Multifunctional deoxybenzoin-based epoxies: Synthesis, mechanical properties, and thermal evaluation

Article abstract of DOI:10.1016/j.polymer.2014.06.043

We describe 2,4,4′,6-tetrahydroxydeoxybenzoin (THDB) as a multifunctional cross-linker in conjunction with bis-epoxydeoxybenzoin (BEDB), affording new resins that combine excellent physical and mechanical properties with low flammability. The char residue and heat release capacity values of the cross-linked epoxies were measured by thermogravimetric analysis (TGA) and pyrolysis combustion flow calorimetry (PCFC), respectively. Resins fabricated from THDB exhibited low total heat release (13 kJ/g) and high char yields (34%), as well as good mechanical properties, making them suitable candidates for consideration in high performance adhesive applications. The desirable heat release and char yield properties of these structures are realized without the presence of any conventional flame retardant, such as halogenated structures or inorganic fillers that are commonly utilized in commercial materials.

Full text of DOI:10.1016/j.polymer.2014.06.043

Polymer 55 (2014) 4441e4446
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Multifunctional deoxybenzoin-based epoxies: Synthesis, mechanical
properties, and thermal evaluation
1
1
1
*
*
Megan W. Szyndler , Justin C. Timmons , Zhan H. Yang , Alan J. Lesser , Todd Emrick
Department of Polymer Science and Engineering, University of Massachusetts, 120 Governors Drive, Amherst, MA 01003, USA
a r t i c l e i n f o
a b s t r a c t
0
Article history:
We describe 2,4,4 ,6-tetrahydroxydeoxybenzoin (THDB) as a multifunctional cross-linker in conjunction
with bis-epoxydeoxybenzoin (BEDB), affording new resins that combine excellent physical and me-
chanical properties with low flammability. The char residue and heat release capacity values of the cross-
linked epoxies were measured by thermogravimetric analysis (TGA) and pyrolysis combustion flow
calorimetry (PCFC), respectively. Resins fabricated from THDB exhibited low total heat release (13 kJ/g)
and high char yields (34%), as well as good mechanical properties, making them suitable candidates for
consideration in high performance adhesive applications. The desirable heat release and char yield
properties of these structures are realized without the presence of any conventional flame retardant,
such as halogenated structures or inorganic fillers that are commonly utilized in commercial materials.
© 2014 Elsevier Ltd. All rights reserved.
Received 11 March 2014
Received in revised form
11 June 2014
Accepted 13 June 2014
Available online 20 June 2014
Keywords:
Deoxybenzoin
Flame retardant
Mechanical properties
1. Introduction
in replacement for conventional bis-phenols such as bisphenol A
(BPA), including as cross-linked epoxides [16]. For example, we
Epoxy resins are prominent among thermoset polymers,
employed as high performance adhesives, surface coatings,
encapsulation matrices, and composites in applications ranging
from aerospace transportation to microelectronics packaging
found that bis-epoxydeoxybenzoin (BEDB), though halogen-free,
gave cross-linked polymer resins in conjunction with aromatic
amines, affording robust adhesives with heat release capacity
(HRC) values that were 20e40% lower than conventional BPA-
based resins [16].
[1e3]. Epoxy resins, like many synthetic polymer materials, require
low flammability for their safe use. While halogenated flame re-
tardants are common additives that impart low flammability to
materials, growing concern and legislation over the potential
health and environmental consequences of such additives drives
the discovery of new routes to achieve flame retardancy. With
respect to epoxy resins, efforts towards non-halogenated materials
center on the integration of phosphorus [2e9], boron [10,11], or
silicon [12e14] into the cross-linked matrix. We seek to achieve
non-flammable resins from purely hydrocarbon systems, without
the need for additives of any sort. This requires hydrocarbons that
have degradation mechanisms that produce few flammable vola-
tiles, yet that also possess suitable processability and performance
characteristics, a challenging combination to achieve. Our prior
work showed that numerous low flammability polymer materials
can be prepared from deoxybenzoin-containing polymers [15e20],
and that bis-hydroxydeoxybenzoin (BHDB) can function as a drop-
This paper describes the preparation of an ‘all-deoxybenzoin’
cross-linked resin consisting of the tetrafunctional THDB, with
BEDB as the epoxide component. The resulting resins possessed
thermal properties similar to, or better than, flame retardant ep-
oxies utilizing halogen or phosphorus-containing additives. More-
over, mechanical measurements, including compression and
fracture toughness, were performed to ensure that these materials
possess comparable properties to commercial epoxy resins. Lap
shear measurements confirmed desirable performance character-
istics from these wholly-deoxybenzoin structures.
2. Experimental
All chemicals were purchased from SigmaeAldrich and used as
received unless otherwise noted.
2.1. Characterization
Nuclear magnetic resonance (NMR) spectra were obtained on a
Bruker DPX300 or Avance 400 spectrometer. Thermogravimetric
analysis (TGA) was performed under nitrogen on a Q500 (TA
*
Corresponding authors.
E-mail address: tsemrick@mail.pse.umass.edu (T. Emrick).
Equal contributions.
1
http://dx.doi.org/10.1016/j.polymer.2014.06.043
032-3861/© 2014 Elsevier Ltd. All rights reserved.
0

4
442
M.W. Szyndler et al. / Polymer 55 (2014) 4441e4446
ꢀ
Instruments) at a heating rate of 10 C/min. Char yields were
function in terms of x, which is the ratio of the length of the pre-
crack to the width of the specimen (a/W).
ꢀ
determined by TGA from the mass residue at 800 C. Differential
scanning calorimetry (DSC) was performed on a Q200 from TA
Instruments with a heating rate of 10 C/min. Specific heat release
À
Á
ꢀ
2
3
4
ð2 þ xÞ 0:886 þ 4:64x ꢁ 13:32x þ 14:72x ꢁ 5:6x
f ðxÞ ¼
rate (HRR, W/g), heat release capacity (HRC, J/(g-K)), and total heat
release (THR, kJ/g) were measured on a microscale combustion
calorimeter (MCC). MCC was conducted over a temperature range
3=2
ð1 ꢁ xÞ
(
2)
ꢀ
ꢀ
3
of 80e750 C at a heating rate of 1 C/s in an 80 cm /min stream of
q
The term K denotes the use of mini-CT samples with a non-
nitrogen. The anaerobic thermal degradation products in the ni-
standard W/B ratio. Each reported fracture toughness value repre-
sents an average of 4e8 measurements.
3
trogen gas stream were mixed with a 20 cm /min stream of oxygen
ꢀ
prior to entering the combustion furnace (900 C). Heat release is
quantified by standard oxygen consumption methods typical to
PCFC [21,22]. During the test, HRR is obtained from dQ/dt, at each
time interval and by the initial sample mass (~5 mg). The HRC is
obtained by dividing the maximum HRR by the heating rate.
2.2.4. Lap shear strength
Single joint lap shear strength measurements were performed
according to the ASTM D1002 standard [27]. Samples were pre-
2
pared with bond area of 12.7 mm using aluminum 2024-T3
(100 mm ꢂ 25 mm ꢂ 1.60 mm) with inclusion of two 36 AWG wires
to maintain constant thickness. The aluminum substrates were
roughened with sandpaper and rinsed with hexane and water prior
to bonding. Three to seven samples were prepared per formulation.
Tests were performed at crosshead speed of 1.3 mm/min on an
Instron universal testing machine (Model 4411) at room tempera-
ture. The lap shear strength was calculated as the ratio of the load at
failure to the overlap bond area.
2
2
.2. Mechanical testing
.2.1. Dynamic mechanical analysis (DMA)
Dynamic mechanical properties of film tension specimens were
measured on a Q800 machine (TA Instruments) at a single frequency
ꢀ
ꢀ
of 1 Hz and a heating rate of 3 C/min with equilibration at ꢁ120 C.
00
The T values were taken as the maxima of the loss moduli (E ).
g
0
2
.3. Synthesis of 4,4 -bishydroxydeoxybenzoin (BHDB)
2
.2.2. Compression testing
Compression bullets were formed by cutting cylinders of resin
0
4,4 -Bishydroxydeoxybenzoin was prepared following pub-
lished procedures [16]. Specifically, desoxyanisoin (100 g,
into sections, with height to diameter ratios of 1:1 and diameters of
approximately 11 mm. The top and bottom faces of the cylinders
were cut parallel to one other, and perpendicular to the sides of the
cylinders. The faces were polished to give smooth surfaces. Di-
mensions for compression testing were measured to the nearest
390 mmol) was demethylated with pyridinium hydrochloride
ꢀ
(180 g, 156 mmol) by stirring at 200 C for 5 h. The resulting
mixture was poured into water, and a yellow precipitate formed.
This precipitate was filtered then recrystallized from acetic acid, to
1
0
.01 mm using calipers. Immediately before compression testing, a
give 68.0 g of BHDB (77% yield). H NMR (DMSO-d6, ppm): 10.38 (s,
surfactant-PTFE film treatment was applied to the top and bottom
surfaces of the compression specimens to produce a low friction
surface that allows for affine deformation over a large range of
strains. The samples are loaded in compression using an Instron
OHeAreCO), 9.27 (s, OHeAreCH ), 7.89 (d, J ¼ 7.0 Hz, 2H, AreH),
2
7.04 (d, J ¼ 8.5 Hz, 2H, AreH), 6.84 (d, J ¼ 9.5 Hz, 2H, AreH), 6.68 (d,
13
J ¼ 8.5 Hz, AreH), 4.10 (s, 2H, AreCH eCOeAr). C NMR (DMSO-d6,
2
ppm): 196.7, 162.5, 156.4, 131.5, 130.9, 128.3, 126.0, 115.7, 115.6, 43.9.
5
2
800 fitted with a 50 kN load cell and controlled using the BlueHill
ꢀ
0
software package. Samples were tested at 20 C, and a constant
2.4. Synthesis of the diglycidyl ether of 4,4 -
bishydroxydeoxybenzoin (BEDB)
ꢁ1
ꢁ3 ꢁ1
true strain rate of 1.0 min (1.67 ꢂ 10
s
) was maintained
during the entire test [23].
The glycidyl ether of BHDB was prepared generally following
published procedures [16]. Specifically, BHDB (68.0 g, 296 mmol)
was added to a roundbottom flask with epichlorohydrin (232 mL,
2
.2.3. Plane-strain fracture toughness
The use of miniature compact tension (mini-CT) specimens for
2.96 mol), isopropanol (115 mL) and water (23 mL). A 20% NaOH
fracture toughness testing of glassy polymers had been reported by
Jones and Lee [24], and Hinkley [25]. In our study, 3 mm thick mini-
CT specimens with 20 mm width (W) were prepared following
ASTM standard D5045 [26]. The thickness (B) satisfied the
solution (23.0 g NaOH in 115 mL water) was added dropwise at
ꢀ
6
5 C and allowed to react for a total of 1.5 h from the first addition
of base. The mixture was cooled and chloroform was added to
extract the product. After several washings with water and brine,
the organic layer was dried over magnesium sulfate. The solution
was poured into hexane, resulting in precipitation of 79.0 g of the
requirement for achieving plane-strain conditions across the crack
2
front, namely B ꢃ 2.5 (K
q
/s
y
) , where K
q
is the measured fracture
toughness and
s
y
is the yield stress estimated from the compres-
ꢀ
1
final diglycidyl ether product (80% yield). mp 105 C. H NMR
(DMSO-d6, ppm): 8.01 (d, 2H, J ¼ 8.0 Hz, AreH), 7.20 (d, 2H,
J ¼ 7.8 Hz, AreH), 6.97 (d, 2H, J ¼ 7.8 Hz, AreH), 6.90 (d, 2H,
sion data [23]. The pre-notches were introduced with a diamond
ꢀ
wafering blade. After conditioning the samples at ꢁ10 C for one
hour, a sharp pre-crack was generated on each specimen by
inserting a fresh razor blade into the pre-notch and tapping lightly
with a hammer. Load-displacement curves were recorded by an
J ¼ 6.9 Hz, AreH), 4.38e4.16 (m, 2H,eOeCH
2
eoxirane), 4.16 (s, 2H,
eoxirane),
.45e3.31(m, 2H, 2(oxirane CH)), 3.01e2.89 (m, 2H, 2(oxirane
AreCH2eCOeAr), 4.07e3.92 (m, 2H, eOeCH
2
3
Instron universal testing machine (Model 4411) at a crosshead
speed of 0.5 mm/min at 20 C. Fracture toughness (K
computed using the following equation:
13
ꢀ
2 2 3
CH )), 2.84e2.74 (m, 2H, 2(oxirane CH )). C NMR (CDCl , ppm):
q
) was
196.5, 162.3, 157.4, 130.9, 130.5, 130.0, 127.5, 114.9, 114.4, 68.9, 66.8,
50.1, 49.9, 44.7, 44.6, 44.4.
P
c
f ðxÞ
1=2
K
q
¼
(1)
0
2.5. Synthesis of 2,4,4 ,6 tetrahydroxydeoxybenzoin (THDB)
BW
where K
is in units of MPa/m1/2, P
W are in cm. The geometric factor f(x) is a dimensionless power
is the critical load in kN, B and
Tetrahyroxydeoxybenzoin (THDB) was prepared following a
reported procedure [28], but at a larger scale. Phloroglucinol
q
c

M.W. Szyndler et al. / Polymer 55 (2014) 4441e4446
4443
(
3
39.5 g, 313 mmol), 4-hydroxyphenylacetonitrile (50.0 g,
synthesized under HoubeneHoesch conditions (Fig. 1), starting
from phloroglucinol and 4-hydroxyphenylacetonitrile. These con-
ditions for synthesizing THDB, known from the natural products
literature [28e30], proved scalable to the 40 g level. Using THDB as
a multifunctional phenol, and BEDB as the epoxy component, epoxy
resins were generated in which the entirety of the polymer resin is
composed of monomer units having the inherently non-flammable
deoxybenzoin as the aromatic structural unit.
76 mmol) and anhydrous ethyl ether (200 mL) were added to a
ꢀ
roundbottom flask. Under a nitrogen atmosphere at 0 C, aluminum
chloride (5.0 g, 39.3 mmol) was added and the flask was capped
quickly. A steady stream of HCl(g), generated by addition of
concentrated HCl over calcium chloride, was piped into the flask.
The HCl stream flowed for 6 h, yielding a pink precipitate. The
ꢀ
precipitate was kept overnight at 0 C. The precipitate was then
filtered and refluxed in water (500 mL) for 2 h. This produced a light
red powder that was filtered and dried. The powder was dissolved
in acetone and charged with charcoal, filtered, and the solvent
removed to afford 42.9 g of pale yellow powder (54% yield). mp
3
.2. Thermal properties
ꢀ
1
We were intrigued to find that the thermal properties of BHDB
2
1
59 C. H NMR (DMSO-d6, ppm): 12.27 (s, 2H,2,6eOH), 10.41 (s,
0
0
0
and THDB, as small molecules, analyzed separately by thermogra-
vimetric analysis (TGA), differed greatly. The TGA curves of Fig. 2
show that BHDB volatilizes completely at ~350 C, while THDB
H, 4eOH), 9.23 (s, 1H, 4 eOH), 7.02 (d, 2H, J ¼ 8.1 Hz, H-2 ,6 ), 6.68
0
0
(
d, 2H, J ¼ 8.1 Hz, H-3 ,5 ), 5.82 (s, 2H, H-3,5), 4.21(s, 2H,
ꢀ
ꢀ
eCOeAr). 1 C NMR (DMSO-d6, ppm): 203.5, 165.2, 164.7,
3
AreCH
56.31, 131.0, 126.4, 115.4, 104.1, 95.2, 48.5.
2
produces a substantial char residue (~40% at 800 C). Since BHDB-
1
containing polymers of many varieties exhibit high char yield, we
expected that THDB-containing structures would be especially
interesting to examine in the capacity of non-flammable polymeric
materials.As a multifunctional structure, THDB is well-suited for
integration into polymer materials as a cross-linker, tested here in
epoxy curing, with subsequent evaluation of mechanical and
thermal properties. TGA thermograms of epoxy resins with and
without deoxybenzoins in the network are shown in Fig. 3, with
clear differences arising from the presence of THDB in the cured
material. For example, resins prepared from the diglycidyl ether of
BPA (DGEBA) by curing with polyetheramine (PEA) displayed
2
.6. Preparation of BEDB/THDB resins
Cross-linked resins were prepared in a stoichiometric ratio of
BEDB to THDB (2 mol BEDB: 1 mol THDB). As the melting points of
the two monomers differ by 150 C, the higher melting point THDB
was solubilized in methanol prior to cross-linking. BEDB was
melted at 150 C and the THDB solution was added slowly while
stirring. The resin was quickly poured into a mold and cured
overnight at 200 C. Teflon molds were used for fracture toughness
specimens and Surfasil-treated test tubes were used for compres-
sion samples. Comparative samples were prepared by reacting the
diglycidyl ether of bisphenol A (DGEBA, D.E.R.332, Dow Chemical
Company) with stoichiometric amount of polyetheramine (PEA,
ꢀ
ꢀ
ꢀ
ꢀ
almost no char residue beyond 450 C. However, replacement of
PEA with THDB gave epoxy polymers with a char residue of 22%.
Impressively, the all deoxybenzoin cured system, prepared from
ꢀ
THDB and BEDB, produced nearly 40% char at 800 C, approxi-
0
Jeffamine D230, Huntsman), 4,4 -diaminodiphenylsulfone (DDS,
mately 3 times that of a conventional DGEBA/DDS system (char
yield ~15%). Such an increase in char yield would be expected to
correlate to reduced heat release values, due to fewer escaping
volatile components upon thermal degradation of the material. In
accord with this expectation, pyrolysis combustion flow calorim-
etry (PCFC) revealed that the heat release capacity (HRC) and total
heat release (THR) values decreased significantly due to the inclu-
sion of deoxybenzoin monomers (defining HRC as the maximum
heat released divided by the heating rate). Describing these sys-
tems in terms of HRC eliminates the reliance on heating rate that is
typical of standard flammability measurements (i.e. heat release
rate), rendering it a material dependent property. THR is the total
heat of complete combustion of the pyrolysis products per initial
mass of the sample. Lower HRC and THR values are thus good in-
dicators of increasing the flame retarding property of the polymer.
For instance, by incorporating deoxybenzoin into resins, the HRC of
Acros Organic), and THDB (2 glycidyl ether: 1 crosslinker). DGEBA/
ꢀ
ꢀ
PEA were cured at 80 C for 3 h, followed by 3 h at 120 C, while
ꢀ
DGEBA/DDS was cured at 200 C for 4 h, followed by an additional
ꢀ
2
h at 220 C. DGEBA/THDB were prepared similarly to the BEDB/
THDB resins.
3
. Results and discussion
3.1. Synthesis of monomers BEDB and THDB
BEDB was prepared by reacting BHDB with an excess of
epichlorohydrin in an aqueous solution of sodium hydroxide, as
shown in Fig. 1. The BEDB synthesis is scalable, prepared in our
laboratories at about 80 g per batch, and BEDB has proven stable
when stored in the laboratory under ambient conditions. THDB was
Fig. 1. Synthesis of BEDB (top) and THDB (bottom).

4
444
M.W. Szyndler et al. / Polymer 55 (2014) 4441e4446
Table 1
Thermal characteristics of deoxybenzoin based epoxy resins.
ꢀ
ꢀ
Formulation
HRC
J/K-g)
THR
(kJ/g)
T
g
( C)
T
d
( C) 5%
Residue %
(
weight loss
DGEBA/PEA
DGEBA/DDS
DGEBA/THDB
BEDB/THDB
850
510
393
229
29
26
20
13
89
349
383
378
322
7
15
22
34
219
164
110
T
g
d
values obtained from DSC. T and residue percent from TGA. HRC and THR ac-
quired on MCC.
DMA were similar to those obtained by DSC, showing DGEBA/THDB
and BEDB/THDB formulations to be intermediate between the
DGEBA/PEA (lowest) and DGEBA/DDS (highest) values. As the
DGEBA/PEA and DGEBA/DDS formulations were prepared with
stoichiometries intended to give similar average molecular weight
between crosslinks, M
tance of backbone structure/stiffness in governing T
c
(Table 2), the results illustrate the impor-
in these ma-
g
Fig. 2. Thermogravimetric analysis (TGA) curves for BHDB, BEDB, and THDB.
terials. For example, while the molecular weight between
crosslinks of BEDB/THDB is essentially identical to the DGEBA/PEA
system (470 g/mol), the higher T
flects the stiffer aromatic backbone structure of the BEDB/THDB
network. Higher T of the deoxybenzoin system is also seen relative
to the network containing PEA. From a thermal perspective, the
BEDB/THDB and DGEBA/THDB formulations fall into a suitable
range for numerous applications. Interestingly, DMA results also
revealed the storage modulus of the BEDB/THDB network to be
g
of the deoxybenzoin system re-
BEDB resins cured with PEA were measured as 480 J/g-K, nearly a
0% reduction from that of DGEBA (850 J/g-K) (see Table 1). Intro-
ducing THDB into DGEBA resins led to further reduction in HRC, to
93 J/g-K, confirming 1) the beneficial effect of deoxybenzoin
monomers for tempering heat release, and 2) their facile integra-
tion into curing chemistry with conventional epoxies. The all-
deoxybenzoin cross-linked resins displayed by far the most favor-
able heat release properties, with THR and HRC values measured as
5
g
3
~
10% higher than the DGEBA/DDS formulation (Fig. 4a and Table 2),
and maintains higher values than the other materials through the
entirety of the glassy region. This property is particularly important
in fiber reinforced composites, in which the composite compressive
strength increases with storage modulus of the resin.
13 kJ/g and 229 J/g-K, respectively. We note that such values are
almost half that of any of the resins we reported previously from
BEDB with conventional cross-linking structures: for example,
BEDB cross-linked with DDS gave THR and HRC values of 17.2 kJ/g
and 420 J/g-K [16]. While we expect that these values could be
lowered further by inclusion of halogenated or inorganic flame
retardant additives, for the purpose of this study we simply convey
the favorable thermal properties of these systems in the absence of
additives of any sort.
When comparing the loss moduli of these epoxy formulations
(
Fig. 4b), the breadth of the beta-transition regime was observed to
broaden more substantially for the deoxybenzoin formulations
BEDB/THDB and DGEBA/THDB) relative to the others. The beta
transition, combined with T values, affect numerous properties
(
g
including the overall fracture toughness of the networks (vide
infra).
Compression tests of the resins were conducted to characterize
their nonlinear viscoelastic response. Compression testing is ad-
vantageous for avoiding premature failure from flaw activation,
thereby allowing for post yield characterization. Fig. 5 shows
representative plots of the true stress vs. neo-Hookian strain for
deoxybenzoin-containing resins and conventional resins, while the
elastic modulus, yield stress, rejuvenated stress, and strain hard-
ening modulus are given in Table 2. The elastic moduli for all of
these formulations are consistent with the storage moduli values
obtained from the DMA tests, with the highest modulus values seen
in the BEDB/THDB system. With respect to yield stress, the DGEBA/
PEA formulation exhibited the lowest values, with DGEBA/THDB
and BEDB/THDB having intermediate values, and the DGEBA/DDS
system measuring the highest among the samples examined. This
trend is expected, since for glassy amorphous polymers the yield
3
.3. Mechanical properties
Cured resin plaques prepared as described above were charac-
terized for their physical and mechanical properties, first employ-
ing dynamic mechanical analysis (DMA). The T values measured by
g
g
stress scales with T for formulations tested at the same tempera-
ture and strain rate. However, the plots of Fig. 5 also show that
differences in post-yield stress drop were noted among the sam-
ples, which can also be seen by comparing the numerical results of
yield stress and rejuvenated stress in Table 2. This is important
since it is known that a more pronounced post-yield stress drop
increases the extent of strain localization that can occur in other
stress states and can lead to premature failure in certain
circumstances.
Fracture toughness tests assessed the relative toughness of
these new deoxybenzoin network formulations; values for fracture
Fig. 3. TGA analysis of THDB-containing resins vs. BPA-based structures.

M.W. Szyndler et al. / Polymer 55 (2014) 4441e4446
4445
Table 2
Mechanical properties of deoxybenzoin resins.
ꢀ
ꢀ
C
1c MPa m0.5
G
1c (kJ/m2)
Resin
M
c
g/mol
T
g, DSC
C
T
g, DMA
Storage
modulus
GPa
Compressive
modulus GPa
Yield
stress
MPa
Rejuvenated
stress MPa
Strain hardening
modulus MPa
K
DGEBA/PEA
DGEBA/DDS
DGEBA/THDB
BEDB/THDB
470
474
480
470
88
92
2.20
1.77
1.46
2.41
2.90
2.61
2.25
3.02
89
138
112
129
70
135
106
98
79
162
119
58
0.82 ± 0.11
0.78 ± 0.05
0.86 ± 0.15
0.78 ± 0.14
0.20 ± 0.004
0.20 ± 0.001
0.28 ± 0.008
0.17 ± 0.005
219
164
108
223
168
116
M
c
was calculated based on 1:1 stoichiometry between epoxide groups and amine hydrogens, assuming full conversion. A Poisson ratio of 0.4 was used for calculating G1c.
Fig. 4. DMA analysis of resins that contain THDB compared with BPA-based resins. a) Storage modulus and b) Loss modulus as a function of temperature.
toughness and energy release rate are shown in Table 2, while Fig. 6
plots fracture toughness against T for each formulation examined.
Fracture toughness generally decreases with increasing T , since
dissipative processes such as yielding and post-yield deformation
can occur, and surface area increases during the fracture process.
Interestingly, each of the four formulations studied exhibited
governs fracture response, and the morphological origins, is
g
beyond the scope of this investigation. Nonetheless, the results
show that both the BEDB/THDB and DGEBA/THDB formulations
have comparable toughness to commonly used commercial sys-
tems with similar crosslink densities.
The utility of these resins as cured adhesives was assessed
through ASTM D1002 lap shear evaluation, performed for each of
the four resin formulations described above. Aluminum (2024-T3)
substrates of 100 mm by 25 mm were prepared by roughening with
sandpaper and rinsing with hexane and water. Resins were applied
to the surfaces with approximately 13 mm of overlap. Two wires
g
comparable fracture toughness, despite the significant breadth
ꢀ
(
spanning 100 C) of T
g
values. This indicates that yield stress is of
secondary importance in determining the relative toughness of
these formulations, and other post-yield characteristics dominate
fracture toughness. Isolating which particular characteristic
Fig. 5. Compression results of deoxybenzoin (THDB and BEDB) resins relative to
Fig. 6. Fracture toughness values of deoxybenzoin-containing resins vs. conventional
conventional BPA-based resins.
BPA-based resins.

4
446
M.W. Szyndler et al. / Polymer 55 (2014) 4441e4446
Table 3
Appendix A. Supplementary data
Lap shear strength values of the THDB and BPA resins.
Resin
Lap shear strength (MPa)
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.polymer.2014.06.043.
DGEBA/PEA
DGEBA/DDS
DGEBA/THDB
BEDB/THDB
13.1 ± 1.4
14.3 ± 1.2
12.6 ± 2.1
10.1 ± 0.9
References
[
[
1] Odian G. Principles of polymerization. 4th ed. Hoboken, NJ: John Wiley & Sons,
Inc; 2004.
2] Hergenrother PM, Thompson CM, Smith JG, Connell JW, Hinkley JA, Lyon RE,
et al. Polymer 2005;46(14):5012e24.
were utilized in to maintain a fixed thickness. Binder clips ensured
the substrates were aligned and maintained contact during curing.
The resins were cured in the oven as per each formulation
[
[
3] Rakotomalala M, Wagner S, Doring M. Materials 2010;3(8):4300e27.
4] Lu SY, Hamerton I. Prog Polym Sci 2002;27(8):1661e712.
[5] Levchik S, Piotrowski A, Weil E, Yao Q. Polym Degrad Stabil 2005;88(1):
ꢀ
ꢀ
requirement (e.g. DGEBA/DDS at 180 C for 4 h, followed by 220 C
for 2 h, see Section 2.6). The substrates were placed in the Instron,
and pulled in tension at a crosshead speed of 1.3 mm/min until
failure. The lap shear strength values, acquired by dividing the load
at failure by the overlap area, are given in Table 3. We were
encouraged to find that the resins with THDB give the same order of
magnitude lap shear strength values as the resins prepared by
crosslinking DGEBA with either PEA or DDS, and are comparable to
values for DGEBA resins found in the literature [31e34].
57e62.
[
6] Lv Q, Huang JQ, Chen MJ, Zhao J, Tan Y, Chen L, et al. Ind Eng Chem Res
013;52(27):9397e404.
7] Mauerer O. Polym Degrad Stabil 2005;88(1):70e3.
2
[
[8] Toldy A, Szabo A, Novak C, Madarasz J, Toth A, Marosi G. Polym Degrad Stabil
008;93(11):2007e13.
9] Weil ED, Levchik S. J Fire Sci 2004;22(1):25e40.
10] Martin C, Ronda JC, Cadiz V. J Polym Sci Pol Chem 2006;44(5):1701e10.
[11] Martin C, Lligadas G, Ronda JC, Galia M, Cadiz V. J Polym Sci Pol Chem
006;44(21):6332e44.
12] Deng LL, Shen MM, Yu J, Wu K, Ha CY. Ind Eng Chem Res 2012;51(24):
178e84.
13] Devaraju S, Vengatesan MR, Selvi M, Song JK, Alagar M. Polym Compos
013;34(6):904e11.
2
[
[
2
[
[
8
2
[
[
[
14] Gao M, Wu WH, Xu ZQ. J Appl Polym Sci 2013;127(3):1842e7.
15] Ryu BY, Emrick T. Macromolecules 2011;44(14):5693e700.
16] Ryu BY, Moon S, Kosif I, Ranganathan T, Farris RJ, Emrick T. Polymer
2009;50(3):767e74.
4
. Conclusions
Multifunctional deoxybenzoin-containing structures, based on
[17] Ellzey KA, Ranganathan T, Zilberman J, Coughlin EB, Farris RJ, Emrick T.
the tetrafunctional THDB and difunctional BEDB were prepared and
utilized in curing chemistry. The adhesive materials reported from
THDB and BEDB-containing resins proved to have suitable me-
chanical properties for consideration in a variety of adhesive ap-
plications, while the flammability properties of the materials
obtained were substantially lower than conventional systems, even
in the absence of anti-flammable additives. Evaluation of me-
chanical properties of these systems revealed a higher elastic
modulus for the BEDB/THDB formulation relative to conventional
materials, along with good fracture toughness compared to for-
mulations of similar molecular weight between crosslinks.
Deoxybenzoin-based materials can thus afford useful epoxy resins
that may be suitable for consideration in a variety of adhesive
applications.
Macromolecules 2006;39(10):3553e8.
[18] Ranganathan T, Cossette P, Emrick T. J Mater Chem 2010;20(18):3681e7.
[19] Ranganathan T, Ku BC, Zilberman J, Beaulieu M, Farris RJ, Coughlin EB, et al.
Polym Sci Pol Chem 2007;45(20):4573e80.
[
20] Ranganathan T, Zilberman J, Farris RJ, Coughlin EB, Emrick T. Macromolecules
006;39(18):5974e5.
[21] Lyon RE, Walters RN. J Anal Appl Pyrol 2004;71(1):27e46.
2
[
22] ASTM. Standard test method for determining flammability characteristics of
plastics and other solid materials using microscale combustion calorimetry.
ASTM D7309-13. West Conshohocken, PA: ASTM International; 2013. p. 11.
[23] Detwiler AT, Lesser AJ. J Mater Sci 2012;47(8):3493e503.
24] Lee CYC, Jones WB. Polym Eng Sci 1982;22(18):1190e8.
[
[
[
25] Hinkley JA. J Appl Polym Sci 1986;32(6):5653e5.
26] ASTM. Standard test method for plain-strain fracture toughness and and
strain energy release rate of plastic materials. ASTM D5045-99. West Con-
shohocken, PA: ASTM International; 1999.
27] ASTM. Standard test method for apparent shear strength of single-lap-joint
adhesively bonded metal specimens by tension loading (metal-to-metal).
ASTM D1002-10. p. 6. West Conshohocken, PA: ASTM International; 2010.
28] Xiao ZP, Shi DH, Li HQ, Zhang LN, Xu C, Zhu HL. Bioorg Med Chem
[
[
[
2
007;15(11):3703e10.
29] Chang YC, Nair MG, Santell RC, Helferich WG. J Agric Food Chem 1994;42(9):
869e71.
[30] Balasubramanian S, Nair MG. Synth Commun 2000;30(3):469e84.
Acknowledgments
1
The authors acknowledge the financial support of the Federal
Aviation Administration (FAA-09-G-13) and the consortium mem-
ber companies of the Center for UMass-Industry Research on
Polymers (CUMIRP) Cluster group on Flammability. The project
utilized the Shared Facilities supported by the Materials Research
Science and Engineering Center (MRSEC) on Polymers at UMass
Amherst (DMR-0820506).
[
[
[
[
31] Ochi M, Tsuyuno N, Sakaga K, Nakanishi Y, Murata Y. J Appl Polym Sci
995;56(9):1161e7.
1
32] Mansourian-Tabaei M, Jafari SH, Khonakdar HA. J Appl Polym Sci 2014;131(6).
http://dx.doi.org/10.1002/app.40017.
33] Mahnam N, Beheshty MH, Barmar M, Shervin M. High Perform Polym
2013;25(6):705e13.
34] Souza JPB, Reis JML. Appl Adhes Sci 2013;1:6. http://dx.doi.org/10.1186/2196-
4351-1-6.