Biobased Epoxy Resin from Canola Oil

Tolibjon S. Omonov, Jonathan M. Curtis

Lipid Chemistry Group, Department of Agricultural, Food, and Nutritional Science, University of Alberta, Edmonton T6G 2P5,
Alberta, Canada
Correspondence to: T. S. Omonov (E - mail: [email protected])

ABSTRACT: Epoxidized canola oil (ECO)-based thermoset epoxy resins were formulated with phthalic anhydride (PA) as the curing
agent for different ratios of ECO to PA (1:1, 1:1.5, and 1:2 mol/mol) at curing temperatures of 155, 170, 185, and 200 C. The gela-
tion process of the epoxy resins and the viscoelastic properties of the systems during curing were studied by rheometry, whereas the
dynamic mechanical and thermal properties of the cured resins were studied by dynamic mechanical analysis (DMA) and differential
scanning calorimetry. We found that the thermomechanical properties of the resins were not strongly dependent on the curing tem-
perature of the resin, although elevated temperatures significantly accelerated the curing process. However, an increase in the curing
agent (PA) amount significantly altered both the reaction rate and the thermomechanical properties of the final resin. Thus, in the
ECO/PA system, the selection of the combination of the curing temperature and the molar ratios of the curing agent could be used
to design thermoset resins with unique thermomechanical properties. V C 2013 Wiley Periodicals, Inc. J. Appl. Polym. Sci. 2014, 131, 40142.

KEYWORDS: biopolymers and renewable polymers; crosslinking; lipids; thermosets

Received 20 June 2013; accepted 30 October 2013

DOI: 10.1002/app.40142

INTRODUCTION Currently, the major components used in most epoxy-based

thermoset polymer formulations originate from petroleum-
Epoxy resins are an important class of materials, which are used
based derivatives. However, resins and composites incorporating
widely in electrical, automobile, and other applications because
renewable resources are increasingly being used in diverse areas
of their excellent chemical resistance, outstanding adhesion to a
of manufacturing because of the heightened awareness of envi-
variety of substrates, high tensile and compressive strengths, elec-
ronmental issues. To date, these biobased resins only represent a
trical insulation properties, and superior fatigue strength.1–3
very small percentage of the entire plastics market, and the
Common applications for epoxy resins were reviewed in detail
widespread application of biobased materials is still limited by
elsewhere.2 Linear or crosslinked epoxy polymers are obtained by
the available materials plus their cost of production. Hence, the
the reaction of epoxy monomers with curing agents (hardeners)
development of cost-effective biobased materials from abun-
and/or initiators. Mechanisms for the polymerization of epoxy
dantly available renewable sources is of great interest to prevent
monomers to form crosslinked polymers via step polymerization,
reliance on finite petroleum resources.
chain polymerization or a combination of mechanisms were
reviewed earlier.3 Epoxy groups can react with many functional For example, processes for creating biobased polyols from vege-
groups, including amines, phenols, mercaptans, isocyanates, and table oils have been described in the literature5–7 as have some
acids.2,3 Typically, the concentration of epoxy groups in the epoxy of the challenges in making biobased polymers from such
monomer used is approximately equal to the concentration of renewable resources.8 A comparison of the life cycles and the
reactive groups in the hardener/initiators because, under these environmental impact of soy- and petro-based polyols has also
conditions, unwanted side reactions are minimized. Crosslinked been discussed.9 Usually, vegetable-oil-based polyol synthesis is
epoxy resins are cohesively very strong, but their brittleness a two-step process, in which the first step is the epoxidation of
reduces their usefulness in many applications, especially those unsaturated vegetable oil fatty acids or fatty esters.10–13 The sec-
requiring a high impact strength. Hence, the development of ond step of the process is the subsequent reaction of the epoxy
epoxy resins in recent years has focused on toughening. This can groups, which usually involves two or more hydroxyl-group-
be achieved by the modification of the chemistry of the raw containing reactants and acid catalysts.11,14–16 In addition to
materials within the formulation4 or by the addition of particu- their broad applications in the making of polyols, biobased-
lates, elastomers, or thermoplastics to the formulation.2 epoxy-resin-epoxidized vegetable oils can also be used in similar

C 2013 Wiley Periodicals, Inc.

V

WWW.MATERIALSVIEWS.COM 40142 (1 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

ways as petrochemical epoxy derivatives. These applications parallel plates under controlled displacement at a frequency of 1
include the use of biobased epoxies as additives in specialty Hz and at constant temperatures of 155, 170, 185, and 200 C to
coatings and as plasticizers,15 as interfiber bonding adhesives in follow the influence of the reaction kinetics at different temper-
glass15,17 and lignocellulosic17 composite materials, and in the atures on the viscoelastic properties of the materials. Rheology
automobile industry.18 Advantage Data Analysis (version 5.7.0) software was used to
analyze the experimental data.
Epoxy resins are most usually cured with primary or secondary
amines to form thermoset polymers. However, it is known that Dynamic Mechanical Thermal Analysis. The dynamic mechan-
the use of anhydrides in the curing reactions of epoxies may ical thermal properties [storage modulus (G0 ), loss modulus
also lead to similar thermoset resins, although the curing reac- (G00 ), and loss factor (tan d)] were measured with a DMA Q800
tion must usually be carried out at elevated temperatures. (TA Instruments) equipped with a liquid nitrogen cooling sys-
Anhydride-cured epoxy resins often serve as matrices for tem in the single cantilever bending mode at a constant heating
advanced, fiber-reinforced composites used in a variety of aero- rate of 2 C/min from 2100 to 100 C. The size of the test sam-
space, structural, and sporting applications, where their high ples was about 17.5 3 11 3 4 mm3 (6 0.5 mm). The measure-
specific strength and stiffness combined with their good thermal ment of the length, width, and thickness of the specimen
stability offer unique design advantages.19 accuracy was 60.01 mm. The measurements were performed
Most commonly, vegetable oil epoxide based materials with according to ASTM E 1640-09 at a fixed frequency of 1 Hz. At
commercial applications have been made from epoxidized soy- least two similar sized specimens prepared by the cutting of the
bean oils (ESOs). For example, ESO is used as plasticizer or sta- bulk polymer sheet were obtained, and the results were aver-
bilizer to modify the properties of plastic resins such as aged. The reported errors are the subsequent standard devia-
thermoplastic poly(vinyl chloride).20 Some researchers have also tions. TA Universal Analysis (version 4.7A) software was used to
investigated the curing process of ESO with anhydrides21 and analyze the experimental data.
the conversion of the ESO into flexible, semiflexible, or rigid
Differential Scanning Calorimetry (DSC). DSC measurements
crosslinked thermoset polymers using various approaches.22,23
were performed with a Q100 series DSC (TA Instruments) with
However, so far, the applications of plant-oil epoxy precursors
a modulation of 61 C every 60 s. Approximately 8.0–10.0 mg
are still limited because of their low crosslinking density and
(60.1 mg) of the sample was placed in an aluminum pan,
mechanical performance and their cost of manufacturing rela-
which was then hermetically sealed. An empty aluminum pan
tive to that of petrochemical epoxides. The development of
was used as a reference, and the experiments were performed
alternatives to petroleum-based epoxy resins for commercial
under a nitrogen flow of 50 mL/min. The sample was trans-
applications remains a challenge for the polymer and composite
ferred to the DSC cell, where it was equilibrated at 250 C for 5
industries. In this study, we explored the use of epoxidized can-
min and then heated to 100 C at a constant rate of 2 C/min.
ola oil (ECO), derived from an oilseed crop widely grown in
The TA Universal Analysis (version 4.7A) software was used to
the Canadian prairie provinces, which is high in monounsatu-
analyze the experimental data.
rated fatty acids and low in saturated fatty acids. Our main
goals were to investigate the curing process of ECO with an Epoxidation of Canola Oil. The epoxidation of canola oil was
anhydride, to evaluate the properties of ECO-based thermoset carried out with performic acid generated in situ from formic
resins, and to assess the commercial feasibility of these resins in acid and hydrogen peroxide. The molar ratio of canola oil to
making cellulosic particle and fiberboards. formic acid to hydrogen peroxide was 1.0:1.0:5.6. The average
MW of canola oil was 885 g/mol.
EXPERIMENTAL
Into a 22-L glass reactor (equipped with a bottom drain and a
Materials water jacket and attached to a recirculating liquid cooler), 4000
Food-grade canola oil (Loblaws, Inc., Canada) was purchased g (4.54 mol) of canola oil was added at room temperature.
from the local supermarket and was used as supplied. Formic Hydrogen peroxide (2470 g, 35%) was loaded into the vessel,
acid (technical grade, 85%) and hydrogen peroxide (technical and the mixture was stirred vigorously with an overhead
grade, 35%) were purchased from Univar Canada. Phthalic mechanical stirrer (350 6 10 rpm) to form a homogeneous mix-
anhydride [PA; molecular weight (MW) 5 148.1 g/mol, ture before we proceeded with addition of formic acid. Formic
density 5 1.53 g/cm3, melting point 5 131 C, reagent grade, acid (246 g, 85%) was added dropwise through an addition
99%] from Lancaster (Alfa-Aesar) was used as a curing agent. funnel into the reactor while the reaction mixture was stirred
ECO was synthesized in our laboratory, as described later. intensively. The formic acid addition was done at a rate of 8–10
g/min to control the temperature rise due to the exothermicity
Methods
of the reaction. After the addition of formic acid was complete,
Rheological Experiments. Rheological experiments were per-
the temperature of the reaction was set to 50 6 5 C. The epoxi-
formed with an AR2000 Advanced Rheometer (TA Instru-
dation process was monitored by liquid chromatography-mass
ments). All rheological experiments were carried out
spectrometry (LC–MS) and was usually complete in 48 h under
immediately after the prepolymer preparation process at low
the conditions described previously.
strain rates (1.25%) and under identical conditions. Disposable
25-mm parallel plates were used for the test geometry. Isother- After the epoxidation process, the product was washed with an
mal time sweep tests were performed with 25-mm disposable equal amount of saturated NaCl solution (4000 mL) at least

WWW.MATERIALSVIEWS.COM 40142 (2 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

three times. The formic acid was neutralized with a 0.6M solu- Table I. Expected Products and Product Yields from the Epoxidation of
tion of NaOH during the second and third washes. Traces of Canola Oil on the Basis of the Fatty Acid Distribution of Canola Oil
water were removed with anhydrous Na2SO4, and the final
product was filtered through filter paper to remove NaCl or/ Number
and Na2SO4 particulates. The resulting ECO was a viscous Expected of epoxy
product at elevated temperatures that tended to crystallize at product of groups Average Expected
room temperature. canola oil per MW yield
epoxidation TAG (g/mol) (mmol/kg)
Epoxy Resin Preparation. Thermoset resins based on ECO and Multifunctional 9 1017 2.3
PA were prepared above the melting point of PA at different epoxy
curing temperatures. The desired amount of PA (0.05, 0.075, or
8 1003 6.9
0.10 mol) was dissolved in ECO (0.05 mol) in a glass flask that
7 989 35.3
was placed in an oil bath at the desired temperature (155, 170,
185, or 200 C). The mixture was thoroughly mixed by a mag- 6 975 101.2
netic stirrer at 500 rpm until all of the PA was melted. The pre- 6 (1 P) 947 2.4
polymer preparation time used was varied depending on the 5 961 215.6
amount of PA added; it was 10 min for the ECO/PA molar ratio 5 (1 P) 933 10.6
of 1.0:1.0, 9 min for the molar ratio of 1.0:1.5, and 8 min for 4 947 254.7
the 1.0:2.0 molar ratio. The majority of the prepolymer formed 4 (1 S) 947 3.4
in this way was transferred into a preheated Teflon mold in an
4 (1 P) 919 40.9
oven for further curing to form sheets with a size of 100 3 100
Triepoxy 3 933 253.0
3 5 mm3. These cured samples were later used for the dynamic
mechanical analysis (DMA) and DSC measurements. 3 (1 S) 933 18.1
3 (1 P) 905 66.5
In addition to the previous treatment, about 3–4 g of the prepol-
Diepoxy 2 (1 P) 891 53.5
ymer was quickly transferred to a small glass flask and cooled
2 (1 2P) 863 3.6
down in an ice–water mixture. These small amounts of prepoly-
mers were used for further isothermal rheological experiments. 2 (1 S) 920 29.3
The curing of the resins in the mold was performed for 6 h at 2 (1 2S) 920 4.5
each temperature, after which the oven was turned off to allow 2 (1 SP) 891 2.3
slow cooling (1.0–1.5 C/min) of the final product. Monoepoxy 1 (1 2P) 849 2.4
1 (1 SP) 877 1.2
RESULTS AND DISCUSSION Saturates PPP 807 1.2
Epoxidation of Canola Oil PPS 835 6.0
The epoxidation of unsaturated fatty acids or vegetable oils SPS 863 2.4
with hydrogen peroxide along with acetic, sulfuric, or formic
SSS 892 1.1
acids is a commonly used procedure that has been described
well elsewhere.24–26 Some advantages of the use of performic
acid, which is formed by the reaction of formic acid with saturated fatty acid and one monounsaturated fatty acid (most
hydrogen peroxide, for the epoxidation process of long-chain, likely oleic acid) and could range up to 9 if the TAG was trili-
unsaturated aliphatic compounds have been described.26 For nolenin. Table I lists the most abundant TAG products from the
example, the formation of peracetic acid has been described as epoxidation of canola oil, including the number of epoxide
significantly slower than the formation of performic acid under groups, the MW, and the expected yield of the epoxidized TAGs
moderate reaction conditions. The exothermic reaction of for- per kilogram of canola oil with the assumption of complete
mic acid with hydrogen peroxide (in oleic acid at 40 C) starts conversion of all double bonds to epoxide groups. The fatty
immediately, and the reaction of performic acid with oleic acid acid composition of canola oil and the MW distribution of can-
is complete within 2–4 h,26 whereas the reaction of peracetic ola oil TAGs used was that reported previously.28
acid (at room temperature) requires more than a week to finish.
As shown in Table I, ECO TAGs can be classified according to
Thus, the epoxidation of canola oil was carried out with an in
the number of epoxide groups per TAG molecule. In addition,
situ generated performic acid, as described in the Experimental
for each number of epoxide groups per TAG, there may be sev-
section.
eral species that have either the same or different MWs because
Because canola oil, like other natural plant oils, is predomi- of the original fatty acyl substitution on the TAG. For example,
nantly comprised of triacylglycerides (TAGs) containing a distri- the most abundant TAG, with four epoxide groups with a nom-
bution of fatty acid types,27,28 we knew that ECO would inal MW of 947, arose from a TAG with no saturated fatty
contain a distribution of epoxidized TAG species. On the basis acids, OOL, or its isomers. In addition, a TAG with four epox-
of the known fatty acid profile of canola oil, the number of ide groups with a nominal MW of 947 also arose from a TAG
epoxide groups per TAG molecule could vary from 0 for TAG containing stearic acid, SLL, SOLn, or their isomers (where S
containing three saturated fatty acids to 1 for TAG with two represents stearic acid, L is linoleic acid, O is oleic acid, and Ln

WWW.MATERIALSVIEWS.COM 40142 (3 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

is linolenic acid). In addition, TAG with four epoxide groups ration procedure. ECO is clear, colorless, and viscous liquid at
arose from palmitic acid, PLL (where P is palmitic acid), POLn, 40 C, but at room temperature, it gradually solidifies. In con-
or their isomers, although these molecules had a nominal MW trast, PA was purchased as irregularly shaped white flakes with a
of 919. melting point of about 131 C. To prepare the thermoset resins,
it was necessary to form a homogeneous mixture of these two
Figure 1(a) shows the distribution of epoxide groups on the
precursors at higher temperatures (above the melting point of
TAGs in ECO on the basis of the data in Table I. In Figure 1(a),
PA) and to maintain and control the curing process of this
TAGs with similar numbers of epoxide groups are grouped
mixture.
together; the group with zero epoxides per TAG contained fully
saturated TAGs. The maximum distribution of the epoxide Generally, the reaction of the epoxy groups with curing agents
groups was between 3 and 4; this was similar to the double- are very complex, and multiple competing reactions, such as the
bound functionality of the canola oil, which was around 3.9 per formation of ethers or esters, can take place,2,29,30 depending on
TAG. the curing agents used. For example, the reaction of epoxides
with anhydrides involves multistep reactions, and some impor-
Figure 1(b) presents the rather broad MW distribution of epoxi-
tant simple examples are given in Scheme 1.
dized TAGs that were present in ECO. The peak maximum
occurred at about 933 g/mol; this was close to the MW of the In reaction A, the anhydride ring is opened by its reaction with
epoxidized triolein (933.43 g/mol) and to the distribution aver- an alcohol to form both monoester and carboxylic acid groups.
age of 942 g/mol. These values were used in later calculations to In reaction B, the carboxylic group so generated then react with
formulate ECO-based thermoset resins. the epoxide; this results in the formation of an ester linkage. In
reaction C, other epoxide groups react with the hydroxyl groups
Thermoset Resin Preparation to produce an ether linkage;2 these hydroxyl groups may be those
A major challenge that was encountered in the development of already existing or those formed via reactions A or B. This gives
biobased resins based on ECO and PA was in the sample prepa- rise to oligomeric structures. It is important to note that the
occurrence and the rate of reactions of reactions A and B depend
on the curing temperature. At low temperatures of curing, both
reactions may occur at a similar rate, whereas higher cure tem-
peratures have been shown to favor reaction B.2,3
On the other hand, epoxidized vegetable oils have internal
epoxide groups that are much less reactive and require different
curing conditions compared to synthetic epoxides with terminal
epoxide groups (e.g., glycidyl ethers).31 A necessary (although
not sufficient) condition for the formation of an epoxy network
is that at least one of the monomers involved in the reaction
should have a functionality higher than two.3
Taking the previous into account, an attempt was made to pre-
pare prepolymers of ECO at temperatures above the melting
point of PA. The objective was to form a homogeneous reactive
mixture that would undergo a controlled amount of

Figure 1. (a) Calculated number of epoxy groups per TAG and (b)
expected MW distribution of TAGs in ECO. The epoxy number 0 corre- Scheme 1. Important reactions of the PA and epoxy precursors occurring
sponds to fully saturated TAGs. during the curing processes.

WWW.MATERIALSVIEWS.COM 40142 (4 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

oligomerization via reactions such as those described in Scheme

1. This mixture could be considered to be the prepolymer prep-
aration; further reaction could be arrested after an appropriate
time (times of 8–10 min were used in this work) by shock cool-
ing in an ice–water bath. We envisioned that this prepolymer
preparation could be finally cured by its heating to a tempera-
ture most suitable for the final polymer or composite applica-
tion. We recently described some aspects of the previous process
for a range of biobased epoxide resin systems.29
In this way, several compositions were prepared at different
molar ratios of ECO to PA (1.0:1.0, 1.0:1.5, and 1.0:2.0). These
particular component ratios were chosen because each PA mole-
cule could potentially react with two epoxides.2 At the same
time, as described previously, on average, each molecule of epoxi-
dized oil had about four epoxide groups. Hence, the expected Figure 2. Viscoelastic properties of the ECO-based resin with an equimo-
stoichiometric ratio of the reactive moieties of these components lar ratio of components cured at 155 C. The complex viscosity
was close to 1:1 at a molar ratio of ECO to PA of 1:2. These (|g*|) of the resin. The vertical dashed line highlights the
three selected compositions were each cured at four different cur- gelation time of the system. The black area corresponds to the prepolymer
ing temperatures above the melting point of PA: 155, 170, 185, preparation time, and the gray area is the equilibration time in the
and 200 C, as described previously in the Experimental section. rheometer.

Depending on the curing conditions and the PA content, clear isothermal curing; these times are shown in Figure 2 as dark
and transparent resins were formed; these ranged from rubber- and gray-filled areas, respectively. Also, the y-axis results are
like flexible thermosets (with low curing temperatures and low presented on logarithmic scales to allow the entire process to be
PA content) to almost rigid semiflexible plastics (with a high presented.
curing temperature and high PA content). The color of the res-
ins varied from light yellow at low curing temperatures to yel- As shown in Figure 2, during the initial period of curing, the
low–light brown at higher curing temperatures. sample behavior could be characterized as liquidlike. This
region was characterized by low values of G0 and G00 , and these
Rheological Experiments and Gelation values remained almost constant with time. Also, in this region,
Gelation is the critical transition in the formation of a network the mixture was in the sol state with a finite value of viscosity.
of constituent units that generates a giant macromolecular Then, as the curing and, consequently, the crosslinking reaction
structure. It is characterized by a physical change in the reaction progressed, the G0 and G00 values of the system started to
medium from liquid to solid where the sample viscosity tends increase. At a certain point (28 min for this particular sys-
to be infinite.3 The determination of the gelation period is a tem), the viscoelastic behavior of the system prevailed.32 This
key parameter in following the curing process of the epoxy pre- point, where a rapid increase of both the moduli and viscosity
cursors; it will provide significant information on the working of the system were observed, was considered to be the gelation
life (pot life) of the epoxy system.2 time for the system, as indicated by a dashed line in Figure 2.
Rheological tools can be used to follow the gelation period of It can be seen from the figure that as the gelation progressed,
epoxy resins. Here, the change in viscosity of the epoxy resin in the viscosity of the system increased dramatically. In an ideal
the early stages of curing was measured as a function of the case, after gelation, the viscosity of the reactive mixture should
resistance to the shear of a sample placed between two plates. tend toward infinite values. The observed finite values of the
However, because the gelation time may depend on the shear viscoelastic properties of this particular mixture within the
rate because of the shear thinning at high rates,17 all of the experimental timeframe could be attributed to the structures,
experiments were carried out at low strain rates (1.25%) and properties, and functionalities of the initial components of the
under identical conditions. In this way, the rheological experi- resin. The structure of ECO (nonlinear chain, moderate MW)
ments on the ECO/PA resins were carried out at different tem- and the curing of ECO with aromatic acid anhydride (PA) led
peratures to characterize the physical transformations during to the formation of flexible, rubberlike thermoset plastics, espe-
the curing process. These rheological experiments were carried cially when they were synthesized at low a concentration of
out immediately after the prepolymer preparation process. Iso- anhydride. In general, at low concentrations of curing agents,
thermal time sweep tests were performed with 25-mm dispos- the molecular structure of the final resin likely consisted of a
able parallel plates under controlled displacement at a frequency mixture of monomeric, oligomeric, and polymeric fractions,
of 1 Hz to monitor the viscoelastic properties of the epoxy resin along with an unreacted epoxidized TAG. Excess epoxide simply
as a function of the curing time. As an example, Figure 2 shows became a diluent or plasticizer within the resin,33 insoluble in
the viscoelastic properties of the 1:1 ECO/PA resin cured at the partially cured, hyperbranched, and crosslinked composi-
155 C. Note that the curing time (x axis) in Figure 2 was cor- tion.34 Hence, the ultimate MW of the heterogeneous resin
rected with the prepolymer preparation time (10 min) and depended on the initial concentration of the curing agent.33
the equilibration time in the rheometer (3 min) before Hypothetically, a perfect thermoset network should have no free

WWW.MATERIALSVIEWS.COM 40142 (5 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

components in the bulk, and the MW should be infinite.35 Table II. Gelation Time (min) of ECO/PA-Based Resins Cured at Different
However, too low of a concentration of curing agent will not Temperatures as Determined from Rheological Experiments
allow the formation of single giant crosslinked macromolecule
of resin because of the limited extent of reactions with the cur- Tcure
ing agent necessary to grow the macromolecule. Also, at a low
ECO/PA (mol/mol) 155 C 170 C 185 C 200 C
curing agent content, the gelation of the system takes an
extended period because of the low concentration of reactive 1.0:1.0 28 18 14 6
groups. Because the final product is a flexible thermoset resin, 1.0:1.5 19 15 12 8
the measurement of viscoelastic properties of the material was 1.0:2.0 13 11 8 6
possible as a result of induced oscillation, even long after the
The italicized numbers were extrapolated from the linear trend.
established gelation period. At higher PA contents, the visco-
Tcure 5 curing temperature.
elastic behavior of the system changed significantly. It can be
seen from Figure 3 and Table II that for samples cured at the
Similarly, increasing the curing temperature at a fixed concen-
same temperature, an increase in the PA content accelerated the
tration of PA led to faster curing rates and shorter gelation
gelation process of the epoxy resin (lower values of gelation
times. As an example of this, Figure 4 shows how G0 of the 1:1
time), and the curing process was completed at a faster rate
ECO/PA resin changed as a function of the curing time at dif-
(increased slope under the linear part of the curve).
ferent isothermal curing temperatures.
At low concentrations of PA and at low curing temperatures,
As shown in Figure 4, the resin cured at 155 C had an extended
the measurement of gelation times for the resins could be
curing period. An increase in the curing temperature accelerated
achieved, as described previously. However, at higher curing
the reaction rate, and the viscoelastic properties of the cured
temperatures, the curing process started before the equilibrium
resin reached their maximum value (under the conditions used)
temperature was achieved, and not surprisingly, this effect was
sooner at high temperatures.
also more pronounced for the resins with high PA contents.
Thus, for the resins with a high PA content cured at high tem- In summary, both the temperature of curing and the concentra-
peratures, the direct measurement of the gelation times was not tion of curing agent were important variables in optimizing the
possible, so an attempt was made to estimate these values by formulation of the ECO/PA epoxy resins. There remains a need
simple extrapolation. to establish which of these variables should best be manipulated
to achieve the desired properties of the final product.
In Table II, the values of the gelation times are highlighted in
gray (italicized numbers) were those estimated from the extrap- DMA
olation of the linear fitting of the experimental data. We In general, epoxy resin formulations might exhibit enhanced or
observed that at lower curing temperatures (155 and 170 C), depressed mechanical properties, depending on the miscibility
increasing the ECO/PA molar ratio from 1:1 to 1:2 decreased and reactivity of the components used. For anhydride-cured
the gelation time by about a factor of 2. epoxy resins, it is expected that the anhydride molecules will
However, at higher curing temperatures, the estimated gelation directly participate in the curing reactions so that the anhydride
time became less dependent on the content of the curing agent. residues will appear as part of the crosslinked structure of the
final thermoset resin. Thus, in the ECO/PA system, the presence
of anhydride (and an increase in its amount up to some limit)

Figure 3. G0 of the ECO and PA-based resin with different molar ratios of
the components and with curing at 155 C as measured with a rheometer. Figure 4. G0 of the ECO and PA-based resin with an equimolar composi-
The ratios of the components are indicated near the respective curves. tion of the components at different temperatures as measured with a rhe-
The black area corresponds to the prepolymer preparation time, and the ometer. The black area corresponds to the prepolymer preparation time,
gray area is the equilibration time in the rheometer. and the gray area is the equilibration time in the rheometer.

WWW.MATERIALSVIEWS.COM 40142 (6 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

should have ultimately increased the formation of the cross- case (Figure 5), an increase in the amount of curing agent led
linked network structure of epoxy and, thus, reduced the poly- to enhanced crosslinking and a smaller quantity of free/side
mer segment mobility and thereby increased the glass-transition chains available for relaxation. Hence, the observed modulus at
temperature (Tg) of the material. low temperatures was lower.
The DMA results in Figure 5 show the temperature dependence With increasing temperature, the large segments of the chain
of G00 and tan d for the ECO/PA resins prepared at different started to move, and the material transformed from a hard glass
molar ratios and cured at a temperature of 155 C. to a viscoelastic state at temperatures above the glass transition.
The area above the Tg (and below the melting point, if it is a
As shown in Figure 5, overall for all of the ECO/PA ratios, the
crystalline polymer) is known as the rubbery plateau, and the
dynamic mechanical behavior of the resins exhibited the typical,
modulus in the plateau region depends on many factors, includ-
classic behavior of G0 . Hence, the DMA plots showed a charac-
ing the ratio of components, the chemical structure of the
teristic glassy plateau at very low temperatures and a dramatic
monomers, the number of crosslinks, and the chain length
decrease in the modulus above the Tg of the resin and reached a
between entanglements.38 In practice, the relative modulus of
rubbery plateau. At low temperatures (below Tg), the resins
the plateau region gives an indication of the relative changes in
showed that G0 was inversely related to the PA content. For the number of crosslinks in a material.40 As shown in Figure 5,
example, G0 of the 1:1 ECO/PA resin was slightly higher at low the resin cured with a low concentration of PA (1:1 mol/mol)
temperatures than that of the other compositions. exhibited a lower modulus in the rubbery plateau region com-
It is well known that changes in the mechanical relaxation of pared to the resin with a high PA content because of the lower
polymer composites and blends can result from both changes in crosslink density. The equimolar system was also characterized
the molecular mobility of the matrix phase due to interactions by the lower values of Tg (see Table III) compared to the other
at the polymer–filler or polymer–polymer interfaces and a compositions studied. The segmental mobility in the highly
mechanical coupling effect between phases.36,37 The transitions crosslinked resins were significantly hindered; hence, they exhib-
below Tg (so called sub-Tg transitions) are associated with the ited a higher modulus in this region.
mechanical performance of materials in the glassy state. These It was clear that in the rubbery plateau region, the dynamic
transitions are described as being due to local motions, bond mechanical properties were mainly determined by the segmental
bending and stretching, and movements in the side chain or mobility of the crosslinked structures in the bulk resin, with the
adjacent atoms in the main chain.38 For instance, this transition assumption that all of the PA reacted with ECO. This assumption
has been attributed to the unreacted molecular segments and/or was justified, as described earlier, because the PA molecule could
inhomogeneities in the epoxy samples arising from dissimilar potentially react with two epoxides, and each molecule of epoxi-
crosslinked regions.39 Thus, higher values of the modulus below dized oil had about four epoxide groups. Hence, even at a high
the Tgs for the ECO/PA system with lower PA contents could content of PA (ECO/PA 5 1:2), the expected stoichiometric ratio
have been due to the relaxation of the unbounded/unreacted of reactive moieties of these components was close to 1:1, and
epoxy monomer, depending on the concentration of the com- the presence and contribution of unreacted PA to the mechanical
ponents. Similar behavior was also described for cured thermo- performance was negligible. However, the presence of the higher
set resins, where increased freedom of movement in the side number of PA units in the material may have also contributed to
chains increased the strength of the transition.38 Thus, in this
the modulus in the rubbery region because of some degree of
change in the chemical structure within the material.
The temperature dependence of tan d of the resins with differ-
ent ratios of the components are also plotted in Figure 5. This
indicates similar trends to those discussed previously. As
expected, the intensity of the tan d peak for equimolar ECO/PA
was much higher than the peak of resins with higher amounts
of PA. The intensity of the tan d peak was directly related to
the initial ECO content in the system as the peak was associated
with the relaxation of ECO units. The intensity of the tan d
gradually decreased with increasing PA content. A significant
decrease in the tan d peak intensity was observed in the resin
with a high content of PA. This was expected because the for-
mation of crosslinked structures restricted the molecular
motions, and this caused the amount of energy that could dissi-
pate throughout the bulk volume to decrease dramatically.17
This could have caused the Tg to shift to higher temperatures
with increasing PA content, as shown in Figure 5.
The Tg is usually determined by the peak temperature of the
Figure 5. G0 and tan d as functions of the temperature for the resin with dynamic G00 or by the temperature at which tan d exhibited the
different ratios of the components and with curing at 155 C. peak.37,41 Here, the Tg’s of the resins were measured from the

WWW.MATERIALSVIEWS.COM 40142 (7 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

Table III. Tgs ( C) of the ECO/PA-Based Resins Cured at Different Temperatures as Determined from DMA and Conventional DSC Experiments

Tcure

DMA (tan d) DSC

ECO/PA (mol/mol) 155 C 170 C 185 C 200 C 155 C 170 C 185 C 200 C
1.0:1.0 23.6 6 1.2 24.4 6 1.4 24.4 6 0.3 23.3 6 0.6 224.1 6 0.3 225.2 6 0.2 224.4 6 0.1 223.9 6 0.1
1.0:1.5 13.1 6 0.4 13.5 6 0.9 15.5 6 1.1 17.7 6 0.3 216.8 6 0.6 217.2 6 1.4 212.4 6 1.8 210.3 6 0.1
1.0:2.0 37.8 6 0.6 40.4 6 0.6 39.0 6 0.1 39.1 6 0.7 16.1 6 1.0 16.8 6 5.5 17.3 6 1.1 18.2 6 0.3

The uncertainties are standard deviations of at least duplicates.

Tcure 5 curing temperature.

tan d, that is, the peak of tan d (Table III). Although, DMA is two techniques arose because of the different processes that
believed more sensitive than DSC to measure the Tg of cross- DSC and DMA measure, as discussed in many previous
linked thermoset materials,42 we also applied conventional DSC works.43,44
experiments (in modulation mode) to measure the Tg values of
Thus, the Tg (and consequently, the mechanical properties) of
the cured materials (DSC traces are not presented here). The
the ECO/PA system largely depended on the amount of curing
data presented in Table III indicates that the Tg of the resins
agent. Hence, we concluded that the curing process could be
varied little with the curing temperature, whereas an increase in
greatly accelerated by an increase in the curing temperature
the PA content significantly increased the Tg of the resin. The
without a significant impact to the Tg. Alternatively, an
influence of the curing temperature and curing agent content
improvement in the Tg of the material at a given temperature
on Tg is illustrated in Figure 6, which shows the Tg values deter-
of curing could be achieved by the variation of the amount of
mined by DMA.
curing agent.29
It is shown in Figure 6(a) that the Tg’s of all of the systems
studied were not dependent or were only weakly dependent on
CONCLUSIONS
the curing temperature for the ECO/PA system over the temper-
ature range studied. However, the influence of the content of A canola-oil-based epoxy resin was successfully produced and
the curing agent on the Tg of the final resin was significant [Fig- applied to the making of thermoset resins with PA as a curing
ure 6(b)]. This strong dependence was observed as a linear agent. The curing process of the resin and the thermomechani-
regression plot of these data on the basis of the average values cal properties of the resin was characterized with different
of Tg at a given composition. techniques.
Although, the values of Tg determined with DSC were signifi- The results show that careful selection of a combination of the
cantly different from those determined by DMA, they followed curing temperature and ECO/PA ratio could be used to create a
exactly the same trends with the ECO/PA ratios and curing thermoset resin with preferred thermomechanical properties.
temperature. The observed difference in Tg measured by the For example, a thermoset resin with a particular modulus or Tg
could be prepared at faster curing rate through the variation of
the curing temperature alone. Alternatively, with fixed curing
conditions, thermoset resins with diverse mechanical properties
could be prepared by the variation of the amount of curing
agent. The combination of these factors could be especially use-
ful for the creation of composites with temperature-sensitive
components, for example, lignocellulosic materials. In the man-
ufacturing of these products (in these cases, there may be upper
temperature limits to prevent the burning of lignocellulosic
materials), the speed of curing and the final mechanical proper-
ties can be enhanced through an increase in the concentration
of curing agent up to a certain amount. However, it should be
noted that extreme curing rates may lead to structures with
localized areas of high stress in the final resin, while excess
amounts of curing agents, from the stoichiometric ratio of reac-
tive groups, may lead to inhomogeneous structure formation.
Such inhomogeneities may significantly affect the impact prop-
Figure 6. (a) Temperature and (b) concentration dependence of the Tg for erties of the materials. An important finding during our resin
ECO/PA resins. The error bars are the standard deviations of at least preparation was that the oligomeric prepolymers of ECO and
duplicates. The cumulative linear regression was plotted with the average PA were relatively stable under ambient conditions and when
values of Tg for the respective compositions. they had been stored for several days without curing. The resins

WWW.MATERIALSVIEWS.COM 40142 (8 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142

ARTICLE WILEYONLINELIBRARY.COM/APP

reported here could also be used as aldehyde-free alternatives 15. Erhan, S. Z. Industrial Uses of Vegetable Oils; American Oil
for aldehyde-containing internal bonding adhesives (melamine, Chemists’ Society: Champaign, IL, 2005.
urea, or phenol formaldehyde) to manufacture lignocellulosic 16. Ionescu, M. Chemistry and Technology of Polyols for Polyur-
fiberboard or particleboard. We demonstrated the applicability ethanes; Rapra Technology: Shrewsbury, United Kingdom, 2005.
of this resin in the making of lignocellulosic fiber- and particle- 17. Wool, R. P.; Sun, X. S. Bio-Based Polymers and Composites;
based biocomposites (not reported here) with a diversity of cel- Academic: Burlington, MA, 2005.
lulosic fibers, fiber mats, and forestry wastes. The increased
18. Hill, K.; Swiecki, B.; Cregger, J. The Bio-Based Materials
diversity in the properties of similar resins could be achieved Automotive Value Chain; Center for Automotive Research:
with other unsaturated plant and animal oils and result from a Ann Arbor, MI, 2012; p 82.
change in functionality. Thus, changing the type and functional-
19. Chian, W.; Timm, D. C. Macromolecules 2004, 37, 8091.
ity of the curing agent or the functionality of the epoxy precur-
sor adds another important variable. For example, the curing 20. Benecke, H. P.; Vijayendran, B. R.; Elhard, J. D. (to Battelle
process could be completed at faster rates with highly functional Memorial Institute). U.S. Pat. 6,797,753 (2004).
epoxidized linseed oil (with an average of 6.5 epoxy groups per 21. Rosch, J.; Mulhaupt, R. Polym. Bull. 1993, 31, 679.
TAG) compared to the moderately functional ECO (with 3.9 22. Crivello, J. V.; Narayan, R. Chem. Mater. 1992, 4, 692.
epoxy groups per epoxidized TAG). 23. Liu, Z. S.; Erhan, S. Z. J. Am. Oil Chem. Soc. 2010, 87, 437.
In summary, thermoset resins with a range of predesigned ther- 24. Gnecco, S.; Pooley, A.; Krause, M. Polym. Bull. 1996, 37, 609.
momechanical properties and reactivities may be produced by 25. La Scala, J.; Wool, R. P. J. Am. Oil Chem. Soc. 2002, 79, 373.
the careful selection of the curing conditions and biobased oil
26. Swern, D.; Billen, G. N.; Findley, T. W.; Scanlan, J. T. J. Am.
epoxide reactants.
Chem. Soc. 1945, 67, 1786.
27. Neff, W. E.; Mounts, T. L.; Rinsch, W. M.; Konishi, H.;
ACKNOWLEDGMENTS Elagaimy, M. A. J. Am. Oil Chem. Soc. 1994, 71, 1101.
The authors acknowledge the support for this work provided by 28. Omonov, T. S.; Kharraz, E.; Curtis, J. M. J. Am. Oil Chem.
the Alberta Crop Industry Development Fund, Ltd., and Natural Soc. 2011, 88, 689.
Sciences and Engineering Research Council (of Canada) (NSERC). 29. Omonov, T. S.; Curtis, J. M. U.S. Prov. Pat. 61,727,357 (2012).
E. Kharraz is acknowledged for his technical assistance. 30. Johnson, D. S. U.S. Pat. 4,398,013 (1983).
31. Wallenberger, F. T.; Weston, N. E. Natural Fibers, Plastics
and Composites; Kluwer Academic: Dordrecht, 2004.
REFERENCES
32. Thomas, R.; Yumei, D.; Yuelong, H.; Le , Y.; Moldenaers, P.;
1. Craver, C. D.; Carraher, C. E. Applied Polymer Science: 21st Weimin, Y.; Czigany, T.; Thomas, S. Polymer 2008, 49, 278.
Century; Elsevier Science: Oxford, United Kingdom, 2000.
33. Nielsen, J. A.; Chen, S. J.; Timm, D. C. Macromolecules
2. Petrie, E. M. Epoxy Adhesive Formulations; McGraw-Hill: 1993, 26, 1369.
New York, 2006.
34. Parzuchowski, P. G.; Kizlinska, M.; Rokicki, G. Polymer
3. Pascault, J.-P.; Williams, R. J. J. Epoxy Polymers: New Mate- 2007, 48, 1857.
rials and Innovations; Wiley-VCH: Weinheim, 2010.
35. Flory, P. J. Principles of Polymer Chemistry; Cornell Univer-
4. Garton, A.; Mclean, P. D. (to Canadian Patents & Develop- sity Press: New York, 1953.
ment, Ltd.). U.S. Pat. 4,851,484 (1989).
36. Gibson, R. F. Principles of Composite Material Mechanics;
5. Benecke, H. P.; Vijayendran, B. R.; Garbark, D. B.; Mitchell, CRC: Boca Raton, FL, 2007.
K. P. Clean-Soil Air Water 2008, 36, 694.
37. Nielsen, L. E.; Landel, R. F. Mechanical Properties of Poly-
6. Guo, A.; Demydov, D.; Zhang, W.; Petrovic, Z. S. J. Polym. mers and Composites; Marcel Dekker: New York, 1994.
Environ. 2002, 10, 49.
38. Menard, K. P. Dynamic Mechanical Analysis: A Practical
7. Petrovic, Z. S.; Guo, A.; Zhang, W. J. Polym. Sci. Part A: Introduction; CRC: Boca Raton, FL, 2008.
Polym. Chem. 2000, 38, 4062.
39. Mikolajczak, G.; Cavaille, J. Y.; Johari, G. P. Polymer 1987,
8. Gandini, A.; Belgacem, M. N. J. Polym. Environ. 2002, 10, 105. 28, 2023.
9. Pollock, J. W. Abstr. Pap. Am. Chem. Soc. 2004, 227, U304. 40. Lobo, H.; Bonilla, J. V. Handbook of Plastics Analysis; Mar-
10. Petrovic, Z. S. Polym. Rev. 2008, 48, 109. cel Dekker: New York, 2003.
11. Curtis, J.; Liu, G.; Omonov, T.; Kharraz, E. WIPO Int. Publ. 41. ASTM E 1640-09; ASTM International: West Conshohocken,
WO 2012/009801 A1 (2012). PA, 2009; p 6.
12. Lee, K. W.; Hailan, C.; Yinhua, J.; Kim, Y. W.; Chung, K. W. 42. Seyler, R. J. Assignment of the Glass Transition; ASTM
Korean J. Chem. Eng. 2008, 25, 474. International: Baltimore, MD, 1994.
13. Mungroo, R.; Goud, V. V.; Pradhan, N. C.; Dalai, A. K. 43. Coombs, C. F. Printed Circuits Handbook; McGraw-Hill:
Asia-Pac. J. Chem. Eng. 2011, 6, 14. New York, 2008.
14. Hwang, H. S.; Erhan, S. Z. J. Am. Oil Chem. Soc. 2001, 78, 44. Gabbott, P. Principles and Applications of Thermal Analysis;
1179. Blackwell: Oxford, 2008.

WWW.MATERIALSVIEWS.COM 40142 (9 of 9) J. APPL. POLYM. SCI. 2014, DOI: 10.1002/APP.40142