J Polym Eng 2014; 34(9): 839–849

S. Kalyana Sundaram* and S. Jayabal

Mechanical properties of potassium hydroxide-

pretreated Christmas palm fiber-reinforced
polyester composites: characterization study,
modeling and optimization
Abstract: The mechanical properties of Christmas palm- 1 Introduction
reinforced polyester composite were improved by chemi-
cal modification of Christmas palm fibers through alkali
The exceptional mechanical properties of synthetic fiber-
treatment. An attempt was made in this investigation to
reinforced polyester composite materials have paved the
identify the possibility of using potassium hydroxide as
way to their wide range of applications, ranging from aero­
alkali solution. The fibers treated with potassium hydrox-
space to automobile parts. But the high cost and nonre-
ide aqueous solution at varying solution concentrations
newability of these composite materials created a growing
and soaking times were reinforced in a polyester matrix
interest in the use of natural cell­ulosic fibers as the rein-
and tested for their mechanical properties. The physico-
forcement in polymeric matrix [1]. Studies on natural fibers
chemical interaction between the treated Christmas palm
such as coir, sisal, banana, hemp and bamboo have shown
fiber and polyester matrix was characterized using scan-
the possibility of using natural fibers as a reinforcement
ning electron microscopy and X-ray diffraction. The sur-
in polymer composites [2–6]. However, the inadequate
face modification of fibers improved the compatibility
mechanical properties of natural fiber-reinforced polyes-
of fibers with polyester matrix, which, in turn, resulted
ter composites have restricted their use in engineering and
in better mechanical properties. A nonlinear regression
commercial applications. Hence it is important to charac-
model was developed to predict the mechanical properties
terize and improve the mechanical properties of natural
of the composites using response surface methodology.
fiber-­reinforced polyester compo­sites [7].
The better treatment conditions for optimum mechanical
The mechanical properties of natural fiber-reinforced
properties were determined using a heuristic optimization
polyester composites depend on the degree of crystallin-
method called particle swarm optimization (PSO). The
ity and cellulose content present in the reinforcing fiber.
optimized value obtained using the PSO technique was
The presence of excess cellulose and noncellulosic com-
validated through a confirmation test, and the result was
ponents such as lignin, wax and pectin makes fiber hydro-
found to be significant.
philic in nature and results in poor inter-laminar bonding
between the reinforcement and the matrix material [8].
Keywords: alkali treatment; Christmas palm; mechanical
It is therefore important to modify the fiber surface to
properties; particle swarm optimization (PSO); response
make fibers compatible with polymer matrix. Alkali treat-
surface; response surface methodology (RSM).
ment is one of the best methods suggested to improve the
bonding nature of natural fiber with matrix materials [9,
DOI 10.1515/polyeng-2014-0084 10]. Most research studies in this field have been moti-
Received March 31, 2014; accepted June 2, 2014; previously pub- vated to develop a mathematical model for predicting and
lished online July 7, 2014
optimizing the mechanical properties of the composites
[11, 12]. Jayabal and Natarajan [13] developed a regression
equation using response surface methodology (RSM) and
optimized the machinability behavior of glass-coir-poly-
*Corresponding author: S. Kalyana Sundaram, Department ester hybrid composite. In addition to this, the heuristic
of Mechanical Engineering, A.C. College of Engineering and
method of optimization can improve the performance of
Technology, Karaikudi 630 004, India,
e-mail: [email protected]
parameters estimation with less computational efforts.
S. Jayabal: Department of Mechanical Engineering, A.C. College of Several research studies in the past reported on the
Engineering and Technology, Karaikudi 630 004, India use of sodium hydroxide for enhancing the mechanical

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
840      S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber

interlock between the fibers and the polymer matrix [14, solution for various periods (12, 24, 36, 48 and 60 h). After
15]. An attempt was made in this investigation to study the each interval of time, the fibers were taken out from the
effect of potassium hydroxide (KOH) pretreatment on the solution and rinsed with dilute hydrochloric acid in order
mechanical properties of Christmas palm fiber-reinforced to remove any excess KOH sticking to the fiber surface.
polyester composites. The role of alkali solution concen-
tration (%) and soaking time (hours) in the modification
of the fiber surface was investigated in this study. The 2.3 C
 omposite fabrication and mechanical
composites were fabricated and tested as per full facto- testing
rial design, and prediction models were developed for
the mechanical properties such as tensile (ts), flexural (fs) The treated Christmas palm fibers of approx. 30  mm in
and impact (is) strength using response surface methodol- length and 30% weight content were randomly reinforced
ogy (RSM). The competency checking of the models was in a polyester resin system consisting of cobalt octoate
carried out using analysis of variance (ANOVA), and the and MEKP mixed at a ratio of 1:0.015:0.015. The prepregs
objective functions developed were optimized using parti- were placed in a stainless steel mold of dimension
cle swarm optimization (PSO). 300 × 300 × 3 mm and compressed by the upper and lower
jaw of the compression molding machine at a pressure of
2.6 MPa and at a temperature of 120°C for 45 min. The com-

2 Materials and methods posite sheets fabricated were cured at room temperature
for 24 h. After curing, the specimens were cut to required
dimensions for tensile, flexural and impact testing as per
2.1 Materials the ASTM D638-08, ASTM D790-07 and ASTM D256-06a
standards, respectively.
Adonidia merrillii, popularly known as Christmas palm The tensile tests were conducted on the prepared com-
fibers (diameter, 0.20–0.25 mm; density, 1060 kg/m3; mean posite samples using a computerized universal testing
breaking load, 37.24 N), was used as the reinforcement machine at a cross-head speed of 5 mm/min at room tem-
materials. The fibers were extracted from the foliage of a perature. Three-point flexural tests were conducted on
Christmas palm tree using a simple water retting process. the fabricated composite sheets using a digital universal
Potassium hydroxide pellets (density, 2.04 g/cm3; melting testing machine. The impact test was carried out on the
point, 360°C) were used for preparing alkali aqueous fabricated composite specimens using a pendulum-type
solution. Unsaturated polyester resin (specific gravity at impact testing machine. The test specimen was supported
27°C, 1.136; melting point, 260°C; viscosity, 470 cps; and as a vertical cantilever beam and broken by a single swing
mass, 449.96 g; GRP Enterprises, Madurai, Tamil Nadu, of a pendulum. To obtain a statistically significant result,
India) was used as the matrix material. The matrix mate- five specimens were tested to evaluate the average value of
rial was selected based on cost and availability [16, 17]. the tensile, flexural and impact strength.
Cobalt octoate and methyl ethyl ketone peroxide (MEKP)
were used as the accelerator and catalyst, respectively. The
accelerator and catalyst present in the resin system help in
2.4 Crystallinity and morphological study
enhancing the reaction rate during composite fabrication
and ensures uniform curing of the composite sheets.
The effect of treatment parameters on the crystallinity of
Christmas palm fibers was determined using an X-ray dif-
fraction technique. The investigation was carried out with
2.2 Fiber treatment Cu-Kα (1.5418 Å) radiation generated at 40 kV and 40 mA.
The diffraction spectra analyzed in this study were col-
The extracted fibers were subjected to chemical treatment lected from a 2θ angle of between 10° and 80° scanned at
using KOH aqueous solution. One gram of KOH was dis- a speed of 0.05°/s. The percentage crystallinity of raw and
solved in 100 mL of distilled water to give a 1% aqueous treated Christmas palm fibers was computed using the fol-
solution. The treatment was carried out as per full facto- lowing formula [18, 19]:
rial design, as shown in Table 1. Five aqueous solution of
I002
varying concentrations (2%, 4%, 6%, 8% and 10%) was %Crystallinity = × 100 (1)
I002 + I am
prepared, and the fibers were soaked in the prepared 

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
 S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber      841

Table 1 Process parameters and levels.

SI No.  Process parameters   Level 1   Level 2   Level 3   Level 4   Level 5

01  Solution concentration (%)  2   4   6   8   10

02  Soaking time (h)   12   24   36   48   60

where I002 and Iam are the crystalline and amorphous the position of each particle (i), which is updated in the
intensities at 2θ scale, respectively. The change in surface search space, can be represented as follows:
topography of the treated Christmas palm fibers and its
xit + 1 = xit + vit + 1 with xi0 ~U ( xmin , xmax ) (3)
physicochemical interaction with polyester matrix were 
characterized using scanning electron microscopy (SEM,
where vit is the velocity vector of the particle that drives
Hitachi S-3000N, Hitachi, Haryana, India) operated at 20
the optimization process and reflects both own and social
kV. To obtain a clear image, the specimens were washed
experience knowledge from all the particles, and U (xmin,
thoroughly, air dried and surface coated with gold to
xmax) is the uniform distribution, where xmin and xmax are
increase the electrical conductivity of the specimens.
its minimum and maximum values, respectively. The
global best PSO was followed in this investigation, where
the position of each particle is influenced by the best-fit
2.5 Response surface methodology
particle in the entire swarm. Each particle in the swarm
keeps track of its coordinate in the search space. The per-
Response surface methodology is a statistical tool that
sonal best position (Pbest, i) corresponds to the position in
uses quantitative data from appropriate experiments to
the search space, where the particle had the largest value
determine the regression equation between the depend-
as determined by the objective function. In addition, the
ent variables and the independent variables [20–22]. The
position yielding the largest value among all the personal
prediction models for determining the tensile, flexural
best (Pbest, i) is called the global best position (Gbest, i). The
and impact strength of pretreated Christmas palm fiber-
velocity of the particle (i) is calculated by
polyester composites over a wide range of treatment con-
ditions were developed using RSM. The Design-Expert v. vijt + 1 = vijt + c1r1tj  Pbest,
t
i
-xijt  + c 2 r2t j Gbest -xijt  (4)
9 modeling software (Stat-Ease, Inc., MN, USA) was used 
in this investigation to formulate the nonlinear regression where vijt is the velocity vector of particle i in dimension j
equation. The tensile strength (ts, MPa), flexural strength at time t; xijt is the position vector of the particle; Pbest
t
,i
is
(fs, MPa) and impact strength (is, kJ/m2) were considered as the personal best position of the particle; Gbest is the global
the process response (y), which is a function of the inde- best position of the particle; c1 and c2 are the positive accel-
pendent variables, namely, solution concentration (sc,%) eration constants, which are used to level the contribution
and soaking time (st, h), which can be expressed as follows: of the cognitive and social components, respectively; and
r1tj and r2t j are random numbers from uniform distribution
y= e+f (sc , st ) (2)
 U (0, 1) at time t. The PSO program was written in and exe-
The term e represents any measurement error on the cuted using the MATLAB 7.1 R2010a software (MathWorks,
response, as well as a statistical error that is assumed to MA, USA), and the control parameters used in this study
distribute normally with zero mean and variance. are listed in Table 2.

2.6 Particle swarm optimization

3 Results and discussion
An optimization technique determines the best-suited
solution to a problem for a given circumstance. The PSO 3.1 E
 ffect of alkali treatment on the
technique was proposed by Kennedy and Eberhart [23] mechanical properties of composites
based on the collective behavior of decentralized, self-
organized swarm systems [24, 25]. In PSO, every particle The experimental results obtained after conducting
flies through the multidimensional space and adjusts its tensile, flexural and impact testing on the fabricated com-
position ( xit ) at every time step (t) with its own experience; posites are shown in Table 3. The effect of the treatment

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
842      S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber

Table 2 PSO control parameters. removed excess lignin and hemicelluloses from the outer
surface of the fibers [26].
SI No.  PSO parameters   Values The flexural strength plot shown in Figure 1B substan-
01  Number of particles   20 tiated the adverse effect of shorter (12 h) and longer (60 h)
02  Particle dimensions   2 (solution concentration, soaking soaking time on the flexural strength of the composites.
time) The Christmas palm fibers soaked in alkali solution for
03  Particle ranges   Solution concentration (%) = [2–10]
12 h also resulted in poor impact strength (Figure 1C). This
    Soaking time (h) = [12–60]
04  Inertia factor   1.2 may be due to the scarcity of lignin and other impurities
05  Number of cycles   5 removed from the fiber surface. Similarly, the lower and
06  Number of iterations  100 higher level of solution concentration (i.e., 2% and 10%,
07  Learning factor   [c1, c2] = [1.5, 1.5] respectively) also resulted in poor mechanical properties.
The tensile, flexural and impact strength of the composites
were found to be finer at solution concentrations of 4%, 6%
parameters, namely, solution concentration and soaking and 8%, and soaking times of 24, 36 and 48 h. However, a
time, on the mechanical properties of treated Christmas better value of the tensile strength (37 MPa) was obtained
palm fiber-reinforced composites is shown in Figure 1. with the composite reinforced with the fibers treated in 6%
The Christmas palm fibers soaked in alkali solution for aqueous solution for 24 h, whereas the flexural strength
60  h resulted in poor tensile strength at all the levels of (81 MPa) and the impact strength (53 kJ/m2) of the compos-
solution concentration considered in this investigation ites were found to be better with the fibers soaked in 4%
(Figure 1A); this is due to glut leaching of the fibers, which aqueous solution for 36 and 48 h, respectively.
The obtained values were compared with the values
of composites fabricated with untreated and sodium
Table 3 Experimental results. hydroxide-pretreated Christmas palm fibers reported in
a previous work [27]. The comparison study shown in
Runs  Solution  Soaking  Tensile  Flexural  Impact Table 4 confirmed that the alkali treatment of Christmas
concentration time (h) strength strength strength
palm fiber using KOH has effectively improved the tensile,
(%) (MPa) (MPa) (kJ/m2)
flexural and impact strength of the composites compared
01   2  12  31.5  67  42 to that of composites reinforced with untreated Christ-
02   2  24  33  73.5  46.5
mas palm fibers and sodium hydroxide-treated Christmas
03   2  36  30  76  49
04   2  48  25.5  74.5  52.5 palm fibers. Moreover, the soaking time required for the
05   2  60  22  67  50 surface modification of Christmas palm fiber using a KOH
06   4  12  34  73  40 aqueous solution was found to be shorter when compared
07   4  24  36  79.5  47.5 to that of sodium hydroxide alkali treatment. This, in turn,
08   4  36  35  81  52
will play a significant role in improving the productivity
09   4  48  31  79  53
10   4  60  24  75  51
during the surface treatment process.
11   6  12  36.5  70  37.5
12   6  24  37  78  44.5
13   6  36  36  80.5  49 3.2 D
 egree of crystallinity of treated
14   6  48  33  77.5  50
15   6  60  26.5  69  47
Christmas palm fibers
16   8  12  35  56  30
17   8  24  36.5  69.5  37 The percentage crystallinity (%Cr) of raw and alkali-treated
18   8  36  34.5  71  41.5 Christmas palm fibers that resulted in better mechani-
19   8  48  32  69  42 cal properties was investigated using an X-ray diffraction
20   8  60  25.5  63  41 technique. Figure 2 shows the diffraction peaks of raw and
21   10  12  32  45  17
KOH-treated Christmas palm fibers. Both raw and treated
22   10  24  35.5  52  28
23   10  36  32.5  64  32
Christmas palm fibers exhibited major crystalline peaks at
24   10  48  29  58  34 2θ = 22°, representing the (002) cellulose crystallographic
25   10  60  27  51  31 plane. The minimum peak intensities identified at 2θ = 18°
Mean   31.62  68.76  41.80 representing the (110) cellulose crystallographic plane
Standard deviation   0.99  2.12  1.03 correspond to amorphous materials present in the cellu-
Coefficient of variation   3.12  3.09  2.41
lose of raw and treated fibers. The shift in the amorphous

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
 S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber      843

A 38 B 90
36 85
80
34
75
32
Tensile strength (MPa)

Flexural strength (MPa)

70
30 65
28 60
26 55
50
24
12 h 45
22 24 h 12 h
40
36 h 24 h
20 35 36 h
48 h
18 30 48 h
60 h
25 60 h
16
20
2 4 6 8 10 2 4 6 8 10
Solution concentration (%) Solution concentration (%)

C 60
55
50
45
Impact strength (kJ/m2)

40
35
30
25
12 h
20 24 h
36 h
15 48 h
10 60 h

5
2 4 6 8 10
Solution concentration (%)

Figure 1 Effect of treatment parameters on the (A) tensile strength, (B) flexural strength and (C) impact strength of the composites.

intensity (I002) found with treated fibers may be due to the fibers was found to decrease with the increase in solu-
increase in the amount of cellulose II formation resulting tion concentration. The fibers treated with 4% aqueous
from alkali treatment. Liu et  al. [28] observed a similar solution for 36 h revealed peaks at 2θ = 37°, 48° and 60°,
result when investigating the X-ray diffraction peaks of which were not observed with the fibers treated with 4%
treated bamboo fibers. The percentage crystallinity of raw aqueous solution for 48 h; this is because the increase in
and treated fibers is shown in Table 5. The crystalline peak soaking time also suppressed the degree of crystallinity of
intensity for raw Christmas palm fibers was found to be the fibers. A similar effect was reported by Borysiak and
higher than that for treated Christmas palm fibers, and, at Doczekalska [29] on studying the effect of alkali treatment
the same time, the intensity of crystalline peak for treated on pine wood. The decrease in crystallinity was due to the

Table 4 Comparison of untreated and treated Christmas palm fiber/polyester composites.

SI No.  Alkali treatment   Tensile   Flexural   Impact strength

strength (MPa) strength (MPa) (kJ/m2)

01  Untreated fiber Christmas palm   23   54.2   35.1

fiber/polyester composites
02  NaOH-pretreated Christmas   35   70.6   42
palm fiber/polyester composites
03  KOH-pretreated Christmas palm   37   81   53
fiber/polyester composites

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
 844      S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber

formation of rough surface with regularly spaced pits.

I002
The rough surface results in an amorphous structure
I110 in the fibers, which leads to better mechanical interac-
tion with the polyester matrix [30]. From the SEM images
of tensile fracture surface shown in Figure 4A, it was
Intensity (arb.u)

Untreated fiber
observed that the rough fiber surface resulting from
the alkali treatment has restricted fiber pullout from
4%-36 h alkali treated
the matrix surface. A similar inference was observed
4%-48 h alkali treated
by Haque et  al. [31] on investigating the effect of alkali
treatment on the mechanical properties of palm and coir
fiber-reinforced polypropylene composites. The fiber-
6%-24 h alkali treated
matrix interaction region highlighted in the SEM image
20 40 60 80
2θ (°)
of tensile fractured specimen substantiated the appreci-
able wetting of treated fibers with polyester resin. The
Figure 2 X-ray diffraction patterns of raw and treated Christmas sticking of the polyester matrix on the fiber surface after
palm fibers. the flexural fracture test, as shown in Figure 4B, shows
the adhesion of the matrix with regularly spaced pits
found over the fiber surface.
swelling of fibers, which, in turn, created better adhesion The SEM images of impact fracture surface shown in
between the fibers and the polyester matrix. Hence it can Figure 4C reveal a uniform breakage of fiber perpendicular
be stated that both solution concentration and soaking to the fiber direction; this is because of the fair interfacial
time have played a prominent role in decreasing the crys- bonding between the treated fibers and the matrix. The
tallinity of the fibers. crack that developed in the fiber indicated the resistance
offered by the fibers to the applied load by observing the
energy in its own fracture. The SEM study confirmed that
3.3 Morphological study the alkali pretreatments of Christmas palm fiber using
KOH have improved the bonding between the fibers and
The SEM images of raw and treated Christmas palm fiber the matrix surface, which, in turn, has diminished void
shown in Figure 3A and B, respectively, confirm the formation, laceration and crack growth in the composites.

Table 5 Percentage crystallinity of raw and treated Christmas palm fibers.

SI No.  Sample   I002 (at 2θ scale)   Iam (at 2θ scale)   % Crystallinity

01  Untreated Christmas palm fiber   570   305   65.1

02  Alkali-treated Christmas palm fiber (4% and 36 h)   380   228   62.5
03  Alkali-treated Christmas palm fiber (4% and 48 h)   368   249   59.6
04  Alkali-treated Christmas palm fiber (6% and 24 h)   120   107   52.8

A B

Figure 3 SEM image of (A) raw Christmas palm fiber and (B) alkali-treated Christmas palm fiber.

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
 S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber      845

A B

Figure 4 SEM image of (A) tensile fractured specimen, (B) flexural fractured specimen and (C) impact fractured specimen.

3.4 Nonlinear regression analysis recapitulated in Tables 6–8 indicate that all factors have
contributed significantly to the models. The coefficient of
The significance of treatment parameters on the tensile, determination (R2) for the tensile (0.95), flexural (0.96) and
flexural and impact strength of the pretreated Christmas impact strength model (0.98) indicates that the second-
palm fiber-reinforced polyester composites was scruti- order quadratic model is adequate to represent the relation-
nized using ANOVA. The ANOVA for the tensile, flexural ship between treatment parameters and the mechanical
and impact strength model is summarized in Tables 6–8, properties. The low value of the coefficient of variation for
respectively. The parameters with 95% confidence the tensile, flexural and impact model supports the high
level were considered as a significant factor. The larger significance of the corresponding models. The low stand-
values of F and smaller values of P (probability > F)  < 0.05 ard deviation value shown in Table 3 indicates that all

Table 6 ANOVA for the tensile strength model.

Source   Sum of squares  df  Mean square  F-value  p-Value (probability  > F)

Model   436.17  5  87.23  89.71   < 0.0001

A – Solution concentration (%)  19.84  1  19.84  20.41  0.0002
B – Soaking time (h)   266.81  1  266.81  274.39   < 0.0001
 A*B   4.62  1  4.62  4.75  0.0420
 A2   61.29  1  61.29  63.03   < 0.0001
 B2   83.60  1  83.60  85.98   < 0.0001
Residual   18.47  19  0.97  –  –
Corrected total   454.64  24  –  –  –
R2 (adjusted) = 0.92   Coefficient of correlation (R2) = 0.95

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
846      S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber

Table 7 ANOVA for the flexural strength model.

Source   Sum of   df   Mean   F-value   p-Value

squares square (probability > F)

Model   2257.34   5   451.47   100.07    < 0.0001

A – Solution concentration (%)   1104.50   1   1104.5   244.81    < 0.0001
B – Soaking time (h)   22.44   1   22.44   4.97   0.0380
 A*B   19.36   1   19.36   4.29   0.0490
 A2   630   1   630   139.64    < 0.0001
 B2   481.03   1   481.03   106.62    < 0.0001
Residual   85.72   19   4.51   –   –
Corrected total   2343.06   24   –   –   –
R2 (adjusted) = 0.94   Coefficient of correlation (R2) = 0.96

data points considered in this investigation tend to be very response variables with respect to treatment parameters.
close to the mean value of the tensile, flexural and impact The 3D response surface plot of the tensile strength model
strength. shown in Figure 6A confirms the increase in the tensile
The perturbation plot shows how the response strength of the composites with the increase in solution
changes, as each factor moves from the chosen reference concentration and soaking time, but, at the same time,
point, when all other factors are held constant at the ref- adverse effect was inferred when the solution concen-
erence value. The reference point was set at the middle tration and soaking time were increased beyond 6% and
of the design space, and the effect of solution concentra- 24 h, respectively. Figure 6B and C shows the 3D response
tion (%) and soaking time (hours) on the tensile, flexural surface plot for the flexural and impact strength model,
and impact strength is shown in Figure 5A–C, respectively. respectively. The flexural and impact strength of the com-
The tensile, flexural and impact strength were found to posites tended to decline when the solution concentra-
decrease as the solution concentration curve (A) tends tion was increased beyond 4%. The increase in soaking
to move towards the positive value from the set refer- time beyond 36 and 48 h also resulted in the decrease in
ence point in the space. The tensile, flexural and impact flexural and impact strength, respectively. The quadratic
strength reached the maximum well before the soaking polynomial equation developed for the tensile, flexural
time curve (B) had approached the reference point in the and impact strength of pretreated Christmas palm fiber-
space but dwindled thereafter. This inference confirmed reinforced polyester composites using the RSM is given in
that a higher solution concentration results in overleach- Equations 5, 6 and 7, respectively.
ing of the fiber surface. t s = +24.395+2.79964sc + 0.30018st +0.00895sc st
The three-dimensional (3D) response surface and
-0.23393sc2 -0.007589st2
counterplot shown in Figure 6 stipulate the variation in (5)

Table 8 ANOVA for the impact strength model.

Source   Sum of  df  Mean square  F-value  p-Value

squares (probability > F)

Model   2010.02  5  402.08  380.77   < 0.0001

A – Solution concentration (%)  1240.02  1  1240.02  1174.31   < 0.0001
B – Soaking time (h)   367.20  1  367.20  347.75   < 0.0001
 A*B   5.06  1  5.06  4.79  0.0412
 A2   226.80  1  226.80  214.78   < 0.0001
 B2   171.29  1  171.29  162.21   < 0.0001
Residual   20.06  19  1.06  –  –
Corrected total   2030.44  24  –  –  –
R2 (adjusted) = 0.95   Coefficient of correlation (R2) = 0.98

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
 S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber      847

A 38
A
40
B
36

Tensile strength (MPa)

35
Tensile strength (MPa)

34 30
A
25
32
A
20
30
A-Solution concentration (%)
28 B-Soaking time (h) 60 10
B 52 8
44
36 6
26 28 4
B:Soaking time (h) 20
12 2 A:Solution concentration (%)
-1 -0.5 0 0.5 1
Deviation from reference point

B B
90
85

Flexural strength (MPa)

80
80
A 70
Flexural strength (MPa)

75 60

B 50
70
B 40

65 A-Solution concentration (%)

B-Soaking time (h) 60 10
60 52 8
A 44
36 6
28 4
55 B:Soaking time (h) 20
12 2 A:Solution concentration (%)
-1 -0.5 0 0.5 1
Deviation from reference point C
60
Impact strength (kJ/m2)

C 55
50

A 40
50 30
Impact strength (kJ/m2)

B
20
45
10

40
60 10
B A-Solution concentration (%) 52
44 8
35 B-Soaking time (h) 36 6
28 4
A B:Soaking time (h) 20
12 2 A:Solution concentration (%)
30

-1 -0.5 0 0.5 1
Figure 6 Three-dimensional surface and contour plot for the (A)
Deviation from reference point
tensile strength model, (B) flexural strength model and (C) impact
strength model.
Figure 5 Perturbation graph for the (A) tensile strength model,
(B) flexural strength model and (C) impact strength model.

represent the solution concentration and soaking time of

fibers in KOH alkali solution.
fs =+45.46+5.99sc + 1.25655 st + 0.018333sc st -0.75sc2
-0.018204st2 (6)
3.5 Optimization using PSO
is =+27.105+2.5725sc + 0.95173st + 0.00937sc st -0.45s 2
c
The objective function developed for the tensile, flexural
  -0.010863st2 (7)
and impact strength using RSM was maximized using the
where ts is the tensile strength, fs is the flexural strength PSO technique. Figure 7A shows the PSO plot for the tensile
and is is the impact strength of the pretreated Christmas strength model. The result converged at the 20th itera-
palm fiber-reinforced polyester composites. sc and st tion, and an optimum tensile strength value of 37.06 MPa

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
848      S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber

A 37 was predicted at a solution concentration of 6.44% and a

soaking time of 23.58 h. A solution concentration of 4.2%
36.5 and a soaking time of 39.6 h resulted in an optimum flex-
ural strength of 81.84 MPa, as shown in Figure 7B. Figure
Tensile strength (MPa)

7C indicates the optimum impact strength value of 53.45

36
kJ/m2 predicted for the fibers soaked in 3.33% KOH solu-
PSO plot for tensile strength model tion for 45.23 h. The flexural and impact results converged
35.5
around the 40th and 50th iteration, respectively. The
obtained results were validated by conducting confirma-
35 tion experiments with the predicted set of treatment con-
ditions. The experimental values and the values predicted
34.5 by PSO were compared, and the percentage of error was
0 10 20 30 40 50 60 70 80 90 100
calculated as shown in Table 9. The minimum error per-
Number of iteration
centage obtained shows the significance of the optimum
B 82
treatment condition predicted by the PSO technique.
81

80
Flexural strength (MPa)

79

78 4 Conclusion
77

76
PSO plot for flexural strength model The alkali treatment of Christmas palm fibers using KOH
alkali solution has played a better role in enhancing the
75
mechanical properties of polyester composites. The mod-
74
eling and optimization technique used in this investigation
73
have effectively predicted the mechanical properties of the
72
0 10 20 30 40 50 60 70 80 90 100 composites. The investigation also witnessed the signifi-
Number of iteration cance of alkali solution concentration and soaking time in
removing the excess lignin and water soluble impurities
C 54
present on the fiber surface. Hence it is proposed that sodium
52 hydroxide alkali solution be predominantly replaced by
KOH alkali solution for promoting better physicochemical
Impact strength (kJ/m2)

50 interaction between natural fibers and the polymer matrix.

Moreover, hybridization of glass fiber composites with
48
PSO plot for impact strength model
treated Christmas palm fibers may open up a new applica-
tion for natural fiber-reinforced polyester composites.
46

Acknowledgments: The author would like to thank Dr. G.

44
Sozhan, Senior Principal Scientist, Central Electrochemi-
cal Research Institute, Karaikudi, Tamil Nadu, India,
42
0 10 20 30 40 50 60 70 80 90 100 and Dr. P. Vijayan, Associate Professor, Department of
Number of iteration
Mechanical Engineering, Government College of Engi-
Figure 7 PSO plot for the optimum value of (A) tensile strength, neering, Salem in Salem, Tamil Nadu, India, for their valu-
(B) flexural strength and (C) impact strength. able suggestion to do this work.

Table 9 Validation of PSO results.

SI No.  Properties   Solution   Soaking   Predicted value   Percentage

concentration (%) time (h) (experimental value) error (%)

01  Tensile strength (MPa)   6.44   23.58   37.06 (37.5)   1.18

02  Flexural strength (MPa)   4.2   39.6   81.84 (81)   1.02
03  Impact strength (kJ/m2)  3.33   45.23   53.45 (53)   0.84

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM
 S. Kalyana Sundaram and S. Jayabal: Potassium hydroxide pretreated Christmas palm fiber      849

References [16] Mishra S, Mohanty AK, Drzal LT, Misra M, Parija S, Nayak SK,
Tripathy SS. Compos. Sci. Technol. 2003, 63, 1377–1385.
[17] Monteiro SN, Terrones LAH, D’Almeida JRM. Polym. Test. 2008,
[1] Harish S, Michael DP, Bensely A, Mohan Lal D, Rajadurai A.
27, 591–595.
Mater. Charact. 2009, 60, 44–49.
[18] Kaith BS, Singha AS, Gupta SK. J. Polym. Mater. 2003, 20,
[2] Satyanarayana KG, Pillai CKS, Sukumara K, Pillai SGK. J. Mater.
195–199.
Sci. 1982, 17, 2453–2462.
[19] Kaushik VK, Kumar A, Kalia S. Int. J. Text. Sci. 2012, 1(6),
[3] Venkata Reddy G, Venkata Naidu ST, Shobha Rani T. J. Reinf.
101–105.
Plast. Compos. 2009, 28, 2035–2044.
[20] Cornell JA. How to Apply Response Surface Methodology, 2nd
[4] Pothan LA, Thomas S, Neelakantan NR. J. Reinf. Plast. Compos.
ed., American Society for Quality Control: Wisconsin, 1990.
2009, 16, 744–764.
[21] Box G, Draper N. Empirical Model Building and Response Sur-
[5] Panthapulakkal S, Sain MM. J. Appl. Polym. Sci. 2007, 103,
faces, Wiley: New York, 1987.
2432–2441.
[22] Cojocaru C, Zakrzewska-Trznadel G. J. Membr. Sci. 2007, 298,
[6] Murali Mohan Rao K, Mohana Rao K. Compos. Struct. 2007, 77,
56–70.
288–295.
[7] Gassan J, Bledzki AK. Compos. Sci. Technol. 1999, 59(9), [23] Kennedy J, Eberhart R. Proceedings of the International
1303–1309. Conference, Neural Networks, Perth, Australia, 1995, pp.
[8] Cao Y, Shibata S, Fukumoto I. Compos. Part A: Appl. Sci. Manuf. 1942–1948.
2006, 37(3), 423–429. [24] Kumar ND, Reddy JM. J. Water Resour. Plann. Manage. 2007,
[9] Kumar R, Obrai S, Sharma A. Der Chem. Sin. 2011, 2(4), 133, 192–201.
219–228. [25] Jarboui B, Ibrahim S, Siarry P, Rebai A. Comput. Ind. Eng. 2008,
[10] Jacob M, Thomas S, Varughese KT. Compos. Sci. Technol. 2004, 54, 526–538.
64, 955–965. [26] Prasad SV, Pavithran C, Rohatgi PK. J. Mater. Sci. 1983, 18,
[11] Jayabal S, Natarajan U. Int. J. Adv. Manuf. Technol. 2010, 51, 1443–1454.
371–381. [27] Kalyanasundaram S, Jayabal S. Appl. Mech. Mater. 2014, 467,
[12] Jayabal S, Natarajan U, Sekar U. Int. J. Adv. Manuf. Technol. 208–213.
2011, 55, 263–273. [28] Liu Y, Hu H. Fibers Polym. 2008, 9, 735–739.
[13] Jayabal S, Natarajan U. Int. J. Mach. Mach. Mater. 2011, 9, [29] Borysiak S, Doczekalska B. Fibers Text East. Eur. 2005, 13,
149–172. 87–89.
[14] Mahjoub R, Yatim JM, Mohd Sam AR, Hashemi SH. Constr. [30] Valadez-Gonzalez A, Cervantes-Uc JM, Olayo R, Herrera Franco
Build. Mater. 2014, 55, 103–113. PJ. Compos. Part B 1999, 30, 309–320.
[15] Ishikawa H, Takagi H, Nakagaito AN, Yasuzawa M, Genta H, [31] Haque MM, Hasan M, Islam MS, Ali ME. Bioresour. Technol.
Saito H. Compos. Interf. 2014, 21, 329–336. 2009, 100, 4903–4906.

Brought to you by | HEC Bibliotheque Maryriam ET J.

Authenticated
Download Date | 6/9/15 2:18 PM