Systematic study on highly efficient Thermal Synergistic Polymerization effect between alicyclic imide moiety and phthalonitrile: Scope, Properties and Mechanism

Article abstract of DOI:10.1016/j.polymer.2016.09.025

Blends of phthalonitrile (CN) and alicyclic imide compounds synthesized from tetrahydrophthalic anhydride (CC), methyl tetrahydrophthalic anhydride (MCC) and hexahydrophthalic anhydride (HCC), respectively, were prepared and characterized. Results from rheological studies demonstrated the scope of a novel Thermal Synergistic Polymerization (TSP) effect between phthalonitrile and the alicyclic imide compounds. FTIR data showed a unique phenomenon, which is, the [sbnd]C[tbnd]N of the phthalonitrile compound could be completely and rapidly consumed without generating any triazine ring. Further works on cured products of CC/CN system showed comparable or even higher thermal oxidation stability, low boiling water absorbance and moderate dielectric constant (ε′ in the range of 4.8～4.5) with almost no voids compared with that of the phthalonitrile resins promoted by traditional curing system of active hydrogen. Model system was designed to study the mechanism of TSP effect of alicyclic imide moiety/phthalonitrile. The systematic characterizations by FTIR, UV–Vis, MALDI-TOF MS, free radical scavenger (DPPH), EPR, etc. showed that the TSP reaction undergoes both copolymerization and homopolymerization with no triazine formation, involving a free radical process. Structural analysis of the oligomers suggested that phthalocyanine rings were a minor component of the thermoset networks, while polyisoindolines were the primary structural motif.

Full text of DOI:10.1016/j.polymer.2016.09.025

Accepted Manuscript
Systematic study on highly efficient Thermal Synergistic Polymerization effect
between alicyclic imide moiety and phthalonitrile: Scope, properties and mechanism
Ping Yuan, Suchun Ji, Jianghuai Hu, Xueping Hu, Ke Zeng, Gang Yang
PII:
S0032-3861(16)30824-2
10.1016/j.polymer.2016.09.025
JPOL 19031
DOI:
Reference:
To appear in: Polymer
Received Date: 10 May 2016
Revised Date: 31 August 2016
Accepted Date: 10 September 2016
Please cite this article as: Yuan P, Ji S, Hu J, Hu X, Zeng K, Yang G, Systematic study on highly
efficient Thermal Synergistic Polymerization effect between alicyclic imide moiety and phthalonitrile:
Scope, properties and mechanism, Polymer (2016), doi: 10.1016/j.polymer.2016.09.025.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Systematic study on highly efficient Thermal Synergistic
Polymerization effect between alicyclic imide moiety and
phthalonitrile: Scope, Properties and Mechanism
Ping Yuan, Suchun Ji, Jianghuai Hu, Xueping Hu, Ke Zeng*, Gang Yang*
State Key Laboratory of Polymer Materials Engineering, College of Polymer Science and
Engineering, Sichuan University, Chengdu 610065, People’s Republic of China
*Corresponding author: Tel: +86-28-85462736, Fax +86-28-85462736
E-mail addrsses:zk_ican@sina.com (K. Zeng);yanggang65420@163.com (G. Yang)
A B S T R A C T
Blends of phthalonitrile (CN) and alicyclic imide compounds synthesized from tetrahydrophthalic
anhydride (CC), methyl tetrahydrophthalic anhydride (MCC) and hexahydrophthalic anhydride (HCC),
respectively, were prepared and characterized. Results from rheological studies demonstrated the scope
of a novel Thermal Synergistic Polymerization (TSP) effect between phthalonitrile and the alicyclic imide
compounds. FTIR data showed a unique phenomenon, which is, the -C≡N of the phthalonitrile
compound could be completely and rapidly consumed without generating any triazine ring. Further
works on cured products of CC/CN system showed comparable or even higher thermal oxidation
stability, low boiling water absorbance and moderate dielectric constant (ε’ in the range of 4.8~4.5) with
almost no voids compared with that of the phthalonitrile resins promoted by traditional curing system
of active hydrogen. Model system was designed to study the mechanism of TSP effect of alicyclic imide
moiety/phthalonitrile. The systematic characterizations by FTIR, UV-Vis, MALDI-TOF MS, free radical
scavenger (DPPH), EPR, etc. showed that the TSP reaction undergoes both copolymerization and
homopolymerization with no triazine formation, involving a free radical process. Structural analysis of
the oligomers suggested that phthalocyanine rings were a minor component of the thermoset networks,
while polyisoindolines were the primary structural motif.
KeyWords
Phthalonitriles; Alicyclic Imide Moiety; Thermal Synergistic Polymerization; Curing Mechanism; Electron
Paramagnetic Resonance; Thermosetting Resins; High Performance Polymers
1. Introduction

ACCEPTED MANUSCRIPT
The phthalonitrile polymers constitute a class of high-temperature polymers that have a
number of outstanding properties such as high glass transition temperatures (Tg), outstanding
thermal and thermo-oxidative stability, excellent mechanical properties, good moisture
resistance and superior fire resistance [1-7]. In addition, phthalonitrile monomers and
oligomers possess a low complex melt viscosity (0.01~1 Pa.s) which enables facile processing by
cost effective, non-autoclavable processing techniques [8-10]. The combination of good
processability and excellent high-temperature properties makes the phthalonitirles attractive
for many military and civilian advanced technological applications [1-11].
However, the polymerization of the neat resin is extremely sluggish, which limits their wide
application [4-5, 8]. To address this problem, considerable research efforts have been expended
on the development for effective additions to accelerate the curing rate, for instance, Lewis
acids such as phenols [7], strong organic acid [12], benzimidazole [8] and hydrogen imide [9], or
Lewis bases such as organic amine [4]. However, these curing agents show many defects, such
as low curing efficiency, high curing temperatures and brittle nature of the cured polymers [4,
8-9]. Thus, the study on the new curing methods becomes an important factor which
determines the further application of the phthalonitrile polymers.
As we know, high performance resins based on these reactive groups of cyanate ester [13],
maleimide [14], phenylethynyl [15-16] and benzocyclobutene [17] have received much
attention in the past decades, in the view of scientific or industrial value. This is due to a fact
that these resins cure via an addition reaction with no volatile by-products evolved during the
processing stage. To further expand their application, functional combination of the different
types of the resins into a new system is a common approach to obtain polymeric materials with
attractive combination of properties. As is known, hetero-atoms such as nitrogen atom play an
important role in tailoring the functionality of the material. In this situation, introducing hetero-
atom by in-situ chemical interaction, a “chemical synergistic effect” between the reactive
groups including but not limited to copolymerization and reciprocal catalyzation, might provide
crucial advantages, for example, producing polymers with versatile structures, good
compatibility, adjustable processabilities and unique functionality. However, few efforts and
breakthroughs were made on combining resins with hetero-atom reactive groups and carbon
unsaturated groups. This is partly due to the difference in the reactivity of the groups by radical
and ionic paths. An example very close to this concept is cyanate-bismaleimide known as B-T
resins possessing attractive thermal, mechanical and dielectric properties, which have been
fully studied and developed, though the underlying reaction mechanism remains unclear [18-
19]. This study offers a significant promise of this new concept by achieving in-situ chemical
interaction between an alicyclic imide moiety and a hetero-atom reactive group, nitrile -C≡N. As
is known, alicyclic moiety widely exists in natural resources, such as turpentine [20]. Therefore,

ACCEPTED MANUSCRIPT
the concept of “chemical synergistic effect” reported in our group has opened an exciting
opportunity for designing new high performance thermosetting polymers.
Along with this new concept, the precious communication in our group reported that methyl
tetrahydrophthalic anhydride end-capped imide compound (MCC) and phthalonitrile (CN)
showed Thermal Synergistic Polymerization (TSP) effect [21]. However, the scope and
mechanism of the TSP effect is not well studied and understood.
On the other hand, Electron Paramagnetic Resonance (EPR) spectroscopy is a very useful tool
for the investigation of paramagnetic species [22-23]. Furthermore, EPR spectroscopy provides
a unique method for determining the sites of attack of reactive free radicals on organic
substrates and biomolecules. Under certain circumstances, it is possible to estimate absolute
free radical concentrations and, from hyperfine splitting, to characterize individual free radicals.
Therefore, structures, concentrations, and dynamics of paramagnetic compounds may be
obtained from EPR measurements in the study of free radical polymerization. However, there
have been no literatures reported a clearer insight into many of the fundamental mechanism of
phthalonitriles, especially free radical mechanisms at high temperatures. To the best of our
knowledge, this is the first time that paramagnetic signals are observed at high temperatures
(>200 °C) especially for the thermosetting resins, such as phthalonitriles.
In this study, a series of alicyclic imide moiety-containing compounds (tetrahydrophthalic
anhydride, methyl tetrahydrophthalic anhydride and hexahydrophthalic anhydride end-capped
imide compound, namely CC, MCC and HCC) were introduced to study the scope of the TSP
effect. Unexpectedly, the results indicated even the carbon saturated HCC also display TSP
effect with CN, which broaden the scope of the TSP effect. The CC/CN system was chosen to
fully characterize the cure behavior and properties of the derived polymers. The cured
polymers showed comparable or even higher thermal oxidation stability with almost no voids
compared with that of the phthalonitrile resins promoted by traditional curing system of active
hydrogen [2, 4, 24-26]. Additionally, water absorbance at boiling condition and dielectric
constant were characterized. Two model compounds (CC1 and CN1) were designed and
synthesized to explore the TSP mechanism. Structural analysis were then investigated by
systematic characterizations by FTIR, UV-Vis, MALDI-TOF MS, free radical scavenger (DPPH: 2,2-
diphenyl-1-picrylhydrazyl) and EPR. The results showed that the TSP reaction undergoes both
copolymerization and homopolymerization, with minor phthalocyanine ring, major aromatic
conjugated rings likely isoindoline and no triazine formation. The TSP reaction between alicyclic
imide moiety and phthalonitrile suggests a free radical process.
2. Experimental
2.1. Materials

ACCEPTED MANUSCRIPT
4, 4'-oxydianiline (ODA) was obtained from Shandong Wanda Chemical Industries Co., Ltd. Cis-
1, 2, 3, 6-tetrahydrophthalic anhydride (97 wt%), cis-1, 2-cyclohexanedicarboxylic anhydride (99
wt%), 2,2-diphenyl-1-picrylhydrazyl (DPPH, 96 wt%) were obtained from Aladdin Industrial
Corporation (Shanghai). 4-methyl-4-cyclohexene-1, 2-dicarboxylic anhydride (95 wt%) was
obtained from Puyang Huicheng electronic material Co., Ltd. Phthalic anhydride (99.7 wt%) was
obtained from Tianjin Bodi Chemical Co., Ltd. 4-nitrophthalonitrile was obtained from Taixing
Shengming Fine Chemical Co., Ltd. 4-phenylphenol (97 wt%) was obtained from Adamas
Reagent, Co., Ltd. (Shanghai) and all of above were used without further purification. Resorcinol
(99.5 wt%), dimethyl acetamide (DMAc), dimethyl sulfoxide (DMSO), Et3N, acetic acid, H2SO4,
NaCl, NaHCO3, K2CO3, CH2Cl2, THF, CH3CN, toluene, acetone and ethanol were obtained from
Chengdu Kelong Chemical Co., Ltd. Other chemicals were used as received unless otherwise
stated.
2.2. Synthesis of alicyclic imide compounds (CC, MCC and HCC)
The tetrahydrophthalic anhydride end-capped imide compound CC (Scheme 1) was prepared
from cis-1, 2, 3, 6-tetrahydrophthalic anhydride (THPA) and ODA in DMAc solvent. To a 250-mL
three-neck round-bottom flask with a Dean-Stark trap was added ODA (9.63 g, 0.048 mol) and
100 mL DMAc. The dissolved solution was stirred and heated to 30 °C, then added THPA (97
wt%, 15 g, 0.096 mol) and the system was quickly dissolved completely. The resultant mixture
was purged with nitrogen and maintained at 35 °C for 5 h. To the solution was added 50 mL
toluene, then heated to reflux at about 133 °C and the water was removed azeotropically over a
period of 5.5 h. After removing the toluene under reduced pressure and cooling to room
temperature (RT), the solution was poured into water, stirred, filtered, thoroughly washed with
methanol and dried. The crude product was purified by rinsing with CH2Cl2 in Buchner funnel
and suction flask, and the filtrate was then dried by evaporation under a vacuum at 80 °C for 6 h
to remove CH2Cl2. The white powder was obtained in 45.9 wt% yield (10.3 g), m. p. peak at 240
°C (sharp) (DSC) at a heat rate of 10 °C/min. FTIR (KBr), 2950, 2925, 2896, 2874 and 2844
1
(alicyclic C-H), 1782 and 1705 (imide C=O), 1238 (Ar-O-Ar), 694 (alicyclic –CH=CH–). H NMR
(400MHz, DMSO-d6): 7.21-7.10 (dd, 8H; Ar H), 5.91 (s, 4H; alicyclic –CH=CH–), 3.26-3.25 (m, 4H;
alicyclic–CH–), 2.46-2.21 (dd, 8H; alicyclic –CH2–).
The methyl tetrahydrophthalic anhydride end-capped imide compound MCC (Scheme 1) was
prepared by 4-methyl-4-cyclohexene-1, 2-dicarboxylic anhydride (MTHPA) and ODA in acetic
acid solvent. To a 250 mL round-bottom flask was added ODA (8.06 g, 0.04 mol) and 143 mL
acetic acid, stirred, then added MTHPA (13.39 g, 0.08 mol) and stirred for 30 min at room
temperature in argon. The suspension was then refluxed at about 130 °C for 2 h, cooled and
poured into water, then filtered and thoroughly washed with water and dried. The crude
product was purified by rinsing with CH2Cl2 in Buchner funnel and suction flask, and the filtrate

ACCEPTED MANUSCRIPT
was then dried by evaporation under a vacuum at 80 °C for 5 h to remove CH2Cl2. The light
yellow powder was obtained in 26.8 wt% yield (5.3 g), m. p. peak at 166 °C (sharp) (DSC) at a
heat rate of 10 °C/min. FTIR (KBr), 2963, 2946, 2921, 2900, 2848 (alicyclic C-H), 1780 and 1709
1
(imide C=O), 1233 (Ar-O-Ar). H NMR (400 MHz, CDCl3): 7.26-7.07 (m, 8H; Ar H), 5.62 (s, 2H;
alicyclic –CH=CH–), 3.26-3.20 (m, 4H; alicyclic –CH–), 2.69-2.56 (m, 4H; alicyclic –CH2–), 2.34-
2.25 (m, 4H; alicyclic –CH2–), 1.77 (s, 6H; –CH3).
The hexahydrophthalic anhydride end-capped imide compound HCC (Scheme 1) was
prepared by cis-1, 2-cyclohexanedicarboxylic anhydride (HHPA) and ODA in acetic acid solvent.
To a 100 mL round-bottom flask was added ODA (5.01 g, 0.025 mol) and 51 mL acetic acid,
stirred, then added HHPA (7.79 g, 0.05 mol) and stirred for 30 min at room temperature in
argon. The suspension was then refluxed at about 135 °C for 2 h, cooled and crystallized, then
filtered and thoroughly washed with 1.0 wt% NaHCO3 solution and water. The filtrate was then
dried by evaporation under a vacuum at 80 °C for 6 h. The white powder was obtained in 83.9
wt% yield (9.9 g), m. p. peak at 250 °C (sharp) (DSC) at a heat rate of 10 °C/min. FTIR (KBr),
1
2941, 2857 (alicyclic C-H), 1778 and 1709 (imide C=O), 1240 (Ar-O-Ar). H NMR (400 MHz,
CDCl3): 7.30-7.25 (m, 4H; Ar H), 7.13-7.09 (m, 4H; Ar H), 3.08-3.00 (m, 4H; alicyclic –CH–), 1.93-
1.85 (m, 8H; alicyclic –CH2–), 1.54-1.52 (m, 8H; alicyclic –CH2–).
2.3. Synthesis of phthalonitrile compound (CN)
To a 500 mL three-neck round–bottom flask was added resorcinol (16.52 g, 0.15 mol), 4-
nitrophthalonitrile (51.94 g, 0.30 mol) and 180 mL DMSO. During the course of the reaction, the
K2CO3 (62.19 g, 0.45 mol) was added in three portions at an interval of 20 min. Then the
resulting mixture was heated at 40 °C for 12 h under nitrogen. After cooling, the product
mixture was poured into water. The light yellow filtrate was collected by suction filtration and
washed with large amount of water until the filtrate was neutral. The crude product (42.0 g)
was purified by rinsing with CH2Cl2 in Buchner funnel and suction flask, and the filtrate was then
dried by evaporation under a vacuum at 80 °C for 5 h to remove CH2Cl2. The yellow-white
powder was obtained in 68.7 wt% yield (28.9 g), m. p. peak at 183°C (sharp) (DSC) at a heat rate
1
of 10 °C/min. FTIR (KBr), 2232 (C≡N), 1284 (Ar-O-Ar), 1248(Ar-O-Ar). H NMR (400MHz, DMSO-
d6): 8.14-8.11 (d, 2H; Ar H), 7.92 (s, 2H; Ar H), 7.63 -7.53(m, 3H; Ar H), 7.15-7.12 (m, 3H; Ar H).
2.4. Synthesis of alicyclic imide model compound (CC1)
The tetrahydrophthalic anhydride end-capped imide model compound CC1 (Scheme 2) was
prepared in two steps with intermediate PA-ODA-NH2. Firstly, PA-ODA-NH2 was prepared in
phthalic anhydride (PA) and ODA in DMAc solvent. To a 250-mL three-neck round-bottom flask
was added ODA (20.02g, 0.1 mol) and 70 mL DMAc, the reaction mixture was soon dissolved,

ACCEPTED MANUSCRIPT
then to the flask with a dropping funnel was added DMAc (70mL) solution of PA (14.81g, 0.1
mol) dropwise (about a drop/8~10 s) at room temperature, and then the reaction mixture was
maintained for 2 h in nitrogen. To the flask was added Et3N (2.04 g, 0.02 mol), and then heated
to reflux at about 160 °C for 2 h. After cooling, the resulting reaction mixture was poured into
1000 mL methanol and stirred for 4 h, filtered to remove most of phthalic imide on both sides,
then the resulting filtrate was evaporated under vacuum at 65 °C for 1.5 h to remove the
methanol. The concentrated solution was poured into 2000 mL water with stirring and
continuously stirred for 1 h, filtered and washed thoroughly with water, giving 26.5 g crude
amine intermediate after drying for 2 h at 80 °C under rotary evaporation. The crude amine
intermediate was dissolved in THF solvent and filtered to remove little insoluble impurities. The
filtrate was then purified by purging dry HCl gas, prepared by dropping H2SO4 into NaCl and then
flowing by passage through a CaCl2 drying tube, to give a lot of gray precipitate (phthalic imide
on one side). After filtered, the precipitate was then dissolved in 1500 mL water and stirred for
2 h and titrated to neutral by NaHCO3 solution, then filtered and washed thoroughly with water,
giving 18.3 g (54.7 wt% yield) mono-capped amine intermediate PA-ODA-NH2 after drying for 2
h at 80 °C under rotary evaporation. 1H NMR (400MHz, DMSO-d6): 7.97-7.88 (m, 4H; Ar H), 7.37-
7.34 (d, 2H; Ar H), 6.99 -6.95 (d, 2H; Ar H), 6.85-6.82 (d, 2H; Ar H), 6.64-6.61 (d, 2H; Ar H), 5.05
(s, 2H; NH2).
The model compound CC1 was then prepared by THPA and PA-ODA-NH2 in acetic acid solvent.
To the 250 mL round-bottom flask was added PA-ODA-NH2 (19.82 g, 0.06 mol) and 196 mL
acetic acid, stirred, then added THPA (97 wt%, 9.42 g, 0.06 mol) and stirred for 2 h at room
temperature in argon. The suspension was then refluxed at about 125 °C for 2 h, cooled and
crystallized, then filtered and thoroughly rinsed with water and dried. The crude product was
purified by rinsing with CH2Cl2 in Buchner funnel and suction flask, and the filtrate was then
dried by evaporation under a vacuum at 60 °C for 5 h to remove CH2Cl2 to give 14.6 g light white
product (52.3 wt%). The product (14.0 g) was further recrystallized from C2H5OH and CH2Cl2 and
dried under rotary evaporation at 60 °C for 5 h to give 9.9 g, m. p. peak at 188 °C (sharp) (DSC)
at a heat rate of 10 °C /min. FTIR (KBr), 2947, 2903 and 2850 (alicyclic C-H), 1785 and 1709
(imide C=O), 1243 (Ar-O-Ar), 697 (alicyclic –CH=CH–).1H NMR (400MHz, DMSO-d6): 7.98-7.88
(m, 4H; Ar H), 7.49-7.46 (d, 2H; Ar H), 7.25 -7.15(m, 6H; Ar H), 5.94 (s, 2H; alicyclic –CH=CH–),
3.31-3.28 (m, 2H; alicyclic –CH–), 2.48-2.24 (dd, 4H; alicyclic –CH2–).1H NMR (400MHz, CDCl3):
7.97-7.78 (m, 4H; Ar H), 7.43-7.40 (d, 2H; Ar H), 7.26 -7.23 (m, 2H; Ar H), 7.16-7.11 (m, 4H; Ar
H), 5.98 (m, 2H; alicyclic –CH=CH–), 3.27-3.25 (m, 2H; alicyclic –CH–), 2.75-2.28 (dd, 4H; alicyclic
–CH2–). Anal. calcd for C28H20N2O5: C 72.41, H 4.34, N 6.03; found: C 72.45, H 4.26, N 6.05.
2.5. Synthesis of phthalonitrile model compound (CN1)

ACCEPTED MANUSCRIPT
To a 250 mL three-neck round-bottom flask was added 4-nitrophthalonitrile (6.92 g, 0.04 mol),
ground K2CO3 (5.89 g, 0.043 mol), 4-phenylphenol (6.8 g, 0.04 mol) and 100 mL DMSO. The
resulting mixture was maintained at 30 °C with stirring for 24 h in nitrogen. After cooling, the
solution was poured into 2000 mL water, stirred and filtered. The precipitated product was then
dried at 80 °C for 4 h under rotary evaporation. The crude product (11.5 g) was recrystallized
from C2H5OH and CH3CN and dried under rotary evaporation at 80 °C for 5 h. The light green
crystals were obtained in 80.5 wt% yield (9.5 g), m. p. peak at 174 °C (sharp) (DSC) at a heat rate
of 10 °C/min. FTIR (KBr), 2231(C≡N), 1246 (Ar-O-Ar). 1H NMR (400 MHz, DMSO-d6): 8.13-8.10 (d,
1H; Ar H), 7.86 (s, 1H; Ar H), 7.80-7.78 (d, 2H; Ar H), 7.70-7.68 (d, 2H; Ar H), 7.50-7.43 (q, 3H; Ar
H),7.39-7.36 (m, 1H; Ar H),7.30-7.28 (d, 2H; Ar H). Anal. calcd for C20H12N2O: C 81.07, H 4.08, N
9.45; found: C 81.33, H 4.07, N 9.54.
2.6. Preparation of Blends, model blends reaction products and fabrication of cured polymers
Blends (CC/CN, MCC/CN, HCC/CN or CC1/CN1) by different molar ratios (3:7, 5:5) were
prepared by dissolving or dispersing the compounds in acetone. After removing the residual
solvent by vacuum evaporation, the mixture was ground into fine powder to insure
homogeneity. Blends of CC1/CN1 with 20 wt% DPPH were ground by mortar and pestle, and
then subjected to isothermal treatment at 300 °C for 15, 30 and 60 min.
Mechanism data were collected on samples stirred for different time at 300 °C in a 100 mL
three-neck round-bottom flask with N2 flow rate of 50 mL/min. Samples were ultrasonicated in
acetone followed by sedimentation and filtration to remove as much as possible the soluble.
The washing process was monitored by Thin Layer Chromatography (TLC) and dried by vacuum
evaporation at 80 °C for 5 h prior to characterization. The panels are for the convenience
marked with the symbol, RT for room temperature, m for mechanism, respectively. The model
compounds (CC1, CN1) and/or blend in equimolar ratio of CC1/CN1 (m55) are investigated by
1HNMR, elemental analysis, FTIR, UV-Vis, MALDI-TOF MS and EPR characterization.
Properties data were collected on samples with two steps, prepolymers and cured polymers.
The specific methods are as follows. Blend prepolymers were fabricated by the following
procedure: on a 12~15-g scale in a 250 mL reaction kettle equipped with a magnetic stirrer,
melting the blends at 200 °C for 4~10 min, stirring the mixture with degassing under 0.06 MPa
for 20 min, quickly taking out the preheated mould from the preheated oven chamber and
pouring the melt into the mould, quickly placing back into the chamber, and heated at 200 °C
for 6 h, followed by 230 °C for 3 h with N2 flow rate of 500 mL/min. After cooling to room
temperature (RT), the prepolymers were removed out from the mould, and then placed in the
muffle furnace to cure with argon flow rate of 500 mL/min. A thermal cure cycle was employed
for cured polymers panel fabrication. The cycle included dwells at 270 °C, 310 °C, 330 °C, 350 °C

ACCEPTED MANUSCRIPT
and 375 °C. The panels are for the convenience marked with the symbol, pc for pre-cure, ppc
for pre- and post-cure, respectively. The cured polymers (55-pc, 55-ppc, 37-pc, 37-ppc for
CC/CN system; M37-ppc for MCC/CN system), prepared from CC/CN and MCC/CN by different
molar ratios (3:7, 5:5), were characterized by FTIR, TGA, DMA, SEM, solubility, boiling water
absorbance and dielectric property.
Procedure (a) for pre-cure: 200 °C for 6 h, 230 °C for 3 h, 270 °C for 6 h and 310 °C for 6 h;
Procedure (b) for pre- and post-cure: pre-cure; and post-cure (330 °C for 3 h, 350 °C for 3 h and
375 °C for 6 h).
2.7. Solution NMR and Elemental Analysis
High-resolution solution nuclear magnetic resonance (NMR) spectra were acquired at room
1
temperature. H NMR (400 MHz) spectra was measured on a Bruker Avance-400 NMR
spectrometer, with CDCl3 or DMSO-d6 as the solvents with tetramethylsilane (TMS) as the
internal standard. Elemental analysis was performed on a vario MICRO select elemental
analyzer (Elementar, Germany).
2.8. Fourier Transform Infrared Spectroscopy (FTIR) and UV-Vis
FTIR spectra of the compounds and cured samples were recorded with a Nicolet FTIR-460
Fourier transform infrared spectrometer by KBr pellet containing KBr (about 150 mg, specpure
grade for FTIR absorption spectroscopy from Tianjin Kemiou Chemical Reagent Co., Ltd.) and
the sample powder (about 1.00 mg) from 4000 to 400 cm−1 at a resolution of 4 cm-1 by 32
scans. UV-Vis spectra were recorded on a Shimadzu UV Pectrophotometer UV2100 at room
temperature with CH2Cl2 as the solvents.
2.9. TGA, DSC and Dynamic Mechanical Analysis (DMA)
TGA was carried out with a TA instrument Q500 thermogravimetric analyzer using about 7
mg of sample in flowing air or nitrogen atmosphere of 60 mL/min at a heating rate of 10
°C/min. DSC was performed on a TA instrument Q200 differential scanning calorimeter using
about 3 mg of the sample at a heating rate of 10 °C/min and a flow nitrogen rate of 50 mL/min.
Dynamic Mechanical Analysis (DMA) were performed using a TA-Q800 under three-point
bending mode. For this, the cured samples were sanded to rectangular specimens of
dimensions (30 mm × 12.5 mm × 3.2 mm). The DMA characterizations were performed for the
temperature range 50~500 °C at a heating rate of 5 °C/min, a frequency of 1 Hz and a low strain
value of 2.5 × 10-4.

ACCEPTED MANUSCRIPT
2.10. Rheological Measurement
The rheological behavior of the imide compounds, phthalonitrile compounds and their
blends was studied by dynamic oscillation employing a TA Instruments AR-2000ex rheometer in
conjunction with an environmental testing chamber for temperature control. Different dynamic
curing experiments between 30 °C and the specified temperature, with a heating rate of 5
°C/min and at the frequency of 1 Hz were performed as well. For the time sweep test, the
heating rate for oscillatory temperature ramp was 50 °C/min, and the measurements were
carried out during isothermal cure at temperatures of 230 °C, 250°C and 270 °C, with an angular
frequency of 1 Hz and an initial strain of 2.5 × 10-4. The variation of the viscoelastic properties
during reaction, such as complex viscosity (|η*|) of imide compounds, phthalonitrile
compounds and their blends was determined in air atmosphere after melting powder samples
between parallel plates in the test chamber of the rheometer as a function of the reaction time.
All rheological measurements were taken at low strain values (2.5 × 10-4) using 25 mm-diameter
parallel plates. Normal force control (1.0 × 10-4 N) was utilized throughout the tests to keep the
samples taut.
2.11. Scanning Electron Microscope (SEM)
The cured samples were hammered, and coated with Au using an auto sputter coater
(SC7620). The morphology of the aggregate was confirmed by means of a field emission
scanning electron microscopy (FE-SEM) microscope (Quanta 250, FEI Corporation) operating at
20 kV.
2.12. Boiling Water Absorption
The cured polymers for boiling water absorption test were immersed in boiling distilled
water at around 98~100 °C for 60 h. The sample was removed from the boiling water, dried
with filter paper, and weighed periodically to determine the amount of boiling water
absorption. The water absorbance was calculated using the following equation:
Boiling Water Absorbance (%) = (m-m0)/m0 × 100
Where, m0 and m are the weight of the cured polymer at dry state and after immersion in
boiling distilled water, respectively.
2.13. Dielectric Measurement
Dielectric characteristics were evaluated with a Broadband Dielectric-Impedance
Spectrometer (Concept 50, Novo-control Technologies, Germany) in a broad range of frequency

ACCEPTED MANUSCRIPT
(0.01 ~ 4 × 107 Hz). Dried samples were fabricated and sanded to wafer specimens of
dimensions (Φ20 mm × 2 mm). Sliver paste was adopted as the electrodes before
measurements. The measured complex impedance data were used to calculate the real (ε’),
and imaginary (ε’’) parts of the complex dielectric function ε (ε = ε’+ i ε’’).
2.14. Matrix-Assisted Laser Desorption Ionization Time-Of-Flight Mass Spectrometry (MALDI-
TOF MS)
MALDI-TOF MS was performed on a Shimadzu Biotech Axima Performance equipped with an
N2 laser (337 nm), with alpha-cyano-4-hydroxycinnamic acid (CCA) as matrix. Samples were
partially dissolved in CH3CN, and then mixed with the matrix. The mixture was spotted onto the
centre of the probe tip and dried under ambient conditions. Spectra were obtained in the linear
positive mode.
2.15. Electron Paramagnetic Resonance (EPR)
EPR spectra were recorded on a Bruker A300-10/12 EPR spectrometer (Bruker BioSpin
GmbH) equipped with an X-band microwave bridge. Temperature control was achieved using a
Bruker A4131VT variable-temperature controller, calibrated using an independently calibrated
thermocouple. Model compounds or their blends were respectively placed in a quartz tube (o.
d. 5 mm × 180 mm), vacuumed (>10-3 Pa), purged with Argon (120 Torr), and then sealed. The
sample tube was put into the cavity of the EPR spectrometer at room temperature. The EPR
cavity was then heated to the required high temperatures from around melting point to 573 K,
with temperatures controlled by purging with high purity dry nitrogen and circulating water
cooling series disc system. A number of spectra of the derived signals were then recorded at
required temperatures after melting via repeated rapid scans typically with a scan time of ca.
41.943 s. The amplitude and frequency of the modulation were respectively 1.00 G and 100.00
kHz, operating at a microwave power of about 20.05 mW and a microwave frequency of about
9.44 GHz. The instrumental conditions were set at a centre field of 3369.450 G and a sweep
width of 100.000 G.
3. Results and discussion
3.1. Rheological Processing Evaluation

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis and preparation of CC, MCC, HCC, CN and their polymers
Fig. 1. Complex viscosity (|η*|) as a function of temperature for compounds CC, MCC, HCC, CN
and their blends

ACCEPTED MANUSCRIPT
Fig. 2.Complex viscosity (|η*|) as a function of time for compounds CC, CN and their blends
with molar ratio 5:5 or 3:7 at 230°C, 250 °C and 270 °C
As shown in Scheme 1 and Fig. 1, rheological characterizations on phthalonitrile (CN) and a
series of alicyclic imide moiety-containing compounds (tetrahydrophthalic anhydride, methyl
tetrahydrophthalic anhydride and hexahydrophthalic anhydride end-capped imide compound,
namely CC, MCC and HCC) were conducted to study the scope of the TSP effect. The results
indicated even the carbon saturated HCC also display TSP effect with CN, broadening the scope
of the TSP effect. Obviously, the activity of the TSP effect depends on the alicyclic imide
structure, with the carbon unsaturated CC/CN system having higher reactivity (Fig. S1). Among
them, the MCC/CN system showed a 130 °C processing window, which is larger than most of
the reported phthalonitrile system.
To systematically study the cure character and properties of the TSP effect and derived
polymers, CC/CN system was chosen to further investigation. Viscosity measurements on
phthalonitrile CN blends were conducted at several temperatures to determine the optimum
CC molar concentration and temperature for composite processing. The thermal polymerization
behaviors of the CC/CN blends (molar ratio 3:7, 5:5) and their respective compounds were
studied by rheological measurements. The complex viscosities change as a function of reaction
time for the blends and their respective compounds at 230 °C, 250 °C and 270 °C are shown in
Fig. 2. As shown in the Fig. 2, all the samples exhibit an initial dynamic viscosity around 0.1 Pa.s
after melting. As we all know, the polymerization of the neat phthalonitrile resins is extremely
sluggish (Fig. 2, CN-270 °C)[8-9], we put the plot of the neat alicyclic imide compound CC in the
Fig. 2, and the data shows that there is no obvious complex viscosity increasing tendency with

ACCEPTED MANUSCRIPT
still below 10 Pa.s after 3 h at 270 °C (Fig. S2, TGA and DSC of CC1, FTIR and 1H NMR of CC1 with
TGA 300 °C & 1 h in N2). Unexpectedly, as shown in the Fig. 2, when these two inertly
compounds were mixed together (CC-CN-37, CC-CN-55), they showed exciting polymerization
behaviors at elevated temperatures. And we named this unique curing behavior as Thermal
Synergistic Polymerization (TSP). The comparison between CC-CN-37 and CC-CN-55 of Fig. 2
reveals that the polymerization rate of CC/CN blends could be related to the
homopolymerization rates of CC and CN and their copolymerization rate. Further, it can be
easily controlled by varying the concentration of the alicyclic imide compound, which is, at a
higher CC molar ratio concentration, the dynamic viscosity of the blend increases at a faster
rate. Additionally, with increasing temperature above 250 °C, the complex viscosity increase
sharply, indicating that the curing temperature of the phthalonitrile blend, which is usually a
high temperature (above 300 °C) process with a long time, is greatly decreased. Also, there is a
large and adjustable process window at relatively low temperatures such as 230 °C with low
alicyclic imide compound concentration, helpful in the fabrication of large size composite
sections [4, 8-9]. This unique TSP effect inspires us to find out its intrinsic mechanism. Their
three possibilities arising from these reactions are (1) homopolymerization of alicyclic imide and
phthalonitrile groups independently, (2) copolymerization of alicyclic imide with the
phthalonitrile groups, (3) alicyclic imide accelerated the polymerization of phthalonitrile groups
like traditional curing system of active hydrogen. The mechanisms of thermosetting resin are
still hard to be figured out, and we will try to make some explanations for this unique
polymerization behavior by the studies of model compounds in the final section of this article.
3.2. Cured Polymers Physical Properties Evaluation
To evaluate physical properties of the cured phthalonitrile blend polymers, a series of the
cured samples were fabricated in the mould (cavity dimensions of 32 mm × 14 mm × 8 mm or
Φ20 mm × 2 mm), hammered to pieces or sanded to rectangular specimens of dimensions (30
mm × 12.5 mm × 3.2 mm) for characterizations of morphology of the cross section, solubility,
DMA, boiling water absorbance, dielectric, thermal and thermal-oxidative properties.

ACCEPTED MANUSCRIPT
Fig. 3. FTIR-DMA-TGA plots of 5:5 and 3:7 molar ratio CC/CN cured polymers at different
duration time and temperatures
3.3. FTIR, DMA, Thermal and Thermal-Oxidative Properties

ACCEPTED MANUSCRIPT
As shown in Fig. 3(a), FTIR data shows that there is no obvious -C≡N absorption (2230 cm-1)
of the cured polymers (55-pc, 55-ppc, 37-pc, and 37-ppc). New absorption peaks at 3285 cm-1
attributed to an N-H stretching and 1011 cm-1 attributed to metal-free phthalocyanine ring
vibration indicated the formation of phthalocyanine polymers [10, 27-29], and the low intensity
peaks assigned to phthalocyanine polymers suggest that they are a minor component of the
thermoset material. Additionally, another new absorption peak at 1092 cm-1 of 55-pc, 55-ppc
and 37-pc, 37-ppc was observed, which is frequently assigned to isoindoline in the reported
literatures [30]. The results are consistent with the UV data from model compounds study.
Significantly, no triazine absorptions are observed. This result is consistent with our previous
report [21]. It has been reported that due to the steric hindrance of the un-reacted -C≡N groups
at the ortho-position of triazine ring, there would be some -C≡N groups left (about 1/3) during
the formation of triazine with strong absorption in FTIR spectra at around 1520 and 1360 cm-1
of conventional phthalonitrile resins [10, 27, 34]. These spectroscopic features imply that the
curing system is very unique and characteristic, which is useful for detection of the curing
mechanism as described later to explain the phenomenon of no occurrence of triazine and -C≡N
absorption.
One of the most fundamental measurements made on high performance polymers (high-
temperature polymers) is the measurement of glass transition, Tg. DMA were performed on the
cured polymers to estimate the Tg. A comparison of storage modulus (G’) vs. temperature and
tan δ vs. temperature plots for the cured polymers at different cure extents with different
molar ratios made from CC/CN blends are shown in Fig. 3(b) and Fig. 3(c), respectively. The
DMA data are shown in Table 1.
Fig. 3(b) shows plots of the storage modulus up to 450 °C or 500 °C for the cured polymers.
The plots present impressive initial storage modulus at about 2515, 3261, 4304, and 3610 MPa
for 55-pc, 55-ppc, 37-pc, and 37-ppc, respectively. When heated from 50 °C to 500 °C, the
storage modulus for 37-ppc changed from 3610 to 1021 MPa. The measurements revealed the
samples that were cured to procedure (a) (310 °C for 6 h) and additional high temperature
treatment, procedure (b) (375 °C for 6 h), exhibited a significant Tg increase from the maximum
of tan δ. Moreover, all of data showed far higher storage modulus than that of resorcinol-based
phthalonitrile polymers (about 1000 MPa before and after DMA characterization) [6].
The multiple transitions as estimated on the basis of tan δ peak for procedure (a) are at
approximately 175 °C, 208 °C, 260 °C for the 5:5 molar ratio samples, this is probably due to the
presence of several different types of oligomeric structures with different molecular weight
with the increased concentration of CC relative to 37-pc. For procedure (b), no obvious Tg are
observed in the range of 500 °C for both molar ratio samples. With longer exposure at 375 °C,
the samples have progressed to afford a resin with a high crosslink density (Table S1) and

ACCEPTED MANUSCRIPT
mechanical properties reflecting its glassy state over the entire temperature range of the
measurements, which shows that the cured polymers have excellent thermal properties,
compared with other phthalonitrile polymers [2, 4, 6, 24-26].
Table 1 The DMA and TGA data of cured polymers
Samples
55-pc
55-ppc
3261 MPa
1059 MPa
32.5%
37-pc
37-ppc
3610 MPa
1021 MPa
28.3%
G’ at 50 °C
G’ at 500 °C
Retention of G’
T5%(N2)
2515 MPa
314 MPa
12.5%
4304 MPa
*488 MPa
*11.3%
371 °C
405 °C
41.9%
363 °C
397 °C
50.2%
486 °C
524 °C
70.5%
508 °C
536 °C
72.2%
T10%(N2)
char yield at 800 °C(N2)
T5%(Air)
379 °C
421 °C
496 °C
529 °C
0.5%
366 °C
413 °C
0.4%
507 °C
538 °C
0.8%
T10%(Air)
char yield at 800 °C(Air) 0.0%
*: data at 400 °C.
The thermal and thermal-oxidative properties of the cured CC/CN blends at different cured
extents with different molar ratios were evaluated up to 800 °C in TGA chamber at a heating
rate of 10 °C/min under nitrogen and air atmosphere. The data (Table 1) showed comparable
thermal and thermal-oxidative stability compared with resorcinol-based phthalonitrile
polymers cured by BAPS or m-APB in N2 or in air atmosphere [6].
Fig. 3(d) shows the thermal stability of the cured polymer samples in N2 atmosphere. Both of
the cured polymers 37-ppc and 55-ppc show good thermal stabilities up to 450 °C and begin to
lose weight at higher temperatures in N2 atmosphere. The significant enhancement trend of
thermal stability from procedure (a) to procedure (b), indicating the stabilities of the cured
polymer samples were depended on the cure cycle and cure temperatures [1-8, 13-17, 24-26].
Fig. 3(e) shows the thermal-oxidative stability of the cured polymer samples in air
atmosphere. The char yields of all samples were found to be 0.0~0.8 % after heated to 800 °C.
On the other hand, both of the cured polymers 55-ppc and 37-ppc did not exhibit obvious
weight loss before the scanning temperature reached 450 °C in air atmosphere, similarly in N2
atmosphere, while undergoing catastrophic weight loss between 600 °C and 750 °C in air
atmosphere, implying superior performance in thermal-oxidative stability [9, 24-26].
3.4. Morphological studies

ACCEPTED MANUSCRIPT
Fig. 4. SEM images of the cured polymers 55-ppc (a, b) and 37-ppc (c, d)
The fractured cross section morphologies of 55-ppc and 37-ppc were studied by FE-SEM. The
fractured cross section morphologies of the cured polymer samples 55-ppc and 37-ppc are
shown in Fig. 4(a, b) and Fig. 4(c, d), respectively. As seen from the comparison of Fig. 4 in the
same magnifications by 500 and 20,000, voids appear in the cured sample 55-ppc. The most
likely reason is oligomers or unreacted reactants such as CC volatilizing during high temperature
treatment in fabrication process. Moreover, no voids or air gaps of the sample 37-ppc were
detected, indicating homogeneity, densification characteristic of this kind of thermosets. These
results could be attributed to the good characteristics of the alicyclic imide and phthalonitrile
blend resins, which can be polymerized into voids-free thermosets.
3.5. Boiling Water Absorption Capability

ACCEPTED MANUSCRIPT
Fig. 5. Boiling water absorption and dielectric property of the cured polymers 55-ppc, 37-ppc
and M37-ppc
Low water absorption capability [31], a key aspect of polymer networks, was determined by
absorption experiments in boiling distilled water over a course of 60 h on several bulk samples
of polymers 55-ppc, 37-ppc and methyl tetrahydrophthalic imide/phthalonitrile polymers, M37-
ppc. Fig. 5(a) shows the water absorption properties of the cured polymers in boiling distilled
water. As shown in Fig. 5(a), the cured polymers 55-ppc, 37-ppc, and M37-ppc showed a
maximum boiling water amount of approximately 2.4%, 1.9% and 2.7%, respectively, after 60 h
immersion, which is lower than that of other phthalonitrile-based polymers (about 3.0%) [32].
The sample 55-ppc was more susceptible than 37-ppc to water penetration under the same
conditions due to voids or air gaps observed in SEM which is probably caused by the
volatilization of unreacted CC. For M37-ppc, the boiling water absorbance increases gradually
with time increase with a maximum boiling water absorbance somewhat higher than 37-ppc.
Thus it appears that the degree of water absorptivity can be controlled by the proper selection
of alicyclic imide moiety.
3.6. Dielectric Property
The dielectric constant (ε’) or relative static permittivity of a material represents the
capacitance of the material relative to a vacuum and is important in many industrial
applications [33]. The variation of the real part of the dielectric function, ε’, as a function of
frequency for samples at room temperature is illustrated in Fig. 5(b). The general trend for the
order of dielectric permittivity observed in the whole frequency range for 55-ppc and 37-ppc is
ε’ (37-ppc)>ε’ (55-ppc). The results suggest that the dielectric properties of the sample 55-ppc
possess better frequency stability, within the range of 4.8~4.5 in frequency from 1 Hz to 1 MHz.
From the graph it is clear that the real part of the dielectric function shows a strong frequency

ACCEPTED MANUSCRIPT
dependence below a certain frequency (1 Hz), while above this frequency the value of ε’
becomes almost constant for the samples. The decrease of ε’ with increase in frequency may be
attributed to electrical relaxation processes.
3.7. Discussion on Polymerization Mechanism
Scheme 2. Synthesis and preparation of model compound CC1, CN1 and their reaction product
This work reported that -C≡N groups can be fully consumed by the addition of alicyclic imide
compound, and this phenomenon has never been reported in the traditional curing system of
active hydrogen which promoted phthalonitrile systems. Thus, to explore the curing mechanism
of TSP effect of the alicyclic imide/phthalonitrile (CC/CN, MCC/CN, and HCC/CN) systems
becomes a significant work. And due to the insoluble and infusible nature and the diversity
structures of products of thermosetting resins, the research on the curing mechanism of
thermosetting resins becomes a challenging work.
In order to explore the curing mechanism of TSP effect, two kinds of representative model
compounds were designed and synthesized (Scheme 2), namely, the mono alicyclic imide end-
capped model compounds (CC1) and mono phthalonitrile terminated model compound (CN1).
Two model compounds were mixed in the molar ratio of 5:5 and isothermal treatment at
300 °C in N2 to explore the structures and curing mechanism by NMR, FTIR, isothermal TGA-
FTIR, UV-Vis, radical scavenger (DPPH), EPR and MALDI-TOF MS. Here, we should have
1
13
discussed the NMR characterization. While as data of H NMR and C NMR were relatively
arguable and need to be further studied, more NMR data such as Deuterated or 13-Carbon
labelled characteristic reactive sites of CC1 will be reported in another paper.

ACCEPTED MANUSCRIPT
The reasons for choosing the 5:5 molar ratio and the temperature 300 °C are as follows: (1) -
C≡N group reacts completely as much as possible, (2) a faster curing reaction rate at the same
temperature, (3) a less vaporization content and a shorter time at the same molar
concentration.
3.8 FTIR and UV-Vis Characterization
Fig. 6. FTIR spectra of the model compound blends and their 300 °C & 1 h reaction products
From FTIR spectra comparison of CC1/CN1 blend m55, m55-a and m55-b (TLC in Fig. S3) in
Fig.6, it is clear that the absorption of -C≡N (2231 cm-1) completely disappeared, without
observing any absorption of triazine ring (1520 and 1360 cm-1), which is consistent with the
result of the cured polymers CC/CN system. Therefore, these results indicate that the
polymerization of alicyclic imide/phthalonitrile system follows some unique mechanism
compared with typical phthalonitrile polymerization style. The characteristic absorptions of
phthalocyanine and isoindoline structures are overlapped with model compounds CC1 and CN1,
which will be confirmed later by other techniques such as UV-Vis and MALDI-TOF MS.
The UV-Vis absorption of the cured products in CH2Cl2 (partially dissolved) are shown in Fig. 7.
Fig. 7 shows two obvious absorption bands at 666 and 700 nm, which are attributed to the
characteristic of phthalocyanine ring [27] though the intensity of characteristic absorption
peaks of phthalocyanine structure was low in the FTIR spectra. Combination of these results,
we can infer that phthalocyanine formation was a minor process in the polymerization.
Additionally, absorption bands from 300 to 550 nm and from 570 to 660 nm indicated
occurrence of reaction products with different aromatic conjugated structures formation such
as isoindoline structures.

ACCEPTED MANUSCRIPT
Fig. 7. UV-Vis spectra of the model compounds and their 300 °C & 1 h reaction products
3.10. MALDI-TOF MS Characterization
Fig. 8. MALDI-TOF MS spectra of m55-a and m55-b
As discussed above, FTIR and UV-Vis spectra analysis imply formation of reaction products
with phthalocyanine and isoindoline structures. However, the polymerization mechanism and
the structures of the products are still unclear. MALDI-TOF MS technology has been widely used
to give much insight into the structure of polymer [35-40]. Here, we tried our best effort to

ACCEPTED MANUSCRIPT
thoroughly study on the structures of the derived CC1/CN1 model system polymers by MALDI-
TOF MS. The MALDI-TOF MS spectra of m55-a and m55-b (before and after ultrasonicated in
acetone) are presented in Fig. 8.
Table 2 The MALDI-TOF MS peaks and the corresponding structures of m55-a and m55-b
m55-a
m55-b
structure
m/z calculated
polymerization style
m/z measured
1186.77
1227.49
1541.88
1855.87
2642.21
2810.11
3428.09
m/z measured
1186.24
1227.08
1541.17
1855.57
2641.18
2810.36
3427.48
(0:4)H+
(2:1)H+
(2:2)Na+
(4:0)H+
(5:1)Na+
(6:0)Na+
(1:10)H+
1185.37
1225.38
1543.45
1857.57
2639.78
2807.83
3426.05
CN1, homopolymerization
copolymerization
copolymerization
CC1, homopolymerization
copolymerization
CC1, homopolymerization
copolymerization
x= the number of CC1, y= the number of CN1, Z= the species of cation ions.
As shown in Fig. 8, m55-a shows a characteristic molecular weight distribution of polymers.
The representative peaks of m/z at 1855.87 and 1186.77 corresponding to tetramer of CC1
(calcd: 1857.57, (4:0)H+), and tetramer of CN1 (calcd: 1185.37, (0:4)H+) confirming the
formation of phthalocyanine (Fig. 7), were observed in the MALDI-TOF MS spectra, indicating
the homopolymerization of CC1 and CN1 independently or the CC1 and CN1 can be reacted into
homopolymers in the presence of each other. In addition to the main peaks of m/z at 1855.87
and 1186.77, there are a series of typical polymer peaks at the range of m/z at 1000~4500, and
these peaks can provide us more details of the TSP effect between tetrahydrophthalic imide
and -C≡N groups (Table S2). As the intensity of phthalocyanine is too low to be observed, m55-a
was ultrasonically washed to get acetone insolubilities m55-b, which was also characterized by
MALDI-TOF MS with data presented in the Table S2. Comparison between the spectra of m55-a
and m55-b authenticated that phthalocyanine formation was a minor process in the
polymerization (Fig. 6, Fig. 7, and Fig. 8). As shown in enlarged spectra of Fig. 8, since there are
many species of polymers and the existence of fragments, cation ions and CC1 precursors, it is
difficult to find out the actual structures of these small peaks and some peaks data are shown in
Table 2. The molecular ion peaks are all labelled in the (x:y)Z format, where x = the number of
CC1 for the reactive product precursors, y = the number of CN1 for the reactive product
precursors, Z = the species of cation ions. And it is important to stress that the polymer

ACCEPTED MANUSCRIPT
fragment ions depends primarily on the extent of precursors ion “breakage” caused by collision
impact with gas and the cation ions such as Na+, which widely exist in the nature world [41-45].
Fig. 9. MALDI-TOF MS spectra of m55 at 300 °C for15, 30 and 60 min

ACCEPTED MANUSCRIPT
Fig. 10. MALDI-TOF MS spectra of m55 with 20 wt% DPPH at 300 °C for 15, 30 and 60 min
On the other hand, reaction products of CC1/CN1 and CC1/CN1 with 20 wt% DPPH at 300 °C
for different time (15, 30 and 60 min) were also characterized by MALDI-TOF MS. The MALDI-
TOF MS spectra of m55-15, 30, 60 and m55&20wt%DPPH-15, 30, 60 are presented in Fig. 9 and
Fig. 10. The polymerization process of CC1/CN1 system are shown in Fig. 9. At an initial stage
(before 30 min), molecular weight is not significantly increased. While in the further stage such
as isothermal for 60 min, large amount of polymeric derivatives (especially copolymerization)
started to emerge with molecular weight increase.
In the past two decades, many researchers have explored the radical reaction mechanism of
the aliphatic or aromatic nitriles [9, 46-49]. Among these studies, H. J. P. de Lijser, et al.
reported that the alicyclic radicals can initiate the reaction of benzonitrile by carbon radical
reactive site [48]. It is rational for us to believe that the curing reaction of alicyclic imide
moiety/phthalonitrile systems is a high temperature free radical process. And Fig. 10 shows the
polymerization process of CC1/CN1 system with free radical scavenger DPPH. The spectra data
exhibit characteristic inhibiting polymerization, implying that the TSP reaction involved a free
radical process.
3.11. EPR Characterization
To further clarify the nature of the reaction intermediates, EPR characterizations of CC1, CN1
and m55 were studied, and a series of EPR spectra at elevated temperatures were reported (Fig.
11, Fig. 12, and Fig. 13). As shown in Fig. 11, a promising result was observed that m55 shows
weak EPR signals at 463 K and the intensity of the signals gradually increased with elevated
temperatures. The signal was obvious at 543 K with centre magnetic field at around 3367.5 G
and its intensity continued to increase within the range of scanning temperature. This result
suggests that some free radical intermediates formed such as polyisoindolines or other
conjugated structures that can delocalize the radicals along the polymer chain, which making
the intermediates stable at high temperatures and further initiate polymerization.
Fig. 12 shows that the EPR signal intensity of m55 increase with scanning time equilibrating
at 543 K, further verifying the existence of the free radical intermediates and their stable
characteristics at high temperatures. As no hyperfine splitting structures were observed in our
following investigation [50-51], the type of the involving free radical intermediates still needs to
be further studied.
Meanwhile, free radical scavenger DPPH characterization on the TSP effect further confirmed
the reaction has some relationships with free radical mechanism (Fig. 10, Fig. S4-S5).

ACCEPTED MANUSCRIPT
Fig. 11. EPR spectra of m55 at elevated temperatures

ACCEPTED MANUSCRIPT
Fig. 12. EPR spectra of m55 at 543 K isothermal for a period of time
Fig. 13. EPR spectra of CC1 and CN1 at elevated temperatures

ACCEPTED MANUSCRIPT
Further, EPR data were also collected in-situ on the model compounds and shown in Fig. 13.
Weak EPR signals were observed in the model compound CN1, which suggests that any radical
mechanism would begin with the phthalonitrile monomer. The sluggish curing characteristic of
neat phthalonitrile resin could be attributed to the relatively lower content of free radicals than
that of m55. No EPR signals were observed in the model compound CC1. These results further
authenticated the mechanism of TSP effect between alicyclic imide moiety and phthalonitrile.
3.12. Mechanism Discussion
Based on the above discussion, a rational explanation for the mechanism of the TSP effect
(Scheme 3) can be proposed and some significant results or evidences can be summarized as
follows:
Scheme 3. Proposed radical polymerization mechanism of alicyclic imide/phthalonitrile system
1. The alicyclic double bond (–CH=CH–) of alicyclic imide based on tetrahydrophthalic
anhydride may only play a minor role in the TSP effect. Carbon saturated alicyclic imide moiety
based on hexahydrophthalic anhydride compound (HCC) also shows the TSP effect with the
phthalonitrile compound (CN) (Fig. 1 and Fig. S6).
2. MALDI-TOF MS indicates that both copolymerization and homopolymerization existed in
the TSP process. In our cases, both of phthalonitrile and alicyclic imide moiety (CC, MCC and
HCC) hardly showed obvious homopolymerization under investigated range of temperatures.
Significantly, for blends of them, both of copolymerization and homopolymerization can be
activated. This characteristic is different from usual thermal copolymerization and epoxy-cure
agent system.

ACCEPTED MANUSCRIPT
3. As is known, hetero-atom containing unsaturated groups, such as C≡N, usually polymerize
by ionic paths. In this paper, EPR technique and free radical scavenger DPPH were used to in-
situ or off-situ discuss this issue. The data firstly demonstrated the TSP effect occurred probably
involving a free radical path.
4. Conclusions
The alicyclic imide moiety based CC, MCC and carbon saturated HCC showed a unique
phenomenon namely Thermal Synergistic Polymerization (TSP) effect between alicyclic imide
moiety and phthalonitrile.
The fully study on the curing behaviors, the thermal and mechanical properties, boiling water
absorption and dielectric properties of CC/CN blends and their cured products showed that the
CC/CN synergistic polymerization system not only possessed outstanding processing
performance, but also showed excellent thermal and mechanical properties, low boiling water
absorbance and moderate dielectric constant (ε’ in the range of 4.8~4.5).
In order to explore the polymerization mechanism of this unique TSP effect, the model
compounds CC1, CN1 and their blends were designed and characterized by FTIR, UV-Vis,
MALDI-TOF MS, free radical scavenger (DPPH), EPR, etc. The high-efficiency of C≡N
transformation of the system undergoes both copolymerization and homopolymerization with
no triazine formation, involving a free radical process. The structures of the oligomers derived
from model compounds contain a minor phthalocyanine ring and major aromatic conjugated
rings likely isoindoline.
This study opens an exciting opportunity for designing new high performance thermosetting
polymers.
Acknowledgements
The authors sincerely thank to the National Natural Science Foundation of China (No. 51203098)
for the financial support.
Appendix A. Supplementary data
Synthesis and Characterization of CC, MCC, HCC, CC1, CN and CN1. Fig. S1−S6 and Table S1-S2.