Kinetics and mechanistic investigation of epoxy-anhydride compositions cured with quaternary phosphonium salts as accelerators

Article abstract of DOI:10.1002/pola.27946

Mechanism and curing kinetics of bisphenol A epoxy resin-iso-methyltetrahydrophthalic anhydride compositions using quaternary phosphonium salts as accelerators were investigated by differential scanning calorimetry (DSC) and electrospray mass-spectrometry (ESI-MS). The DSC method was applied to investigate curing kinetics and apparent activation energy values for the overall curing process. The DSC results showed that some of the phosphonium salts lead to a lower activation energy, that means they are more effective accelerators for the curing of epoxy-anhydride systems. The mechanism of curing was studied by ESI-MS using the model reaction of epichlorohydrin (E) with phthalic anhydride (PA) in the presence of phosphonium salts or 2-methylimidazole. Products containing the alkyl moiety of the phosphonium salt in form of alkyl esters could be identified. This suggests that the phosphonium salts activate the anhydride by electrophilic attack.

Full text of DOI:10.1002/pola.27946

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
Kinetics and Mechanistic Investigation of Epoxy–Anhydride Compositions
Cured with Quaternary Phosphonium Salts as Accelerators
Lyaysan R. Amirova,1 Alexander R. Burilov,2 Liliya M. Amirova,1 Ingmar Bauer,3
Wolf D. Habicher3
1Kazan Federal University, Kremlyovskaya Str. 18, Kazan 420008, Russia
2A.E. Arbuzov Institute of Organic and Physical Chemistry, Kazan Scientific Center, Russian Academy of Sciences,
Arbuzov Str. 8, Kazan 420088, Russia
3TU Dresden, Institute of Organic Chemistry, Dresden, Bergstr. 66, 01069, Germany
Correspondence to: L. R. Amirova (E-mail: Lyaisanamirova@mail.ru)
Received 5 August 2015; accepted 5 October 2015; published online 20 November 2015
DOI: 10.1002/pola.27946
ABSTRACT: Mechanism and curing kinetics of bisphenol
A
reaction of epichlorohydrin (E) with phthalic anhydride (PA) in
the presence of phosphonium salts or 2-methylimidazole. Prod-
ucts containing the alkyl moiety of the phosphonium salt in
form of alkyl esters could be identified. This suggests that the
epoxy resin–iso-methyltetrahydrophthalic anhydride composi-
tions using quaternary phosphonium salts as accelerators were
investigated by differential scanning calorimetry (DSC) and
electrospray mass-spectrometry (ESI-MS). The DSC method
was applied to investigate curing kinetics and apparent activa-
tion energy values for the overall curing process. The DSC
results showed that some of the phosphonium salts lead to a
lower activation energy, that means they are more effective
accelerators for the curing of epoxy–anhydride systems. The
mechanism of curing was studied by ESI-MS using the model
phosphonium salts activate the anhydride by electrophilic
C
V
attack.
2015 Wiley Periodicals, Inc. J. Polym. Sci., Part A:
Polym. Chem. 2016, 54, 1088–1097
KEYWORDS: cure kinetics; curing mechanism; epoxy–anhydride;
differential scanning calorimetry; quaternary phosphonium
salts
cal efficiency of the curing process catalysts are usually
administered.1,4
INTRODUCTION Development of polymeric composite materials
with high technological and operational properties is an impor-
tant challenge. Epoxy polymers surpass many other classes of
synthetic polymers from the viewpoint of such important prop-
erties as adhesion, strength and chemical resistance.1,2 This
makes them indispensable as a basis for adhesives, coatings,
compounds, and binders for reinforced plastics. In a number of
applications, the common two-component systems (epoxy
resin–hardener) are insufficient. Therefore one-component com-
positions where epoxy resin and curing agent coexist with long
pot life at ambient conditions are needed. Such systems are suit-
able for applications as coatings, adhesives or binders for pre-
pregs in order to meet the industrial needs, for example, aircraft
manufacturing.3,4
Various types of compounds (e.g., tertiary amines, quaternary
ammonium compounds, imidazole derivatives, and metal
complexes) have been used as curing catalysts for epoxy res-
ins.6–11 However, tertiary amines react with both epoxy
oligomers and anhydrides, so they cannot be combined. One-
pot epoxy anhydride compounds based on imidazole cata-
lysts have insufficient storage life for many applications. Cur-
ing of epoxy–anhydride compounds in the presence of metal
complexes occurs at high temperatures.12 As all of the
above-mentioned compounds have various drawbacks, devel-
opment of new accelerators for curing reactions is of great
importance.
The use of anhydride hardeners in such compositions
gives materials with improved physical and mechanical
properties, for example, enhanced heat resistance.1,5 How-
ever, in the absence of catalysts, curing of epoxy–anhy-
dride compositions is very slow and occurs at extremely
high temperatures. To improve the energy and technologi-
Phosphorus compounds are of great interest in terms of
modifying the properties of epoxy composites and perform-
ance characteristics of materials based on these composites.
However, the influence of phosphorus compounds on the
curing rate of epoxy–anhydride composites has been ana-
lyzed only in a few studies.13,14 Similar to the nitrogen-
Additional Supporting Information may be found in the online version of this article.
C
V
2015 Wiley Periodicals, Inc.
1088
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
FIGURE 1 Chemical structures of ED-22, IMTHPA, PA, E and 2MI.
containing compounds the quaternary phosphonium salts
can also be assumed to be effective catalysts for the curing
of epoxy resins. The mechanism of epoxy–anhydride curing
using tertiary amines has been extensively studied and
described elsewhere,1,15,16 while the interaction of phospho-
nium salts with epoxides and anhydrides has been investi-
gated only in a sole work.13 However, thermal methods have
not been used in this study and the mechanical properties of
the obtained polymers still remained unexplored.
tives) were obtained from Acros Organics and were used as
received.
Synthesis of Phosphonium Salts
A number of phosphonium salts (F1–F5) was synthesized
according to known methods (Scheme 1).17–20 All phospho-
nium compounds were obtained in high yield. The structures
were identified by means of 1ꢀ and 31ꢁ NMR spectroscopy
(see Supporting Information), IR spectroscopy, mass spec-
trometry, and elemental analysis.21
In the present work, we investigated the kinetics of epoxy–
anhydride curing with phosphonium salts by DSC measure-
ments and intended to get new insights into the mechanism
of the curing process by mass-spectrometry.
METHODS
The NMR spectra of CDCl3 solutions were recorded with
Bruker AVANCE-500 instrument, observing H at 500.1 MHz.
1
Elemental analyses were carried out on a Carlo-Erba EA
1108 elemental analyzer. MALDI mass spectra were obtained
by ULTRAFLEX III mass spectrometer. Measurements were
conducted using plastic and metal plates. 2,5-Dihydroxyben-
zoic acid was used as a matrix.
EXPERIMENTAL
Materials
DGEBA type epoxy dianic resin ED-22 (epoxy number
23.5%) from Chimex Limited (Russia) was used as epoxy
oligomer. Anhydride of iso-methyltetrahydrophthalic acid
(IMTHPA) from Chimex Limited (Russia) was used as a hard-
ener. To compare the results we used also a commercial cat-
alyst 2-methylimidazole (2MI) from Chimex Limited (Russia).
For conducting the model reaction epichlorohydrin (E) and
phthalic anhydride (PA) from Alfa Aesar were used. The
chemical structures of ED-22, IMTHPA, E, PA, and 2MI are
shown in Figure 1. All chemicals used for the synthesis of
phosphonium salts (triphenylphosphine, bromine deriva-
The mechanism of curing was investigated using ESI mass
spectrometry performed on a Bruker Esquire spectrometer
using cone voltages in the positive and negative mode (610,
25, 50, 75, 100 V).
The curing reaction was investigated with a Netzsch DSC
204F1 Phoenix differential scanning calorimeter.
RESULTS AND DISCUSSION
Kinetics Investigation Using DSC Method
The effect of the synthesized phosphonium salts in compari-
son with 2-methylimidazole on the acceleration of epoxy–
anhydride curing was investigated by the DSC method.
The epoxy resin and hardener were employed in a 100:85
ratio. The catalyst was used as additive in 1–5 wt % relative
to epoxy oligomer.
The DSC curves of the ED-22–IMTHPA system with a series
of phosphonium salts and 2-methylimidazole are shown in
SCHEME 1 Synthesis of phosphonium salts IIa–e.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097
1089

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
FIGURE 2 The DSC curves of ED-22—IMTHPA—2 wt % accelera-
tor systems. Accelerators: F3 (1), F4 (2), F2 (3), 2MI (4), F1 (5), F5 (6).
FIGURE 3 The DSC curves of ED-22—IMTHPA—F3 system for
accelerator concentrations, wt %: 1 (1), 2 (2), 5 (3).
Figure 2. The general properties of accelerators as well as
heat flow, onset, and peak temperatures of these formula-
tions are collected in Table 1.
ers.25–30 The kinetics of curing reaction of epoxy anhydride
compounds catalyzed by tertiary amines was examined in
depth elsewhere31 using model-fitting non-isothermal DSC,
which is simple and convenient method of calculation of the
apparent activation energy values. The authors plotted the
heating rate versus the exothermic peak temperature, and
calculated the apparent activation energy from eq 1:
One can see that systems with the phosphonium salts F3
and F4 have the lowest peak temperature. Other systems
provide broadened curves with a higher curing temperature.
Moreover, the system containing F5 differs from all other
systems by the highest Tpeak value, as well as by the pres-
ence of a shoulder at the beginning of the curing.
lnðb=Tp2eak
1=Tpeak
Þ
52
Ea
R
(1)
It is well known that, the lower the peak temperature is the
more technological is the curing process. As the F3 and F4
salts showed the best results, further investigations were
carried out using these accelerators as promising candidates
for future applications.
where b is the sample heating rate, Tpeak is the exothermic
peak temperature at the appropriate heating rate, R is the
universal gas constant.
As was noted in Ref. 31 for tertiary amines, the type of
accelerator had little effect on the values of apparent activa-
tion energy while for the same accelerator a higher concen-
tration led to a lower activation energy.
The influence of accelerator concentration on the curing
temperature was also investigated for a number of epoxy–
anhydride compositions (e.g., with F3 accelerator, Fig. 3).
The curing kinetics of epoxy compositions can be studied
using various research methods.22–24 DSC is one of the most
popular tools for studying curing kinetics of epoxy oligom-
We conducted similar studies with the ED-22–IMTHPA sys-
tem using phosphonium salts F3 and F4. The DSC curves for
the ED-22–IMTHPA system with F3 and F4 phosphonium
TABLE 1 Parameters of the Accelerators, the Main Characteristics of the DSC Curves for the ED-22—IMTHPA—Accelerator (2 wt %)
Systems
The structure of compound
Notation
m.p. (8C)
289
ꢂon (8C)
113
ꢂpeak (8C)
155
Q (mW/mg)
966
F1
F2
209
111
139
970
F3
F4
244
285
111
112
137
138
994
999
F5
161
143
125
110
162
145
986
956
2MI
1090
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
TABLE 2 Values of the Apparent Activation Energy for the
ED-22—IMTHPA—Accelerator System for Various Amounts of
Accelerator
Heating
Content
(wt %)
rate,
Tpeak
ða*
Accelerator
F3
(8C/min)
(8C)
(kJ/mol)
1
2
5
1
2
5
1
2
5
5
143
150
161
169
137
149
156
162
125
136
143
148
144
153
162
171
138
150
156
162
126
138
144
150
149
162
176
181
145
155
162
168
140
149
154
161
62.31
62.31
62.31
65.97
65.97
65.97
70.06
70.06
70.06
10
15
20
5
10
15
20
5
10
15
20
5
F4
10
15
20
5
10
15
20
5
10
15
20
5
2MI
10
15
20
5
FIGURE 4 The DSC curves of ED-22—IMTHPA—accelerator, 2
wt % systems (accelerators: F3 (a), F4 (b), and 2MI (c)) at heat-
ing rates, 8C/min: 5 (1), 10 (2), 15 (3), 20 (4).
10
15
20
5
10
15
20
salts and 2-methylimidazole as catalysts are shown in Figure
4. Obviously, a greater amount of catalyst causes a lower a
peak temperature Tpeak. In contrast, the apparent activation
energy values are insensitive to accelerator content.
The curves of lnðb=Tp2eakÞ versus 1=Tpeak are demonstrated in
Figure 5. The calculated values of the apparent activation
energy are tabulated in Table 2.
FIGURE 5 The curves of ln(b/Tp2eak ) against 1/Tpeak of ED-22—
IMTHPA – accelerator (2 wt %) systems.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097
1091

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
SCHEME 2 Curing of epichlorohydrin and phthalic anhydride initiated by water.
Our results show that for various amounts of the catalyst the
values of the activation energy are practically equal. In con-
trast, the type of catalyst greatly influences the activation
energy.
that the curing mechanism for phosphonium salts and for
2-methylimidazole (and tertiary amines) is different.
Therefore, the next step was to investigate the curing mecha-
nism of epoxy–anhydride systems.
Another interesting result is that, compared to phosphonium
salts, 2-methylimidazole cures epoxy–anhydride composi-
tions at higher temperatures, despite the fact that its melting
point is much lower than Tpeak of these salts. We assume
Investigation of Curing Mechanism
The proposed curing mechanisms of epoxy–anhydride com-
positions using tertiary amines and imidazoles as catalysts
FIGURE 6 ESI mass spectra of the non-catalyzed E–PA polymerization in positive (a) and negative mode (b).
1092
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
SCHEME 3 Curing of epichlorohydrin and phthalic anhydride initiated by 2-phenylimidazole.
are well known and have been extensively described in the
literature.16,32–34
hydrin may act as an accelerator giving rise to anhydride
ring opening and subsequent nucleophilic attack of the acid
at the epoxide (Scheme 2).1
The only mention of a (unproved) mechanism for epoxy–
anhydride curing with phosphonium salts by Smith dates
back to 1973.13
The experiment was conducted by heating the mixture of
phthalic anhydride with epichlorohydrin at 110–113 8C for
2–3 hours. Then the reaction mixture was dissolved in
dichloromethane with a concentration of 1-2 mg/ml and
investigated by ESI mass spectrometry.
Aiming at the elucidation of the curing mechanism of epoxy
compounds in the presence of acid anhydrides catalyzed by
phosphonium compounds we decided to run a series of
model reactions between monomeric compounds monitored
by mass spectrometry.
The MS of the reaction mixture in positive mode [Fig. 6(a)]
displayed a peak at m/z 5 281 which can be assigned to the
[M1Na]1 ion of phthalic monoester A (Scheme 2). Structure
A was also confirmed by the [M–H]– peak at m/z 5 256.9
obtained in the negative mode [Fig. 6(b)]. Both peaks show
the typical isotopic pattern for compounds containing one
First, we investigated the curing of epichlorohydrin with
phthalic anhydride in the absence of any amine or phospho-
nium catalyst. In this system, the water present in epichloro-
FIGURE 7 Mass spectrum of the curing of epichlorohydrin and phthalic anhydride initiated by 2-phenylimidazole.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097
1093

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
SCHEME 4 Mechanism of the curing of E–PA with phospho-
nium salts by Smith.13
FIGURE 8 ESI mass spectrum of the curing of epichlorohydrin
and phthalic anhydride initiated by 2-phenylimidazole.
chlorine atom. In addition, the peak at m/z 5 515.0 corre-
sponds to the [2M2H]– ion of A. In the positive mode mass
spectrum we identified a series of peaks (m/z 5 373.1,
613.2, 853.3, 1093.2) with a mass difference of 240 charac-
teristic for the E–PA unit. They can be attributed to the
[M1Na]1 ions of copolymerization products starting from
compound B (m/z 5 350.0) via C (m/z 5 590.1) and so forth.
Peaks in the negative mode MS can be attributed to dimeric
ions as follows: 515.0 [2A–H]–, 606.8 [A1B2H]2, 698.9
[2B2H]2, and 790.9 [B1D2H]2.
In summary, the ESI MS results account for two parallel
processes: (a) homopolymerization giving polymers consist-
ing mostly of epichlorohydrin units; (b) copolymerization
producing polymers with an equal distribution of epoxide
and anhydride units.
FIGURE 10 Mass spectrum of E–PA–F3 system in positive
mode and proposed structures.
FIGURE 9 ESI mass spectrum of E-PA-F3 system in positive mode.
1094
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
SCHEME 5 Curing reaction of the E–PA–F3 system.
The basic idea of the curing using imidazole as accelerator
includes an initial nucleophilic ring opening of the anhydride
[Scheme 3(a)] or epoxide (Scheme 3b) by the nitrogen atom
and subsequent reaction of the carboxyl anion with the epox-
ide [Scheme 3(a)] or reaction of the alkoxide with the anhy-
dride [Scheme 3(b)].22,35,36 As this mechanism has long been
suggested but has not been proven,29 the process of curing
using imidazole as catalyst was also under consideration in
our work. 2-Phenylimidazole (2PI) was used as accelerator
(2 wt %) and, as in the previous case, the mixture was
heated at 110 8C for 2 hours.
ence of peaks differing by the mass of the E–PA unit (m/
z 5 819.4 and m/z 5 1059.3) confirms the proceeding of the
polymerization.
In a more detailed study of the mass spectra, the structure of
an epoxide–anhydride reaction product without imidazole moi-
ety was identified. Figure 8 shows the negative mode mass
spectrum for the epoxide–anhydride curing using imidazole as
accelerator. The peak at m/z 5 589.2 corresponds to the
[M2H]2 ion of structure E depicted in Figure 8. In the negative
mode the molecule can also add a chlorine anion. Thus, the
peak at m/z 5 625.2 can be assigned to the [M1Cl]2 ion.
However, in the ESI mass spectrum of the mixture no peaks
corresponding to the proposed imidazole-containing inter-
mediates have been found (Fig. 7). Nevertheless, the pres-
Even though we could not find ions by ESI MS which corre-
spond to imidazole-containing polymerization products, we
SCHEME 6 Structures derived from the mass spectrum of E–PA–F3 curing.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097
1095

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
cannot rule out this mechanism completely. It seems not
unlikely that the imidazole functions are split off for example
by elimination reactions from the alkylated imidazole [Scheme
3(b)] of by hydrolysis of the carbonylimidazole obtained
according to Scheme 3(a). The imidazole-accelerated polymer-
ization may occur in parallel to the curing process initiated by
water through the mechanism depicted in Scheme 2.
to the epoxide, the anhydride may function as a nucleophile
with the most basic position at the carbonyl oxygen atom
(Scheme 5). Concomitantly, a water molecule can attack the
second carbonyl group producing the carboxylic acid func-
tion in compound H which subsequently reacts with a mole-
cule of epichlorohydrin to give I. Further reaction with one
molecule of phthalic anhydride and one molecule of epichlor-
ohydrin affords J. Repetition of this process provides product
G the mass of which was confirmed by ESI MS in form of
the [M1NH4]1 and [M1Na]1 ions. Alternatively, reaction of
J with ester H leads to the oligomer F which has also been
identified by its [M1NH4]1 and [M1Na]1 ions in ESI MS.
In-depth study of the curing process was undertaken using
butyltriphenylphosphonium bromide (F3) as accelerator.
The ESI mass spectrum of the E–PA–F3 system after heating
proved that the curing reaction had occurred (Fig. 9). This is
demonstrated by the presence of peaks for ions with a mass
difference of 240 (m/z 5 781.4, 1021.4, 1261.3) characteris-
tic for the E–PA unit.
The third group of peaks in the ESI mass spectrum (Fig. 10)
corresponding to a structure with four units of epichlorohy-
drin still remains to be explained. We propose that it results
from structure K (Scheme 6) which is formed by water-
initiated curing of the E–PA system according to Scheme 2.
Subsequent elimination of water, either during the reaction
or under MS conditions, affords olefin L with a molecular
mass of m/z 5 812.1. The ESI MS peaks at 830.3 and 835.3
refer to the [M1NH4]1 and [M1Na] ions of L, respectively.
The only mechanistic proposal for the epoxide–anhydride
curing with phosphonium salts was reported by Smith.13
This includes anhydride ring opening by attack to an epoxide
that is activated by the phosphonium salt by hydrogen
bonding or protonation (Scheme 4, eq. a). The carbon adja-
cent to the phosphorus acts as a nucleophile (may be in
form of the corresponding ylide) which reacts with the car-
bonyl group. Alternatively, the anhydride can be directly
opened by reaction with the phosphonium salt (Scheme 4,
eq. b) exploiting the acidity of the phosphonium species in
a-position and the nucleophilicity of the a-carbon.
Similar results were found for another phosphonium salt (F2)
that confirm the suggested mechanism of curing acceleration by
phosphonium salts in epoxy–anhydride systems. In contrast, no
phosphonium-containing products have been identified which
would support the mechanism proposed by Smith with the phos-
phonium salt playing the role of a nucleophile after preceding
deprotonation (Scheme 4).13 Thus, it seems more likely that the
phosphonium accelerators act via an electrophilic mechanism
(Scheme 5) similar to those proposed by Park et al. for the homo-
polymerization of epoxides.
We envisaged to find proves for this mechanism. Thus, we
reacted epichlorohydrin and phthalic anhydride in the presence
of butyltriphenylphosphonium bromide and followed the reac-
tion by ESI mass spectrometry. However, no intermediates of
this mechanism could be identified in the mass spectra.
CONCLUSIONS
A zoomed part of the mass spectrum of the E–PA–F3 system
is presented in Figure 10. A pairwise arrangement of peaks
with a mass difference of 5 was observed in the spectrum.
Thus, these peaks can be attributed to [M1NH4]1 (ꢃ118)
and [M1Na]1 (ꢃ123) ions, respectively. This allows to cal-
culate the molecular mass of the parent compounds as m/
z 5 758.4 for the first group of signals, m/z 5 794.4 for the
second and m/z 5 812.3 for the third set of signals. More-
over, based on the distribution of chlorine isotope peaks, it
was established that the first group of peaks corresponds to
a structure containing two units of epichlorohydrin, the sec-
ond three units, and the third four units of epichlorohydrin.
A series of triphenylphosphonium salts was for the first time
used as accelerators for curing epoxy–anhydride composi-
tions. The acceleration of the curing process in the presence
of phosphonium salts was demonstrated by DSC, determin-
ing concentration and temperature dependences of apparent
activation energy values. The activation energy for the over-
all curing process (the diffusion stage) of the epoxy–anhy-
dride system with 2-methylimidazole as a catalyst is greater
than those with two phosphonium salts. Thus, the phospho-
nium salts are more effective accelerators for epoxy–anhy-
dride curing.
Accordingly, the masses of 758.4 and 794.4 have been tenta-
tively assigned to structures F and G which both contain the
butyl fragment from the phosphonium salt (Fig. 10). This
confirms the direct participation of the phosphonium salt in
the curing reaction.
The curing reaction was run with epichlorohydrin and
phthalic anhydride as model compounds, and the products
were analyzed by mass spectrometry (ESI). For the
imidazole-initiated curing the imidazole moiety was not
found to be embedded in the polymer network. However, it
might be split off from the polymer by hydrolysis or elimina-
tion. Conversely, a competitive curing pathway was docu-
mented which is initiated by water.
In agreement with the mechanism of the phosphonium-
initiated homopolymerization of epoxides proposed by Park
37
et al.
the formation of the butyl ester can be explained by
a nucleophilic substitution reaction at the a-carbon atom of
the phosphonium salt releasing triphenylphosphine. Similar
A novel mechanism of epoxy–anhydride curing with phos-
phonium salts as accelerators is proposed as an alternative
1096
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
Br2]1. Butane-1,4-bis(triphenylphosphonium) dibromide (IIb).
Reagents and conditions: PPh3 (0.11 mol), 1,4-dibromobutane
(0.046 mol), n-PrOH. A white solid; m.p. 287–2898ꢄ, 76% yield.
1ꢀ NMR (500 MHz, CDCl3, d): 2.14-2.20 (m, 4H, CH2ꢄꢀ2), 3.99
(dt, J 5 13.61 Hz, 4H, ꢄꢀ2PPh3); 7.59–7.69 (m, 15ꢀ, Ar-H); 7.83–
7.89 (m, 15ꢀ, Ar-H). 31P NMR (400 MHz, CDCl3, d): 24.49. Anal.
Calcd for C40H38Br2P2: C, 64.52; H, 4.98; Br, 21.50; P, 8.26;
found: C, 64.86; H, 5.14; Br, 21.62; P, 8.38. MS (MALDI), m/
z5579 [(M-HBr-Br2)]1, 290 [(M-2Br2)/2]1. Butyltriphenylphos-
phonium bromide (IIc). Reagents and conditions: PPh3
(0.19 mol), 1-bromobutane (0.012 mol), toluene. white solid;
m.p. 243–2458ꢄ, 73% yield. 1ꢀ NMR (500 MHz, CDCl3, d): 0.75
(t, J 5 7.03 Hz, 3H, ꢄH3); 1.45–1.53 (m, 4H, ꢄH2); 3.53 (dt,
J 5 12.9 Hz, 2H, ꢄꢀ2PPh3); 7.55–7.60 (m, 6ꢀ, Ar-H); 7.65–7.70
(m, 9ꢀ, Ar-H). 31P NMR (400 MHz, CDCl3, d): 23.98. Anal. Calcd
for C22H24BrP: C, 65.74; H, 5.78; Br, 20.11; P, 7.65; found: C,
66.16; H, 6.02; Br, 20.05; P, 7.77. MS (MALDI), m/z 5 319
[M-Br2]1. Benzyltriphenylphosphonium bromide (IId). Reagents
and conditions: PPh3 (0.015 mmol), bromomethylbenzene
(0.012 mol), toluene. white solid; m.p. 284–2888ꢄ, 87% yield. 1ꢀ
NMR (500 MHz, CDCl3, d): 5.28 (d, J 5 14.36 Hz, 2H, ꢄꢀ2PPh3);
7.02–7.08 (m, 4ꢀ, Ar-H); 7.14–7.19 (m, 1ꢀ, Ar-H); 7.55–7.60 (m,
6ꢀ, Ar-H); 7.63–7.74 (m, 9ꢀ, Ar-H). 31P NMR (400 MHz, CDCl3, d):
23.13. Anal. Calcd for C25H22BrP: C, 68.97; H, 4.98; Br, 18.52; P,
7.24; found: C, 69.28; H, 5.08; Br, 18.48; P, 7.16. MS (MALDI), m/
z 5 353 [M-Br2]1. (Ethoxycarbonylmethyl)triphenylphosphonium
bromide (IIe). Reagents and conditions: PPh3 (0.038 mol), ethyl
bromoacetate (0.053 mol), ethanol. white solid; m.p. 159–1638ꢄ,
68% yield. 1ꢀ NMR (500 MHz, CDCl3, d): 0.74 (t, J 5 7.14 Hz, 3H,
ꢄH3); 3.72 (q, J 5 7.14 Hz, 2H, ꢄH2CH3), 5.11 (d, J 5 13.88 Hz, 2H,
ꢄH2PPh3); 7.40–7.45 (m, 6ꢀ, Ar-H); 7.53–7.64 (m, 9ꢀ, Ar-H). 31P
NMR (400 MHz, CDCl3, d): 20.41. Anal. Calcd for C22H22BrO2P: C,
60.96; H, 4.88; Br, 18.52; P, 7.15; found: C, 61.54; H, 5.13; Br,
18.65; P, 7.23. MS (MALDI), m/z 5 349 [M–Br2]1.
to the previous one by Smith.13 In the mass-spectra molecu-
lar ions were detected containing the alkyl part of the phos-
phonium salt in form of alkyl esters. Thus, we propose that
the phosphonium salt accelerates the curing via electrophilic
attack at the phthalic anhydride.
ACKNOWLEDGMENT
This work is performed according to the Russian Government
Program of Competitive Growth of Kazan Federal University.
REFERENCES AND NOTES
1 H. Lee, K. Neville, Handbook of Epoxy Resins; Mcgraw-Hill:
New York, 1967.
2 X. M. Chen, B. Ellis, Chemistry and Technology of Epoxy Res-
ins, Blackie A & P, Chapman & Hall: London, 1993.
3 L. A. White, J. W. Weber, L. J. Mathias, Polym. Bull. 2001, 46,
339–344.
4 V. J. Eddy, J. E. Hallgren, R. E. Colborn, J. Polym. Sci. Part a:
Polym. Chem. 1990, 28, 2417–2426.
5 J. F. Zhang, X. Sun, Biomacromolecules 2004, 5, 1446–1451.
6 E. M. Woo, J. C. Seferisi. J. Appl. Polym. Sci. 1990, 40, 1237–
1256.
7 I. Hamerton, B. J. Howin, P. Jepson. Coord. Chem. Rev. 2002,
224, 67–85.
8 A. V. Kurnoskin, J.M.S.-Rev. Macromol. Chem. Phys. C 1996,
36, 457–599.
9 S. Montserrat, C. Flaqut, M. Calafell, G. Andreu, J. Malek,
Thermochim. Acta 1995, 269/270, 213–229.
22 J. Rocks, L. Rintoul, F. Vohwinkel, G. George, Polymer 2004,
45, 6799–6811.
10 N. Chantarasiri, N. Sutivisedsak, C. Pouyuan. Eur. Polym. J.
2001, 37, 2031–2038.
23 C. Jubsilp, S. Damrongsakkul, T. Takeichi, S. Rimdusit, Ther-
mochim. Acta 2006, 447, 131–140.
11 M. S. Fedoseev, L. F. Derzhavinskaya, V. N. Strelnikov, Russ.
J. Appl. Chem. 2010, 83, 1408–1412.
24 R. Biju, C. P. Reghunadhan Nair, C. Gouri, K. N. Ninan, J.
Therm. Anal. Calorim 2012, 107, 693–702.
12 Z. Zhang, C. P. Wong, J. Appl. Polym. Sci. 2002, 86, 1572–
1579.
25 P. Navabpour, A. Nesbitt, T. Mann, R. J. Day, J. Appl.
Polym. Sci. 2006, 104, 2054–2063.
13 J. D. B. Smith, J. Appl. Polym. Sci. 1979, 23, 1385–1396.
14 I. Glavchev, K. Petrova, I. Devedjiev, Polym. Testing 2002,
21, 243–245.
26 M. Harsch, J. Karger-Kocsis, M. Holst, Eur. Polym. J. 2007,
43, 1168–1178.
15 L. Matejka, J. Lovy, S. Pokorny, K. Bouchal, K. Dusek, J.
Polym. Sci., Polym. Chem. Ed. 1983, 21, 2873–2885.
27 P. P. Adroja, R. Y. Ghumara, P. H. Parsania, J. Appl. Polym.
Sci. 2013, 130, 572–578.
16 X. Fernandez-Francos, X. Ramis, A. Serra, J. Polym. Sci.
Part A: Polym. Chem. 2014, 52, 61–75.
28 X. Huang, B. Patham, J. Appl. Polym. Sci. 2013, 127, 1959–
1966.
17 A. Mondon, Liebigs Ann. Chem. 1957, 603, 115–129.
29 M. S. Heise, G. C. Martin, J. Appl. Polym. Sci. 1990, 39, 721–738.
30 J. G. Gao, X. Zhang, L. Huo, Iran Pol. J. 2010, 19, 731–742.
18 J. C. Chang, W. Y. Ho, I. W. Sun, Y. K. Chou, H. H. Hsieh, T.
Y. Wu, Polyhedron 2011, 30, 497–507.
31 T. Yang, C. Zhang, J. Zhang, J. Cheng, Thermochim. Acta
2014, 577, 11–16.
19 M. Delmas, Y. Le Bigot, A. Gaset, J. P. Gorrichon. Synth.
Comm. 1981, 11, 125–132.
32 L. Matejka, J. Lovy, S. Pokorny, K. Bouchal, K. Dusek, J.
Polymer Sci. Polymer Chem. Ed. 1983, 21, 2873–2885.
20 J. J. Kiddle,. Tetrahedron Lett. 2000, 41, 1339
21 (4-Bromobutyl)triphenylphosphonium
bromide
(IIa).
33 J. Berger, F. Lohse, J. Appl. Polym. Sci. 1985, 30, 531–546.
Reagents and conditions: PPh3 (14 mmol), 1,4-dibromobutane
(0.14 mmol). white solid; m.p. 208–211 8ꢄ, 92% yield. 1ꢀ NMR
(500 MHz, CDCl3, d): 1.74-1.84 (m, 2H, ꢄH2ꢄꢀ2PPh3); 2.22 (p,
J 5 6.64 Hz, 2H, ꢄH2ꢄꢀ2Br); 3.5 (t, J 5 5.56 Hz, 2H, ꢄH2 Br); 3.8
(dt, J 5 13.1 Hz, 2ꢀ, ꢄꢀ2PPh3,); 7.45–7.68 (m, 6ꢀ, Ar-H); 7.73-
7.94 (m, 9ꢀ, Ar-H). 31P NMR (400 MHz, CDCl3, d): 24.32. Anal.
Calcd for C22H23Br2P: C, 54.94; H, 5.08; Br, 33.36; P, 6.35; found:
C, 55.23; H, 4.81; Br, 33.47; P, 6.48. MS (MALDI), m/z5398 [M-
34 A. R. Mahendran, G. Wuzella, A. Kandelbauer, N. Aust, J.
Therm. Anal. Calorim. 2012, 107, 989–998.
35 R. F. Fischer, J. Polym. Sci. 1960, 44, 155–172.
36 P. Musto, M. Abbate, G. Ragosta, G. Scarinzi, Polymer.
2007, 48, 3703–3716.
37 S. J. Park, G. Y. Heo, D. H. Suh, J. Polym. Sci. Part A:
Polym. Chem. 2003, 41, 2393–2403.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2016, 54, 1088–1097
1097