Synthesis, characterization, and cross-linking strategy of a quercetin-based epoxidized monomer as a naturally-derived replacement for BPA in epoxy resins

Article abstract of DOI:10.1002/cssc.201600392

The natural polyphenolic compound quercetin was functionalized and cross-linked to afford a robust epoxy network. Quercetin was selectively methylated and functionalized with glycidyl ether moieties using a microwave-assisted reaction on a gram scale to a

Full text of DOI:10.1002/cssc.201600392

DOI: 10.1002/cssc.201600392
Full Papers
Synthesis, Characterization, and Cross-Linking Strategy of
a Quercetin-Based Epoxidized Monomer as a Naturally-
Derived Replacement for BPA in Epoxy Resins
[
a]
[b]
[a]
[b]
Samantha L. Kristufek, Guozhen Yang, Lauren A. Link, Brian J. Rohde,
[
b]
[a]
Megan L. Robertson, and Karen L. Wooley*
The natural polyphenolic compound quercetin was functional-
ized and cross-linked to afford a robust epoxy network. Quer-
cetin was selectively methylated and functionalized with gly-
cidyl ether moieties using a microwave-assisted reaction on
a gram scale to afford the desired monomer (Q). This querce-
tin-derived monomer was treated with nadic methyl anhydride
tion (T ) temperatures] and comparable mechanical properties
d
(i.e., Young’s Modulus, storage modulus) to that of DGEBA-
NMA. However, it had a lower tensile strength and higher
flexural modulus at elevated temperatures. The application of
naturally derived, sustainable compounds for the replacement
of commercially available petrochemical-based epoxies is of
great interest to reduce the environmental impact of these
materials. Q-NMA is an attractive candidate for the replace-
ment of bisphenol A-based epoxies in various specialty engi-
neering applications.
(NMA) to obtain a cross-linked network (Q-NMA). The thermal
and mechanical properties of this naturally derived network
were compared to those of a conventional diglycidyl ether bi-
sphenol A-derived counterpart (DGEBA-NMA). Q-NMA had sim-
ilar thermal properties [i.e., glass transition (T ) and decomposi-
g
Introduction
Epoxy resins are one of the most widely used thermosetting
systems across a vast number of industries including transpor-
tation (e.g., automobiles and airplanes), coatings, wind turbine
search is devoted to finding a DGEBA replacement that is safe
and sustainable without greatly sacrificing the mechanical in-
tegrity or thermal properties of the epoxy resin.
[
1]
blades and electronics. It has been estimated that the epoxy
industry was worth nearly 6 billion USD in 2013 and will grow
There are many ways to introduce sustainability into epoxy
[5]
resins, such as the addition of natural toughening agents, in-
corporation of natural cross-linkers or synthesis of naturally de-
[
2]
to more than 9 billion USD by 2019. This enormous industry
is one of the many polymer/plastic markets that still largely
rived monomers through epoxidation or glycidyl ether func-
[3]
[1,3]
rely on nonrenewable petrochemicals as starting materials.
tionalization.
A natural-product-based substitute can be se-
Replacing these starting materials with sustainable and renew-
able resources would be beneficial for both environmental and
lected depending on the petrochemical-based monomer that
needs to be replaced, its properties and the intended applica-
tion of the material. Currently, natural product derivatives such
as itaconic acid, isosorbide and furan derivatives are platform
[
1]
financial reasons. One common industrially used monomer
for epoxy resins is the diglycidyl ether of bisphenol A (DGEBA),
which provides excellent mechanical and thermal properties.
Not only is this monomer derived from nonrenewable petro-
chemicals, but there have also been several health concerns
associated with the precursor, bisphenol A (BPA), including its
[6]
chemicals for the synthesis of epoxy resins. Although the use
of natural toughening agents and cross-linkers reduces the pe-
troleum-derived content of epoxy resins, in some systems,
harmful BPA still remains as a building block in the DGEBA-
based system. As plant-based substitutes of BPA, epoxidized
[
4]
effect as an estrogen hormone disruptor. A central area of re-
[
7–9]
natural products, derived from sources such as plant oils,
[10–12]
[13,14]
[15–18]
sugars,
ly derived materials,
mers. Alternatively, functional groups inherent to the natural
lignin,
rosin acids
1,3]
and many other natural-
[a] S. L. Kristufek, Dr. L. A. Link, Prof. K. L. Wooley
[
Department of Chemistry
have been explored as epoxy mono-
Department of Chemical Engineering
Departments of Materials Science & Engineering
Texas A&M University
[19,20]
product, such as the hydroxyl groups on tannic acid,
have
been used for ring opening of epoxides in cross-linked net-
works. Replacing a commercially available product such as
a DGEBA-derived epoxy resin requires the new material to
have the desired thermal, mechanical and other properties for
the specific applications. It has been proposed that a class of
natural polyphenolic compounds may be utilized as the start-
College Station, Texas 77842-3012 (United States)
E-mail: wooley@chem.tamu.edu
[
b] G. Yang, B. J. Rohde, Prof. M. L. Robertson
Department of Chemical and Biomolecular Engineering
University of Houston
Houston, Texas 77204-4004 (United States)
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cssc.201600392.
[1]
ing materials for monomers in epoxy networks.
ChemSusChem 2016, 9, 1 – 9
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Quercetin is one such natural polyphenol that has the po-
tential to be functionalized and act as a monomer in cross-
linked epoxy networks. The first consideration is its wide avail-
a novel, quercetin-based monomer, 5,3’-diglycidyl ether-3,7,4’-
trimethylquercetin (2), was explored using traditional and mi-
crowave synthesis conditions. This monomer was then cross-
linked with nadic methyl anhydride (NMA) to form an epoxy
resin (Q-NMA), which was characterized thermally and mechan-
ically and compared to a DGEBA-based system prepared with
the same cross-linker. The motivation for the synthesis of these
quercetin-based materials was to investigate their potential as
direct replacements for the BPA-based epoxy resins used in
many applications.
[
21]
ability in commonly ingested fruits and vegetables, which
has led to numerous studies investigating its health benefits
[22,23]
including anti-inflammatory and antioxidant properties.
Methylated derivatives of quercetin have also been isolated in
[
24]
nature and studied for their biological properties. Quercetin
is not as cost efficient as petrochemical-based monomers at
current market prices but extraction methods have been inves-
tigated as the demand for quercetin has increased over the
past decade, in part owing to its numerous therapeutic appli-
[
25]
cations. Isolation methods from various fruits and vegetables Results and Discussion
[
26]
have recently been reported. A promising green method has
been used to extract quercetin glycosides from onion waste
using hot water extraction followed by enzymatic hydrolysis to
Synthesis and characterization of epoxy monomer and
cross-linked networks
[
27]
afford quercetin. Quercetin has been used as a cross-linker
in epoxy chemistry in which the hydroxyl groups were reacted
with sorbitol polyglycidyl ether and wood flour to form bio-
The epoxy monomer was prepared by selective alkylation of
the hydroxyl groups in two steps. The regioselectivity of quer-
cetin has been thoroughly studied and has shown to allow
good selectivity for alkylation of the 3, 7 and 4’ positions over
[
28]
composites. To date, quercetin has not been explored as a re-
placement for DGEBA in epoxy resins, except as a co-monomer
[30,31]
the 5 and 3’ hydroxyl groups,
which provided a convenient
[
28]
together with sorbitol, or as a cross-linker in an epoxidized
route for the stepwise selective methylation and installation of
the epoxide-containing glycidyl moieties to form the epoxy
monomer (Scheme 1). Methylation using methods similar to
[29]
natural rubber/polycaprolactone, both in biocomposites. The
rigid tricyclic structure of quercetin is expected to impart desir-
able thermal and mechanical properties to the resulting epoxy
resins. In addition, quercetin undergoes unique chemical reac-
tivity with selective alkylation of the hydroxyl groups in the
[32]
those reported previously, but on a larger scale, afforded
3,7,4’-trimethylquercetin (1). Briefly, quercetin was allowed to
dissolve in N,N-dimethylformamide (DMF) at room temperature
under an inert atmosphere, and was then treated with potassi-
um carbonate followed by the addition of methyl iodide. The
reaction was monitored by TLC and allowed to proceed for
24 h. The resultant mixture contained observable amounts of
remaining starting material along with the di-, tri- and tetra-
methylated quercetin products, which were separated by
column chromatography and recrystallized to afford the de-
sired tri-methylated product in 42% yield. The product was
[
30,31]
order 4’>7>3>3’>5 (Figure 1),
allowing for direct func-
tionalization at positions 3’ and 5 with glycidyl ether groups,
followed by reaction with a cross-linking agent to synthesize
networks that have the potential to replace BPA-based epoxy
resins.
This research entails the detailed evaluation of natural-prod-
uct-based alternatives to DGEBA epoxy resins. The synthesis of
1
13
confirmed using H NMR, C NMR, IR and ESI-MS (Figure S1
and S2, Supporting Information). The regioselectivity was con-
firmed by single crystal X-ray analysis (Figure S6). The motiva-
tion for this specific selective protection was three-fold. Firstly,
1
has previously been extracted as the natural product, ayanin
[33]
from Croton schiedeanus, a plant found across Central and
South America. The bioactivity of this natural product has
[34–36]
been studied against several cell lines.
Secondly, the alky-
lation of the three hydroxyls leaves the two remaining hydrox-
yl sites on quercetin available for functionalization with glycid-
yl ether moieties, mimicking the two epoxide functionalities
on DGEBA. Finally, limiting the number of free hydroxyl groups
Figure 1. The numbering of the hydroxy groups for the selective alkylation
of quercetin.
Scheme 1. Synthesis of the pre-monomer and monomer used for cross-linking Q-NMA networks.
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
present on the monomer would limits the number of reactive
functionalities present during the cross-linking reaction and re-
duces the number of side reactions.
conditions were found to be effective at directing the reaction
towards the desired product. In particular, limiting the opening
of the epoxide or accelerating the closing of the ring to reform
the epoxide could assist in promoting the forward reaction
and increasing the yield. The Pchelka group has demonstrated
that it is possible to activate ionic species in the analogous re-
actions for the preparation of 3-aryloxy-1,2-epoxypropane
The second step, O-alkylation to synthesize the epoxy mo-
nomer (2), was first attempted using traditional synthesis con-
[
1]
ditions for the addition of an oxirane to a phenol (Table 1, en-
tries 1–2) utilizing a strong base (NaH) or an aqueous base so-
[40]
using microwave-assisted reaction conditions. In conjunction
with changing the heating method, potassium carbonate was
used in addition to sodium hydroxide to further increase the
yield. The conditions used in the studies of the 3-aryloxy-1,2-
epoxypropanes were applied to the formation of the querce-
tin-based monomers.
Table 1. Experimental conditions used for the attempted synthesis of the
monomer, 3’,5-diglycidyl ether-3,7,4’-trimethylquercetin, (2).
Although the O-alkylation of simple phenols to afford
3-aryloxy-1,2-epoxypropanes has been studied extensively, the
specific conditions for the alkylation of 1 required optimization
owing to its unique characteristics. The most significant chal-
lenge was the limited reactivity of the 5-OH in quercetin and
its analogs owing to hydrogen bonding with the ketone in the
C ring (Figure 1), which was strongly suggested to be occur-
[
a]
Entries
Reagent Base+Additive
Solvent T [8C] t [h] Yield [%]
1
2
3
4
3a
3a
3a
3b
3b
3c
3c
3a
3a
NaH
NaOH+TBAB
DMF
0–25 120
–
–
–
H
2
O
80
80
120
120
K
K
K
K
K
2
CO
2
CO
2
CO
2
CO
2
CO
3
3
3
3
3
DMF
DMF
DMF
[41]
80
80
120 12
120
ring, as shown in the crystal structure of 1 (Figure S6). Addi-
tionally, the 3’-position of 1 is sterically affected by the ortho-
[
b]
b]
5
6
7
8
9
–
DMF
DMF
80
80
110
110
120 34
120 25
0.5 49
0.5 38
[40]
[
methoxy substituent. By adapting the literature conditions
c
to the current reaction, 2 was prepared in reasonable yield in
gram-scale quantities (Table 1, entry 9, 38%). The conditions
were altered by extending the reaction time and increasing
the number of equivalents of epichlorohydrin (3a). TBAB was
also replaced with tetrabutylammonium iodide (TBAI). More
importantly, this method significantly reduced the reaction
time from 120 h to 0.5 h. To further confirm the regioselectivity
NaOH+K
NaOH+K
2
CO
CO
3
3
+TBAI EtOH
+TBAI EtOH
[
b,c]
2
[
a] Standard conditions were performed on a 100 mg scale; [b] Performed
on a gram scale; [c] Performed under microwave conditions; TBAB: tetra-
butylammonium bromide, TBAI: tetrabutylammonium iodide.
1
lution (NaOH) with a phase transfer catalyst tetrabutylammoni-
um bromide (TBAB); the reactions did not provide the desired
product after 5 days, as monitored by TLC, even at elevated
temperatures (808C). Subsequently, traditional conditions pre-
viously used to functionalize quercetin using K CO as a solid
of the di-substituted product was confirmed by H and
13
C NMR spectroscopies and high-resolution mass spectrometry
(NMR spectra are shown in Figure S3 and S4). To further con-
firm the regioselectivity of the final product structure, single
[42]
crystal X-ray analysis of 2 was performed (Figure S7).
2
3
[
32,37]
base and the solvent, DMF
a 12% yield of the desired product using epibromohydrin
3b). However, when the reaction was scaled up to a gram
(Table 1, entries 3–5) delivered
Using the prepared monomer, cross-linked networks were
synthesized by allowing the oxirane groups of 2 to undergo
reaction with nadic methyl anhydride (NMA), a commercially
available cross-linker. The catalyst chosen to accelerate the
(
scale, the only isolated product was the mono-reacted glycidyl
ether, 3’-glycidyl ether-3,7,4’-trimethylquercetin (Table 1, en-
tries 4–5). When tosylate (ꢀOTs) was used as the leaving group
of the oxirane starting material (3c), the large-scale reaction
proceeded in five days with a poor yield (Table 1, entries 6–7).
Additionally, this reaction required the extra step of converting
glycidol to the (ꢁ)-oxiran-2-ylmethyl 4-methylbenzenesulfo-
ꢁ
cross-linking reaction was ANCAMINE K54 (2,4,6-tris(dimethyla-
minomethyl)phenol) (Figure S5) at a concentration of 3 phr
(parts per hundred resin). This tertiary amine was used to pro-
mote alternating co-polymerization of NMA and an epoxy-
based monomer through chain growth anionic polymeri-
[43,44]
zation.
Side reactions, such as homopolymerization of the
[
38,39]
nate (3c) following a literature procedure,
which proceed-
oxirane ring and etherification, can occur to further induce
cross-linking. Although the natural-based monomer was ex-
pected to undergo the same types of reactions as previously
studied commercially available monomers, the physical proper-
ties of 2 led to a unique curing process.
There were several challenges in developing a curing proto-
col for Q-NMA, including the high melting point of the mono-
mer (1608C), evaporation of the cross-linker (>908C) (Fig-
ure S10) and rapid reaction (less than 1 min) of the starting
materials in the presence of the catalyst at elevated tempera-
tures. First, to determine an appropriate curing protocol for Q-
NMA, the reaction kinetics were examined using non-isother-
ed in excellent yield but increased the overall number of linear
steps for the synthesis of 2. A variety of different conventional
heating methods for the addition of multiple glycidyl ether
functionalities did not allow the reaction to be performed at
a large scale to afford sufficient materials for cross-linking in
the form of mechanical testing bars.
Considering the two competing mechanisms that can take
place to install the epoxide functionality using epichlorohydrin
(
3a), 1) the S 2 reaction with the alkyl halide and 2) the S 2’ re-
N N
action, which involves opening of the epoxide followed by ref-
ormation of the epoxide to replace the halogen, alternative
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ꢀ
1
mal heating at 108Cmin from 40–2008C using differential
scanning calorimetry (DSC) (Figure 2) and ending at 2008C to
avoid degradation of the material. The DSC data indicated
curing over a broad temperature (ꢂ125–2008C), with peak re-
activity located at approximately 1808C. As the quercetin-
perature, until no noticeable changes occurred (Figure S8). The
required post-curing time was 30 min at 1608C. In summary,
the final curing protocol chosen for Q-NMA included a heated
mixing step followed by a two-stage curing that involved an
initial curing step at 908C for 2 h and a post-curing at 1608C
for 30 min.
The quercetin-network was characterized by IR spectroscopy
(
Figure 3). The IR spectrum of NMA (Figure 3b) shows peaks at
ꢀ
1
ꢀ1
1
852 cm (asymmetrical C=O stretch) and at 1775 cm (sym-
metrical C=O stretch), each of which disappeared upon forma-
tion of Q-NMA (Figure 3c), indicating the consumption of the
Figure 2. Assessment of the conditions used to cure 2 and NMA to produce
Q-NMA, with evaluation of heat flow versus temperature by DSC.
based epoxide monomer (2) was not soluble in NMA at room
temperature, the first step in the protocol was the melting of
2
(1608C) and mixing it with NMA in the absence of the cata-
lyst for 3 min at 1608C. Without the K54 catalyst, the mixture
did not appear to visibly gel (observed up to 2 h at 1608C).
After mixing the two monomers at 1608C for 3 min, the mass
loss was 0.60ꢁ0.07% (Figure S11). Given the low expected
conversion of a thermoset proceeding via chain growth poly-
[
43]
merization without the presence of an initiator, it is not an-
ticipated that significant conversion of the monomer during
the 3 min mixing time. The mixture was subsequently cooled
to room temperature (258C) (during which it remained homo-
ꢁ
geneous), followed by addition of ANCAMINE K54 and heating
to the desired mixing/curing temperature. Mixing tempera-
tures ranging from 30–908C were tested at 108C increments.
At low temperatures the mixture was too viscous for stirring;
therefore, the mixing could only proceed efficiently at 908C to
ensure homogeneous distribution of the catalyst. After the ad-
ꢁ
dition of ANCAMINE K54, the mixture was stirred for no
longer than 3 min to avoid gelation of the materials in the
vials. Following the mixing process, the mixture was cast into
molds preheated at 908C to ensure a continuous cross-linking
reaction. The networks were allowed to undergo reaction for
2
h at 908C. Thermogravimetric analysis (TGA) confirmed that
NMA did not evaporate at this temperature (Figure S10). The
curing time was determined by measuring the glass transition
temperature (T ) of the bar every 30 min during the curing at
g
9
08C until the T was constant. Subsequently, the resin was
g
post-cured at a temperature higher than the network T and at
g
the melting point of the monomer to help drive the reaction
to completion. The post-curing time was also determined by
Figure 3. IR spectra of (a) monomer 2, (b) cross-linker (NMA) and (c) cured
monitoring the T as the resin was held at the post-curing tem-
network (Q-NMA).
g
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
anhydride functionality. The anhydride absorbances were re-
Table 2. Thermal and mechanical properties of the cross-linked networks
Q-NMA and DGEBA-NMA.
ꢀ
1
placed by a characteristic ester peak at 1736 cm
(C=O
stretch) indicating the formation of the ring-opened product.
The Q-NMA spectrum also contained a strong ester peak at
Property
Q-NMA
DGEBA-NMA
ꢀ
1
[a]
1
157 cm (CꢀO stretch) for the CꢀO stretching vibration and
Tg [8C]
134
266
37ꢁ8
1.3ꢁ0.4
3.5ꢁ0.2
1.72
3.49
7.7
6.67ꢂ10
137
284
66ꢁ15
2.8ꢁ0.9
3.0ꢁ0.2
1.75
ꢀ
1
5%
T
d
[8C]
broad, albeit weak peak at 3675–3152 cm (OꢀH stretch) indi-
cating the presence of the ꢀOH group of the unreacted chain
ends. The combination of the IR peaks indicates that the reac-
tion proceeded and cross-linking of the Q-NMA network oc-
curred to afford the desired cross-linked materials.
[
b]
tensile strength [MPa]
elongation at break [%]
Young’s modulus [GPa]
storage modulus [GPa]
flexural modulus [GPa]
rubbery modulus, e’ [MPa]
[
b]
[b]
[
b]
[
b]
3.48
[
c]
14.0
1.21ꢂ10
To directly compare Q-NMA and DGEBA cross-linked with
NMA (DGEBA-NMA), the DGEBA-NMA network was optimized
by applying the same catalyst loading (3 phr) and curing tem-
peratures. The reaction time was optimized for each monomer.
Although the DGEBA-NMA curing kinetics have been previous-
ꢀ
3
ꢀ4
ꢀ3
n
c
[molcm
]
[
a] Measured by DSC; [b] storage modulus, flexural modulus, tensile
strength, % elongation at break and Young’s Modulus were measured at
58C; [c] rubbery modulus was measured at 1908C.
2
[
45,46]
ly reported,
Figure 2 for a comparison to the quercetin-based network
Scheme S1). The same process was applied to optimize the
the trace for the mixture curing is shown in
(
with high tensile strength and modulus, which was later ex-
plored.
curing protocol times for the DGEBA-NMA network. In this
case, the DGEBA monomer was a liquid at room temperature,
therefore, there was no need for a complex mixing process.
The DGEBA/NMA/K54 mixture was stirred at 908C for 5 min to
ensure a homogeneous mixture before casting into the hot
molds heated to 908C. The optimal conditions for the DGEBA-
NMA reaction were determined to be curing for 1 h at 908C
followed by 2 h at 1608C, by monitoring the change in Tg
during the second step. IR spectroscopy confirmed the pres-
ence of a cross-linked network, with the appearance of an
Mechanical properties of the Q-NMA and DGEBA-NMA net-
works
The dynamic moduli were measured for each system in ten-
sion and in three-point-bending mode using DMA (Figure 4A
and 4B). The tensile storage modulus behavior with respect to
temperature for the two systems was similar. Under tension,
the temperature range at which the network is expected to
ꢀ1
ester peak at 1737 cm (C=O stretch) and a strong ester peak
ꢀ
1
at 1152 cm (CꢀO stretch) as well as the disappearance of the
anhydride peaks from NMA. The faint peak at 3645–3140 cm
ꢀOꢀH stretch) is indicative of the hydroxyl group of the ring-
behave as a glassy solid (T<T
which the network is expected to behave as a rubbery solid
(T <T<T ) are equivalent for the two systems. Conversely,
g
) and the temperature range at
ꢀ1
(
g
d
opened side reaction, similar to the one that appears in the Q-
NMA system. The IR peaks of DGEBA-NMA were similar to
those for Q-NMA, with the appearance of the hydroxyl and
ester peaks and the disappearance of the anhydride peak (Fig-
ure S13). Following confirmation of the synthesis of the epoxy
resins, the thermal and mechanical properties of the two sys-
tems were measured and compared.
under flexural stress, DGEBA-NMA began to transition to its
rubbery state at approximately 708C, a lower temperature
than the Q-NMA transition (ꢂ1208C). From approximately 70–
1708C, Q-NMA exhibited a flexural modulus which was greater
than that of DGEBA-NMA. At 1308C Q-NMA exhibited a flexural
modulus which was two orders of magnitude greater than
that of DGEBA-NMA. Therefore, for elevated temperature appli-
cations (50–1508C), Q-NMA is expected to be more resistant to
bending (flexural stress/strain) compared to DGEBA-NMA.
Using the theory of rubber elasticity, the cross-link density (nc),
defined as the number of moles of elastically effective network
chains per cubic centimeter of sample, can be calculated using
the following equation:
Thermal properties of the Q-NMA and DGEBA-NMA net-
works
The two systems had similar thermal properties. The decompo-
sition temperatures (T ) of Q-NMA and DGEBA-NMA were 266
d
0
and 2848C, respectively, measured at the 5% mass loss by TGA
e
ð1Þ
uc ¼
(
Table 2). The glass transition temperatures, T , of Q-NMA and
3RT
g
DGEBA-NMA measured by DSC were 1348C and 1378C, respec-
tively, (Table 2). Using dynamic mechanical analysis (DMA) in
tensile mode, the temperatures at which the maximum tan d
where R is the gas constant, T is the temperature and e’ is the
rubbery modulus at 1908C measured using DMA in tension
mode. DGEBA-NMA has a higher cross-link density than that
of the natural-product-based counterpart. Although there is
a considerable difference in cross-link density between the dif-
[47]
occurred (which are proportional to T ), were determined to
g
be 1708C and 1718C for Q-NMA and DGEBA-NMA, respectively.
Given that the two epoxy resins have similar thermal proper-
ties, it is expected that replacement of DGEBA-NMA with Q-
NMA could occur in applications covering a wide range of
ferent networks, the T , flexural modulus and storage modulus
g
are similar, which is possibly owing to the rigidness of querce-
tin.
temperatures. High T values of the systems tend to correlate
g
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Young’s Modulus, storage modulus and flexural modulus, are
comparable to that of DGEBA-NMA, indicating the quercetin-
based epoxy monomer is appropriate for the replacement of
DGEBA in many epoxy resin applications.
Conclusions
A novel monomer from a natural polyphenolic starting materi-
al was synthesized in two steps. The monomer has the poten-
tial to replace the diglycidyl ether of bisphenol A (DGEBA) and,
therefore, to eliminate the use of bisphenol A (BPA) in epoxy
networks. The functionalization of the monomer with diglycid-
yl ether groups was performed using microwave synthesis con-
ditions, which led to an increase in the yield and the scale, as
well as a greatly reduced reaction time. The quercetin-based
monomer was cross-linked with nadic methyl anhydride (NMA)
to afford the cross-linked network (Q-NMA). The thermal and
mechanical properties of Q-NMA were compared to that of
DGEBA cross-linked with NMA (DGEBA-NMA), at optimized
curing conditions. The two systems have similar decomposition
and glass transition temperatures and many comparable me-
chanical properties including Young’s modulus, flexural modu-
lus and storage modulus at 258C. At elevated temperatures, Q-
NMA exhibited a flexural modulus which was greater (by two
orders of magnitude at 1308C) than that of DGEBA-NMA. This
naturally derived monomer shows tremendous potential to re-
place BPA in practical applications.
Experimental Section
Materials and methods
Quercetin was purchased from Cayman Chemicals. Epichlorohydrin
and nadic methyl anhydride (NMA) were purchased from Sigma–
Aldrich. Sodium hydroxide was purchased from EMD Millipore. Po-
tassium carbonate and tetrabutylammonium iodide were pur-
chased from Alfa Aesar. DER 331 (DGEBA) was received from Dow
Chemical and ANCAMINE K54 (2,4,6-tris(dimethylaminomethyl)phe-
nol) was received from Air Products & Chemicals. All chemicals
were used as received without further purification. The monomer
was synthesized in a CEM Intellivent Explorer microwave reactor
and purified by medium pressure liquid chromatography (MPLC)
Figure 4. (a) Tensile storage modulus as a function of temperature. (b) Flexu-
ral modulus as a function of temperature. (c) Representative stress–strain be-
havior (n)=5–6.
using a CombiFlash R (Teledyne Isco).
f
Spectroscopic characterization
The tensile stress/strain behavior of the Q-NMA and DGEBA-
NMA networks measured through tensile testing at 258C is
shown in Figure 4c and relevant parameters are reported in
Table 2. The average tensile strength, or maximum stress, was
1
13
H (500 MHz) and C (125 MHz) NMR spectra were recorded on
a Varian 500 spectrometer interfaced to a UNIX computer using
VnmrJ software. Chemical shifts were referenced to the residual
proton resonance of the solvent (DMSO-d : 2.50 ppm or CDCl :
6
3
3
7ꢁ8 and 66ꢁ15 MPa for Q-NMA and DGEBA-NMA, respec-
7
.26 ppm). IR spectra were obtained on a Shimadzu IR Prestige at-
tenuated total reflectance Fourier-transform infrared spectrometer
ATR-FTIR). Spectra were analyzed using IRsolution software pack-
tively. Both materials exhibited brittle fracture at less than 3%
elongation. The Young’s modulus, calculated from the linear
portion of the stress–strain curve was 3.5ꢁ0.2 and 3.0ꢁ
(
age (Shimadzu) and were exported to Origin for analysis. The mini-
mum number of scans was 32 with a minimum resolution of 16
and the apodization was set to SqrTriangle.
0
.2 GPa for Q-NMA and DGEBA-NMA, respectively (Table 2, Fig-
ure 4C). Overall, the mechanical properties of Q-NMA, such as
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Thermal characterization
DSC studies were performed on a DSC822 (Mettler-Toledo), with
Synthesis of 5,3’-diglycidyl ether-3,7,4’-trimethylquercetin
(2)
e
ꢀ
1
a heating rate of 108Cmin to determine the melting point (m.p.)
of the monomers, as the peak onset of the transition. The thermal
degradation temperature was analyzed by TGA, using a TA Instru-
ments Q500 analyzer. The sample was heated from 30–8008C at
3,7,4’-trimethylquercetin (2.00 g, 5.81 mmol), sodium hydroxide
(460 mg, 11.5 mmol), potassium carbonate (6.40 g, 46.3 mmol), tet-
rabutylammonium iodide (420 mg, 1.14 mmol), epichlorohydrin
(4.80 mL, 61.4 mmol) and ethanol (3.0 mL) were combined in an
80 mL microwave vessel including a magnetic stirring bar. The so-
lution was heated to 1108C for 30 min with stirring in a microwave
reactor. The solution was filtered and the solid was washed with
ethyl acetate (15 mL). The solvent was removed in vacuo. The
crude product was purified using MPLC/flash column chromato-
graph (gradient of 0–100% hexanes/ethyl acetate). The solution
was concentrated to a yellow solid and washed with tetrahydrofur-
ꢀ1
a rate of 108Cmin in an argon environment (the balance argon
ꢀ
1
purge flow was 40 mLmin and the sample purge flow was
ꢀ
1
6
0 mLmin ). The T and curing protocols were characterized by
g
a TA Instruments Q2000 differential scanning calorimeter, calibrat-
ed with an indium standard, with a nitrogen flow rate of
ꢀ
1
5
0 mLmin . The sample was placed in the calorimeter (in a Tzero
aluminum pan) and was heated from 40–2008C, cooled back to
ꢀ1
4
08C, and subsequently heated to 2008C at a rate of 108Cmin .
an and dried in vacuo to afford 5,3’-diglycidyl ether-3,7,4’-trimethyl-
1
The value of T was determined at the midpoint of the slope of
quercetin as a white solid (1.02 g, 38%). H NMR (500 MHz, CDCl ):
g
3
the second heating using the TA Universal Analysis software.
d=7.77–7.70 (m, 2H), 6.98 (app d, J=8.5 Hz, 1H), 6.52 (app d, J=
2
1
.2 Hz, 1H), 6.38 (app d, J=2.1 Hz, 1H), 4.38 (dd, J=11.3, 2.6 Hz,
H), 4.34 (dd, J=11.3, 3.4 Hz, 1H), 4.15 (dd, J=11.3, 4.3 Hz, 1H),
Mechanical characterization
4.09 (dd, J=11.3, 5.6 Hz, 1H), 3.95 (s, 3H), 3.89 (s, 3H), 3.84 (s, 3H),
.46-3.41 (m, 2H), 3.19 (m, 1H), 2.94 (m, 2H), 2.80 ppm (m, 1H);
3
Strain-to-failure measurements were performed on an Instron 5966
tensile tester with a 2 kN load cell at a strain rate of 10 mmmin
13
C NMR (125 MHz, CDCl ): d=173.7, 163.7, 159.6, 158.6, 152.5,
ꢀ
1
3
1
9
5
1
51.4, 147.6, 147.5, 141.2, 123.3, 122.7, 113.9, 111.3, 109.7, 97.6,
7.6, 93.3, 93.3, 70.4, 69.0, 59.99, 59.96, 55.99, 55.96, 55.8, 55.8,
0.2, 50.1, 45.1, 45.0 ppm; FTIR (ATR): n˜ =3105-3009, 2931-2847,
to obtain tensile strength and elongation at break of both Q-NMA
and DGEBA-NMA at room temperature. Test specimens were dog-
bone-shaped testing bars (following ASTM D638, bar type 5, thick-
ness 2.0 mm). Pneumatic grips were used to affix the sample in the
testing frame at a compressed air pressure of 50 psi. Each measure-
ment was repeated with 5–6 test specimens.
ꢀ1
624, 1607, 1211, 1018, 822 cm ; HRMS (ESI+, m/z): calculated for
+
+
[M+H] : [C H O ] 457.1499, found 457.1482; m.p. 1568C. Regio-
24 25 9
chemistry was confirmed by X-ray analysis (Figure S7).
DMA was performed on a Mettler Toledo TT-DMA system. Samples
with dimensions of 3.32ꢂ2.03ꢂ5.10 mm for Q-NMA and 3.06ꢂ
Synthesis of 5,3’-diglycidyl ether-3,7,4’-trimethyl quercetin-
nadic methyl anhydride (Q-NMA)
1
.75ꢂ5.18 mm for DGEBA-NMA were tested in tension mode. Dy-
namic measurements were recorded at a frequency of 1 Hz, a dy-
namic force of 1 N and a static/dynamic force ratio of 1.5 from 25–
5
,3’-diglycidyl ether-3,7,4’-trimethylquercetin (2) (1.50 g, 3.25 mmol)
ꢀ1
2
008C at a rate of 38Cmin with data sampling interval of 10 s.
and nadic methyl anhydride (1.17 g, 6.50 mmol) were combined in
DMA data were obtained using Triton Laboratory software and ex-
ported to Origin Pro 9.0 for analysis. Three-point bending tests
were performed on bars with dimensions of approximately 2.00ꢂ
a glass vial, heated to 1608C and stirred for 3 min. The mixture
was allowed to cool to room temperature (258C) and ANCAMINE
K54 (45.0 mg) was added. The mixture was heated to 908C, stirred
for 3 min and cast into molds preheated to the same temperature
ꢁ
4
1
.05ꢂ10 mm. The measurements were recorded at a frequency of
Hz, a dynamic force of 1 N and a static/dynamic force ratio of 1.5
(
3
908C). The bars were cured for 2 h at 908C then at 1608C for
0 min to afford the Q-NMA networks. Td =2668C, Td =2978C,
ꢀ
1
from 0–2008C at a rate of 38Cmin with data sampling interval of
0 s.
5%
10%
1
Tg =1348C; FTIR (ATR): n˜ =3676–3152, 3047–3000, 3000–2808,
1
8
780, 1736, 1623, 1436, 1429, 1349, 1325, 1157, 1108, 1020,
18 cm .
ꢀ1
Synthesis of 3,7,4’-trimethylquercetin (1)
The compound was prepared by a procedure adapted from Moalin
[
32]
Synthesis of diglycidyl ether of bisphenol A-nadic methyl
anhydride (DGEBA-NMA)
et al. A round bottom flask was charged with quercetin (20.0 g,
6.2 mmol), potassium carbonate (20.4 g, 148 mmol), methyl
6
iodide (16.0 mL, 257 mmol) and DMF (500 mL) under nitrogen. The
reaction was stirred at room temperature for 24 h and the DMF
was then removed in vacuo, resulting in a brown solid product.
The brown solid was purified using MPLC by eluting with 10%
ethyl acetate in chloroform and recrystallized from chloroform to
DGEBA (10.0 g, 29.4 mmol), nadic methyl anhydride (10.5 g, 58.8)
and ANCAMINE K54 (300 mg) were combined in a round bottom
flask, heated to 908C and stirred for 5 min. The mixture was cast
into six molds preheated to 908C. The bars were cured for 1 h at
ꢁ
9
08C then at 1608C for 2 h to afford the DGEBA-NMA network.
afford 3,7,4’-trimethylquercetin as a yellow solid (9.5 g, 42% yield).
5%
10%
Td =2848C, Td =3378C, T =1368C; FTIR (ATR): n˜ =3645–3124,
3
g
1
H NMR (500 MHz, DMSO-d ): d=12.65 (s, 1H, ꢀOH), 9.44 (s, 1H,
ꢀ1
6
124–3022, 2917–2850, 1737, 1225, 1152, 1106, 1032, 827 cm .
ꢀ
OH), 7.59–7.55 (m, 2H, Ar), 7.11 (d, J=8.4 Hz, 1H, Ar), 6.71 (d, J=
2
.2 Hz, 1H, Ar), 6.37 (d, J=2.2 Hz, 1H, Ar), 3.87 (s, 3H, ꢀCH ), 3.86
3
13
(
s, 3H, ꢀCH ), 3.80 ppm (s, 3H, ꢀCH ); C NMR (125 MHz, DMSO-
3
3
Acknowledgements
d ): d=178.5, 165.5, 161.3, 156.7, 156.0, 150.7, 146.8, 138.6, 122.9,
6
1
20.8, 115.5, 112.2, 105.6, 98.2, 92.6, 60.2, 56.5, 56.1 ppm; FTIR
We gratefully acknowledge the financial support from the Na-
tional Science Foundation (CHE-1057441, CHE-1410272 and
CMMI-1334838) and the Welch Foundation (A-0001). Use of the
TAMU/LBMS and Dr. Yohannes Rezenom and Vanessa Santiago
(
ATR): n˜ =3579–3140, 3047–3000, 3000–2847, 1643, 1597, 1196,
ꢀ1
+1
9
18, 895, 818 cm ; HRMS (ESI+, m/z): [M+H]
calculated for
+
[
C H O ] , 345.0974, found 345.0970; m.p. 1738C. Regiochemistry
18 17 7
was confirmed by X-ray analysis (Figure S6).
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
are acknowledged for obtaining the mass spectrometry spectra.
The authors would like to thank the laboratory of Daniel Romo
[18] H. H. Wang, X. Q. Liu, B. Liu, J. W. Zhang, M. Xian, Polym. Int. 2009, 58,
1
435–1441.
19] M. Shibata, N. Teramoto, K. Makino, J. Appl. Polym. Sci. 2011, 120, 273–
78.
[
(TAMU) for the use of their microwave reactor. The authors
2
would also like to thank Khoi N. Van for the many insightful con-
versations. The X-ray diffractometers, small angle scattering in-
strumentation and crystallographic computing systems in the X-
ray Diffraction Laboratory at the Department of Chemistry, Texas
A & M University were purchased with funds provided by the Na-
tional Science Foundation (CHE-9807975, CHE-0079822 and CHE-
[20] M. Shibata, K. Nakai, J. Polym. Sci. Part B 2010, 48, 425–433.
[21] K. H. Miean, S. Mohamed, J. Agric. Food Chem. 2001, 49, 3106–3112.
[
[
22] J. V. Formica, W. Regelson, Food Chem. Toxicol. 1995, 33, 1061–1080.
23] A. W. Boots, G. R. M. M. Haenen, A. Bast, Eur. J. Pharmacol. 2008, 585,
325–337.
[24] H. Matsuda, T. Morikawa, M. Yoshikawa, Pure Appl. Chem. 2002, 74,
1301–1308.
[
[
25] G. D’Andrea, Fitoterapia 2015, 106, 256–271.
26] S. G. Dmitrienko, V. A. Kudrinskaya, V. V. Apyari, J. Anal. Chem. 2012, 67,
0215838). For access and training on the TGA instrument, we
thank Dr. Ramanan Krishnamoorti and Daehak Kim. This work
was supported in part by the Norman Hackerman Advanced Re-
search Program of the Texas Higher Education Coordinating
Board (Grant No. 003652-0022-2013).
2
99–311.
[27] C. Turner, P. Turner, G. Jacobson, K. Almgren, M. Waldebꢄck, P. Sjçberg,
E. N. Karlsson, K. E. Markides, Green Chem. 2006, 8, 949–959.
[
28] M. Shibata, S. Yoshihara, M. Yashiro, Y. Ohno, J. Appl. Polym. Sci. 2013,
28, 2753–2758.
1
[
[
[
[
29] A. Masek, M. Zaborski, C. R. Chim. 2014, 17, 944–951.
30] L. Jurd, J. Am. Chem. Soc. 1958, 80, 5531–5536.
31] L. Jurd, J. Org. Chem. 1962, 27, 1294–1297.
32] M. Moalin, G. P. F. van Strijdonck, M. Beckers, G. J. Hagemen, P. J. Borm,
A. Bast, G. R. M. M. Haenen, Molecules 2011, 16, 9636–9650.
33] P. Puebla, J. L. Lꢃpez, M. Guerrero, R. Carrꢃn, M. L. Martꢅn, L. San Romꢆn,
A. San Feliciano, Phytochemistry 2003, 62, 551–555.
Keywords: biomass · materials science · quercetin · renewable
resources · sustainable chemistry
[
[
[
[
[
[
[
[
[
1] R. Auvergne, S. Caillol, G. David, B. Boutevin, J. P. Pascault, Chem. Rev.
014, 114, 1082–1115.
2] Epoxy Resin Market by Application (Paints & Coating, Wind Turbine,
Composites, Construction, Electrical & Electronics, Adhesives) and Geog-
raphy, Trends & forecasts to 2019 http://www.marketsandmarkets.com/
Market-Reports/epoxy-resins-market-762.html 2014 (accessed February
2
34] M. Kaulich, F. Streicher, R. Mayer, I. Mꢇller, C. E. Mꢇller, Drug Dev. Res.
2
003, 59, 72–81.
35] Z.-H. Shi, N.-G. Li, Y.-P. Tang, L. Wei, Y. Lian, J.-P. Yang, T. Hao, J.-A. Duan,
Eur. J. Med. Chem. 2012, 54, 210–222.
36] Y. Wang, M. Hamburger, J. Gueho, K. Hostettmann, Phytochemistry
2, 2016).
1
989, 28, 2323–2327.
37] M. Bouktaib, S. Lebrun, A. Atmani, C. Rolando, Tetrahedron 2002, 58,
0001–10009.
38] N. Toelle, H. Weinstabl, T. Gaich, J. Mulzer, Angew. Chem. Int. Ed. 2014,
3, 3859–3862; Angew. Chem. 2014, 126, 3940–3943.
39] N. Toelle, H. Weinstabl, T. Gaich, J. Mulzer, Angew. Chem. 2014, 126,
940–3943.
[
[
3] C. Ding, A. S. Matharu, ACS Sustainable Chem. Eng. 2014, 2, 2217–2236.
4] L. N. Vandenberg, M. V. Maffini, C. Sonnenschein, B. S. Rubin, A. M. Soto,
Endocr. Rev. 2009, 30, 75–95.
1
[5] W. Liu, R. Zhou, H. L. S. Goh, S. Huang, X. Lu, ACS Appl. Mater. Interfaces
5
2014, 6, 5810–5817.
[6] S. Ma, T. Li, X. Liu, J. Zhu, Polym. Int. 2016, 65, 164–173.
3
[7] J. M. Pin, N. Guigo, L. Vincent, N. Sbirrazzuoli, A. Mija, ChemSusChem
[
[
40] B. K. Pchelka, A. Loupy, A. Petit, Tetrahedron 2006, 62, 10968–10979.
41] CCDC 1451608 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre.
42] CCDC 1451607 contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre.
43] Y. Tanaka, H. Kakiuchi, J. Polym. Sci. Part A 1964, 2, 3405–3430.
44] J. Leukel, W. Burchard, R.-P. Krꢇger, H. Much, G. Schulz, Macromol. Rapid
Commun. 1996, 17, 359–366.
2015, 8, 4149–4161.
[
[
8] J. M. Pin, N. Sbirrazzuoli, A. Mija, ChemSusChem 2015, 8, 1232–1243.
9] G. Yang, B. J. Rohde, M. L. Robertson, Green Mater. 2013, 1, 125–134.
[
[
[
10] C. S. Kovash, Jr., E. Pavlacky, S. Selvakumar, M. P. Sibi, D. C. Webster,
ChemSusChem 2014, 7, 2289–2294.
11] Z. Rapi, B. Szolnoki, P. Bakꢃ, P. Niedermann, A. Toldy, B. Bodzay, G. Kegle-
vich, G. Marosi, Eur. Polym. J. 2015, 67, 375–382.
12] J. Hong, D. Radojcic, M. Ionescu, Z. S. Petrovic, E. Eastwood, Polym.
Chem. 2014, 5, 5360–5368.
[
[
[
[
[
13] S. Zhao, M. M. Abu-Omar, Biomacromolecules 2015, 16, 2025–2031.
14] A. Llevot, E. Grau, S. Carlotti, S. Grelier, H. Cramail, Macromol. Rapid
Commun. 2016, 37, 9–28.
[
45] B. J. Rohde, M. L. Robertson, R. Krishnamoorti, Polymer 2015, 69, 204–
214.
[
[
46] M. K. Antoon, J. L. Koenig, J. Polym. Sci. A 1981, 19, 549–570.
47] L. W. Hill, Prog. Org. Coat. 1997, 31, 235–243.
[15] C. Mantzaridis, A. L. Brocas, A. Llevot, G. Cendejas, R. Auvergne, S. Cail-
lol, S. Carlotti, H. Cramail, Green Chem. 2013, 15, 3091–3098.
[16] A. L. Brocas, A. Llevot, C. Mantzaridis, G. Cendejas, R. Auvergne, S. Cail-
lol, S. Carlotti, H. Cramail, Des. Monomers Polym. 2014, 17, 301–310.
[17] Q. Q. Ma, X. Q. Liu, R. Y. Zhang, J. Zhu, Y. H. Jiang, Green Chem. 2013, 15,
Received: March 23, 2016
Revised: April 27, 2016
1300–1310.
Published online on && &&, 0000
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
S. L. Kristufek, G. Yang, L. A. Link,
B. J. Rohde, M. L. Robertson, K. L. Wooley*
&& – &&
Synthesis, Characterization, and Cross-
Linking Strategy of a Quercetin-Based
Epoxidized Monomer as a Naturally-
Derived Replacement for BPA in Epoxy
Resins
A natural solution: A diglycidyl mono-
mer prepared in two steps using a natu-
ral polyphenolic starting material, quer-
cetin, is reacted with nadic methyl anhy-
dride to afford epoxy cross-linked net-
works. The thermal and mechanical
properties of these networks are com-
pared to similarly synthesized bisphenol
A-based epoxy materials for potential
replacement in advanced engineering
applications.
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ