Materials and Design 32 (2011) 22602268

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/matdes

Thermo-mechanical correlations to erosion performance of short carbon bre

reinforced vinyl ester resin composites
Sandeep Kumar a, Bhabani K. Satapathy a,, Amar Patnaik b
a
Centre for Polymer Science and Engineering, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110 016, India
b
Department of Mechanical Engineering, National Institute of Technology, Hamirpur, Himachal Pradesh 177 005, India

a r t i c l e i n f o a b s t r a c t

Article history: Thermo-mechanical properties and erosion performance of short carbon bre reinforced vinyl ester resin
Received 10 July 2010 based isotropic polymer composites with four different bre weight fractions have been investigated. The
Accepted 10 November 2010 storage, loss and damping characteristics were analysed to assess the energy absorption/viscous recover-
Available online 3 December 2010
able energy dissipation and reinforcement efciency of the composites as a function of bre content in
the temperature range of 0140 C. The composite with 30 wt.% of short carbon bres has been observed
Keywords: to exhibit superior thermo-mechanical response with highest energy dissipation/damping ability accom-
Polymer matrix
panied with a constant storage modulus without any substantial decay till 60 C. The erosion rates (Er) of
Mechanical
Wear
these composites are evaluated at different impingement angles (3090), bre loadings (2050 wt.%),
impact velocities (4376 m/s), stand-off distances (5585 mm) and erodent sizes (250600 lm) follow-
ing the erosion test schedule in an air jet type test rig. An optimal parameter combination is determined
and subsequently validated for erosion rate minimization following Taguchi method and by conducting
conrmation experiments. A correlation between the loss-modulus inverse and the erosion rate has been
observed which conceptually establishes a possible mechanistic equivalence between erosion and
dynamic mechanical loading modes. The morphologies of eroded surface are examined by the scanning
electron microscopy to investigate the nature of wear-craters, material damage mode and other qualita-
tive attributes responsible for promoting erosion.
2010 Elsevier Ltd. All rights reserved.

1. Introduction esters and phenolic matrices have been reported to be successful

Short-bre-reinforced polymers (S-FRP) are a unique class of
composite materials since they not only provide superior mechan-
ical properties, but they can also be easily processed via premixing
of the short bre and the liquid resin prior to being injection or
compression molded giving rise to isotropic composites with
spatially distributed short bres in many complicated shapes for
a varied class of structural, industrial, automotive and marine
component applications unlike their continuous bre analogues.
Additionally, such composites with spatially distributed short
bres as reinforcing elements in the polymer matrix potentially
offers a balanced set of properties, in the form of improved
through-the-thickness stiffness/strength ratio combined with
enhanced fracture resistance [1]. For example, glass bre, carbon
bre and aramid bre (unidirectional continuous bres or woven
fabrics) reinforced thermosetting composites (reaction injection
molded/compression molded) based on epoxy, polyesters, vinyl
Corresponding author. Tel.: +91 11 26596043; fax: +91 11 26591421.
E-mail addresses: bhabaniks@gmail.com, bhabani@polymers.iitd.ernet.in (B.K.
Satapathy).
in enhancing not only the mechanical properties such as elastic
modulus, fracture and fatigue toughness but also the tribological
properties in many cases. However, the potential of short bre
reinforced polymer composites has not been explored despite their
wide success in thermoplastic matrix based composites, such as in
automobile, aerospace, marine and energy applications [2]. Due to
the seriously complicated and challenging operational require-
ments the application of many such composites remain largely
restricted to light-duty components with exposure to low-stress
situations. Typically the performance requirements of composites
in harsh/extreme and dusty/slurry environments are very complex.
In this regard solid particle erosion resistant polymeric composites
may be of high relevance from several industrial and marine appli-
cations point of view. Literature widely reports on solid particle
erosion behaviour of both thermoplastic and thermosetting resin
based composites barring the rarely reported class of short bre
reinforced vinyl ester resin composites, despite the fact that vinyl
ester resin is a promising binder with many inherently functional
properties associated with it [3]. Erosion performance has been
performed under various experimental conditions (erosive particle
speed, erodent size, impact velocity, etc.) on various compositions
of composite where it has been reported that erosive wear of

0261-3069/$ - see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.matdes.2010.11.019
 S. Kumar et al. / Materials and Design 32 (2011) 22602268
polymer composites is usually higher than that of the un-rein-
forced polymer matrix and poorer than metallic materials [25].
Solid particle erosion manifests itself in many facets such as
thinning of components, surface roughening, surface degradation,
macroscopic scooping appearance and reduction in functional life
of structures. Hence, solid particle erosion has been considered
as a serious problem being responsible for many failures in engi-
neering applications. The inuence of various experimental vari-
ables (impingement angle, impact velocity, erodent type, size,
shape and hardness) and target material variables/properties
(strength, ductility, crystallinity, cross link density, reinforcement
content and arrangement) on solid particle erosion of polymers
and their composites was comprehensively reviewed recently by
Barkoula and Karger-Kocsis and previously by Harsha et al. [6,7].
However, the erosive wear behaviour of short bre reinforced
polymer composites is an area which is ill addressed till date de-
spite the success of their corresponding long bre composites.
Hence, the present paper comprehensively makes an attempt to
study the erosive performance of short carbon bre reinforced vi-
nyl ester resin composites since carbon bre offers the dual advan-
tage of enhanced toughness and thermal resistance compared to
glass and aramid bres, on one hand, and on the other, vinyl ester
resin offering the best performance combination of epoxy and
unsaturated polyester resins in terms of stiffness and chemical
resistance.
2. Experimental
2.1. Preparation of composites
Short carbon bres (density: 1.81 g/cm3) (Zoltek) of 6 mm
length are used to prepare the composites. The curing of vinyl ester
resin is done by incorporation of: 2% methyl ethyl ketone peroxide
(MEKP) catalyst, 0.3% cobalt naphthenate (accelerator) is also
added together with 0.05% of 2,4-pentanedione (2,4-P) used as a
retarder for the extension of the reaction gel time. The resin is sup-
plied by Ciba-Geigy having density 1.05 g/cm3 and 40 Barcol Hard-
ness is used as the matrix material. The matrix tensile modulus and
strength are about 2.63.4 GPa and 65 MPa, respectively; exural
modulus and strength are about 3.6 GPa and 125 MPa respectively;
and heat distortion temperature is 105 C. Composites having dif-
ferent bre content were prepared by varying the weight fraction
of bres from 20 wt.% to 50 wt.%. It is then mixed with vinyl ester
resin and the mixture is poured into various molds conforming to
the requirements of various testing and characterization standards.
The entrapped air bubbles (if any) are removed carefully with a
roller and the mold is closed for curing at a temperature of 30 C
for 24 h at a constant pressure of 10 kg/cm2. The details of compos-
ite designation and composition are given in Table 1.
2.2. Dynamic mechanical properties
The thermo-mechanical properties of the composites were
measured using Dynamic Mechanical Analysis (DMA). Dynamic
Table 1
Details of composite designation and composition.
Composite designation Composition (wt.% fraction)
Vinyl ester resin Carbon bre
VCF0 100 00
VCF20 80 20
VCF30 70 30
VCF40 60 40
VCF50 50 50
2261
Mechanical Analysis (DMA) was conducted in a nitrogen atmo-
sphere at a xed frequency of 1 Hz, heating rate of 5 C/min, a tem-
perature range of 0150 C and a strain of 1% on rectangular
samples with approximate dimensions of 38  12.5  3.5 mm
using a Q800 DMA instrument in bending mode [8].
2.3. Test apparatus for erosion experiment
The short bre reinforced composites are molded into dimen-
sion of 150 mm  150 mm  3 mm from which test samples of
approximately 30 mm  30 mm are cut using a diamond cutter
for erosive performance evaluation on an erosive wear test rig.
The erosion experiments are carried out as per ASTM G76 on the
erosive wear test rig shown schematically in Fig. 1, which is com-
posed of an air compressor, an air drying unit, a conveyor belt-type
particle feeder and an air particle mixing and accelerating cham-
ber. The dried and compressed air is then mixed with the silica
sand which is fed constantly by a conveyor belt feeder into the
mixing chamber and then accelerated by passing the mixture
through a convergent brass nozzle of 3 mm internal diameter.
The set up is capable of creating reproducible erosive situations
for assessing erosion wear resistance of the composite samples.
The erodent particles impact the specimen which can be held at
different angles with respect to the direction of erodent ow. The
velocity of the eroding particles is determined using standard dou-
ble disc method [9]. The apparatus is equipped with a heater which
can regulate and maintain the erodent temperature at any pre-
determined xed value during an erosion trial. In the present
study, dry silica sand (assumed to be of pyramidal shaped) of dif-
ferent particle sizes (250 lm, 355 lm, 420 lm and 600 lm) are
used as erodent. Before the erosive performance tests all speci-
mens were cleaned with acetone, balanced at electronic balance
with the accuracy of 0.1 mg. Great care is given to ensure clean sur-
face before and after the tests. Sand and dust particles are cleaned
after erosion test with air blasting and then balanced carefully.
2.3.1. Erosion performance evaluation
The erosive performance were determined following the Tagu-
chis technique as a multi-parametric optimization model since
the erosive performance may be a functional variable of ve inde-
pendent parameters and their simultaneous interaction. Using this
technique: (a) rstly the control factors have been identied and
(b) subsequently steady-state erosion characteristics of the com-
posites were determined at three selected level of optimally con-
trolled operating variables.
2.3.1.1. Parametric optimization by design of experiment (DOE). The
evaluation of erosion performance is carried out by adopting a
DOE approach following the Taguchi technique to minimize overall
testing time and the experimental costs for parametric optimiza-
tion followed by the steady state erosive wear rate determination
at the optimized parameter levels. The Taguchi method, straight-
forward and consistent, involves laying out the experimental con-
ditions using specially constructed tables known as orthogonal
arrays. The L16 (45) inner orthogonal array permits the analysis
of ve four-level factors. The array consists of 16 rows, where each
row represents a trial condition with a particular combination of
factor level settings. The static and the dynamic variables chosen
for the test are given in Table 2 and four levels of chosen control
factors are given in Table 3. The controllable factors are assigned
to a particular column of the orthogonal array. Accordingly, the
present study assigns the impact velocity (factor A), bre loading
(factor B), impingement angle (factor C), stand-off distance (factor
D) and erodent size (factor E) to Columns 1, 2, 3, 4 and 5 of the
adopted L16 (45) inner orthogonal array, respectively (Table 3)
and each trial will be replicated three times with different level
2262 S. Kumar et al. / Materials and Design 32 (2011) 22602268

Fig. 1. Schematic diagram of an erosion test rig.

Table 2 mean-square deviation (MSD) for the lower-the-better characteris-

Parameters of the setting. tic can be expressed as [10]:
Control factors Symbols Fixed parameters
1 Xm
Velocity of impact Factor A Erodent Silica sand MSD T2 2
Fibre loading Factor B Erodent feed rate (g/min) 10.0 1.0
m i1 i
Impingement angle Factor C Nozzle diameter (mm) 3
Stand-off distance Factor D Length of nozzle (mm) 80 where m is the number of tests and Ti is the value of experimental
Erodent size Factor E result of the ith test.
Furthermore, a statistical analysis of variance (ANOVA) is
performed to identify the process parameters that are statistically
Table 3 signicant. With the S/N and ANOVA analyses, the optimal combi-
Levels for various control factors. nation of the process parameters can be predicted to a useful level
Control factor Levels of accuracy. Finally, a conrmation experiment is conducted to
verify the optimal process parameters obtained from the parame-
I II III IV Units
ter design.
A: Velocity of impact 43 54 65 76 m/s
B: Fibre loading 20 30 40 50 %
C: Impingement angle 30 45 60 90 C 2.3.2. Steady-state erosion rate
D: Stand-off distance 55 65 75 85 degree
The steady-state erosion experiments will be carried out at the
E: Erodent size 250 355 420 600 lm
level of optimal control factors which would be estimated from
Taguchi optimization technique. The steady-state erosion test
determines the material performance as a function of impinge-
settings conforming to the orthogonal arrays. The operating condi- ment angle, particle size and impact velocity. The weight loss is re-
tions under which erosion tests have been carried out are given in corded for subsequent calculation of erosion rate. The ratio of the
Table 3. In practice, these factors can be assigned arbitrarily to any weight loss to the weight of the eroding particles causing the loss
of the arrays columns, provided that all combinations are included. is then computed as a dimensionless incremental erosion rate. The
After assigning appropriate level settings, the S/N analysis (S/N: process is repeated till the erosion rate attains a constant value
signal-to-noise ratio) is needed to evaluate experiment results. In called steady-state erosion rate.
S/N analysis, the greater the S/N, the better the experimental
results:
2.4. Scanning electron microscopy (SEM)
g 10 log MSD 1
The eroded composite samples was analyzed using ZEISS EVO
where MSD is the mean-square deviation for the output character- Series Scanning Electron Microscope Model EVO 50 after coating
istic (erosive wear rate). with gold by BIO-RAD POLARAN sputter coater to avoid the electro-
As mentioned in Taguchi method, there are three categories of static charging and poor image resolution. To make the samples
quality characteristics, i.e. lower-the-better, higher-the-better, electrically conducting, during mounting the use of colloidal silver
and nominal-the-better. To obtain optimal performance, lower- paste coating to one side of samples was applied. The worn sam-
the-better characteristic for erosion rate must be taken. The ples are mounted on stubs to take photomicrographs.
 S. Kumar et al. / Materials and Design 32 (2011) 22602268 2263

3. Results and discussion

3.1. Thermo-mechanical properties of composites

Dynamic Mechanical Analysis (DMA) of the composites have

been carried out to investigate the variation of storage modulus
(E0 ), loss modulus (E00 ) and damping factor (tan d) as a function of
temperature to characterize the thermo-mechanical response in
the entire range of bre loadings as shown in Figs. 24. The storage
modulus (E0 ) representing the stiffness of a visco-elastic material
have been observed to remain constant without any substantial
decay till 30 C irrespective of the compositions. However, on fur-
ther increasing the temperature the composites undergo a sharp
decay in the magnitude of E0 in the temperature range 3090 C
irrespective of the carbon bre content. At lower temperature re-
gime the trend of storage modulus remained in the order E0VCF30 >
E0VCF40 > E0VCF50 > E0VCF20 which is attributed to reinforcement effects
imparted by the bres allowing enhanced stress transfer ability
across the interface [11,12]. Such an observation is well in agree- Fig. 3. Variation of the loss modulus (E00 ) as a function of temperature.
ment to the composite theory predictions referring bre weight
fraction, bre orientation, bre lengths and brematrix adhesion
as the determinants of storage modulus. However, in the temper-
ature range of 3050 C the composite with 30 wt.% carbon bre
exhibited maximum E0 whereas the other investigated composites
showed temperature dependent uctuations in the magnitude of E0
as indicated by the cross-over of E0 versus temperature plots across
each other. Interestingly, above 5060 C, the composites suf-
fered signicant drop in the magnitude of storage moduli indicat-
ing storage-energy decay followed by the absence of any further
changes above 100 C irrespective of the bre content. The loss
modulus (E00 ) represents the energy dissipation ability of the mate-
rial that has theoretical correspondence to the toughness of the
composites. The variation of E00 as a function of temperature is
shown in Fig. 3 where the maximum E00 followed the trend
E00VCF30 > E00VCF40 > E00VCF50 > E00VCF20 . Apparently E00VCF30 indicates a higher
viscous energy dissipation ability than the other composites. Fur-
ther a broadened peak-width for the composites with 30 wt.%
(VCF30) and 40 wt.% (VCF40) carbon bre theoretically indicates
a higher entanglement density. On the other hand a narrow peak
for composites with 20 (VCF20) and 50 wt.% (VCF50) of carbon - Fig. 4. Variation of the damping parameter (tan d) as a function of temperature.

bre indicates composition related structural in-homogeneities.

Such structural in-homogeneities are dominant in situations such
as: (a) inadequate wetting of the bres when the bre weight frac-
tion is very high, (b) inadequate resin inltration into the bre-
bunch and (c) inadequate mixing/distribution/dispersion of the
Fig. 2. Variation of the storage modulus (E0 ) as a function of temperature.
hardener/reaction accelerator/curing agent in the highly viscous
resin. However, for intermediate bre weight fraction the curing
related irregularities either get statistically averaged or are over-
come via reinforcement effects and thus consequentially leading
to enhanced loss-modulus performance e.g. E00VCF30 . The damping
factor (tan d) indicates the recoverable energy in terms of mechan-
ical damping or internal friction in a visco-elastic system. The var-
iation of the tan d of the composites as a function of temperature is
shown in Fig. 4. A maximum in the tan d has been observed for the
composite with a carbon bre content of 30 wt.% (VCF30) indicat-
ing enhanced damping performance which is closely followed by
the composite (VCF40) with 40 wt.% of carbon bres. Further it
was observed that the tan d peaks were observed over a narrow
temperature-width barring the composite (VCF40) with 40 wt.%
of short carbon bre. Such an observation may be due to homoge-
neous stress dissipation mechanism in all the composites barring
the composite (VCF40) with 40 wt.% of carbon bre. The trend in
damping properties of composites at/above 60 C remained as
(tan d)VCF30 > (tan d)VCF40 > (tan d)VCF50 > (tan d)VCF20. The maximum
in the damping parameter (tan d) of the composite with 30 wt.% of
carbon bre (VCF30) may be attributed to possible deformation of
the resin molecules in-between the cross-links. However, there
was no change after 120 C in damping properties irrespective of
the bre weight fraction in the composites.
2264 S. Kumar et al. / Materials and Design 32 (2011) 22602268

3.2. Erosive performance of the composites By combining like-terms, the equation reduces to

3.2.1. Taguchi analysis and response optimization

From Table 4, the overall mean for the S/N ratio of erosion rate is
found to be 62.82 dB. The analysis is made using the popular soft-
ware specically used for design of experiment applications known
as MINITAB 14. Before any attempt is made to use this simple mod-
el as a predictor, the measures of performances must be consid-
ered. The response table (S/N ratio) for erosion rate is mentioned
in Table 4. In the main effect of control factors on wear rate plot
if the line for a particular factor is near horizontal, then the factor
has no signicant effect. On the other hand, a factor for which the
line has the highest inclination will have the most signicant
effect. It is very much clear from Fig. 5 that factors A1, B2, C2, D3
and E4 have been observed to be signicantly inuential with
regard to the effects derived on the basis of smaller-the-better
norm.
g 1 A2 B3 C 4 D2 E3  4T 4
A new combination of factor levels A2, B3, C4, D2 and E3 is used to
predict wear rate through prediction equation and it is found to
be g 1 57:99 dB:
For each performance measure, an experiment is conducted for
a different factors combination and compared with the result ob-
tained from the predictive equation as shown in Table 6.
The resulting model seems to be capable of predicting wear rate
to a reasonable level of accuracy. An error of 3.56% for the S/N ratio
of wear rate is observed. However, the error can be further reduced
if the numbers of measurements are increased. This validates the
development of the mathematical model for predicting the mea-
sures of performance based on knowledge of the input parameters.
3.3. Steady-state erosion

3.2.2. ANOVA and the effects of factors
In order to nd out the statistical signicance of various factors
like impact velocity (A), bre loading (B), impingement angle (C),
stand-off distance (D) and erodent size (E) on erosion rate, analysis
of variance (ANOVA) is performed on experimental data. Table 5
shows the results of the ANOVA with the erosion rate. This analysis
is undertaken for a level of condence of signicance of 95%. The
last column of the table indicates that the main effects are highly
signicant (all have very small p-values) [13].
From Table 5, one can observe that impact velocity (p = 0.004),
stand-off distance (p = 0.013), erodent size (p = 0.374) and bre
loading (p = 0.400) have great inuence on wear rate but impinge-
ment angle (p = 0.889) has less signicant contribution on erosion
rate.
3.2.3. Conrmation experiment
The optimal combination of control factors has been deter-
mined in the previous analysis. However, the nal step in any de-
sign of experiment approach is to predict and verify improvements
in observed values through the use of the optimal combination le-
vel of control factors. The conrmation experiment is performed by
taking an arbitrary set of factor combination A2B3C4D2E3 on erosion
rate as evident from Table 6. The estimated S/N ratio for wear rates
can be calculated with the help of following predictive equation:
g 1 T A2  T B3  T C 4  T D2  T E3  T 3
g 1 is the predicted average, T is the overall experimental average
A2 ; B3 ; C 4 ; D2 and E3 is the mean response for factors at designated
levels.
Among numerous factors inuencing the erosion performance
of materials, the most important are impingement angle, impact
velocity and particle size as shown by Taguchi methods. In order
to quantify the extent of the damage, the erosion rate of weight
loss is usually expressed as the weight of material removed by unit
weight of impacting particles.
3.3.1. Inuence of impingement angle
It is well known that impingement angle is one of the most
important parameters in erosion behaviour. Fig. 6 shows the inu-
ence of impingement angle on the steady-state erosion rate of
short carbon bre reinforced vinyl ester resin composites at a con-
stant impact velocity of (54 m/s), erodent size of 355 lm and
stand-off distance of 65 mm. It can be seen that steady-state ero-
sion rate is maximum at 60 impingement angle for all the com-
posites irrespective of the bre loading. In the literature,
materials are broadly classied as ductile or brittle, based on the
dependence of their erosion rate with impingement angle. The
ductile behaviour is characterized by maximum erosion rate at
low impingement angle, typically 15 < a < 30. On the other hand,
if the maximum erosion rate occurs at normal impact (a = 90), the
behaviour of material is brittle [14,15]. Reinforced composites have
been found to exhibit semi-ductile behaviour with the maximum
erosion rate at intermediate impingement angles, typically
45 < a < 60 [3,16,17]. However, the above classication is not
absolute as the erosion behaviour of materials strongly depends
upon the experimental conditions and composition of target mate-
rials. Although there is a dispute about this classication-failure,
Barkoula and Karger-Kocsis [5] argued that the erosive wear

Table 4
Experimental design using L16 orthogonal array.

Expt. no. A (m/s) B (%) C (degree) D (mm) E (lm) Er (104 mg/kg) S/N ratio (dB)
1 43 20 30 55 250 6.0833 64.3172
2 43 30 45 65 355 2.9167 70.7023
3 43 40 60 75 420 2.5000 72.0412
4 43 50 90 85 600 1.2500 78.0618
5 54 20 45 75 600 2.5625 71.8267
6 54 30 30 85 420 3.8750 68.2346
7 54 40 90 55 355 22.625 52.9082
8 54 50 60 65 250 20.250 53.8715
9 65 20 60 85 355 4.5263 66.8851
10 65 30 90 75 250 1.6842 75.4721
11 65 40 30 65 600 5.1053 65.8396
12 65 50 45 55 420 13.2632 57.5471
13 76 20 90 65 420 36.7143 48.7033
14 76 30 60 55 600 32.4762 49.7687
15 76 40 45 85 250 11.5238 58.7681
16 76 50 30 75 355 30.7143 50.2532
 S. Kumar et al. / Materials and Design 32 (2011) 22602268 2265

Fig. 5. Effect of control factors on erosion rate.

Table 5
ANOVA table for erosion rate.

Source DF Seq SS Adj SS Adj MS F P

A 3 572.37 572.37 572.37 13.83 0.004
B 3 32.03 32.03 32.03 0.77 0.400
C 3 0.85 0.85 0.85 0.02 0.889
D 3 372.82 372.82 372.82 9.01 0.013
E 3 35.89 35.89 35.89 0.87 0.374
Error 0 413.99 413.99 41.40
Total 15 1427.95

Table 6
Results of the conrmation experiments for erosion rate.

Optimal control parameters

Prediction
Level A2B3C4D2E3
S/N ratio for erosion rate (dB) 57.99
behaviour strongly depends on the experimental conditions and
the composition of the target material. The neat vinyl ester resin,
a relatively ductile matrix however, showed a peak erosion rate
at 60 impingement angle. The ductility of the matrix is not re-
ected in erosion behaviour. Also in the present study short carbon
bres were used as reinforcement for the vinyl ester resin matrix,
which are typically brittle materials. Hence, erosion peak shifts to
larger value of impingement angle (i.e. 60). The erosion resistance
of matrix signicantly decreases with increase in short carbon bre
content. Composite having 30 wt.% short carbon bre content
(VCF30) in the matrix showed better erosion resistance as com-
pared to other investigated composites. Hence, erosion rate de-
pends on the matrix material, bre content by weight fraction
and type of bre used in the composite. Composite having
50 wt.% bre loading (VCF50) shows poor erosion resistance as
compared to other composites but comparatively better than neat
resin (VCF0). To characterize the morphology of as-received and
eroded surfaces and the mode of material removal, the eroded
samples are observed under scanning electron microscope.
Experimental
A2B3C4D2E3
55.93
Fig. 6. Effect of impingement angle on the erosion wear rate of the composites (at
constant impact velocity: 54 m/s, stand-off distance: 65 mm and erodent size:
355 lm).
3.3.2. Inuence of impact velocity
As seen in Fig. 7, wear rates of the samples were remarkably
higher at higher impact velocity conforming to the theoretical
expectations. Particles have a higher kinetic energy at higher veloc-
ity, which results in greater impingement-effect and results in in-
creased wear. In order to study the effect of impact velocity on
erosion rate, erosion tests are performed by varying the impact
velocity from 43 to 76 m/s and bre loading from 20 to 50 wt.%
for constant impingement angle (60), stand-off distance (65 mm)
and constant erodent size (355 lm). It is quite clear from Fig. 7 that
steady-state erosion rate of neat vinyl ester resin (VCF0) and its
composites increases with increase in impact velocity. The velocity
of the erosive particles has a very strong effect on erosion rate.
3.3.3. Inuence of erodent size
Similarly the erosion rate is also observed to be increased by
varying erodent size from 250 lm to 600 lm and bre loading
2266 S. Kumar et al. / Materials and Design 32 (2011) 22602268
Fig. 7. Effect of impact velocity on the erosion wear rate of the composites (at
constant impingement angle: 60, stand-off distance: 65 mm and erodent size:
355 lm).
Fig. 8. Effect of erodent size on the erosion wear rate of the composites (at constant
impact velocity: 54 m/s, impingement angle: 60 and stand-off distance: 65 mm).
Fig. 9. Correlation between erosion rate (Er) and loss modulus (E00 ) of carbon bre reinforced vinyl ester composites.
(2050 wt.%) at constant impact velocity (54 m/s), stand-off dis-
tance (65 mm) and impingement angle (60) as shown in Fig. 8.
The increase in the erosion rate with the increase in the erodent
size may be attributed to two factors. Firstly, with the increase in
the size of the erodent the momentum with which the particle hits
the composite target is increased causing increased indentation
damage to the target and secondly, the repeated hitting of such
particles with larger erodent sizes may progressively damage the
subsurface lying underneath the already indented surface of the
composite and thereby promoting the fatigue induced erosive
wear of the composites. The interplay of momentum of the ero-
dent, indentation efciency and fatigue assisted erosive wear
modes have also been reported by several other researchers [18].
3.4. Thermo-mechanical correlations to erosion performance
The steady-state erosion performance assessment as a function
of impingement angle, impact velocity and erodent sizes has re-
vealed that the erosive wear rate minimization of short carbon -
bre reinforced vinyl ester resin composites may be accomplished
by resorting to an impingement angle of 60, impact velocity of
65 m/s, erodent size of 420 lm and at a short carbon bre content
of 30 wt.% (VCF30) in the composite. A correlation plot of erosive
wear rate/loss-modulus inverse (E001 ) as a function of bre concen-
tration is shown in Fig. 9 which revealed a striking resemblance in
the trend of the erosion performance and inverse of loss-modulus
dependence of the bre content irrespective of the optimum
impingement angle, impact velocity and erodent size. This concep-
tually establishes a mechanistic correlation between the response
of the composites to erosion conditions and the response of the
composites to thermo-mechanical loading situations despite the
quantitative differences in the evaluation modes. However, the
equivalence of the mechanistic pathway needs to be further inves-
tigated since the dynamic loading intensity and the erodent bom-
bardment frequency are fundamentally different modes of physical
loading.
3.5. Worn surface morphology
From SEM observations of eroded surfaces, it appears that com-
posites under consideration exhibit several stages of the erosion
 S. Kumar et al. / Materials and Design 32 (2011) 22602268 2267

and material removal process. First, there is local removal of resin the bres to the erosive environment, as seen in Fig. 10a and b.
material from the impacted surface; which results in exposure of The erosion rate data shown in Table 4 (Experiment no. 1) supports

0
30 1.5kX 300 3.0kX

(a) (b)

450 150X 450 150X

(c) (d)

900 1.5kX 900 1.5kX

(e) (f)

900 150X 900 150X

(g) (h)
Fig. 10. SEM micrograph of carbon bre reinforced vinyl ester resin composites.
2268 S. Kumar et al. / Materials and Design 32 (2011) 22602268
this observation. The micrograph further reveals that the impact of
solid sand particles onto the matrix renders the matrix prone to
breakage via the formation of cracks transverse to bre-length.
These cracks across the bre are caused by bending due to the im-
pact of these particles on the unsupported bres. Bending is possi-
ble because the matrix resin surrounding and supporting the bres
have been removed. But on further increase in impact velocity
(54 m/s) under similar bre loading (20 wt.%), it can be seen from
Fig. 10c and d that, when impacting at low angles, the hard erodent
particles can penetrate the surfaces of the samples and cause mate-
rial removal by micro-cutting and/or micro-ploughing. The matrix
is randomly grooved and cratered with local material removal un-
der 54 m/s impact velocity (see Table 4, Experiment no. 5).
Further continuation of erosion by sand particle impacts re-
sulted in damage to the interface between the bres and the resin
matrix. This damage is characterized by the separation and detach-
ment of broken bres from the resin matrix, as shown in Fig. 10e
and f. The observed behaviour of these materials can be attributed
to the following mechanism. It is well known that the bres in
composites subjected to particle ow, undergo breakage when
subjected to bending. Thus, there is a local removal of resin mate-
rial from impacted surface, which results in the exposure of the -
bres (see Table 4, Experiment no. 7). This causes intensive de-
bonding and breakage of the bres, which are not supported by
the matrix. The observed erosion damage is characterized by expo-
sure of carbon bres, brematrix de-bonding, multiple bre
cracking and material removal. Plastic deformation, melting of ma-
trix, pit formation and dislodging of bre are also seen on the sur-
face of the investigated polymeric composites. The impingement of
the erodent caused roughening of the surface of the material, espe-
cially in higher contact angles. A characteristic feature of more cut-
ting with chip formation is reected (Fig. 10e and f). Erosion along
the bres and clean removal of the matrix to expose carbon bres
is also seen (Fig. 10f). The matrix shows multiple fractures and
material removal. The exposed bres are broken into fragments
and thus can be easily removed from the worn surfaces. Fig. 10g
and h shows the SEM micrographs of surface features in carbon -
bre (30 wt.%) reinforced vinyl ester resin composites (VCF30)
eroded at impact angles of 90. Interestingly both the ductile dam-
age and the brittle cracks were viewed simultaneously not only at
the lower but also at the higher impact angles. It is indicated from
Fig. 10g that damage processes of the composite is only on the sur-
face without any broken-off bres (see Table 4, Experiment no. 10).
4. Conclusions
The thermo-mechanical properties and the solid particle ero-
sion performance of short carbon bre reinforced vinyl ester resin
based composites have been evaluated. It was observed that the
composite with 30 wt.% of carbon bre exhibited higher storage
modulus in the temperature range of 3050 C whereas above
50 C the storage modulus shows a sharp decay. The same compos-
ite also has exhibited maxima in the loss-modulus and tan d indi-
cating enhanced energy dissipation ability of the composite with
30 wt.% of carbon bre. Such specic dependence of the dynamic
mechanical performance on an optimum amount of bre content
i.e. 30 wt.% of short carbon bre may be attributed to adequate -
bre-resin adhesion and a better state of bre distribution in the
bulk of the composite. The solid particle erosion of such randomly
oriented/distributed short carbon bre reinforced vinyl ester resin
composites have revealed that impact velocity, impingement angle
and size of the erodent had a pronounced effect on the erosion
resistance of the composites besides the bre content. The opti-
mum conditions for maximization of erosion resistance have been
found to be at an impingement angle of 60, impact velocity of
65 m/s and erodent size of 420 lm whereas from material compo-
sition point of view the maximum wear resistance is observed in
the composite with 30 wt.% of carbon bres. Design of experiment
(DOE) approach based on Taguchi method has been found to be
successful for parametric optimization of erosion performance
with respect to experimental/operational and material variables.
The successful predictions have also been reconrmed by conduct-
ing analysis of variance (ANOVA) and conrmation experiments on
a randomly selected set of operational variables. Wear surface
morphology has revealed that erosion process involved different
mechanisms depending on the type and arrangement of bre con-
tent in the matrix. For short carbon bre composite, severe deteri-
oration of both bre and matrix, micro-ploughing in the matrix,
transverse shearing and stripping have been observed whereas
brillation, composite de-bonding, bre pull-out, and bre-frac-
ture were found out to be the characteristic features of carbon bre
damage. A direct correlation between erosive wear rate and loss-
modulus inverse (E001 ) irrespective of the bre concentration has
been observed. Such a resemblance of the two different material
performance aspects conceptually demonstrates a mechanistic
interrelation between the response of the composites to erosion
conditions and to thermo-mechanical loading situations.
References
[1] Kari S, Berger H, Gabbert U. Numerical evaluation of effective material
properties of randomly distributed short cylindrical bre composites. Comput
Mater Sci 2007;39:198204.
[2] Tewari US, Harsha AP, Hger AM, Friedrich K. Solid particle erosion of
unidirectional carbon bre reinforced polyetheretherketone composites. Wear
2002;252:9921000.
[3] Roy M, Vishwanathan B, Sundararajan G. The solid particle erosion of polymer
matrix composites. Wear 1994;171:14961.
[4] Hager A, Friedrich K, Dzenis YA, Paipetis SA. Study of erosion wear of advanced
polymer composites. In: Street K, Whistler BC, editors. Proceedings of the
ICCM-10. Cambridge (Canada): Woodhead Publishing Ltd.; 1995. p. 15562.
[5] Barkoula N-M, Karger-Kocsis J. Effects of bre content and relative bre-
orientation on the solid particle erosion of GF/PP composites. Wear
2002;252:807.
[6] Barkoula N-M, Karger-Kocsis J. Review-processes and inuencing parameters
of the solid particle erosion of polymers and their composites. J Mater Sci
2002;37:380720.
[7] Harsha AP, Tewari US, Venkataraman B. Solid particle erosion behaviour of
various polyaryletherketone composites. Wear 2003;254:693712.
[8] ASTM Standard D 4065-01. Standard practice for plastics: dynamic mechanical
properties: determination and report of procedures. West Conshohocken, PA,
USA: American Society for Testing and Materials; 2004.
[9] Ruff AW, Ives LK. Measurement of solid particle velocity in erosive wear. Wear
1975;35(1):1959.
[10] Glen SP. Taguchi methods: a hands on approach. Addison-Wesley; 1993.
[11] Dipa R, Sarkar BK, Das S, Rana AK. Dynamic mechanical and thermal analysis of
vinylesterresinmatrix composites reinforced with untreated and alkali-
treated jute bres. Compos Sci Technol 2002;62:9117.
[12] Yusriah L, Mariatti M, Abu Bakar A. Mechanical properties of particulateller/
woven-glass-fabric-lled vinyl ester composites. J Vinyl Add Technol
2010;16(1):98104.
[13] Wiederhorn SM, Hockey BJ. Effect of material parameters on the erosion
resistance of brittle materials. J Mater Sci 1983;18:76680.
[14] Aquaro D, Fontani E. Erosion of ductile and brittle materials. Meccanica
2001;36:65161.
[15] Montgomery DC. Design and analysis of experiments. 5th ed. John Wiley &
Sons, Inc.; 2005. p. 36382.
[16] Kulkarani SM, Kishore K. Inuence of matrix modication on the solid particle
erosion of glass/epoxy composites. Polym Polym Compos 2001;9:2530.
[17] Tewari US, Harsha AP, Hger AM, Friedrich K. Solid particle erosion of carbon
bre and glass breepoxy composites. Comput Sci Technol 2003;63:54957.
[18] Stachowiak G, Batchelor AW. Engineering tribology. Butterworth; 2005. ISBN-
13: 978-0-7506-7836-0.