Synthesis and Properties of Cardanol-Based Epoxidized

Novolac Resins Modified with Carboxyl-Terminated

Butadiene–Acrylonitrile Copolymer

Ranjana Yadav, Deepak Srivastava

Department of Plastic Technology, H. B. Technological Institute, Kanpur 208 002 (Uttar Pradesh), India

Received 4 March 2008; accepted 9 August 2008

DOI 10.1002/app.29376
Published online 24 June 2009 in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: Cardanol-based, novolac-type phenolic res- also confirmed by proton nuclear magnetic resonance
ins were synthesized with a cardanol-to-formaldehyde and matrix-assisted laser desorption/ionization time-of-
molar ratio of 1 : 0.7 with different dicarboxylic acid cat- flight mass spectroscopy analysis. The molecular weights
alysts, including oxalic and succinic acids. These novolac of the prepared novolacs and their epoxidized novolac
resins were epoxidized with a molar excess of epichloro- resins were determined by gel permeation chromatogra-
hydrin at 120 C in a basic medium. The epoxidized phy analysis. The blend samples, in both cases, with 15
novolac resins were separately blended with different wt % CTBN concentrations showed the minimum cure
weight ratios of carboxyl-terminated butadiene–acryloni- times. These blend samples were also the most thermally
trile copolymer (CTBN) ranging between 0 and 20 wt % stable systems. The blend morphology, studied by scan-
with an interval of 5 wt %. All of the blends were cured ning electron microscopy analysis, was, finally, corre-
at 120 C with a stoichiometric amount of polyamine. The lated with the structural and property changes in the
formation of various products during the synthesis of blends. VC 2009 Wiley Periodicals, Inc. J Appl Polym Sci 114:

the cardanol-based novolac resin and epoxidized novolac 1670–1681, 2009

resin and the blending of the epoxidized novolac resin
with CTBN was studied by Fourier transform infrared Key words: blends; differential scanning calorimetry
spectroscopy analysis. Furthermore, the products were (DSC); FT-IR; gel permeation chromatography (GPC)

INTRODUCTION (ECH) to duplicate the performance of such pheno-

lic-type novolacs.5 Having several outstanding char-
Cashew nut shell liquid (CNSL), an agricultural
acteristics, epoxy resins show low impact resistance
byproduct abundantly available in the country, is
in their cured state6–11 which limits the applications
one of the few major and economic sources of natu-
of epoxy resins. To alleviate this deficiency, epoxy
rally occurring phenols and can be regarded as a
resins are modified by the incorporation of reactive
versatile and valuable raw material for polymer pro-
liquid rubber without significant losses in other
duction. Cardanol, a natural alkyl phenol from
properties, particularly, the mechanical proper-
CNSL and a potential natural source for biomono-
ties.12,13 In this way, the carboxyl-terminated copoly-
mers, cannot even today be said to have found its
mer of butadiene and acrylonitrile (CTBN) has been
niche in terms of an appropriate industrial applica-
used by various workers6,7,13 with diglycidyl ether
tions. By far, the greatest amount of work on poly-
of bisphenol A (DGEBA) epoxy resin and epoxidized
meric materials derived from CNSL or cardanol
phenolic novolac resins. However, CTBN has hardly
have been done on their use in the manufacture or
been used with cardanol-based epoxy resins. There-
modification of phenolic resins.1–4 The phenolic na-
fore, we tried to produce modified epoxy matrices
ture of the material makes it possible to react it
based on cardanol by physical blending with CTBN
under a variety of conditions to form both base-cata-
and studied the effect of CTBN addition on the ther-
lyzed resoles and acid-catalyzed novolacs. Cardanol-
mal and morphological changes in the blends.
based, novolac-type phenolic resins may further be
modified by epoxidation with epichlorohydrin
EXPERIMENTAL
Correspondence to: D. Srivastava (dsri92@gmail.com).
Contract grant sponsor: Council of Science and Materials
Technology, UP, Lucknow.
Cardanol (M/s Satya Cashew Pvt., Ltd., Chennai,
Journal of Applied Polymer Science, Vol. 114, 1670–1681 (2009) India), formaldehyde (40% solution), oxalic and suc-
C 2009 Wiley Periodicals, Inc.
V cinic acids, sodium hydroxide, ECH (all from M/s
CARDANOL-BASED NOVALAC RESIN 1671

TABLE I with n-hexane and charged to silica gel column

Physical Characteristic of Cardanol chromatographic purification was adopted mainly to
Calculated remove unreacted components, impurities, and so
Distilled From the From the on from the methylolated cardanol. Purification was
Property cardanol open market literature effected with the eluent mixture of ethyl acetate and
benzene (60 : 40). The purified resin was analyzed
Viscosity (P) 41.5 40.2 32
Specific gravity (g/m3) 0.900 0.854 0.873
by infrared (IR) spectroscopic, nuclear magnetic res-
Iodine value 285.6 280.4 279.8 onance (1H-NMR) spectroscopic, mass spectroscopic,
Moisture content 3.2 2.8 2.6 and gel permeation chromatography (GPC) analysis.
at 100 C (wt %) Two novolac-type phenolic resins, CF71 and CF72,
with catalyst oxalic and succinic acids, respectively,
were prepared.

Thomas Baker Chemicals, Ltd., Mumbai, India),

polyamine (M/s Ciba Specialty Chemicals, Ltd., Epoxidation of the cardanol-based, novolac-type
phenolic resins and preparation of the blends of
Mumbai, India) with amine values of 1240–1400 mg the epoxidized novolac resin and CTBN
of KOH/g, and CTBN (Hycar 1300  13) were used
during the investigation. CTBN was kindly supplied The previously prepared cardanol-based, novolac-
by M/s Emerald Performance Materials, LLC (Hong type phenolic resins were epoxidized by a method
Kong) with a number-average molecular weight similar to a method given in literature.16 Approxi-
(Mn) of 3500 and acrylonitrile and carboxyl contents mately, 1.0 mol of the novolac resin was placed in
of 27 and 32 wt %, respectively. a 500-mL, three-necked, round-bottom flask, and
6.0 mol of ECH was added to it with stirring. Then,
a 40% sodium hydroxide solution was added drop-
Analysis of cardanol wise to this mixture for a period of 5 h at 120 C.
Cardanol, obtained from the distillation of commer- The reaction mixture was then subjected to distilla-
cial CNSL under reduced pressure (1 mmHg) and tion in vacuo to remove unreacted ECH. The result-
collected at 206–208 C, was subjected to extensive ing viscous product was stored for further analysis.
analysis, namely, for iodine value, viscosity, and Two epoxidized novolac resins, ECF71 and ECF72,
specific gravity (see Table I), as per the procedure with novolac resins CF71 and CF72, respectively,
mentioned in IS standard 840-1964. were prepared.
The epoxy resin was mixed physically with vari-
ous concentration of CTBN ranging between 0 and
Synthesis of the cardanol-based, novolac-type 20 wt % with an interval of 5 wt %. All of the sam-
phenolic resins ples were designated according to Table II.
Novolac resins with a cardanol-to-formaldehyde
molar ratio of 1 : 0.7 were prepared with dicarbox-
Characterization techniques of the
ylic acids, namely, oxalic and succinic acids, as cata- prepared samples
lysts by a method similar to that adopted by Knop
and Schieb14 for phenol-based novolac. The catalyst Fourier transform infrared (FTIR) spectra of the pre-
(1% based on cardanol) was first dissolved in metha- pared samples were recorded on a PerkinElmer
nol at 60 C. Half of the catalyst solution was added
to cardanol (about 30 g), which was charged in a TABLE II
three-necked, round-bottom flask fitted with a Lei- Sample Designations
big’s condenser and mechanical stirrer. The remain- Sample Sample
ing half of the methanolic solution of the catalyst number Epoxy (wt %) CTBN (wt %) code
was added to formaldehyde (40%), and this was Cardanol-based epoxidized novolac resin (prepared from
added to the cardanol dropwise within 1 h once the novolac resin CF71)
temperature of the reaction kettle was maintained at 1 95 5 BECF711
120 C. The initial pH of the reaction mixture was 2 90 10 BECF712
3 85 15 BECF713
6.0, which decreased to a value of 2.0 after 5 h of
4 80 20 BECF714
reaction at 120 C. Free-formaldehyde and free-phe- Cardanol-based epoxidized novolac resin (prepared from
nol contents were checked after every 45 min to novolac resin CF72)
determine the completion of the methylolation reac- 5 95 5 BECF721
tion.15 The reaction product was cooled and dried in 6 90 10 BECF722
7 85 15 BECF723
vacuo at 60 C overnight before purification by col-
8 80 20 BECF724
umn chromatography. A resin solution prepared

Journal of Applied Polymer Science DOI 10.1002/app

1672 YADAV AND SRIVASTAVA

Figure 1 FTIR spectra of (a) cardanol, (b) novolac resin CF71, (c) novolac resin CF72, (d) epoxidized novolac resin ECF72,
(e) uncured blend sample BECF723, and (f) cured blend sample BECF723.

(USA) (model 843) IR spectrophotometer with KBr and stationary phase, and injected into the
pellets in the wavelength range 500–4000 cm 1. instrument.
1
H-NMR spectra of the cardanol-based novolac The mass spectra of the prepared samples were
and epoxidized novolac resins were recorded on a recorded on Micromass TofSpec2e matrix-assisted
Bruker (Germany) DRX-300 NMR spectrophotometer laser desorption/ionization time-of-flight (MALDI-
in the temperature range 90 to 80 C. TOF) (Waters Micromass UK Ltd., England) and fast
GPC was used (the instrument procured from E. atom bombardment (FAB) mass spectrophotometer.
Merck (Germany) consisted of a pump, an L-7350 The elemental composition of the prepared novolac
column oven, and an L-7490 refractive-index detec- and epoxy resins were determined with a Heraeus
tor) to determine the number-average molecular Vario EL III Carlo Erba 1108 elementals analyzer
weight of the synthesized cardanol-based novolac (Haraeus GmbH, Germany).
resins. A small quantity of the resin was dissolved The cure temperature of the prepared samples
in tetrahydrofuran; which acted both as a mobile were observed by the placement of a very small

Journal of Applied Polymer Science DOI 10.1002/app

CARDANOL-BASED NOVALAC RESIN 1673

1
Figure 2 H-NMR spectra of (a) cardanol, (b) novolac resin CF71, (c) novolac resin CF72, and (d) epoxidized novolac resin
ECF72.

quantity of the blend samples into a shallow alumi- matrix and the interaction of these domains with the
num pan sealed by an aluminum cover of a differen- epoxy matrix specimen surface were observed by
tial scanning calorimeter (TA Instruments, USA, scanning electron microscopy (SEM). For this, the
model DSC Q20). This was placed in the sample cell fractured samples were coated with a thin layer of
of the instrument. The starting temperature, pro- gold–palladium alloy by sputtering to provide con-
grammed rate, and final temperature were taken at ductive surfaces.
a heating rate of 10 C/min. Dynamic scans were
obtained, which were used to determine the cure
temperature. RESULTS AND DISCUSSION
The curing of the blend samples with a stoichio-
metric amount of polyamine was performed in an Synthesis of the cardanol–formaldehyde
novolac-type phenolic resins
air oven (M/s Indian Equipment Corp., Mumbai,
India) at 120 C. The methylolation of cardanol was carried out with
The thermal stability of the blend samples was formaldehyde in the presence of dicarboxylic acids,
determined by a comparison of the onset degradation namely, oxalic and succinic acids. The initial pH of
temperature (up to 5% weight loss) of the cured sam- the reaction mixture in both cases was found to be
ples with a thermogravimetric analyzer (TA Instru- 6.0. This indicated that these acid catalysts imparted
ments, model Q50 TGA) at a heating rate of 10 C/min alkaline pH under these formulations of reaction
in a nitrogen atmosphere from 50 to 650 C. mixtures as per the norms of pH of phenolic res-
The morphological changes due to the addition of ins.17 The formylation reactions were carried out
CTBN into the epoxy matrix were studied with a with a cardanol-to-formaldehyde molar ratio of 0.7.
Jeol scanning electron microscope (England) (model Therefore, under these experimental conditions, the
JSM 5800). The rubber domains distributed in the complete formylation might have been expected to

Journal of Applied Polymer Science DOI 10.1002/app

1674 YADAV AND SRIVASTAVA

Figure 3 (a) FAB mass spectrum of cardanol, (b) MALDI-TOF mass spectrum of novolac resin CF71, and (c) MALDI-TOF
mass spectrum of novolac resin CF72.

yield resins with high ortho–ortho linkages for phe- condensation of methylolated cardanol but also the
nolic novolac resins. The completion of the methylo- degree of ortho and para substitution.22 A shift of a
lation reaction was checked by periodic withdrawal peak from 1075 to 1090 (CF71) and 1104 cm 1 (CF72)
of the reaction mixture to analyze formaldehyde and the appearance of a peak near 1708 cm 1 [Fig.
with the hydroxylamine hydrochloride method.18 1(b,c)] were observed in methylolated cardanol
The final pH of the reaction mixture was found to because of the C¼ ¼O stretching from CH2OH. Also,
be 2.0. The decrease in pH in the methylolated car- the intensity of the peaks at 1594 cm 1 (C¼ ¼C,
danol might have been due to the formation of stretching), 3010 cm 1 (CAH stretching of alkene),
monohydroxyl-substituted cardanol. Sperling19 and 778 cm 1 (CAH out-of-plane deformation)
found that formylated phenol favored hydrogen remained almost unaffected, which indicated that
bonding with ortho-methylol and hydroxyl groups, the polymerization took place through the substitu-
which released hydrogen ions and led to an increase tion of CH2OH and not through the double bonds in
in the acidity of the reaction mixture. the side chain. The bands at 3361 and 3342 cm 1 for
samples CF71 and CF72, respectively, might have
been due to the presence of hydroxyl groups in the
Spectral analysis of the cardanol–formaldehyde methylolated cardanol. The most notable differences
novolac resins CF71 and CF72
were the changes in the intensities of the absorption
The FTIR [Fig. 1(a)], 1H-NMR [Fig. 2(a)], and mass
[Fig. 3(a)] spectra of cardanol clearly indicated that
it was a monoene meta-substituted phenol. There-
fore, the empirical formula was taken as
C21H34O,20,21 and the structure of cardanol is pro-
posed as shown in Scheme 1.

FTIR spectroscopic analysis of CF71 and CF72

The IR spectral analysis of the cardanol–formalde-
hyde resins (CF71 and CF72) revealed not only the Scheme 1 Structure of cardanol.

Journal of Applied Polymer Science DOI 10.1002/app

CARDANOL-BASED NOVALAC RESIN 1675

1
H-NMR spectroscopic analysis of CF71 and CF72
In the 1H-NMR spectra of the CF71 [Fig. 2(b)] and
CF72 [Fig. 2(c)] novolac resins, the peak near 6.6–6.8
d might have been due to the aryl protons of ben-
zene nuclei. The peaks around the region 7.1–7.3 d
appeared to be due to the phenolic hydroxyl in the
novolac resins. The peak at 5.3 d might have been
due to the methylene protons, whereas the peaks
between 0.8 and 2.8 d appeared to be due to the
presence of a long alkyl aliphatic side chain, origi-
nally observed in cardanol [Fig. 2(a)], which was fur-
ther confirmed by the appearance of a strong peak
at 1.3 d. The terminal methyl group of the alkyl side
chain was also seen, as there appeared a small peak
at 0.8 d. The peak at 3.7 d indicated the presence of
methylene protons of C6H5ACH2AC6H5 for the
bridge between the phenyl rings.25–27 All these spec-
tral data indicated that the condensation of methyo-
lated cardanol was completed under the
experimental conditions [see Fig. 2(b,c)] and was
fully consistent with the proposed structure (Scheme
1) resulting from the reaction mechanism shown in
our recent article16 for the cardanol-based novolac
resin prepared with a citric acid catalyst.

Determination of molecular weights of CF71 and

CF72 resins by GPC and mass spectroscopic analysis
The number of phenolic units (p) per molecule of
novolac resins, CF71 and CF72, were determined
from the ratio of aromatic methylene ([CH2]) to aro-
matic protons [[AR]; the areas are indicated in the
1
H-NMR spectra for the novolac resins CF71 and
CF72 in Fig. 2(b,c), respectively], as shown by the fol-
Figure 4 GPC trace of (a) novolac resin CF71 and (b) lowing equation:28
novolac resin CF72.
½CH2 Š=½ARŠ ¼ ð2p 2Þ=ð3p þ 2Þ (1)

bands near 1458 and 994 cm 1, which might have The values of p were calculated to be 2.20 and 2.25,
been related to the ortho- and para-substituted aro- whereas the GPC traces [Fig. 4(a,b)] resulted in Mn
matic rings. The appearance of peaks near 725 and values of 679 and 693 g/mol for samples CF71 and
869 cm 1 [Fig. 1(b,c)] also indicated ortho and para CF72, respectively. The values of Mn were further
substitution, respectively, at the benzene nuclei. The confirmed by mass spectroscopic analysis [Fig.
preceding spectral data was found to be identical to 3(b,c)], which yielded values of 680 and 693 g/mol
that given in the literature.19,23,24 for samples CF71 and CF72, respectively.

Scheme 2 Proposed structure of the cardanol-based novolac resin.

Journal of Applied Polymer Science DOI 10.1002/app

1676 YADAV AND SRIVASTAVA

Scheme 3 Proposed structure of the epoxidized novolac resin.

In the mass spectrum of CF71 [Fig. 3(b)], peaks resulted in glycidyl ether. The structure of the epoxy
were observed at m/z values of 680 and 850, which resin is proposed as shown in Scheme 3.
corresponded to n ¼ 1 and n ¼ 2, respectively, As shown in Figure 1(d), the FTIR spectrum of the
whereas for resin CF72 [Fig. 3(c)], the peaks were at uncured cardanol-based epoxidized novolac resins,
m/z values of 693 and 850, which corresponded to n that is, sample ECF72, the characteristic band of the
¼ 1 and n ¼ 2, respectively, as shown in Figure 3(c). oxirane ring was observed near 911 and 856 cm 1.
The GPC curves clearly indicated the existence of The 1H-NMR spectrum [Fig. 2(d)] further confirmed
dimers and trimers in the final reaction mixture. So, the formation of the cardanol-based epoxidized
molecular weights, as obtained from mass spectra novolac resins.
[Fig. 3(b,c)], of 680 and 693 g/mol for resins CF71
and CF72, respectively, were considered for this dis-
cussion. These values were very close to the values FTIR analysis of the uncured and cured blend
samples
obtained from the previously discussed 1H-NMR
spectra. Therefore, the values of n in the structure of The FTIR spectral analysis of the uncured blend
the novolac resins, shown in Scheme 2, were calcu- sample containing 15 wt % CTBN in the epoxy resin
lated to be 0.20 and 0.25 for resins CF71 and CF72, ECF72 is shown in Figure 1(e). The peaks related
respectively. These values resembled the values of to oxirane functionality appeared near 911 and
peak-to-peak mass increments (Dm) in the mass 856 cm 1 [Fig. 1(d)]. When CTBN was added to the
spectra [Fig. 3(b,c)]. pure epoxy resin, these peaks disappeared, and new
Finally, the structure of the cardanol-based novo- peaks appeared near 914 and 850 cm 1 [Fig. 1(e)].
lac resins is proposed as shown in Scheme 2. The peaks appearing near 911 and 856 cm 1 [Fig.
1(d)] might have overlapped these peaks. The peaks
near 1719 cm 1 due to carbonyl stretching, at 1440
and 972 cm 1 due to CAH bending,30–32 and a sharp
Epoxidation of the novolac prepolymer
peak near 2239 cm 1 due to the ACBN group of the
The novolac-based epoxy resins were synthesized by CTBN molecule were also seen in the spectrum of
reaction with ECH. The number of glycidyl groups the uncured blend system [Fig. 1(e)]. These observa-
per molecule in the resin was dependent on the tions clearly indicated that there was no chemical
number of phenolic hydroxyls in the starting novo- interaction between the oxirane group of the epoxy
lac, the extent to which they reacted, and the extent and the carboxyl group of CTBN. The epoxy resin
to which the lowest molecular species were poly- and CTBN remained as a discrete phase in the
merized during synthesis. Theoretically, all of the uncured stage. However, the addition of CTBN and
phenolic hydroxyls may react, but, in practice, all of polyamine to the epoxy caused chemical interaction
them do not react because of steric hindrance.29 The between the oxirane ring and the carbonyl function
reaction between ECH and novolac resin may be of the CTBN, which resulted in the complete dimi-
thought to proceed in a similar fashion, as in the nution of the peaks at 911 and 856 cm 1 in the cured
study by Lee and Neville.29 The epoxide group of blend sample BECF723 [Fig. 1(f)]. The ACBN group
ECH reacted with phenolic hydroxyls in the alkaline was also not observed in the cured blend. This was
medium and formed the chlorohydrin ether, which perhaps due to the lower volume fraction of CTBN
underwent a dehydrochlorination reaction and in the blend system. Another possibility is that it

Journal of Applied Polymer Science DOI 10.1002/app

CARDANOL-BASED NOVALAC RESIN 1677

Scheme 4 Proposed carboxyl-epoxide esterification reaction.

could have also been used in network modification. oxy matrices in Table III. It was clear from the
The blend also showed the appearance of results that the peak exotherms were shifted to
new stretched peaks between 1258 and 1635 and lower temperatures because of an enhanced reaction
1046 cm 1 and peak broadening at 1608 cm 1 due to rate, which, finally, reduced the cure time of the
CAC multiple stretching. The appearance of these CTBN-modified blend systems (see Table III). These
peaks at different positions resembled a discussion peaks appeared during the first heating run but
on the FTIR analysis of DGEBA/CTBN blend sys- were completely absent during the second heating
tems given in the literature.33 The carboxyl–epoxide cycle; this indicated the completion of the curing
esterification reactions are illustrated in Scheme 4. reaction. The initial addition of CTBN in the epoxy
The epoxide group present in P0 may have also resin decreased the cure time sharply, and this trend
combined with the carboxyl group of the CTBN, remained up to 15 wt % CTBN addition and
which is shown in Scheme 5. increased thereafter. The enhanced rate behavior
The products formed by the reactions shown in could be interpreted in terms of an intermolecular
Schemes 4 and 5 reacted with the polyamine curing transition state for the epoxy–amine reaction.
agent in the same way as those that that reacted According to this mechanism,35–37 strong hydrogen
with the epoxy resin34 to form an epoxy network. bonding species, such as acids and alcohols, stabi-
The FTIR spectrum of the cured blend sample lized the transition state and strongly accelerated the
shown in Figure 1(f) confirmed this. curing reaction. Also, the reaction products contain-
ing the carboxylic acid component favored the gela-
tion conditions, which might have created
Differential scanning calorimetry (DSC) analysis of fluctuations in the concentration and induced phase
the blend samples
separation.38 The curing reactions led to an increase
Figures 5 and 6 show the dynamic DSC scans of the in viscosity and, eventually, to the gelation of the
cardanol-based epoxidized novolac resins without entire resin mixture, as evidenced from the decrease
and with CTBN (i.e., samples ECF72 and BECF723) at in onset temperatures with increased addition of
a heating rate of 10 C/min. Also, the effect of CTBN CTBN in the epoxy matrices (Table III). Furthermore,
concentration on the cure parameters of different ep- the reduction in molecular mobility lowered the rate
oxy matrices is compared to that of nonmodified ep- of the reaction39 and demised the epoxy–amine–

Scheme 5 Proposed reaction between epoxy groups and CTBN.

Journal of Applied Polymer Science DOI 10.1002/app

1678 YADAV AND SRIVASTAVA

TABLE III
DSC Results for the Unmodified and CTBN-Modified Cardanol-Based Epoxidized Novolac Resins
Cured with Polyamine
Sample Ti ( C) Tonset ( C) TP ( C) Tstop ( C) DH (J/mol) tcure (min)a

ECF71 58.7 115.1 126.2 165.5 64.7 240

BECF711 58.1 113.5 123.0 164.8 68.7 223
BECF712 57.9 102.5 124.8 163.8 65.3 211
BECF713 57.6 76.3 120.1 163.0 76.9 180
BECF714 58.6 79.6 125.2 164.3 72.8 191
ECF72 65.7 67.7 134.5 195.4 34.1 360
BECF721 65.3 66.7 134.0 194.1 39.7 331
BECF722 61.0 64.3 135.4 186.5 45.6 317
BECF723 46.9 58.2 133.4 184.8 58.6 300
BECF724 53.8 62.3 134.6 188.8 53.8 330

DH ¼ heat of curing; tcure ¼ cure time; Ti ¼ kick-off temperature at which the curing started; Tonset ¼ temperature at
which the first detectable heat was released; TP ¼ temperature of the peak position of the exotherm; Tstop ¼ temperature
of the end of the curing exotherm.
a
Obtained by the curing of the sample in an air oven at 120 C.

CTBN systems40 up to 15 wt % CTBN addition. The affect DH, which indicated that there was not much
decrease of cure time could also be explained by the influence on the crosslink density of the epoxy ma-
fact that, during the reaction of CTBN with the ep- trix. The DH values ranged from 64.7 J/mol for the
oxy resin, some of the exothermic energy released unfilled ECF71/polyamine to 76.9 J/mol (expressed
during epoxy crosslinking might have been con- in terms of the mass of the epoxy/amine mixture)
sumed by CTBN, which resulted in a decrease in the for 15 wt % CTBN and 34.1 J/mol for the unfilled
cure time.41 The high cure rate might have resulted ECF72/polyamine to 58.6 J/mol for 15 wt % CTBN
in the formation of spherical domains with fairly (i.e., blend sample BECF723). This confirmed that the
uniform particle size. The particle size became final reaction state was not significantly affected by
smaller with increasing curing rate. These domains the presence of CTBN. DH was previously found to
might have become bigger and more closely packed be 99.9 J/mol for aniline/DGEBA, 98.9 J/mol for
and reached phase-inversion conditions42 beyond 15 DDM/DGEBA, 99.1 J/mol for aniline/DGEBA, and
wt % CTBN, and so, the curing rate dropped, which 100–118 J/mol for phenyl glycidyl ether type epoxy–
increased the cure time. amine reactions, as tabulated in a review by Rozen-
The heat of polymerization (DH) values (Table III) berg.37 The DH values did not compare with those
related to the cure process were determined from found in our systems, that is, for the blend systems
the area of the exotherm peak obtained from DSC of the cardanol-based epoxidized novolac resins,
analysis (Figs. 5 and 6) taken in dynamic mode. In CTBN, and polyamine and, hence, the cured product
contrast, the presence of CTBN did not significantly was not the same as that for the aforementioned
systems.

Figure 5 DSC scan of sample ECF72. Figure 6 DSC scan of blend sample BECF723.

Journal of Applied Polymer Science DOI 10.1002/app

CARDANOL-BASED NOVALAC RESIN 1679

TABLE IV
Results of the TG/DTG Traces of the Unmodified and CTBN-Modified Cardanol-Based Epoxidized Novolac Resins
Cured with Polyamine
First step Second step
Blend sample To Tmax Tf ML (%) To Tmax Tf ML (%) Total CY (%)

ECF71 223 326 381 42.6 381 442 500 53.2 4.2
BECF711 198 318 353 62.6 353 439 514 34.4 3.2
BECF712 168 325 357 90.7 357 439 499 7.0 2.3
BECF713 155 318 359 61.4 359 439 508 38.5 .1
BECF714 149 322 355 55.6 355 441 505 40.5 3.9
ECF72 228 317 374 39.8 374 386 508 51.9 8.3
BECF721 204 314 369 60.3 369 382 508 32.2 7.5
BECF722 176 316 356 54 356 376 504 41.9 4.1
BECF723 160 301 349 80.2 349 360 475 13.3 6.5
ECF724 155 312 360 51 360 367 498 41.1 7.9

CY ¼ char yield; ML ¼ mass loss.

Thermal stability were lower than those of the blend systems cured
with 5–15 wt % CTBN. Mass loss in the whole
The cured polymer samples (ECF72 and BECF723)
decomposition process was found to be almost inde-
were analyzed for their thermal stability by ther-
pendent of CTBN addition in the blends. The mass
mogravimetry (TG)/differential thermogravimetry
loss for the blend systems prepared from the epoxy
(DTG) traces resulting from thermogravimetric anal-
resin ECF72 was found to be more than that for the
ysis (TGA) at a heating rate of 10 C in a nitrogen
blend systems prepared from the epoxy ECF71. Also,
atmosphere. The onset temperature of degradation
the char yield was found to be higher in the blend
(To), the temperature of the maximum rate of mass
sample prepared from the epoxy resin ECF72 than
loss (Tmax), and the extrapolated final decomposition
that in the blend systems prepared from ECF71. This
temperature (Tf) were noted from the TG traces and
was attributed to the presence of more crosslinks in
are presented in Table IV. We compared the relative
the ECF72–CTBN–amine system than in the ECF71–
thermal stability of the cured blend resins by deter-
CTBN–amine system. The decreased thermal stabil-
mining the percentage char yield at 500 C. A two-
ity, as observed by To, confirmed this.43–46
step decomposition behavior was observed for all of
the unmodified and CTBN-modified samples (Figs. 7
and 8). A major mass loss was observed in the sec-
ond stage of the decomposition (Table IV), whereas SEM analysis
the mass lost only 37–39% in the second step of the Figure 9(a,b) shows, respectively, the SEM micro-
decomposition in the temperature range 320–500 C. graphs of fractured surface of the pure epoxy and
These two stages of mass loss were attributed to the CTBN-modified epoxy matrix. SEM of the CTBN-
decomposition of (1) the epoxy and (2) the epoxy– modified system showed the presence of precipi-
CTBN–amine network. The values of To and Tmax tated, discrete rubber particles, which were

Figure 7 TGA trace of blend sample ECF72. Figure 8 TGA trace of blend sample BECF723.

Journal of Applied Polymer Science DOI 10.1002/app

1680 YADAV AND SRIVASTAVA

1. The proposed mechanism for the curing reac-

tion of the blend of cardanol-based epoxidized
novolac resins and CTBN in the presence of
polyamine was found to be well suitable for
such systems, as confirmed by IR analysis.
2. DSC studies showed the exothermal heat of
reaction of the epoxy crosslinking due to the
addition of rubber into the epoxy matrix.
3. The thermal stability of the cardanol-based ep-
oxy systems improved with the addition of
CTBN, where CTBN might have acted as a ther-
mal stabilizer.
4. The morphological study of the cured system
revealed a two-phase region where the liquid
rubber particles were distributed in the epoxy
matrix. A further increase in the concentration
of the elastomer led to a phase-inversion
morphology.

The authors thank M/s Satya Cashew Pvt. for providing car-
danol and M/s Emerald Performance Materials for provid-
ing CTBN.

References
1. Attanasi, O. A.; Bunatti, S. B. Chim Ind 1996, 78, 693.
2. Prabhakaran, K.; Narayan, A.; Pavithran, C. J. Eur Ceram Soc
Figure 9 SEM micrographs of prepared samples: (a) 2001, 21, 2873.
ECF72 and (b) BECF723. 3. Pillai, C. K. S.; Prasad, V. S.; Sudha, J. D.; Bera, S. C.; Menon,
A. R. R. J Appl Polym Sci 1990, 41, 2487.
4. Bhunia, H. P.; Jana, R. N.; Basak, A.; Lenka, S.; Nando, G. B. J
dispersed throughout the epoxy matrix; that is, they Appl Polym Sci 1998, 36, 391.
revealed the presence of two-phase morphological 5. Kim, Y. H.; An, E. S.; Park, S. Y.; Song, B. K. J Mol Catal B
features. The soft elastomeric phase was separated 2007, 45, 39.
from the hard epoxy matrix during the early stage 6. Kinloch, A. G.; Reiw, C. K. Rubber-Toughened Plastics;
Advances in Chemistry 22; American Chemical Society: Wash-
of the cure. The fractured surface of most rubber-
ington, DC, 1985; p 67.
toughened epoxy systems has a rigid continuous ep- 7. Kinloch, A. G.; Young, R. J. Fracture Behaviour of Polymers;
oxy matrix with a dispersed rubbery phase as iso- Applied Science: London, 1983.
lated particles.47–49 Some cavitations of the rubber 8. Huang, J.; Kinloch, A. G. Polymer 1992, 33, 1330.
particles accompanied by stress-whitening zones 9. Huang, J.; Kinloch, A. G. J Mater Sci 1992, 27, 2763.
10. Riew, C. K.; Rowe, E. H.; Siebert, A. R. Presented at the ACS
were also observed in the scan [Fig. 9(b)]. This
Symposium on Toughness and Brittleness of Plastics (Division
stress-whitening effect may have been related to of Organic Coatings and Plastics), Atlantic City, NJ, 1974.
location deformation at the crack tip. Furthermore, 11. Frigone, M. E.; Masica, L.; Aciermo, D. Eur Polym J 1995, 31,
the cavitations were followed by the onset of a shear 1021.
localization process,50 which resulted in the 12. Bascom, W. D.; Cottington, R. L.; Jones, R. L.; Peyser, P. J
Appl Polym Sci 1975, 19, 2545.
observed increase in thermal properties. The SEM
13. Shi, H.; Fang, Z.; Gu, A.; Tong, L.; Xu, Z. J Appl Polym Sci
analysis also proved that there occurred a chemical 2007, 106, 3098.
reaction between the molecules of the epoxy and 14. Knop, A.; Schieb, W. Chemistry and Application of Phenolic
rubber, as already shown by the FTIR analysis, Resins; Springer-Verlag: New York, 1979.
which led to strong interactions between the two 15. Devi, A.; Chandra, K.; Srivastaava, D. Proc Natl Therm Anal
Symp 2004, 14, 22.
phases at lower CTBN concentrations.
16. Devi, A.; Srivastava, D. Mater Sci Eng A 2007, 458, 336.
17. Mishra, A. K.; Pandey, G. N. J Appl Polym Sci 1985, 30, 969.
18. Urabanski, J.; Czer Winkski, W.; Janika, K.; Majewsta, F.;
Zowall, H. Handbook of Analysis of Synthetic Polymers and
CONCLUSIONS Plastics; Ellis Horwood: Chichester, England, 1977.
19. Sperling, G. R. J Am Chem Soc 1954, 76, 1190.
The following conclusions were drawn from the pre- 20. Sathiyalekshmi, K. Bull Mater Sci 1993, 16, 137.
viously discussed results: 21. Chakrawarti, P. B.; Mehta, V. Ind J Technol 1987, 25, 109.

Journal of Applied Polymer Science DOI 10.1002/app

CARDANOL-BASED NOVALAC RESIN 1681

22. Mythili, C.; Retna, A.; Gopalkrishnan, S. Bull Mater Sci 2004, 38. Fisher, W.; Hofman, W. J Polym Sci 1954, 12, 497.
27, 235. 39. Tmomas, R.; Durix, S.; Sinturel, C.; Omonov, T.; Goossens, S.;
23. Tyman, J. H. P. Chem Soc Rev 1979, 8, 499. Groeninckx, G.; Moldenaers, P.; Thomas, S. Polymer 2007, 48,
24. Antony, R.; Pillai, C. K. S. J Appl Polym Sci 1993, 49, 2129. 1695.
25. Huang, J.; Xu, M.; Lin, M.; Lin, Q.; Chem, Y.; Chu, J.; Dai, H.; 40. Wise, C. W.; Cook, W. D.; Goodwin, A. A. Polymer 2000, 41,
Zou, Y. J Appl Polym Sci 2005, 97, 652. 4625.
26. Kuriaposa, A. P.; Manjooran, S. K. B. Surf Coat Technol 2001, 41. Calabrese, L.; Valenza, A. Compos Sci Technol 2003, 63, 851.
145, 132. 42. Butta, E.; Levita, G.; Marchetti, A.; Lazzeri, A. Polym Sci Eng
27. Gardziella, A.; Pilato, L. A.; Knop, A. Phenolic Resins; 1986, 26, 63.
Springer-Verlag: New York, 2000. 43. Maity, T.; Samanta, B. C.; Dalai, S.; Banthia, A. K. Mater Sci
28. Lin-Gibson, S.; Baranauskas, V.; Riffle, J. S.; Sorathia, V. Poly- Eng A 2007, 464, 38.
mer 2000, 43, 7389. 44. Kaji, M.; Nakahara, K.; Endo, T. J Appl Polym Sci 1999, 74,
29. Lee, H.; Neville, K. Hand Book of Epoxy Resins; McGraw- 690.
Hill: New York, 1982. 45. Menon, A. R. R.; Agibodion, A. I.; Pillai, C. K. S.; Mathew, A.
30. Evtushenko, Y. M.; Jvanov, V. M.; Zaitsev, B. E. J Anal Chem G.; Bhagawan, S. S. Eur Polym J 2002, 38, 163.
2003, 58, 47. 46. Srivastava, K.; Kauushik, M. K.; Srivastava, D.; Tripathi, S. K.
31. Smith, A. Prikladnaya Ik-Spektroscopiya; Mir: Moscow, 1982. J Appl Polym Sci 2006, 102, 4171.
32. Nigam, V.; Setua, D. K.; Mathur, G. N. J Appl Polym Sci 1998, 47. Chan, L. C.; Gillham, J. K.; Kinloch, A. G.; Shaw, S. J. In Rub-
70, 537. ber-Modified Epoxies: Morphology, Transitions and Mechani-
33. Ramos, V. D.; da Costa, H. M.; Soares, V. L. P.; Nascimento, cal Properties; Riew, C. K.; Gillham, J. K., Eds.; American
R. S. V. Polym Test 2005, 24, 219. Chemical Society: Washington, DC, 1984; Vol.208, p 274.
34. Chem, D.; Pascault, J. P.; Bertsch, R. J.; Drake, R. S.; Siebert, A. 48. Douglass, S. K.; Beaumont, P. W. R.; Ashby, M. F. J Mater Sci
R. J Appl Polym Sci 1994, 51, 1959. 1980, 15, 1109.
35. Horie, K.; Hiura, H.; Sawada, M.; Mita, I.; Kambe, H.J Polym 49. Tripathi, G.; Srivastava, D. Adv Polym Sci 2008, 26, 258.
Sci Part A-1: Polym Chem 1970, 8, 1357. 50. Sue, H. J.; Garciameitin, E. I.; Pickelman, D. M. In Polymer
36. Smith, I. T. Polymer 1961, 2, 95. Toughening; Arands, C. B., Ed.; Marcel Dekker: New York,
37. Rozenberg, B. A. Adv Polym Sci 1985, 75, 113. 1996; p 131.

Journal of Applied Polymer Science DOI 10.1002/app