International Journal of Biological Macromolecules 83 (2016) 385–395

Contents lists available at ScienceDirect

International Journal of Biological Macromolecules

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

Production and characterization of cellulose reinforced starch (CRT)

films
K. Sudharsan a , C. Chandra Mohan a , P. Azhagu Saravana Babu a , G. Archana a , K. Sabina a ,
M. Sivarajan b , M. Sukumar a,∗
a
Centre for Biotechnology, A.C. Tech., Anna University, Chennai 600 025, India
b
Chemical Engineering Division, Central Leather Research Institute, Chennai 600 020, India

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

Article history: Starch from Tamarind seed is considered to be a nonedible and inexpensive component, with many
Received 28 July 2015 industrial applications. Extraction and characterization of tamarind seed starch was carried out for the
Received in revised form 16 October 2015 synthesis of biopolymer. Tamarind seeds were collected, cleaned and further roasted, decorticated, and
Accepted 12 November 2015
pulverized to get starch powder. Total starch content present in each tamarind seed is estimated to
Available online 26 November 2015
be around 65–70%. About 84.68% purified starch can be recovered from the tamarind seed. Defatted
Tamarind seed starch has an amylose content of 27.55 wt.% and 72.45 wt.% of amylopectin. Morphological
Keywords:
(SEM) and X-ray diffraction were used to evaluate crystallinity. Likewise, TGA and DSC of starch have also
Tamarindus indica
Defatted starch
been analyzed. Thermal properties of starch obtained from tamarind seeds showed good thermal stability
Biopolymer material when compared to other starch sources such as Mesquite seed and Mango kernel. This study proved
that the tamarind seed starch can be used as a potential biopolymer material. Thermo-stable biofilms
were produced through initial optimization studies. Predictive response surface quadratic models were
constructed for prediction and optimization of biofilm mechanical properties. Correlation coefficient
values were calculated to me more than 0.90 for mechanical responses which implies the fitness of
constructed model with experimental data.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction The ‘biodegradability’ of plastics not only depends on raw materi-

Synthetic polymer-based composite materials are well estab-
lished all over the world and used for a wide variety of applications
because of their high specific strength and modulus. When com-
pared to conventional materials such as metals and alloys, it
possesses long durability. The main types of plastics that are cur-
rently used in all applications are derived from non-renewable
petroleum resources which are unstable and expensive [1].
Petroleum-based plastics do not degrade and remain accumulated
in the environment in the form of solid waste for longer period
leading to increased GHG Emissions.
As usage of plastics becomes more rapid in recent years,
microorganisms capable of synthesizing and designing new
enzyme structures capable of degrading synthetic polymers are
required [2]. Bio plastics would play a big role in the future as a
viable alternative for petroleum-based plastics. Many biopolymers
are biodegradable, hydro-biodegradable or photo-biodegradable.
∗ Corresponding author.
E-mail address: sukumaractech@gmail.com (M. Sukumar).
als used for production but also on the chemical structure of the
material and constitution of the final product [3].
Biodegradable polymers can either be naturally produced from
renewable resources (such as starch) or it can be artificially syn-
thesized. Renewable biodegradable polymers are produced from
polysaccharides (starch, cellulose, lignin), proteins (gelatin, wool,
silk), lipid (fats and oil), and polyesters by plant or microorgan-
isms (PHA); polyesters are derived from bio derived monomers
(polylactic acid), and miscellaneous polymers like natural rubbers,
composites. However, non-renewable synthetic biodegradable
plastics are petroleum based.
In the family of renewable source-based polymeric materials,
starch is the most inexpensive, abundant natural resource and
increasingly used as a packaging material [4]. Starch is degraded
to carbon dioxide and water in a relatively short period of time
when compared to most of the synthetic polymers. A range of
new applications is possible due to the attractive characteristics of
granular starch such as natural availability, biodegradability, low
cost and simple physico-chemical modifications [5–7]. Currently,
research is carried out on the materials synthesized from starch,
because of its biodegradability and renewable nature [8,9]. Raw

http://dx.doi.org/10.1016/j.ijbiomac.2015.11.037
0141-8130/© 2015 Elsevier B.V. All rights reserved.
386 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395
materials such as corn, potato, and wheat starches were mostly
used as fillers in traditional plastics [7]. Structure, properties, and
possibility of starch-based materials have been influenced by the
amylose/amylopectin ratio [10,11]. Superior strength and tough-
ness can be achieved in polymers by increasing the amylose content
of starch [12]. Both tensile modulus and yield stress of starch were
also increased with the increasing amylose content of the raw
material.
Tamarind seed gum or tamarind kernel powder (TKP) is derived
from the seeds of Tamarindus indica Linn which belongs to the
dicotyledonous subfamily Caesalpinioideae (Leguminosae). Asian
countries like India, Bangladesh, Sri Lanka, Thailand, and Indonesia,
and also in the African and American continents [14]. Tamarind
fruits are edible but their seeds are generally discarded, and there-
fore it can be a source of starch. Starch obtained from tamarind
seeds is nonedible. Starch extracted from tamarind seeds is cheap
with many industrial applications. This starch is used mainly in
textile and plywood industry and also in industries manufacturing
starch-based adhesives. Cellulose is also one of the commercially
important polysaccharides, capable of forming hydrocolloids in a
suitable solvent system. Bio-plastics made from cellulose have been
reported to have higher resistant to water and microwave heating
[36]. Addition of cellulose increases tensile strength and rigidity of
biofilms. Dias et al. reported increase in stability of starch-based
films in combination with cellulose fibres [37].
Response surface methodology (RSM) is an effective tool for
modelling complex and nonlinear problems in various scientific
and technological fields. RSM has been used effectively in many
fields for optimization studies.
The aim of the present study is to extract and character-
ize tamarind seed starch and also to analyze their feasibility on
starch-based biopolymer production. Morphological, structural,
and thermal properties of starch have also been characterized and
reported.
2. Materials and methods
2.1. Raw materials
Tamarind seeds were collected from nearby villages, in and
around Chennai, Tamil Nadu, India. Microcrystalline cellulose (98%
pure) extracted from sugar cane bagasse was purchased from Jeyam
Scientific’s Chennai. All the chemicals and solvents were procured
from Merck (Germany) Purity-EMPARTAACS .
2.2. Extraction of starch from tamarind seed
The process of synthesizing Tamarind seed starch involves
roasting followed by pulverization. The tamarind seeds are roasted
in oil-fired roaster and decorticated to remove skin. Seeds are then
broken into smaller pieces in a grinder. These broken pieces are
then finally pulverized to make starch powder. Tamarind seed
starch has been defatted using methanol for 10 h at 65 ◦ C fol-
lowed by n-hexane for another 10 h at 69 ◦ C in a Soxhlet extractor.
Defatted starch was blended with 70% ethanol and 0.1 M NaOH for
5 min in each solvent, respectively. The filtrate was centrifuged at
11,000 × g for 15 min. The starch powder residue thus obtained is
mixed with water, re-filtered twice using 200 mesh screen and
Whatman analytical grade No. 5 filter paper with 2.5 ␮m pore
size, and then washed successively with 0.1 M NaOH and deion-
ized water [15]. Finally, starch powder was dried in an oven at a
temperature of 50 ◦ C till it reaches constant weight. The prepared
starch was then packed and stored in polythene lined gunny bags
at −5 ◦ C for further studies.
2.3. Characterization of starch
2.3.1. Thermal analysis
TGA analyses were performed in TA Instruments STD Q500
(New Castle, DE, USA). Samples of starch were extracted from
tamarind seed and other nonedible plants such as Prosopis glan-
dulos (Mesquite seed) and Mangifera indica (Mango kernel). About
10 (±3) mg of the sample were placed in alumina pans in sepa-
rate experiments. The samples were heated at a rate of 5 ◦ C/min
from ambient temperature until it reached 800 ◦ C in a nitrogen
atmosphere with a sample flow of 60 mL/min and a balance flow
of 40 mL/min, to avoid the production of unwanted by-products in
the presence of oxygen during the analysis. The experiments were
conducted in duplicates.
Differential Scanning Calorimetry (DSC) analysis was carried out
using a DSC Q200 apparatus of TA Instruments (New Castle, DE,
USA) for studying out gelatinization temperature of starch. A sam-
ple for DSC was weighed (about 6 mg) on a small aluminium pan.
The pans were hermetically sealed and placed on mixing rollers
overnight to allow homogeneous hydration of the sample before
carrying out the analysis. An empty pan was used for a reference.
The samples were heated from 30 to 220 ◦ C at a rate of 20 ◦ C/min.
DSC scans were made under an inert atmosphere with a nitrogen
flow of 25 mL/min.
2.3.2. Proximate analysis
Total starch content was analyzed using the modified method
of Association of Analytical Communities [16]. 3,5-Dinitrosalicylic
acid, popularly known as DNS was chosen as the reagent instead
of glucose oxidase–peroxidase–aminoantipyrine buffer mixture
mentioned in the official method. The protein and ash content were
determined by Nitrogen-Ammonia-Protein Modified Kjedahl and
firestone Methods [17,18] respectively.
The amylose content was determined by a colorimetric method
(EN ISO 6647 parts I and II), which is based on amylose–iodine
complex formation. The absorbance was measured at 620 nm using
an UV-Vis-NIR Spectrophotometer Shimadzu, Japan and the amy-
lopectin content was calculated by difference.
2.3.3. Particle size distribution
Particle size distribution of starch was measured in water with
a laser diffraction system (H1140, Sympatec Inc., New Jersey, USA).
Deionized water was used as the medium for wet size measurement
of the starch sample. Sonic measurements were performed to avoid
the aggregation of starch.
2.3.4. Morphological properties
The micrographs were obtained through scanning electron
microscope (SEM) with EDS (HITACHI-S3400N, Japan) in the
Department of Mechanical Engineering, Anna University, Chennai.
The starch grains were scattered on the surface of a double-sided
tape which is attached to a stub and coated with gold. The images
were taken at an accelerating potential of 10 kV with 100× magni-
fication.
Wide angle X-ray diffraction measurements were carried out
using a D8 Advance Bruker AXS X-ray powder diffraction using
a one-dimensional Debye–Scherr camera, Cu Ka radiation (wave-
length 0.1542 nm) operating at 40 kV and 35 mA. The scattering
angle (2h) covered the range from 3◦ to 30◦ (h is the Bragg angle)
with a step of 0.02◦ and a sampling interval of 10 s.
The specimens were stored in a desiccator over a saturated solu-
tion of NaCl in a constant humidity atmosphere of 75% for 1 week
before the spectra were collected.
 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395 387

2.3.5. Fourier transform infrared spectroscopy Linear regression was used to estimate the slope of the g s−1
Fourier transform infrared spectroscopy (FTIR) was used to plot. p was calculated according Eq. (2), where pvap sat was the
determine the functional groups, chemical bonds and chemical saturated vapour pressure of pure water and equals to 3160 Pa at
nature of the sample. The FTIR analysis was performed using 25 ◦ C. RH was the relative humidity gradient between the cell and
Perkin-Elmer Co., USA model spectrum with samples dispersed in the surroundings.
the pellets of KBr. The dried active fraction of T. indica seed starch
samples (0.3–0.5 mg) were ground in about 80 mg of spectral-grade 2.8. Predictive modelling
KBr (sigma) and pressed into pellets under about 5–6 tonnes cm−2
pressure with the help of a hydraulic press. The spectral measure- RSM was done using Design Expert Software package (version
ments were carried out in an absorbance range of 400–4000 cm−1 7). Tamarind starch concentration, Cellulose concentration, Glyc-
with plain KBr pellet as the background reference. erol concentration and Acetic acid concentrations were taken as
fixed input factors. Mechanical parameters were taken as main
2.4. Film preparation method parameters for predictive modelling. Thus mechanical properties
such as tensile strength and elongation at break were taken as
Initial studies were conducted for concentration optimization responses of RSM. Central composite design experimental trail was
of film constituents. 12% (w/v) Starch extracted from the Tamarind used as input data for construction of Response Surfaces models
seed and 4% (w/v) cellulose powder was homogenized with 100 ml [19]. Different equation models were selected for predictive mod-
of deionized water using Homogenizer (REMI Elektrotechnik Ltd. elling of responses in accordance to fitness of experimental data.
Model RQT-127-A). 2.5% Glycerol (v/v) and 1.5% acetic acid 1.5% RSM–ANOVA parameters were calculated through following equa-
(v/v) were added to this homogenized mixture which was gela- tions.
tinized at 105 ◦ C for 15–20 min. The above mixture was made to Predicted residual sum of square:
spread over a Glass tray with the help of spacer. These glass trays n
were then incubated at 50 ◦ C for 6 h. After incubation the dried thin Press = (ei,−i )2 (3)
i=1
films were peeled off from the trays and stored at 64% RH, 20 ◦ C for
further use. where, ei,−i = yi − ȳi,−i ; yi is actual residual values; ȳi,−i is predicted
residual values.
2.5. Tensile property Correlation coefficient (R2 ):
n
(xi − x̄)(yi − ȳ)
The tensile strength of the CTR films was investigated using Uni-
2
R =  i=1
 (4)
i=1 i=1
versal Testing Machine (UTM, H10KS, Tinius Olsen, UK) according
n
(xi − x̄)2 n
(yi − ȳ)2
to the standard of ASTM D638. The samples had a dimension of
150 mm × 25 mm × 0.048 mm with a gauge length of 25 mm at a where xi – experimental values; yi – predicted values; x̄ – mean of
cross head speed of 10 mm/min. The tensile strength was expressed experimental values; ȳ – mean of predicted values.
in MPa. Adjusted R2 value:
n − 1
2
2.6. Oxygen transmission rate (OTR) Radj = (1 − R2 ), (5)
n−p

The OTR of the CTR film was calculated by OXTRAN Oxygen per- where n is number of experiments of runs; p is number of model
meability Tester (MOCON, Minneapolis, MN) at 23 ◦ C under the parameters.
condition of 0% RH at 1 atm according to the standard of ASTM Predicted R2 value:
 Press

D3985. The specimens were conditioned at ambient conditions. The 2
Rpred =1− , (6)
oxygen permeability was calculated using the following mathemat- SStotal − SSblock
ical formula.
where SStotal is total sum of squares; SSblock is block sum of squares;
OTR Press is predicted residual sum of squares.
Oxygen permeability =
Film thickness × O2 partial pressure Adequate precision:

max(ŷ) − min(ŷ)
2.7. Water vapour permeability test (WVP) Ad. Precision =  , (7)
V̄ (ŷ)
Water vapour permeability was performed according to ASTM

(1996) and Debeaufort, Martin-Polo, and Voilley (1993) [35]. Film where V̄ ŷ = p 2 /n; p is number of model parameters;  2 is resid-
samples were used to seal a hole on the top of a plastic cell. The ual mean square from ANOVA; n is number of experiments or runs.
plastic cell, containing distilled water, was loaded into a desiccator.
The desiccator was placed into a chamber at 25 ◦ C for 24 h. At least 3. Results and discussion
three weights of the cell were taken during the 24 h of storage.
Water vapour permeability was calculated with Eq. (1). 3.1. Characterization of tamarind seed starch

m · l 3.1.1. Thermo gravimetric analysis

WVP = (1)
A · t · p Thermal stability of tamarind seed starch was assessed by
where, m/t – weight of moisture loss per unit of time (g s−1 ), thermo gravimetric analysis (TGA) under inert atmosphere (Nitro-
A – film area exposed to moisture transfer (m2 ), l – film thickness gen). Three major weight loss phases have been observed as
(m), p – water vapour pressure difference between the two sides discussed by previous authors [15,27]. Evaporation and dehydra-
of the film (Pa). tion of water occurred in the first phase at a temperature interval
 RH  of 58–179 ◦ C as shown in Fig. 1. Total mass loss of 12% w/w was
p = pvap sat (2) observed during this interval and it was majorly due to loss of mois-
100 ture content of starch. The second and third phases correspond
388 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395
Fig. 1. Thermal analysis of tamarind seed starch powder (blue continuous line indicates DTA; green continuous line indicates TGA). (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
to the thermal decomposition occurring at a temperature inter-
val of 310–630 ◦ C. Major loss of mass (45%, w/w) was observed at
345 ◦ C, which has been confirmed from DTA curve. It can be noted
that the onset of decomposition temperature does not affect the
initial water content since water evaporation from the samples
occurred before reaching its decomposition temperature. Finally,
43% residue remained after the dehydration and decomposition
process.
TGA results of mango kernel (Fig. 2) demonstrated that the
weight loss of a material is the function of temperature. The samples
were analyzed with a heating rate of 10 ◦ C min− 1 under nitrogen
atmosphere. Fig. 2 shows the degradation of first phase occurred
at a temperature range of 0–220 ◦ C with 27% water loss. Second
phase took place between 270 ◦ C and 633 ◦ C with a major weight
loss of 58%. De-polymerization started after this temperature and
continued till 550 ◦ C resulting in a major weight loss. Next phase of
degradation occurs between 650 and 800 ◦ C with very less weight
loss. About 0.1199 mg of the sample remained out from the total
sample of 6.0520 mg. Remaining of 42% of the sample at the end of
the experiment proves its good thermal stability.
TGA results of Mesquite seed atmosphere are shown in Fig. 3.
The results proves that first phase occurs between 0 and 250 ◦ C
with 21% water loss; second phase took place between the tem-
perature range of 270 ◦ C and 550 ◦ C with major weight loss of 47%.
De-polymerization started after this temperature and continued till
the temperature reached 550 ◦ C. Next phase of degradation took
place from 550 to 800 ◦ C showing very minor weight loss of 3%.
The weight of the sample was 3.868 mg at the end of the experi-
ment out of the total sample of 13.3430 mg. Remaining of 30% of
the sample at the end of the experiment proves its good thermal
stability.
When we compare the above results of TGA, tamarind seed
starch had 43% remaining mass content when compared with
mesquite seed (30%) and mango kernel starch (42%). This implies
good thermal stability of tamarind seed starch. Phase 1 degrada-
tion may be due to water evaporation during the heating process,
Table 1
Proximate values of defatted tamarind seed starch.
Parameter Weight (%)
Starch 84.68%
Protein 0.97%
Ash 3.5%
Crude fibre 3.2%
Others 7.65%
while the second phase of degradation may be attributed to thermal
decomposition as evidenced in other studies [9,28].
3.1.2. Differential scanning calorimetry
Differential scanning calorimetry (DSC) was used to study the
thermal characteristics of the tamarind seed. DSC provided quan-
titative measurements of heat flow associated with gelatinization,
where the endothermic peaks are indicative of the melting point.
Endothermic peak observed at 107.6 ◦ C (Fig. 4) corresponds to the
occurrence of starch gelatinization. The gelatinization temperature
is characteristic of the starch type and depends on the glass transi-
tion of the amorphous fraction of the starch [4].
3.1.3. Proximate analysis
Proximate composition of Tamarind starch has been analyzed;
major initial components such as total starch, protein, ash, crude
fibre, and other impurities were 65%, 12.7%, 3.5%, 3.2%, and 15.69%,
respectively. Because of the complex nature of protein binding
with amylase and amylopectin, the separation of protein was diffi-
cult. Washing agents such as water and sodium hydroxide (0.1 N
NaOH) were used for reducing the protein content as much as
possible as previously reported [21]. Composition of the major
components after defatting the Tamarind starch is shown in
Table 1.
Native granular starches are a mixture of amylose, a linear
structure of ␣-1,4 linked glucose units and amylopectin, a highly
branched structure of short ␣-1,4 chains linked by ␣-1,6 bonds.
 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395 389

Fig. 2. Thermal analysis of Mango kernel starch powder (blue continuous line indicates DTA; black continuous line indicates TGA). (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)

Almost all native starches are semi crystalline with crystallinity of
about 20–45% [22]. Level of starch crystallinity can be determined
by the ratio of amylose/amylopectin present in it [13,20]. Starch
is classified into three groups based on amylose content, amylose-
rich starch (amylose content >30 wt.%), moderate amylose starch
(amylose content 10–30 wt.%), and waxy starch (amylose content
<10 wt.%). Defatted Tamarind seed starch has an amylose content
of 27.55 wt.% and amylopectin content of 72.45 wt.%. The compo-
sitions reported for other known starches were 38–40% (amylose)
and 60–62% (amylopectin) in the case of pehuen starch and 24–26%
(amylose), 74–76% (amylopectin) in the case of defatted cashew nut
shell starch [23].

Fig. 3. Thermal analysis of Mesquite seed starch powder (blue continuous line indicates DTA; black continuous line indicates TGA). (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
390 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395

Fig. 4. Differential scanning colorimetric analysis of tamarind seed starch.

3.1.4. Particle size distribution

Four different particle size fractions have been observed (Fig. 5).
About 70–80% of the granules were observed in size ranges from
90-130
60 to 90 ␮m and 90 to 130 ␮m. The remaining 20–30% were in the
Size range (μm)

range of 0–30 ␮m and 30–60 ␮m. Particle size ranging from 3 to

60-90
8 ␮m, 3 to 24 ␮m, 2 to 35 ␮m, 5 to 70 ␮m and 5 to 100 ␮m have been
reported for rice, maize, wheat and potato starches, respectively [8].
30-60

0-30
3.1.5. Scanning electron microscopy
Shape and size of defatted tamarind seed starch was examined 0 50 100 150 200 250
using SEM as shown in Fig. 6a and b. The starch granules were Number of parcles
appeared to have rough surface with irregular shape [24]. Multiple
layers observed in each granule correlates with the results observed Fig. 5. Particle size distribution of tamarind seed starch powder.

previously [21,25,26]. The diameter of the granules ranged between

4 ␮m and 130 ␮m (for 500 granules).

Fig. 6. (a, b) SEM pictures depicting starch granules which appear to have a rough surface with irregular shape powder.
 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395 391

Table 2
Design summary of fixed factors for central composite design experimental trails.

Factor Name Units Low actual (−1) Median actual (0) High actual (+1)

A Starch % (w/v) 11 12 13
B Cellulose % (w/v) 3 4 5
C Glycerol % (v/v) 2 2.5 3
D Acetic acid % (v/v) 1 1.5 2

polymorphs, which can be referred to as C-type. The peak at 2

values of 12◦ and 22.9◦ is characteristic of A-type pattern, whereas
at 5.6◦ is an indicative of B-type pattern [25,29–31] all found a
C-type pattern, in agreement with the results observed. Double
helix found in maize starch is a characteristic of A-type structure,
whereas tuberous plants (such as potato) contain B-type starches.
C-type structure is more rare and intermediate form, probably
due to the combination between A and B occurring in some plant
sources such as pea starch [28,31]. Six-fold double-helical structure
has been observed in Native A- and B-type crystal lattices. The dif-
ference between the A-type and B-type crystallinity is the packing
density of the double helices in the unit cell.

3.1.7. Fourier transform infrared spectroscopy

Fig. 7. X-ray diffraction of tamarind seed starch powder.
3.1.6. X-ray diffraction studies
The molecular organization of granules was investigated using
X-ray diffraction. Native starches were biosynthetically assembled
as semi crystalline granules. Depending upon on the nature of
starch source, crystalline structure may be labelled and classified
as A-, B-. As shown in Fig. 7, diffraction peaks obtained for 2 values
at 5.6◦ , 12.0◦ , and 22.9◦ evidenced a mixture of the A- and B-type
The IR showed broad intense peak at 3340 cm−1 (Fig. 8) which
is characteristic for hydroxyl and amine group, absorption at
2933 cm−1 stretch is characteristic of CH2 groups and a peak
at 1147–1380 cm−1 suggests the presence of carboxyl group.
Protein-related bonds were observed at 1650 cm−1 . The absorp-
tion peaks at 1064 and 1250 cm−1 indicated the C O stretching
in ether or alcohol groups. Our results corresponds to the pre-
vious reports that the four bands appearing in 3409.14 cm−1 ,
2932.47 cm−1 , 1725.84 cm−1 and 1379.63 cm−1 corresponds to the
stretching vibrations of hydroxyl, C H, carboxyl, and C H bending
of aliphatic CH2 and an infrared band at 1643.05 cm−1 (Amide I)
and 1554.37 cm−1 respectively [3].

Table 3
Central composite design experimental trail for RSM quadratic predictive model.

Run B:Starch C:Cellulose A:Glycerol D:Acetic acid Tensile strength Elongation at break
% (w/v) % (w/v) % (v/v) % (v/v) MPa %

1 12 3 2.5 1 2.98 394.32

2 11 5 2 1 5.08 365.25
3 11 5 3 2 7.68 494.41
4 13 4 2 1.5 2.05 265.36
5 13 4 3 2 1.95 321.24
6 12 3 2 1.5 3.29 312.24
7 13 4 2.5 2 1.96 345.85
8 11 3 3 1 5.36 536.16
9 13 5 2 1 1.68 150.74
10 13 5 3 2 2.98 312.21
11 11 4 2 1 4.38 272.52
12 12 2 2.5 1.5 4.58 306.62
13 12 4 3 1 4.48 452.27
14 11 5 2 2 3.28 364.98
15 11 3 3 2 5.15 534.42
16 11 3 3 1 4.25 521.46
17 12 4 3 1 4.48 452.34
18 11 5 2 2 4.39 421.62
19 13 3 2 1.5 1.81 251.42
20 13 5 3 2 2.68 314.38
21 11 4 2.5 1 4.70 513.54
22 12 4 2.5 1.5 4.05 354.47
23 13 4 2.5 1 1.95 271.84
24 13 3 3 2 2.97 364.86
25 11 4 2.5 2 3.36 373.38
26 12 3 2.5 1 3.05 321.71
27 11 4 3 1.5 7.16 543.18
28 12 5 2.5 1 4.36 305.34
29 13 3 2 2 1.98 304.75
30 13 5 2.5 2 1.34 314.61
392 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395

Table 4
Analysis of variance for response surface quadratic models of mechanical properties.

Parameters Sum of squares Residual sum Mean sum of Residual mean F=MSS/RMS Probability of
(df) of squares (df) square (MSS) square (RMS) F < Fcritical

Tensile strength 66.98 (14) 3.87 (15) 4.78 0.26 18.55 <0.0001
Elongation at break 256,500.00 (10) 27,083.47 (19) 25,645.27 1425.45 17.99 <0.0001

60 Table 5
Coefficients of polynomial equation models for prediction of mechanical properties.

50 Equation terms (coefficients in ˇ) Coefficients of polynomial equation models

Tensile strength Elongation at break

40
Constant (ˇ0 ) −62.63 −1807.71
A (ˇ1 ) 9.26 96.76
2933
% Transmittance

30 B (ˇ2 ) 4.47 318.43

C (ˇ3 ) 3.27 1101.07
D (ˇ4 ) 4.74 −264.70
20 AB (ˇ5 ) −0.61 −22.05
AC (ˇ6 ) −0.79 −59.65
1620 AD (ˇ7 ) 0.29 45.82
10 3340 BC (ˇ8 ) 0.48 −21.37
BD (ˇ9 ) −0.42 2.01
0 CD (ˇ10 ) 0.88 −106.06
A2 (ˇ11 ) −0.28 –
B2 (ˇ12 ) 0.36 –
-10 1127
C2 (ˇ13 ) 0.99 –
D2 (ˇ14 ) −3.06 –
4000 3500 3000 2500 2000 1500 1000 500
-1
Wavenumbers (cm )
with coded coefficients. ANOVA for polynomial models parameters
Fig. 8. Fourier transform infra-red spectral bands of tamarind seed starch. were calculated and shown in Table 4. All the constructed polyno-
mial models were found to be fit with probability (F values < Fcritical
values) less than 0.0001 for acceptance level 0.05. Degree of polyno-
3.2. Mechanical properties of tamarind starch films mial model equations was selected according to significance of its
coefficients. Thus, Quadratic (2◦ ) polynomial equation model was
Film formation was found be improper when cast with only 12% selected for predictions of Tensile strength values. 2FI polynomial
tamarind seed starch (w/v) mixed with 2.5% glycerol (v/v) as a plas- equation model was selected for predictions of elongation at break
ticizer and 1.5% acetic acid (v/v) as the ion exchange polymerizer values. Actual coefficients of selected polynomial equation mod-
[32]. Good interfacial adhesion between starch and cellulose was els are tabulated in Table 5. These coefficients can be substituted
expected because both contained hydroxyl groups which can form in following polynomial quadratic equation model to get separate
hydrogen bonds between interfaces [33]. Thus the mixture of 12% equations for predictions of mechanical parameters.
tamarind seed starch (w/v) mixed with 2.5% glycerol (v/v) and 1.5%
acetic acid (v/v) was reinforced and homogenized with 4% cellulose Y = ˇ0 + ˇ1 A + ˇ2 B + ˇ3 C + ˇ4 D + ˇ5 AB + ˇ6 AC + ˇ7 AD
(w/v) for preparation of films and also to improve the mechanical
+ ˇ8 BC + ˇ9 BD + ˇ10 CD + ˇ11 A2 + ˇ12 B2 + ˇ13 C 2 + ˇ14 D2
properties of the film. These cellulose reinforced tamarind starch
films (CRT) were analyzed for its mechanical properties (tensile
strength and elongation at break). The tensile strength and elon- Equations derived for mechanical parameters of tamarind starch
gation at break of cellulose reinforced tamarind starch films were films were used to predict corresponding parameter values (for
found to be 4.05 MPa and 354.47%, respectively. experimental trials). Model parameters such as R2 , Adjusted R2 ,
Predicted R2 and adequate precisions denote the fitness of the
3.2.1. Predictive model for mechanical properties of tamarind constructed model equation to experimental data. Correlation coef-
starch films ficient (R2 ) values were calculated by comparing predicted and
Response surface polynomial predictive models were used for experimental values of tensile strength and elongation at break
modelling of mechanical properties of Tamarind starch films. parameter. It is the measure of amount of variation around the
mean explained by the model. Radj 2 (Adjusted R2 value) is the mea-
Tamarind starch concentration, cellulose concentration, glycerol
concentration and acetic acid concentrations were taken as fixed sure of the amount of variation around the mean explained by the
input factors. These factors were taken in different levels for con- model, adjusted for the number of terms in the model. The adjusted
struction of central composite design experimental trials. Fixed R2 value decreased as the number of terms in the model. Press
levels of input factors are given in design summary of central com- (Predicted Residual Error Sum of Squares) is a measure of how the
posite experimental trials (Table 2). model fits each point in the design. Press value helps in calculation
2
of predicted R2 values. Rpred (Predicted R2 value) is the measure of
Totally, 30 experimental trials of central composite design were
conducted for different levels of input factors as shown in Table 3. amount of variation in new data explained by the model. Deviation
Tamarind starch films were prepared separately for the differ- between adjusted and predicted R2 values should be within 0.20,
ent concentration of film constituents. Mechanical parameters of otherwise there may be a problem with data or model constructed.
tamarind starch films such as tensile strength and elongation at Predicted values are tabulated with corresponding model
break were also analyzed separately for each film. parameters of fitness in Table 6. R2 value of Tensile strength
These experimental trial analysis data were taken as input for (0.9454) implied that constructed (2◦ ) polynomial equation mod-
construction of predictive polynomial models. Actual coefficients els were fit with deviation (less than 0.20) between adjusted R2
of each polynomial model equations were analyzed for its variance and predicted R2 values. R2 value of elongation at break (0.9045)
 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395 393

Design-Expert® Software

Tensile strength
7.68

1.34
6.8
X1 = A: Starch
X2 = B: Cellulose
5.65
Actual Factors

Tensile strength
C: Glycerol = 2.50
D: Acetic acid = 1.50 4.5

3.35

2.2

5.00
13.00
4.50 12.50
4.00 12.00
B: Cellulose 3.50 11.50 A: Starch
3.00 11.00

Fig. 9. RSM 3D response of tensile strength for change in concentration of cellulose.

Design-Expert® Software

Tensile strength
7.68

1.34
7.2
X1 = A: Starch
X2 = C: Glycerol
5.875
Actual Factors
Tensile strength

B: Cellulose = 4.00
D: Acetic acid = 1.50 4.55

3.225

1.9

3.00 13.00
2.75 12.50
2.50 12.00
2.25 11.50
C: Glycerol A: Starch
2.00 11.00

Fig. 10. RSM 3D response of tensile strength for change in concentration of glycerol.

implied the fitness of the corresponding (2FI) polynomial model
with fewer deviations between adjusted and predicted R2 values.
Adequate precision values represent the signal to noise ratio in the
given model. For adequate model discrimination this value must be
greater than 4. Ad. Precision values of models for Tensile strength
and elongation at break were found to be 16.80 and 16.55. This
implies adequate discrimination of data and minimal error in con-
structed polynomial models.
3D responses of mechanical properties of tamarind starch films
were calculated and they were depicted in Figs. 9–12. Fig. 9 shows
the change in tensile strength properties of tamarind starch films
with a change in concentration of cellulose. Tensile strength of films
was found to be increasing with every increase in the concentra-
tion of cellulose. This may be due to increase in the total solid
concentration of film mixture which in turn may affect the ten-
sile strength. Deviations of tensile strength were few for change in
concentration of tamarind seed starch. Fig. 10, depicts the change in
tensile strength properties of tamarind starch films with a change
in the concentration of glycerol. Tensile strength of films was found
to be increasing for every increase in the concentration of glycerol
with corresponding tamarind starch concentration.
Fig. 11 depicts change in the elongation at break of films for
change in concentration of cellulose. Elongation at break was found
to be slightly increasing with increasing concentration of cellulose.
When cellulose concentrations were low, increase in the concen-
tration of starch showed lesser influence in Elongation at break.
But when cellulose concentrations were high, increase in the con-
centration of starch shows higher influence on Elongation at break
values. This increase in Elongation at break at higher concentration
of cellulose may be due to high polymerizing capacity of cellulose.
Fig. 12, shows changes in the elongation at break with a change
in concentration of glycerol. Elongation at break was found to be
394 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395

Design-Expert® Software

Elongation at break
543.18

150.74
460
X1 = A: Starch
X2 = B: Cellulose
410
Actual Factors

Elongation at break
C: Glycerol = 2.50
D: Acetic acid = 1.50 360

310

260

5.00 13.00
4.50 12.50
4.00 12.00
3.50 11.50
B: Cellulose A: Starch
3.00 11.00

Fig. 11. RSM 3D response of elongation at break for change in concentration of cellulose.

Design-Expert® Software

Elongation at break
543.18

150.74
550
X1 = A: Starch
X2 = C: Glycerol
472.5
Actual Factors
Elongation at break

B: Cellulose = 4.00
D: Acetic acid = 1.50 395

317.5

240

3.00 13.00
2.75 12.50
2.50 12.00
2.25 11.50
C: Glycerol A: Starch
2.00 11.00

Fig. 12. RSM 3D response of elongation at break for change in concentration of glycerol.

increasing with every increase in the concentration of glycerol. This
may be due to the plasticizing property of glycerol. Change in the
concentration of starch showed lesser influence on elongation at
break.
All the depictions of predictive models were purely based on
statistical analysis of experimental values. Thus, these depictions
may not be reflected on experimental analysis of films.
3.2.2. Mechanical properties of optimized CRT films
Optimized concentration of the film constituents: 11% tamarind
seed starch (w/v), 5% cellulose (w/v), 3% glycerol (v/v) and 2%
acetic acid (v/v) were homogenized and this homogenized mixture
was used for formation of CRT films as shown in Fig. 9. This film
was analyzed for its mechanical properties (tensile strength and
elongation at break). The tensile strength, elongation at break and
Young’s Modulus of optimized CRT films were found to be 7.68 MPa,
494.41% and 1.55 MPa respectively. The improvement in tensile
strength may be due to increased concentration of cellulose pow-
der. Oxygen transfer rate (OTR) of optimized CRT film was found
to be 1034.45 ± 2.57, 682.76 ± 3.41 and 395.23 ± 3.75 cc m−1 day−1
for film thickness of 0.25 mm, 0.38 mm and 0.47 mm, respectively.
Water Vapour Permeability of optimized CRT film was found to be
0.91, 0.73 and 0.65 g s−1 m−1 Pa−1 for film thickness of 0.25 mm,
0.38 mm and 0.47 mm, respectively. Increase in thickness of CRT
film found to reduce its OTR and WVP properties. Low WVP of opti-
mized CRT film may be due to rich amylose content of Tamarind
seed starch. Increase in amylose content of starch tends to reduce
WVP of starch based films. The reason for such behaviour may be
that during recrystallization, amylose forms a B-type crystalline
structure whereas amylopectin has an amorphous structure. Diffu-
sion of moisture is easier in amorphous systems than in crystalline
ones [34,35].
 K. Sudharsan et al. / International Journal of Biological Macromolecules 83 (2016) 385–395 395

Table 6 Acknowledgment
RSM predicted values for mechanical properties of tamarind seed starch films.

Runs Different mechanical parameters The authors sincerely acknowledge their greatest grat-
itude to Department of Science and Technology (DST)
Tensile strength Elongation at break
MPa % DST/TSG/NTS/2011/186, India for the financial support.
1 2.88 367.63
2 3.86 314.28
3 2.48 339.33 References
4 1.56 251.45
5 4.47 443.03 [1] K.G. Satyanarayana, G.G.C. Arizaga, F. Wypych, Prog. Polym. Sci. 34 (9) (2009),
6 4.57 348.62 982-102.
7 4.34 353.37 [2] A.A. Shah, F. Hasan, A. Hameed, S. Ahmed, Int. Biodeterior. Biodegrad. 60
(2007) 109–115.
8 1.9 247.91
[3] M. Mitrus, A. Wojtowicz, L. Moscicki, Thermoplastic Starch, Wiley-VCH Verlag
9 3.05 343.77
GmbH & Co. KGaA, Weinheim, 2009, ISBN 978-3-527-32528-3.
10 4.05 423.19 [4] D.S. Rosa, D.R. Lopes, M.R. Calil, J. Mater. Sci. 42 (2007) 2323–2328.
11 4.47 443.03 [5] J.L. Guimaraes, F. Wypych, C.K. Saul, L.P. Ramos, K.G. Satyanarayana,
12 2.8 311.03 Carbohydr. Polym. 80 (2010) 130–138.
13 4.78 440.94 [6] O. Martin, L. Avérous, G. Della Valle, Carbohydr. Polym. 53 (2003) 169–182.
14 3.76 386.73 [7] R. Zullo, S. Iannace, Carbohydr. Polym. 77 (2009) 376–383.
15 6.82 532.21 [8] R.L. Shogren, G.F. Fanta, W.M. Doane, Starch/Stärke 45 (1993) 276–280.
16 1.5 307.74 [9] A. Oya, Y. Kurokawa, H. Yasuda, J. Mater. Sci. 35 (2010) 1045–1050.
17 4 360.23 [10] A.L. Chaudhary, P.J. Torley, P.J. Halley, N. McCaffery, D.S. Chaudhary,
18 4.94 553.81 Carbohydr. Polym. 78 (2009) 917–925.
19 5.35 349.44 [11] A. Taguet, M.A. Huneault, B.D. Favis, Polymer 50 (24) (2009) 5733–5743.
20 7.45 512.62 [12] M.R. Almeida, R.S. Alves, L. Nascimbem, R. Stephani, R.J. Poppi, L. De Oliveira,
Anal. Bioanal. Chem. 397 (2010) 2693–2701.
21 3.31 275.79
[13] L. Yu, G. Christie, J. Mater. Sci. 40 (2005) 111–116.
22 3.05 343.77
[14] K. El-Siddig, H.P.M. Gunasena, B.A. Prasad, D.K.N.P. Pushpakumara, K.V.R.
23 3.76 386.73
Ramana, P. Vijayanand, J.T. Williams, Southampton Centre for Underutilized
24 1.42 325.35 Crops, Southampton, UK, 2006, pp. 188.
25 4.94 553.81 [15] C. Fabian, A. Ayucitra, S. Ismadji, Y.H. Ju, J. Taiwan Inst. Chem. E 42 (2011)
26 2.8 311.03 86–91.
27 2.57 255.84 [16] AOAC Official Method 996.11, Starch (total) in cereal products:
28 1.69 175.51 amyloglucosidase-amylase method, in: W. Horwitz (Ed.), Official Methods of
29 5.07 481.02 Analysis of AOAC International, AOAC International, Gaithersburg, MD, USA,
30 1.73 318.3 1996.
R2 0.9454 0.9045 [17] AOCS Official Method Ba 4a-38, Nitrogen-ammonia-protein modified Kjedahl
Press 18.35 82,698.98 method, in: D. Firestone (Ed.), Official Methods and Recommended Practices
Radj
2
0.8944 0.8542 of the AOCS, American Oil Chemists’ Society, Champaign, IL, USA,
1997.
Rpred
2
0.7410 0.7083
[18] AOCS Official Method Ba 5a-49, Ash, in: D. Firestone (Ed.), Official Methods
Ad. Precision 16.806 16.55 and Recommended Practices of the AOCS, American Oil Chemists’ Society,
Champaign, IL, USA, 1997.
[19] A. Rajendran, A. Palanisamy, V. Thangavelu, Chin. J. Biotechnol. 24 (3) (2008)
4. Conclusion 436–444.
[20] M. Yuliana, L.H. Huynh, Q.P. Ho, C.T. Truong, Y.H. Ju, Carbohydr. Polym. 87
In this study, Tamarind seed starch has been defatted, isolated, (2012) 2576–2581.
[21] L. Copeland, J. Blazek, H. Salman, M.C. Tang, Food Hydrocoll. 23 (2009)
and purified and about 84.68% of starch has been recovered. The 1527–1534.
main constituents of tamarind starch was observed to be amy- [22] J. Castano, R. Bouza, S. Rodriguez-Llamazares, C. Carrasco, R.V.B. Vinicius,
lose (27.55 wt.%) and amylopectin (72.45 wt.%). Average particle Carbohydr. Polym. 88 (2012) 299–307.
[23] B. Panda, P. Jain, L. Sharma, N. Mallick, Bioresour. Technol. 97 (2006)
size distribution of tamarind starch granules were in the range
1296–1301.
of 60–90 ␮m. Tamarind starch was found to have C-type crys- [24] Y.J. Wang, W.J. Liu, Z.H. Sun, J. Mater. Sci. Lett. 22 (2003) 57–59.
tallinity. Thermo gravimetric analysis results confirms that the [25] L. Xia, G. Wenyuan, J. Qianqian, H. Luqi, L. Changxiao, Starch-Starke 63 (2011)
24–32.
tamarind seed starch (43% residue) had higher thermal stabil-
[26] K.N. Waliszewski, M.A. Aparicio, L.A. Bello, J.A. Monroy, Carbohydr. Polym. 52
ity than Mesquite seed starch (30% residue) and Mango kernel (2003) 237–242.
starch (42% residue). Thus these unwanted tamarind seeds can [27] X. Ma, J. Yu, Starch-Starke 56 (2004) 545–555.
be effectively processed to produce starch-based biopolymers. [28] F.M. Deng, T.H. Mu, M. Zhang, O.K. Abegunde, Starch-Starke 65 (2013)
162–171.
Cellulose can be used as a reinforcing material with tamarind [29] L.A. Bello-Perez, A. De Francisco, E. Agama-Acevedo, F. Gutierrez-Meraz, F.J.L.
starch to produce CRT biopolymer films. Predictive modelling Garcia-Suarez, Food Sci. Technol. Int. 11 (2005) 367–372.
was used for optimization of CRT film constituents and the opti- [30] J. Jane, Y.Y. Chen, L.F. Lee, A.E. McPherson, Cereal Chem. 76 (1999)
629–637.
mized film was found to be having good mechanical properties [31] P. Myllärinen, A. Buleon, R. Lahtinen, P. Forssell, Carbohydr. Polym. 48 (2002)
(T.S – 7.68 M pa and E.B – 494.41%). OTR of optimized CRT film 41–48.
was found to be 395.23 ± 3.75 cc m−1 day−1 for film thickness of [32] E.S. Dragan, New Trends in Ionic (Co) Polymers and Hybrids, Nova Publishers,
2007.
0.47 mm. WVP of optimized CRT film was found to be 0.91, 0.73 [33] M.B. Agustin, B. Ahmmad, E.R.P. De Leon, J.L. Buenaobra, J.R. Salazar, F. Hirose,
and 0.65 g s−1 m−1 Pa−1 for film thickness of 0.25 mm, 0.38 mm and Polym. Compos. 34 (2013) 1325–1332.
0.47 mm, respectively. Results observed in this study can be helpful [34] T.D. Phan, F. Debeaufort, D. Luu, A. Voilley, J. Agric. Food Chem. 4 (2005)
973–981.
in blending tamarind seed starch with other biopolymer materials
[35] S. Santacruz, C. Rivadeneira, M. Castro, Food Hydrocoll. 49 (2015) 89–94.
such as Polylactic acid (PLA), Polycaprolactone (PCL), and also with [36] A. Mishra, A.V. Malhotra, J. Mater. Chem. 19 (2009) 8528–8536.
other cellulose nanocrystals to produce potential biodegradable [37] A.B. Dias, C.M.O. Muller, F. Larotonda, J.B. Laurindo, LWT Food Sci. Technol. 44
(2011) 535–542.
polymer which may be more economical and eco-friendly.