Downloaded by [Dalhousie University] at 21:15 26 September 2012
 J. Adhesion Sci. Technol., Vol. 17, No. 10, pp. 1299– 1320 (2003)
Ó VSP 2003.
Also available online - www.vsppub.com

The in uence of surface modiŽ cations of glass

on glass Ž ber/ polyester interphase properties
Downloaded by [Dalhousie University] at 21:15 26 September 2012

V. CECH ¤ , R. PRIKRYL, R. BALKOVA, J. VANEK and A. GRYCOVA

Institute of Materials Chemistry, Brno University of Technology Purkynova 118,
CZ-612 00 Brno, Czech Republic

Received in Ž nal form 19 April 2003

Abstract—Unsized glass Ž bers and planar glass substrates were subjected to low temperature plasma
or wet-chemical process to modify the Ž ber or substrate surface and thus in uence the interphase
properties of the glass/ polyester system. Plasma-polymerized thin Ž lms (interlayers) of organosilicon
monomers (hexamethyldisiloxaneand vinyltriethoxysilane)were deposited in an RF helical coupling
plasma system on the glass surface. Commercial silane coupling agent (vinyltriethoxysilane) was
coated onto an unmodiŽ ed glass surface from an aqueous solution. Bonding at the glass/ interlayer
interface was analyzed by employing a micro-scratch tester together with an optical polarizing
microscope for the planar samples. The results revealed that the adhesion bonding could be
controlled by plasma process parameters. Scanning electron and atomic force microscopies enabled
characterization of the Ž lm surface morphology. Chemical composition and chemical structure
of prepared interlayers were characterized using X-ray photoelectron and infrared spectroscopies.
Microcomposites (macrocomposites) were tested to evaluate the interfacial shear strength (short-beam
strength) of the glass Ž ber/ polyester interphase using the microbond test (short-beam shear). Our
study indicated that the most efŽ cient interphase could be prepared by plasma polymerization or wet-
chemical process using the vinyltriethoxysilanemonomer. The short-beam strength was 110% higher
than that for untreated Ž bers in both cases.

Keywords: Composite; glass Ž ber; interface/ interphase; plasma polymerization.

1. INTRODUCTION
Interfaces are created when high strength and high stiffness Ž bers are combined
with a matrix material. The growing number of Ž ber-reinforced composites uti-
lized in many engineering applications has increased the interest in an interface, or
more properly termed interphase [1], in the design and manufacture of composite
components. It is well known that properties of an interphase are governed largely

¤ To
whom correspondence should be addressed. Phone: +4205 4114 9304. Fax: +4205 4114 9361.
E-mail: cech@fch.vutbr.cz
 1300 V. Cech et al.
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 1. Schematic illustration of a composite interphase.

by chemical / morphological nature and physical/ thermodynamic compatibility be-

tween the Ž bers and the matrix. Fibers generally are surface treated and modiŽ ed in
order to protect the Ž ber surface from damage during handling and further process-
ing and, of course, in order to improve the interphase bond quality for efŽ cient stress
transfer and fracture resistance / damage tolerance of inherently brittle composites
without sacriŽ cing other important mechanical properties [2].
The extent of interfacial bonding in uences the material characteristics of the
interphase. Strong bonding is often demanded for composite materials under
loading. However, a strong interfacial bonding increases the strength of composite
material at the expense of the material toughness. A functional interlayer deposited
between the Ž ber and the matrix can improve compatibility between both materials.
A proper thin interlayer with adhesion bonding to both components results in an
improvement in the material strength and toughness simultaneously [3]. It must
be said that it is not yet completely known, as to which coating materials are
suitable for a speciŽ c composite system and also a wide industrial application of
chemical surface treatments has been limited because of increasing concern about
environmental pollution problems [4]. These are the main reasons that have led
researches to Ž nd new types of Ž ber surface coatings.
Let us look at the composite interface / interphase. There are great differences be-
tween the interface (2D surface) and the interphase (3D region). The concept of the
interphase is schematically illustrated in Fig. 1 [5]. We can distinguish two inter-
faces in the interphase region. One of them at the Ž ber surface (Ž ber /interphase) is
relatively sharp and the other at the matrix (interphase/ matrix) is a diffused one. If
the Ž ber surface is modiŽ ed by a coating (interlayer), there is the third inner inter-
face between the interlayer and the modiŽ ed matrix.
Glass Ž ber-reinforced polyester composites combine materials of different physi-
cal and chemical properties. With respect to the mechanical properties, the modulus
of glass Ž bers (GF) (73 GPa) differs from that of polyester matrix (3.7 GPa) and
 Glass on glass Ž ber/ polyester interphase properties 1301

similarly the Ž ber strength (3.5 GPa) is different from the matrix (76 MPa). Silane
coupling agents [6, 7] are commonly applied onto glass Ž bers surfaces from liquid
solutions in order to produce sizing, i.e. a functional Ž lm for improving the transfer
of stress from the matrix to the Ž ber by enhancing Ž ber wettability, adhesion, com-
patibility, etc. However, it is known that the molecules of silane coupling agents
have a tendency towards self-condensation, forming siloxane oligomers rather than
complete bonding with the glass surface [8, 9] and the density of Ž nally formed
siloxane bonds is low and the bonds are hydrolytically unstable [10]. These facts
work against the usage of commercially sized glass Ž bers as reinforcements for
polymer composites because transfer stress from the matrix to the Ž bers is reduced.
An alternative to the silane-coupling agents is a plasma polymer. Plasma
Downloaded by [Dalhousie University] at 21:15 26 September 2012

polymer Ž lms prepared during plasma polymerization are pinhole free, highly
cross-linked, insoluble and heat-resistant, and generally have excellent chemical,
mechanical, electrical and thermal properties. Further, the materials formed
by plasma polymerization are vastly different from conventional polymers and
constitute a new kind of material [11]. Plasma treatment of Ž bers may be used
to modify their surface properties and surface morphology.
The quality of a surface coating can be assessed by several experimental tech-
niques. Fiber-matrix interphase bond quality on a microscopic scale can be mea-
sured by the so-called single Ž ber microcomposite tests, including single-Ž ber
compression, Ž ber fragmentation, Ž ber pull-out, Ž ber push-out (or indentation),
and slice compression tests [12]. The Ž ber-matrix interphase is measured in
terms of interface fracture toughness or interfacial shear (bond) strength (IFSS).
These tests are considered to provide direct measurements of interphase proper-
ties vis-a-vis the test methods based on bulk composite specimens, which measure
interlaminar / intralaminar properties. Apart form the short-beam shear test, which
measures the short-beam strength, many other tests are available for translaminar or
in-plane strength measurements. These include Iosipescu shear test, [§45± ]s ten-
sile test, [10± ] off-axis tensile test, rail-shear test, cross-beam sandwich test and
thin-walled tube torsion test [13].
Unsized glass Ž bers were plasma treated and plasma coated by polymer Ž lms
to modify surface functional groups, surface roughness, wettability, thickness and
mechanical properties of the interlayer. To investigate the in uence of surface
modiŽ ed Ž bers on interphase properties of the glass Ž ber / polyester composite
system was the aim of this study.

2. VOLUME FRACTION OF THE INTERPHASE

A bond between a Ž ber and a polymer matrix may be formed by the interdiffusion of
atoms or molecules across the interface. The interphase formed thereby possesses
chemical, physical and mechanical properties that are different from those of either
bulk Ž ber or matrix acting alone. Thus, the interphase represents a transition region
the properties of which may vary continuously between the Ž ber and the matrix.
 1302 V. Cech et al.

In case of a GF/polyester composite system the interphase region created is

complicated by the large number of components and the interactions that occur
between the silane and the glass Ž ber surface, and between the silane and the
polyester matrix. Several attempts have been made to evaluate the width of the
interphase region. The results have shown similar trends; however, the interphase
width determined varied depending on the technique employed for its evaluation.
Ikuta et al. [14] have investigated the interphase region in the GF /vinylester
composite system using a microscopic FTIR spectroscopy. They scanned the region
while recording spectra, which were compared with that of the bulk vinylester
resin, and evaluated the interphase width as 80 ¹m. The interface region in
GF/polymer composites was investigated using nanoindentation and nanoscratch
Downloaded by [Dalhousie University] at 21:15 26 September 2012

techniques that seem to be more efŽ cient and sensitive for such investigations [15].
The nanoindentation test involved indentations as small as 30 nm in depth in order
to detect the material properties in the transition region between the Ž ber and the
matrix. The nanoscratch test involves moving a sample while being in contact with
a diamond tip. The coefŽ cient of friction, deŽ ned as the ratio of the tangential force
from the tip-sample contact to the normal force, also indicated a transition zone
of different widths. The nanotechniques revealed an interphase width of 2 ¹m
for GF/polyester and 6 ¹m for GF/phenolic composite systems. The nominal
diameter of glass Ž bers was 20 ¹m. The same authors extended their measurements
to GF/polymer composites degraded during water aging [16] and found that the
interphase region increased with water aging depending on the polymer matrix
(polyester resin 5 ¹m, phenolic resin 10 ¹m after 10 weeks of water aging). Kim
et al. [17] used nanoindentation and nanoscratch tests and the thermal capacity
jump measurements for investigation of the interphase region in GF /vinylester
composites with glass Ž bers of diameter 9 ¹m. The effective thickness of interphase
was estimated from the consideration of differential heat capacity jump between
the Ž ber-reinforced and unŽ lled resins in the glass transition region. The results
obtained based on the heat capacity jump measurements agreed well with those
obtained from the nanoscratch test. The measured interphase width varied between
0.8 and 1.5 ¹m depending on the type and concentration of silane coupling.
We can introduce volume fractions of Ž bers, matrix and interphase as

vf D V f =Vcomp ; vm D Vm =Vcomp ; vi D Vi =V comp ; (1)

where

Vcomp D Vf C Vm C Vi ; (2)

Vf ; V m ; Vi ; Vcomp represent volumes of Ž bers, matrix, interphase and composite,

respectively. Then, we can write

vf C vm C vi D 1: (3)
 Glass on glass Ž ber/ polyester interphase properties 1303
Downloaded by [Dalhousie University] at 21:15 26 September 2012

(a) (b)

(c)
Figure 2. Interphase region in a unidirectional Ž ber composite, (a) Ž ber diameter and interphase
width, (b) well-ordered Ž bers regularly surrounded by the interphase, and (c) randomly distributed
Ž bers.

The ratio between the interphase volume fraction and the Ž ber volume fraction can
be expressed as
vi Vi .Df C 2wi /2 ¡ Df2
D D ; (4)
vf Vf Df2
where Df is the Ž ber diameter and wi is the interphase width, see Fig. 2a.
Here, we are interested in equation (3) for the case of vm D 0, which implies that
vf C vi D 1: (5)
A solution of equation (5) using relation (4) gives
Df2
vf D : (6)
.Df C 2wi /2
 1304 V. Cech et al.

The Ž ber volume fraction expressed in (6) means a critical value at which Ž bers are
surrounded only by the interphase (there is no bulk matrix in the system) for Ž bers
of diameter Df and the interphase width wi .
Thus, we can determine the Ž ber volume fraction using equation (6) using the in-
formation on the interphase width, which results from experimental measurements.
We assume only a slight dependence of the interphase width on the Ž ber diameter.
For a glass Ž ber with a diameter of 16 ¹m, which is frequently used in industrial
fabrication of composites, the critical volume fraction of Ž bers is 0.64 if wi D 2 ¹m
(0.33 if wi D 6 ¹m). A Ž ber volume fraction of 0.64 is used for pultruded com-
posites and 0.33 is commonly used for all other types of polymer composites. In
this case, Ž bers are fully surrounded by the modiŽ ed matrix and there is no bulk
Downloaded by [Dalhousie University] at 21:15 26 September 2012

matrix in this composite system. Of course, the modiŽ ed matrix has quite different
chemical and physical properties.
We did our calculations with the assumption that the Ž ber was regularly sur-
rounded by the interphase as one can see in Fig. 2b. But a real case could look like
that in Fig. 2c and so the determined values of critical volume fraction are overesti-
mated. However, we can expect that the interphase volume fraction is high enough
to in uence the performance of composite materials signiŽ cantly. Therefore, the
interphase can really be the most important part of composite materials.

3. EXPERIMENTAL

3.1. Materials

Hexamethyldisiloxane (CH3 )3 Si O Si (CH3 )3 (HMDSO, 99.5% purity,

Aldrich) and vinyltriethoxysilane CH2 CH Si( O CH2 CH3 )3 (VTES, purity
>98%, Fluka) monomers were used for the surface modiŽ cation of glass Ž bers and
planar substrates. Argon (99.999%) gas was employed for the cleaning procedure
and surface treatment of glass Ž bers.
Planar glass substrates were special microscope slides without  aws (1:0 £ 26 £
76 mm3 , refractive index n D 1:518, Knittel Glaser, Germany). Infrared-transparent
silicon wafers (0:8 £ 10 £ 10 mm3 , Terosil Co., Czech Rep.) were used as planar
substrates as well.
Glass Ž ber bundles (E-glass, 1200 tex, mean diameter 19 ¹m) with commercial
sizing based on ° -methacryloxypropyltrimethoxysilane and those without sizing
were supplied by Saint-Gobain Vertex, Czech Rep.
The matrix material used in this study was unsaturated polyester resin VIAPAL
VUP 4649 E(M) on the base of isophthalic acid (Vianova Kunstharz, Austria) with
a density (½m ) of 1.13 g/cm3 . The resin was mixed with hardener (styrene), low-
temperature initiator Perkadox 16 (Akzo Nobel Chemicals, The Netherlands), high-
temperature initiator Norpol No. 62 (Norac Andos, Sweden), UV stabilizer (3V
Sigma S.p.A., Italy), and internal lubricant INT-PUL24 (AXEL, USA).
 Glass on glass Ž ber/ polyester interphase properties 1305

3.2. Plasma process

A new helical coupled plasma system [18] (Fig. 3) was used for surface treatment
and modiŽ cation of glass Ž bers and planar substrates as well. The system was
optimized for glass Ž ber bundles.
The bundles of glass Ž bers are sensitive to mechanical damage and so the bundle
cannot be drawn through differentially pumped vacuum chambers using vacuum
feedthroughs. Therefore the apparatus is equipped with cylindrical chambers
100 mm in diameter and 250 mm in length (1) (Fig. 4) on both ends of the
reactor. Supply (2) and take-up spools are placed in the chambers with spindles
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 3. Schematic diagram of a helical coupling plasma system for continuous surface modiŽ cation
of glass Ž ber bundle.

Figure 4. Schematic illustration of RF plasma system for continuous surface treatment and
modiŽ cation of glass Ž bers. Details are described in the text.
 1306 V. Cech et al.

perpendicular to the axis of the chamber. Both stainless steel chambers are equipped
with  anges of inner diameter 63 mm for mounting the spool spindles and the
 anges of inner diameter of 100 mm for spool manipulation. The spool diameter
is 60 mm, high enough not to damage Ž bers during winding the bundle up. The
bundle is directed into the reactor axis by a guide wheel (3). The spool spindles
are sealed by mechanical feedthroughs (5) and are turned by a pulley and a low
revolution direct current motor via a mechanical clutch. The plasma system is
evacuated by a vacuum system (6) composed of the rotary oil pump (5 m3 /h) as the
Ž rst stage and the diffusion oil pump (110 l/ s). The LN2 cold trap (7) is employed
to eliminate residual oil and water vapors. A ball valve is used for limiting the
pumping rate (8). Pressure measurements in the range 0.01– 130 Pa are realized
Downloaded by [Dalhousie University] at 21:15 26 September 2012

by a capacitron (Leybold InŽ con, USA). Pirani and Penning type gauges enable
to determine the pressure ranging from 10¡5 Pa to the atmospheric pressure. The
plasma energy is supplied by the radio-frequency power supply 2 –50 W, 13.56 MHz
with the possibility of continuous or pulsed regime, or by another power supply
10– 500 W, 13.56 MHz in the continuous regime. Both power supplies are equipped
with an analog remote control. The plasma-generating helical coil (9) is coaxial with
the plasma chamber and is connected through a matching network.

3.2.1. Plasma treatment. Two process gases Ar and O2 for cleaning and plasma
treatment of Ž bers are dosed by two  owmeters (Bronkhorst High-Tech, The
Netherlands) in the  ow range 2–100 sccm.

3.2.2. Plasma polymerization. Monomer vapors are supplied to the plasma

chamber by a gas dosing and shut-off valve EVN 116 (Pfeiffer Vacuum, Germany).
Monomer itself is a liquid and is placed in the glass vessel of a thermostated bath
enabling temperature regulation in the range from ¡10 to C60 ± C with an accuracy
of 0:1 ± C. The thermostat is equipped with a Peltier thermo-electrical module and
PID controller with fuzzy logic circuitry. The bath temperature is measured by
platinum measuring resistor (Pt100) sensor. The pressure of monomer-saturated
vapors is checked by a pressure gauge. Plasma process was monitored using in-
situ mass spectrometry. The samples prepared were stored in a desiccator to avoid
contamination before measurements.

3.3. Wet-chemical process

The VTES monomer is one of the organofunctional silane coupling agents recom-
mended for surface modiŽ cation of glass Ž bers employed as reinforcements for
unsaturated polyester resin. The molecule is a multifunctional one, which reacts at
one end with the glass surface and at the other end with the polymer matrix. The
ethoxy group ( O CH2 CH3 ) represents a hydrolyzable unit and so the silane is
hydrolyzed to the corresponding silanol in the aqueous solution to which the glass
substrate is exposed. These silanol molecules compete with water molecules to
 Glass on glass Ž ber/ polyester interphase properties 1307

Figure 5. Flow diagram for continuous surface modiŽ cation of glass Ž bers using the wet-chemical
process.

form hydrogen bonds with the hydroxyl groups present at the glass surface. When
Downloaded by [Dalhousie University] at 21:15 26 September 2012

the glass substrate is dried, the free water is driven off and condensation reactions
then occur, both at the silanol / glass interface and between neighboring silanol mole-
cules. The result is a polysiloxane layer bonded to the glass surface, presenting an
array of vinyl groups (CH2 CH ) oriented outwards.
5 ml of VTES was hydrolyzed for 60 min in 250 ml of deionized water using a
magnetic stirrer. The initial VTES concentration was about 2% by weight. The pH
value of the solution was adjusted to about 4 by adding acetic acid. The aqueous
solution was poured into a bath, which is a part of the automated apparatus for
continuous surface modiŽ cation of glass Ž ber bundles using the prepared solution
(Fig. 5). The surface modiŽ ed bundles were dried at a temperature of 120 ± C for
12 h. Thus prepared samples were stored in a desiccator to avoid contamination
before measurements.

3.4. Analysis of thin Ž lms on planar substrates

3.4.1. Chemical composition. The elemental composition in the surface region
(top 6 – 8 nm) of the deposited layers was determined by X-ray photoelectron
spectroscopy (XPS) on an ADES 400 VG ScientiŽ c photoelectron spectrometer
using Mg K® (1253.6 eV) or Al K® (1486.6 eV) photon beams at normal emission
angle. The binding energy scale was calibrated with respect to the Au 4f7=2 core
level (83.8 eV). All spectra were charge-corrected with respect to the lowest binding
energy component of the Au 4f line (evaporated dots of Au). Atomic concentrations
were determined semi-quantitatively assuming that the model dealt with a solid
which was homogeneous in composition, taking into account photoelectron cross
section asymmetry parameters and the inelastic mean free paths. The exposure of
polymer materials to X-ray irradiation results in their degradation. Specimens were
exposed to X-ray radiation for a maximum period of 30 min and, therefore, the
degradation induced amounted to less than 4% [19].

3.4.2. Chemical structure. Infrared measurements were carried out using a

Nicolet Impact 400 Fourier transform infrared (FTIR) spectrophotometer in an
H2 O-purged environment. Deuterated triglycine sulphate (DTGS) detector was
used at ambient temperature in the wavelength range from 400 to 4000 cm¡1 . The
 1308 V. Cech et al.

Happ-Genzel apodization function was used in the whole region. The spectral
resolution was 2 cm¡1 . Approximately 300 scans were recorded to achieve the
signal-to-noise ratio shown. Coherent re ections in Si substrates were eliminated
by decreasing the instrumental resolution to 8 cm¡1 . An absorption subtraction
technique was employed to remove the spectral features of Si wafers. The
refractive index difference at the Ž lm-substrate interface gives rise to broad, shallow
interference fringes in the transmission spectra.

3.4.3. Surface morphology. The thickness of polymer Ž lms was measured by

a Talystep proŽ lometer (Taylor– Hobson) using a scratch in the layer as deep as
Downloaded by [Dalhousie University] at 21:15 26 September 2012

the substrate. The Ž rst observation on the Ž lms was done by an optical polarizing
microscope (BX-P 50 Olympus) with a digital microphotocamera and details of the
surface morphology were observed by a scanning electron microscope, SEM (JEOL
JXA-840A electron probe microanalyzer) and an atomic force microscope, AFM
(Veeco). AFM together with SEM were used to observe the surface morphology
and to evaluate the roughness of prepared Ž lms.

3.4.4. Surface free energy. Sessile drop method employing an OCA 10 goniome-
ter (DataPhysics) was used to measure the contact angles and the surface free energy
was evaluated using the approach of Owens– Wendt [20] and Wu [21]. The surface
free energy, as well as the dispersive and polar components, of pp-HMDSO and
pp-VTES layers deposited on glass slides were determined from contact angle mea-
surements using distilled water, ethylene glycol, glycerol and methylene iodide.

3.4.5. Film adhesion. Scratch test is widely used to evaluate Ž lm adhesion since
it can provide fast qualitative, semi-quantitative, or quantitative information. The
test consists in drawing a stylus (Rockwell ball or diamond tip) with a known radius
of curvature over a Ž lm under increasing normal (vertical) load and the value of the
load at which adhesion failure (delamination between the Ž lm and the substrate)
is detected is known as the critical load, LC . The acoustic emission signal and
the tangential (friction) force are also monitored as a function of the progressively
increasing load. The scratch lines are also observed with optical and scanning
electron microscopes to determine the critical load.
Soft polymeric Ž lms produce negligible acoustic emission during scratching and
so the signal cannot be utilized for evaluation of the critical load. So a special low-
load micro-scratch tester has to be used for adhesion measurements of soft Ž lms.
Such a micro-scratch tester was developed in our Institute. The description of the
instrument has been reported in Ref. [22]. The apparatus is equipped with three
Rockwell diamond tips with an apex radius of 50, 100 and 200 ¹m. A special
instrument arrangement enabled the normal load to range from 20 mN to 10 N at a
scratch speed ranging from 1 to 20 mm /min.
 Glass on glass Ž ber/ polyester interphase properties 1309

3.5. Fiber/ matrix adhesion

3.5.1. Microbond test. The microbond test [23], also called microdroplet test,
involves pulling the Ž ber out of a thin disk or a small droplet of resin placed on and
around the Ž ber surface (Fig. 6). Specimens for this test were prepared by touching
the test Ž ber with a resin-coated Ž ber, and thus small droplets of resin were formed.
Six or eight droplets with a length of 120 to 320 mm were formed on each Ž ber
and cured for 30 min at 100 ± C and for 1 h at 140 ± C. The length of the droplets
and the diameter of the Ž bers were measured after curing with an optical polarizing
microscope (BX-P 50, Olympus) using a digital video camera. The same optical
system was used for monitoring microdroplet fracture process.
Fibers with cured droplets were tested in tension at a low crosshead speed of 0.5
Downloaded by [Dalhousie University] at 21:15 26 September 2012

mm /min to minimize premature breakage on a universal testing machine ZWICK

Z010/ TH2A. Assuming that the measured force is equal to the average shear force
applied uniformly along the entire interface, the interfacial shear strength (IFSS),
¿b , was determined by
¿b D Fp =.¼ Df l/; (7)
where Fp is pull-out force, Df is Ž ber diameter, and l is embedded Ž ber length.

3.5.2. Composite fabrication. Hand lay-up composites, reinforced with un-

treated Ž bers and those coated from the VTES solution, Ar-plasma treated, and
coated with plasma polymers using HMDSO and VTES monomers, were fabricated
in our laboratory. Six specimens 12 mm wide and 6 mm thick were prepared for a
particular Ž ber surface sizing or treatment, cured for 30 min at 100 ± C, then for 1 h
at 140 ± C, and postcured for 12 h at 140 ± C. The beams were polished with emery
paper of 220 ¹m particle size and stored in a desiccator before further testing.

3.5.3. Short-beam strength. Short-beam composites were tested in a three point

bending (Fig. 7) according to the conŽ guration designated in ASTM Standard D
2344/ D 2344 M-00 [24], where R1 D 6:0 mm and R2 D 3:0 mm. Span length

Figure 6. Schematic diagram of microbond test.

 1310 V. Cech et al.
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 7. Schematic representation of three-point short beam shear test.

to specimen thickness ratio was 4. Five specimens were tested for each kind of
reinforcement and storage environment. In the short-beam shear test, the maximum
value ¿max (short-beam strength) of the shear stress along the thickness direction
is related to the maximum applied load Pmax , specimen width b and thickness t,
according to the classic short-beam shear relationship
¿max D 3Pmax =.4bt/: (8)
The test speed was set at a crosshead movement of 1.0 mm /min and the force
applied on the specimens was monitored on a universal testing machine ZWICK
Z010/TH2A.

4. RESULTS AND DISCUSSION

Plasma-polymerized hexamethyldisiloxane (pp-HMDSO) and vinyltriethoxysilane
(pp-VTES) in a form of thin Ž lms were prepared on planar substrates (glass, silicon)
using RF helical coupling plasma system. Typical deposition parameters used were
as follows: a mass  ow rate of 0.5 sccm, a process gas pressure of 1 Pa (a base
pressure of 2£ 10¡3 Pa) and the effective power ranging from 0.05 to 100 W. Planar
substrates were surface treated by Ar plasma.
Glass Ž ber bundles were surface modiŽ ed by pp-HMDSO and pp-VTES interlay-
ers during a continuous movement of the bundle in the plasma chamber at an effec-
tive power of 25 W and a movement rate of 1 mm /s (monomer  ow rate 0.5 sccm,
process pressure 1 Pa). The mean Ž lm thickness was expected to range from 100
to 150 nm on the basis of the deposition rate. The Ž bers were plasma treated by Ar
plasma using a power of 50 W.

4.1. Analysis of thin Ž lms on planar substrates

4.1.1. Chemical composition. The elemental composition in the surface region
(top 6– 8 nm) of plasma polymer Ž lms was determined by X-ray photoelectron
 Glass on glass Ž ber/ polyester interphase properties 1311

Table 1.
Elemental concentrations (at.%) and elemental ratios evaluated from XPS spectra of plasma-
polymerized (pp) and polycondensed (pc) Ž lms

C O Si C/ Si O/ Si
pp-HMDSO 61 16 23 2.7 0.7
pp-VTES 60 27 13 4.6 2.1
pc-VTES 60 24 16 3.8 1.5

spectroscopy (XPS). The results are presented in Table 1. It is evident that

surface atomic concentrations of plasma-polymerized Ž lms prepared from different
Downloaded by [Dalhousie University] at 21:15 26 September 2012

organosilicon monomers differ in oxygen and silicon concentrations. The oxygen

to silicon atomic ratio is three-times higher for pp-VTES Ž lms than that for pp-
HMDSO Ž lm.
Plasma polymers deposited at high power are unstable in a chemical and physical
sense showing the aging effect, which is unsuitable for most applications. However,
the plasma polymer rich in oxygen could be used for preparation of sufŽ ciently
cross-linked material employing even a lower RF power (<50 W). Therefore,
VTES monomer was used for deposition of sufŽ ciently tough (modulus of 109 Pa)
but not brittle material, suitable as an interlayer for surface modiŽ cation of glass
reinforcements. The high concentration of oxygen in the plasma polymer is
auspicious for good adhesion at the glass / interlayer interface due to siloxane bonds.
The RF power can be diminished using the pulsed plasma: The plasma was
switched on (switched off) for the time denoted as ton .toff /. The sum of these
represents the total period, P D ton C toff . Duty cycle (DC) means a portion of the
active process expressed in percentage, DC D .ton =P / £ 100%. The effective power
used for thin Ž lm deposition can be determined as follows: Weff D Wtotal £ ton =P ,
where Wtotal D 50 W.

4.1.2. Chemical structure. Infrared spectra of monomers and plasma-polymeriz-

ed Ž lms together with the poly-condensed Ž lm are compared in Fig. 8. In spite
of different monomers the most important absorption band at 1100– 1000 cm¡1 ,
corresponding to Si O Si vibrations, is very similar for both plasma polymer
Ž lms and evidences high cross-linking of the polymer material. Other authors
have obtained similar results, e.g. see Ref. [25]. The effective power plays a very
important role in cross-linking of the plasma polymer. The main absorption band
at 1100– 1000 cm¡1 broadens with the increasing power and the same is true for
bands at 850 cm¡1 and 800 cm¡1 , which correspond to Si O and Si C vibrations,
respectively [26]. So, the material crosslinks with the increasing power forming a
continuous, random carbosiloxane network at the expense of side methyl groups
(band at 1260 cm¡1 corresponds to the deformation of Si CH3 bonds).

4.1.3. Surface morphology. Plasma polymer Ž lms of thickness ranging from

80 nm to 1.5 ¹m were deposited on planar substrates. The mean deposition rate was
 1312 V. Cech et al.
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 8. FTIR spectra of the monomers and their corresponding plasma polymers.

determined from the measured thickness and the corresponding deposition time and
the rate varied from 0.1 to 10 ¹m /h depending on the RF power.
Scanning electron micrograph in Fig. 9a shows a rough surface of the pp-VTES
Ž lm prepared at a power of 100 W on a polished silicon substrate. Grains of different
sizes up to 1 ¹m are randomly dispersed across the surface and some grains are
sticking together forming an agglomerate that looks like cauli ower. The height of
the grains reaches up to 30 nm as determined from measurements of surface proŽ les
by the Talystep. The grain structure observed at plasma polymer surfaces [27– 29]
depends on the RF power. Pp-VTES Ž lms without surface grains and with a low
roughness ranging from 0.5 nm (0.05 W) up to 3 nm (25 W) can be deposited using
pulsed plasma regime.

4.1.4. Surface free energy. Wettability and surface morphology of the interlayer
are very important for adhesion bonding at the interlayer / polymer interface [4].
Contact angle measurements were used to determine the surface free energy of the
plasma polymers (Table 2). The measured contact-angle data were analyzed to
determine the dispersive and polar components of surface energy according to the
Owens–Wendt– Rabel– Kaelble method [20, 30]. The work of adhesion determined
using the Young– Dupré equation is higher for pp-VTES than for glass and even pp-
HMDSO, see Table 2.

4.1.5. Film adhesion. The scratch test involves drawing a stylus with a known
radius of curvature over the Ž lm under increasing normal loads and the value of
 Glass on glass Ž ber/ polyester interphase properties 1313
Downloaded by [Dalhousie University] at 21:15 26 September 2012

(a)

(b)
Figure 9. Surface morphology of pp-VTES Ž lms, (a) scanning electron micrograph of the Ž lm
prepared at a power of 50 W (continuous plasma), (b) atomic force micrograph corresponding to
the Ž lm prepared at an effective power of 5 W (pulsed plasma).
 1314 V. Cech et al.

Table 2.
Surface free energy and its components (in mJ / m2 ) for plasma-polymerized Ž lms and the work of
adhesion (in mJ / m 2 ) evaluated by contact angle measurements using the polyester resin

Plasma Total Dispersive Polar Work of

polymer adhesion
pp-HMDSO 27 25 2 53
pp-VTES 40 32 8 60
E-glass 58 23 35 44
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 10. Critical load as a function of the effective power (duty cycle) for pp-VTES Ž lm on
microscope slide as determined from the scratch test.

the load at which adhesion failure is detected is known as the critical load. The
Rockwell diamond tip with an apex radius of 50 ¹m was employed to characterize
the adhesion bonding at the interlayer / glass interface. The critical load versus the
effective power for the pp-VTES Ž lm with a thickness of 1 ¹m deposited on planar
glass substrate is depicted in Fig. 10. The critical load increased with increasing
power up to a value of 5 W and then reached a plateau value of 0.4 N. The critical
load for pp-HMDSO and polycondensed (pc) VTES Ž lms was about 0.2 N.

4.2. Fiber/ matrix adhesion

4.2.1. Microbond test. The quality of bond between the Ž ber and the matrix was
investigated using the microbond test. Images of the resin droplet on untreated Ž ber
before and after test are shown in Fig. 11a– c. The droplets were either completely
 Glass on glass Ž ber/ polyester interphase properties 1315
Downloaded by [Dalhousie University] at 21:15 26 September 2012

(a)

(b)

(c)
Figure 11. Photographs of (a) cured droplet on untreated glass Ž ber before microdroplet test,
(b) completely debonded droplet on the Ž ber, and (c) droplet fractured into two parts.
 1316 V. Cech et al.
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 12. Variation of the debonding force with the embedded length of glass Ž bers.

debonded or broken into two parts with the larger part completely debonded and the
smaller part still bonded to the Ž ber. The latter effect could be attributed to tensile
stresses in the meniscus area. Debonding due to premature failure of the matrix
could trigger interfacial failure, thus resulting in a lower value of debonding force.
Failure mode is also strongly dependent on the distance between the supported knife
and the Ž ber. The distance should be as small as possible.
There is a large scatter in the values of debonding force in dependence on the
droplet length, as is seen form Fig. 12. This feature is accompanied by uneven
wettability of polyester resin along the Ž ber.
The in uence of surface modiŽ cations on the quality of bond between the Ž ber
and the matrix is shown in Fig. 13. The IFSS with pp-HMDSO Ž lm seems to be
higher than that with pp-VTES Ž lm, but the error bars overlap.

4.2.2. Short-beam strength. The storage modulus, E, of hand lay-up composites

(short beams) was measured on a Dynamic Mechanical Analyzer DMA 2980 (TA
Instruments) in a single cantilever mode. Samples were loaded using a frequency
of 100 Hz and amplitude of 5 ¹m at 40 ± C for 2 h. Two beams each of a particular
composite material were tested.
The quality of the composite interphase on a macroscopic scale was evaluated
using the three-point short-beam shear test. The short-beam strength for plasma
modiŽ ed glass Ž bers was compared with those untreated and modiŽ ed by silane
coupling agent (VTES) (Fig. 14). The best interlayer was that using the VTES
monomer (both the pp-VTES Ž lm and pc-VTES Ž lm) and the corresponding short-
 Glass on glass Ž ber/ polyester interphase properties 1317
Downloaded by [Dalhousie University] at 21:15 26 September 2012

Figure 13. Interfacial shear strength determined from microbond test for untreated, plasma treated
and plasma polymer coated glass Ž bers.

Figure 14. Short-beam strength of glass Ž ber/ polyester composite for plasma modiŽ ed, wet-chemical
modiŽ ed, and untreated Ž bers.

beam strength was 110% higher with respect to the untreated glass Ž bers. The
high strength given by the pp-VTES and pc-VTES interlayers could be a result of
synergism. Reasons for better adhesion at both interfaces for pp-VTES interlayer
 1318 V. Cech et al.

Table 3.
Fiber volume fraction, vf , and storage modulus, E, of short beams reinforced by surface modiŽ ed
glass Ž bers

ModiŽ cation vf (%) E (GPa)

Untreated 56:5 § 1:5 7:3 § 1:3
Ar-plasma 51:4 § 2:1 9:4 § 1:3
pp-HMDSO 51:3 § 4:7 10:6 § 0:5
pp-VTES 52:4 § 1:2 10:7 § 1:0
pc-VTES 52:3 § 2:6 11:7 § 1:0
Downloaded by [Dalhousie University] at 21:15 26 September 2012

than for the pp-HMDSO interlayer were evident from scratch test (critical load)
and contact angle measurements (work of adhesion). A strong chemical bonding at
the interlayer / polyester interface could be possible due to vinyl groups situated at
the interlayer surface but the vinyl groups in the pp-VTES Ž lm were not conŽ rmed
by FTIR spectroscopy (Fig. 8). The short-beam strength corresponding to the pp-
HMDSO interlayer (Ar-plasma treatment) was 56% (77%) higher in comparison
with the untreated glass Ž bers.
The Ž ber volume fraction, vf , was calculated based on weight content of Ž bers
and resin, determined by a Thermogravimetric Analyzer TGA 6 (Perkin Elmer),
and Ž ber (½f D 2:54 g /cm3 ) and matrix (½m D 1:13 g/ cm3 ) densities. Specimens
after three-point bending test were heated from 40 to 550 ± C at a heating rate of
10 ± C /min. Fiber volume fraction and storage modulus of short beams are compared
in Table 3. The Ž ber content is similar for all the beams and so enables a comparison
among the short-beam strength values shown in Fig. 14.

5. CONCLUSION

HMDSO and VTES monomer vapors were used for plasma polymerization of
thin Ž lms and the VTES monomer was also used for preparation of thin Ž lms
using the wet-chemical process. Deposited Ž lms were characterized and tested
as interlayers for glass Ž ber/ polyester composite. Thin Ž lms were deposited on
microscope slides, silicon wafers and glass Ž bers. Infrared spectroscopy validated
that organosilicon monomers resulted in highly cross-linked materials mainly due
to the siloxane bonds. Analyses of XPS spectra revealed a high concentration of
oxygen with respect to silicon in pp-VTES Ž lms. Oxygen is favorable to adhesion
bonding at the interlayer / glass interface (chemical bonding via siloxane bonds),
which was conŽ rmed by the scratch test. Adhesion of the pp-VTES Ž lm to glass
substrate increased with increasing effective power and was higher than that for
pp-HMDSO. Wettability of the pp-VTES interlayer by liquid polyester resin was
better than those of pp-HMDSO and untreated glass, and a high value of the work
of adhesion was favorable for good bonding at the interlayer / polyester interface.
 Glass on glass Ž ber/ polyester interphase properties 1319

So, the performance of the glass Ž ber /polyester composite with the Ž bers modiŽ ed
by pp-VTES was the best within plasma modiŽ cations, and the short-beam strength
was 110% higher than that for untreated Ž bers as well as for glass Ž bers modiŽ ed
by VTES aqueous solution.

Acknowledgements
The authors express their thanks to M. Sirovy and Saint-Gobain Vertex for providing
glass Ž bers. This work was supported by contracts GACR 104/ 00/ 0708 (Grant
Agency of the Czech Republic), OC 527.110, ME 597 (Ministry of Education of
the Czech Republic).
Downloaded by [Dalhousie University] at 21:15 26 September 2012

REFERENCES
1. L. T. Drzal, M. J. Rich and P. F. Lloyd, J. Adhesion 16, 1 (1983).
2. J. K. Kim and Y. M. Mai, Engineered Interfaces in Fiber Reinforced Composites. Elsevier,
Amsterdam (1998).
3. M. Labronici and H. Ishida, Composite Interfaces 2, 199 (1994).
4. R. Li, L. Ye and Y. W. Mai, Composites Part A 28A, 73 (1997).
5. V. Cech, in: Proc. 8th Int. Conf. Fibre Reinforced Composites, A. G. Gibson (Ed.), p. 246.
Woodhead, Newcastle, UK (2000).
6. K. L. Mittal (Ed.), Silanes and Other Coupling Agents. VSP, Utrecht (1992).
7. K. L. Mittal (Ed.), Silanes and Other Coupling Agents, Vol. 2. VSP, Utrecht (2000).
8. G. M. Nishioka, J. Non-Cryst. Solids 120, 102 (1990).
9. W. Wang and A. T. DiBenedetto, J Adhesion 68, 183 (1998).
10. E. P. Plueddemann, Silane Coupling Agents. Plenum Press, New York (1991).
11. S. H. Lee and D. C. Lee, Thin Solid Films 83, 325 (1998).
12. L. T. Drzal, P. J. Herrera-Franco and H. Ho, in: Comprehensive Composite Materials, A. Kelly
and C. Zweben (Eds), Chapter 5.05. Elsevier, Amsterdam (2000).
13. D. F. Adams, in: Comprehensive Composite Materials, A. Kelly and C. Zweben (Eds), Chapter
5.06. Elsevier, Amsterdam (2000).
14. N. Ikuta, A. Yanagawa, Y. Suzuki and S. Ochiai, Composite Interfaces 7, 511 (2001).
15. A. Hodzic, Z. H. Stachurski and J. K. Kim, Polymer 41, 6895 (2000).
16. A. Hodzic, J. K. Kim and Z. H. Stachurski, Polymer 42, 5701 (2001).
17. J. K. Kim, M. L. Sham and J. Wu, Composites Part A 32A, 607 (2001).
18. R. Prikryl, O. Salyk, J. Vanek and V. Cech, Czech. J. Phys. 52, D816 (2002).
19. T. R. Gengenbach and H. J. Griesser, J. Polym. Sci., Part A: Polym. Chem. 36, 985 (1998).
20. D. R. Owens and R. C. Wendt, J. Appl. Polym. Sci. 13, 1741 (1969).
21. S. Wu, J. Polym. Sci. C 34, 19 (1971).
22. R. Prikryl, O. Salyk, L. Kripal and V. Cech, Czech. J. Phys. 52, D824 (2002).
23. B. Miller, P. Muri and L. Rebenfeld, Composites Sci. Technol. 28, 17 (1987).
24. ASTM D2344, Standard Test Method for Short-Beam Strength of Polymer Matrix Composite
Materials and Their Laminates.
25. A. M. Wrobel and M. R. Wertheimer, in: Plasma Deposition, Treatment, and Etching of
Polymers, R. d’Agostino (Ed.), p. 163. Academic Press, New York (1990).
26. R. Balkova, J. Zemek, V. Cech, J. Vanek and R. Prikryl, in: 8th Int. Conf. Plasma Surface
Engineering, held in Garmisch-Partenkirchen,Germany, p. 470 (2002).
27. H. Nagai, Y. Nakata, M. Suzuki and T. Okutani, J. Mater. Sci. 33, 1897 (1998).
 1320 V. Cech et al.

28. M. Suzuki, Y. Nakata, H. Nagai, K. Goto, O. Nishimura and T. Okutani, Mater. Sci. Eng. A246,
36 (1998).
29. N. M. Mackie, D. G. Castner and E. R. Fisher, Langmuir 14, 1227 (1998).
30. S. Wu, Polymer Interface and Adhesion. Marcel Dekker, New York (1982).
Downloaded by [Dalhousie University] at 21:15 26 September 2012