Synthesis and properties of cationic photopolymerizable hyperbranched polyesters with terminal oxetane groups by the couple-monomer polymerization of carboxylic anhydride with hydroxyl oxetane

Article abstract of DOI:10.1016/j.polymer.2010.05.044

The hyperbranched polyesters with terminal oxetane groups were synthesized via couple-monomer methodology based on carboxylic anhydride and hydroxyl oxetane. Two competitive reactions, ring-opening reaction of carboxylic group with oxetane group and ester

Full text of DOI:10.1016/j.polymer.2010.05.044

Polymer 51 (2010) 3402e3409
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Synthesis and properties of cationic photopolymerizable hyperbranched
polyesters with terminal oxetane groups by the couple-monomer polymerization
of carboxylic anhydride with hydroxyl oxetane
,
Fu Zhan a, Anila Asif b, Jianhua Liu a, Hailong Wang a, Wenfang Shi a
*
a CAS Key Laboratory of Soft Matter Chemistry, Department of Polymer Science and Engineering, University of Science and Technology of China, Hefei, Anhui 230026, PR China
b Interdisciplinary Research Centre in Biomedical Materials, COMSATS Institute of Information Technology, Defence Road, Off Raiwind Road, Lahore 54000, Pakistan
a r t i c l e i n f o
a b s t r a c t
Article history:
The hyperbranched polyesters with terminal oxetane groups were synthesized via couple-monomer
methodology based on carboxylic anhydride and hydroxyl oxetane. Two competitive reactions, ring-
opening reaction of carboxylic group with oxetane group and esterification of carboxylic group with
hydroxyl group, occurred synchronously during the polymerization. The results showed that the
hyperbranched polyesters, poly(SA-EHO) and poly(SA-EHPO), were synthesized successfully when suc-
cinic anhydride (SA) was selected to react with 3-Ethyl-3-(hydroxymethyl)oxetane (EHO) and 3-ethyl-3-
((4-hydroxymethyl)phenoxymethyl)oxetane (EHPO), respectively. The degree of branching was deter-
mined to be about 0.7 for poly(SA-EHO) from the 1H NMR result, which was higher than those reported
in literature, indicating that the esterification was dominant. The glass transition temperatures (Tg)s were
measured as ꢀ11.5 ꢁC for poly(SA-EHO) and 14.3 ꢁC for poly(SA-EHPO) by DSC. The number average
molecular weights were 1914 g/mol and 2108 g/mol with the polydispersity indices of 2.39 and 2.28 for
poly(SA-EHO) and poly(SA-EHPO), respectively. Both poly(SA-EHO) and poly(SA-EHPO) were added to
bisphenol A epoxy resin, EP828, with different contents and cured under UV exposure. From the DMTA
and tensile results, the UV cured poly(SA-EHPO) showed higher Tg and tensile strength compared with
poly(SA-EHO) because of the existence of phenyl group in poly(SA-EHPO) chain. Moreover, both the Tgs
and mechanical strength decreased, whereas the elongation percents at break increased with poly(SA-
EHO) and poly(SA-EHPO) increased. This indicates that the flexibility of cured films was improved by the
addition of poly(SA-EHO) and poly(SA-EHPO) in the formulations.
Received 4 January 2010
Received in revised form
16 May 2010
Accepted 24 May 2010
Available online 1 June 2010
Keywords:
Hyperbranched polyester
Hydroxyl oxetane
Carboxylic anhydride
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
“AA’þCB2” approach [4]. Li and co-workers reported a series of
hyperbranched poly(ester-amide)s prepared from commercially
Recently, a new approach called “couple-monomer method-
ology” was developed for preparing hyperbranched polymers
based on the non-equal reactivity of functional groups in special
monomer pairs [1e4]. This methodology does not only resolve the
source of monomers but also avoid gelation usually occurring in the
“A2þB3” approach. There have been many hyperbranched polymers
synthesized through the couple-monomer methodology. Yan and
Gao reported the preparation of hyperbranched poly(sulfone
amine)s, poly(ester amine)s, poly(urea urethane)s, and poly(amide
amine)s, and so on [5]. van Benthem group synthesized the bis(2-
hydroxypropyl) amide-based hyperbranched polyesteramides via
available aliphatic carboxylic anhydrides or dicarboxylic acids and
multihydroxyl primary amines [6,7]. Wang and co-workers
synthesized hyperbranched polysiloxysilanes from A2 and CB2 type
monomers [8,9].
UV-induced photocuring, mainly classified as cationic and
radical photocuring according to the curing mechanism, has
obtained important applications in many fields. The cationic pho-
tocuring possesses some advantages in contrast to the radical one,
such as low shrinkage, absence of oxygen inhibition, better adhe-
sion [10,11]. Oxetane monomers were introduced in UV curing by
Crivello and Sasaki [12]. They found out that the oxetanes are
generally more reactive than their structurally similar epoxy
counterparts and the formulation with oxetane and epoxide
possesses higher reactivity toward photoinitiated cationic poly-
merization [13,14]. Similar results were also reported in some other
* Corresponding author. Tel.: þ86 551 3606084; fax: þ86 551 3606630.
E-mail address: wfshi@ustc.edu (W. Shi).
0032-3861/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2010.05.044

F. Zhan et al. / Polymer 51 (2010) 3402e3409
3403
literatures and oxetane group is characterized to be more reactive
because of not only the ring strain similar to that of epoxide but also
the higher nucleohilicity than epoxy group [15e17]. Many hyper-
branched polymers with cationic curable functional groups have
been prepared and investigated as additives to improve the prop-
erties of the cured materials [18e21]. However, most of these
hyperbranched additives were synthesized based on modification
of hyperbranched polymers. Only a few cationic curable hyper-
branched polymers were synthesized through polymerization of
the monomers with cationic curable functional groups.
In this paper, we present the synthesis of hyperbranched poly-
ester with terminal oxetane groups via couple-monomer method-
ology based on carboxylic anhydride and hydroxyl oxetane. The
ring-opening reaction and esterification occur synchronously
during the polymerization. The results showed that the linear or
little branched polymers were obtained when cyclohexene-1,2-
dicarboxylic anhydride and phthalic anhydride were selected.
However, as succinic anhydride was chosen to react with hydroxyl
oxetanes, the hyperbranched polyesters were synthesized
successfully. The obtained hyperbranched polyesters, as additives,
were added to cationic systems and cured under UV exposure. Both
dynamic mechanical thermal properties and mechanical behaviors
of the cured films were investigated.
modulus, R is the ideal gas constant, and T is the temperature in
Kelvin. The mechanical properties were measured with an Instron
Universal tester (model 1185, Japan) at 25 ꢁC with a crosshead
speed of 25 mm/min. The dumb-bell shaped specimens were
prepared according to ASTM D412-87. Five samples were analyzed
to determine an average value in order to obtain the reproducible
result.
2.3. Synthesis of 3-ethyl-3-((4-hydroxymethyl)phenoxymethyl)
oxetane (EHPO)
A mixture of 3-ethyl-3-(bromomethyl)oxetane (33 g, 0.184 mol),
4-hydroxybenzenemethanol (23 g, 0.184 mol), potassium carbonate
(127 g, 0.920 mol) and 220 mL of DMF was poured into a glass flask
equipped with a mechanical stirrer and reacted at 25 ꢁC for 1 h and
then at 70 ꢁC for 5 h. The reaction mixture was concentrated, filtered,
poured into water and then extracted for several times with
dichloromethane. The combined organic layer was washed again
with water, dried with anhydrous sodium sulfate, and filtrated.
Finally the solvents were removed under reduced pressure. 3-ethyl-
3-((4-hydroxymethyl)phenoxymethyl)oxetane(EHPO)wasobtained
as a pale yellow liquid with a yield of 79%. 1H NMR (300 MHz,
CDCl3):
d (ppm) 7.30 (ArH), 6.92 (ArH), 4.62 (HOeCH2eAre),
2. Experimental section
4.57e4.46 (OxetaneeH), 4.08 (eAreOeCH2eOxetane), 1.92e1.84
(CH3eCH2eOxetane), 0.93 (CH3eCH2eOxetane).
2.1. Materials
3-Ethyl-3-(hydroxymethyl)oxetane (EHO) was supplied by Per-
storp AB, Sweden. Cyclohexene-1,2-dicarboxylic anhydride (CHA)
was supplied by SigmaeAldrich. 3-Ethyl-3-(bromomethyl)oxetane
was synthesized according to the procedure reported in the liter-
ature [22]. Bisphenol A epoxy resin, EP828, used as a cationic
curable resin, was supplied by Shell Chemical Co. Irgacure 250, used
as a cationic photoinitiator, was supplied by CibaeGeigy. 4-
Hydroxybenzenemethanol was supplied by Tianma Chemical Co.,
Suzhou, China. Acetic anhydride, 4-dimethylamino-pyridine
(DMAP), succinic anhydride (SA), phthalic anhydride (PA), trime-
thylolpropane (TMP) and other chemicals were purchased from
Shanghai First Reagent Co. All the chemicals were used as received
without further purification except for EHO, which was dried with
2.4. General polymerization procedure
A calculated amount of EHO, an equimolar amount of dicar-
boxylic anhydride and 2 mol% of DMAP were charged into a dry
glass flask equipped with a mechanical stirrer, gas-outlet and -inlet
tubes, and stirred continuously at 100 ꢁC under N2 atmosphere
until the 1H NMR signal peak at 3.79 ppm for HOCH2-oxetane dis-
appeared. Then the reaction temperature was raised to 150 ꢁC. A
vacuum system (about 2000 Pa) was applied to exclude the water
formed during the esterification for driving the polymerization
toward high conversion. The reaction was performed until the
mechanical stirring became difficult. The reactant was then cooled
down, and dissolved in acetone, and precipitated into deionized
water twice. The precipitate was dissolved in CH2Cl2 again, dried
with anhydrous sodium sulfate, filtrated and condensed by distil-
lation, and finally dried at 50 ꢁC in vacuum for 48 h to give
a yellowish resin or solid.
ꢀ
4-A molecular sieves before use.
2.2. Measurements
The 1H NMR spectra were recorded with a Bruker 300-MHz or
400-MHz NMR spectrometer according to the research demand.
CDCl3 and tetramethylsilane were used as a solvent and internal
reference, respectively. The average molecular weight and its
polydispersity index were determined with a gel permeation
chromatography (GPC) equipped with a refractive-index detector
and calibrated with the standard linear PSt. DMF was used as an
eluent with a rate of 1.0 mL/min. The differential scanning calo-
rimetry (DSC) curves were recorded with a Shimazhu DSC-60
apparatus. All the samples were heated from ꢀ100 to 50 ꢁC with
a rate of 10 ꢁC/min under nitrogen for the first scan, then cooled to
ꢀ100 ꢁC at 40 ꢁC/min, and immediately heated at 10 ꢁC/min from
ꢀ100 ꢁC to 50 ꢁC again. The tensile storage modulus (E0) and tensile
Poly(PA-EHO) 1H NMR (300 MHz, CDCl3):
d (ppm) 7.82e7.31
(AreH), 4.58e4.45(Oxetane-H), 4.45e4.09(AreCOOCH2e), 3.62e3.20
(eCH2eOH), 1.81e1.29 (CH3eCH2e), 1.01e0.62 (CH3eCH2e).
Poly(CHA-EHO) 1H NMR (300 MHz, CDCl3):
d (ppm) 4.58e4.45
(Oxetane-H), 4.31e3.85 (eCH2eOOCe), 3.65e3.04 (eCH2eOH,),
2.98e2.45 (eOOCeCH(hexamethylene)), 2.20e1.89 (-CH2(hexa-
methylene)-), 1.89e1.14 (eCH2(hexamethylene)e, CH3eCH2e),
1.01e0.75 (CH3eCH2e).
Poly(SA-EHO) 1H NMR (400 MHz, CDCl3):
(OxetaneeH), 4.26e4.20
(eCOOCH2e), 3.59e3.42 (eCH2eOH), 2.74e2.52 (eOOCeCH2e
d
(ppm) 4.51e4.38
(eCOOCH2eOxetane-), 4.10e3.96
CH2eCOOe),
1.78e1.70
(CH3eCH2eOxetanee),
1.54e1.26
(CH3eCH2e), 0.95e0.74 (CH3eCH2e).
loss factor (tan
d) were measured using a dynamic mechanical
Poly(SA-EHPO) 1H NMR(400 MHz, CDCl3):
d
(ppm) 7.43e7.20 (AreH),
thermal analyzer (Diamond DMA, PE Co., USA) at a frequency of
2 Hz and a heating rate of 5 ꢁC/min in the range of 0e220 ꢁC on the
sheet of 25 ꢂ 5 ꢂ 1 mm3. The crosslinking density (Ve) as the molar
number of elastically effective network chain per cube centimeter
7.07e6.78 (AreH), 5.12e4.96 (eOOCeCH2eAre), 4.63e4.44
(OxetaneeH), 4.23e4.09 (eCOOCH2e), 4.09e4.04 (eAreOe
CH2eOxetanee),
3.88e3.77(eAreOeCH2eC(C2H5)(CH2)(CH2)e),
2.73e2.49 (eOOCeCH2eCH2eCOOe),
3.62e3.35 (eCH2eOH),
of the film was calculated from the storage modulus in the rubbery
2.49e2.12 (eCH2eOH), 1.91e1.79 (CH3eCH2eOxetane-), 1.63e1.40
(CH3eCH2e), 0.96e0.73 (CH3eCH2e).
E0
plateau region according to: ne
¼
3RT, where E0 is the elastic storage

3404
F. Zhan et al. / Polymer 51 (2010) 3402e3409
O
2.5. End-capping of poly(SA-EHO)
O
HOOC
4 g Poly(SA-EHO) was firstly dissolved in 40 mL of chloroform in
a glass flask. 2.5 g Acetic anhydride was added, and stirred for 12 h at
room temperature. Then chloroform and most of superfluous acetic
anhydride were removed by distillation. The crude product was dis-
solved in chloroform again and washed with saturated NaHCO3
aqueous solution thrice and subsequently with deionized water twice.
The organic layer was collected and dried with anhydrous sodium
sulfate, filtrated and condensed by distillation, and finally dried at 50 ꢁC
in vacuum for 48 h to give a yellowish resin. 1H NMR (400 MHz, CDCl3):
O
R
O
R
O
+
HO
O
(EHO)
O
O
HOOC
R
O
O
R
O
O
O
d
(ppm) 4.51e4.38 (OxetaneeH), 4.26e4.20 (eCOOCH2eOxetanee),
O
OH
4.10e3.96 (eCOOCH2e), 2.74e2.52 (eOOCeCH2eCH2eCOOe),
2.08e2.02 (eOOCCH3), 1.78e1.70 (CH3eCH2eOxetanee), 1.54e1.40
(CH3eCH2e), 0.95e0.74 (CH3eCH2e).
Esterification
and Ring-Open
Ring-Open Only
O
OH
O
2.6. Copolymerization with TMP
O
O
HO
O
2.5 g SA (0.025 mol), 2.9 g EHO (0.025 mol) and 0.061 g DMAP
(0.0005 mol) were charged into a dry glass flask equipped with
a mechanical stirrer and gas-outlet and -inlet tubes, and stirred
continuously at 100 ꢁC under N2 atmosphere until the 1H NMR
signal peak at 3.79 ppm for HOCH2eoxetane disappeared. The
product was collected as a monomer for copolymerization.
0.1865 g TMP (0.00139 mol) was charged into a dry glass flask
equipped with a mechanical stirrer and an isobarical dropping
funnel, and placed into a pre-heated (150 ꢁC) oil bath. Then
a vacuum was applied. The monomer obtained above was dropped
with a rate of about 2 g/h through the isobarical dropping funnel
and reacted for another 3.5 h. Then the cooled mixture was dis-
solved in acetone and precipitated into deionized water twice. The
precipitate was dissolved in CH2Cl2, dried with anhydrous sodium
sulfate, filtrated and condensed by distillation, and finally dried at
50 ꢁC in vacuum for 48 h to give a yellowish resin.
O
O
O
OH OH OH
OH
OH
O
OH
O
Linear Polymer
Hyperbranched Polymer
O
O
O
O
O
O
R
O
O
=
O
O
O
O
(CHA)
(PA)
(SA)
Scheme 1. Schematic illustration for ideal polymerization route of carboxylic anhy-
dride and hydroxyl oxetane.
Copoly(SA-EHO) 1H NMR (400 MHz, CDCl3):
(OxetaneeH), 4.26e4.20 (eCOOCH2eOxetanee),
(eCOOCH2e), 3.59e3.42 (eCH2eOH), 2.74e2.52 (eOOCeCH2e
CH2eCOOe), 1.78e1.70 (CH3eCH2eOxetanee), 1.54e1.26
d
(ppm) 4.51e4.38
4.10e3.96
produced, considered as a latent ABB’ type monomer. After raising
the reaction temperature and applying vacuum to the vessel, the
ring-opening reaction occurred further, leading to the formation of
a dimer with one carboxylic acid group, one hydroxyl group and
one oxetane group. Then both oxetane groups and hydroxyl groups
may react with carboxylic acid groups. Two kinds of reactions, ring-
opening reaction of carboxylic group with oxetane group and
esterification of carboxylic group with hydroxyl group, occur
during the polymerization. If only the former ring-opening reaction
happened, a linear polymer with pendant primary hydroxyl groups
can be prepared, which is similar to the result reported by Nishi-
kubo. However, if the esterification occurs synchronically,
a branched polymer with both oxetane and hydroxyl groups can be
obtained. When the rate of esterification is much higher than that
of ring-opening reaction, most of the hydroxyl groups formed
during ring-opening reaction can be reacted with carboxylic acid
groups rapidly, resulting in the formation of hyperbranched poly-
mer with high degree of branching (DB). This may overcome the
disadvantage of low DB of hyperbranched polymers prepared via
one-pot method of ABx type monomers.
Alicyclic, aromatic, and aliphatic dicarboxylic anhydrides were
selected to prepare a series of polymers, as shown in Scheme 1.
Fig. 1 shows the 1H NMR spectra of poly(PA-EHO) and poly(CHA-
EHO). As mentioned above, some residual oxetane rings still exist in
the system whereas most of the hydroxyl groups were consumed
when a hyperbranched polymer was obtained. However, the signal
peaks attributed to proton on the oxetane rings of both poly(PA-
EHO) and poly(CHA-EHO) at 4.58e4.45 ppm are very weak.
Moreover, the signals corresponding to the methylene proton next
to hydroxyl group are obvious. From the spectrum of poly(CHA-
EHO), the integrations of the proton attributed to oxetane ring,
(CH3eCH2e), 0.95e0.74 (CH3eCH2e).
3. Results and discussion
3.1. Polymerization of carboxylic anhydride and hydroxyl oxetane
Nishikubo and co-workers have done systematic and extensive
investigation on the reactions of oxetanes with protonic and aprotic
compounds, including carboxylic acid, phenol, thiol, acyl chloride,
thioester, phosphonyl dichloride, silyl chloride and chloroformate
[23e27]. Several molecular structure-controlled polymers have
been prepared through these reactions. The linear and hyper-
branched polyesters with pendant primary hydroxyl groups were
prepared by the reaction of bis(oxetane) with dicarboxylic acid and
1,3,5-benzenetricarboxylic acid via an A2 þ B3 approach, respec-
tively. The ester bond in the polymer chain was formed via ring-
opening polyaddition of carboxylic acid and oxetane group,
generating pendant primary hydroxyl group, simultaneously. The
formed hydroxyl group failed to react further with carboxylic group
under the condition.
However, in our opinion, the esterification might easily occur as
removing water from the vessel, or adding a catalyst. Thus, every
reacted hydroxyl group can be considered as a branched point,
resulting in the formation of highly branched polymers.
As shown in Scheme 1, as the equimolar amount of dicarboxylic
anhydride reacted with EHO at lower temperature, an intermediate
compound with one carboxylic group and one oxetane group was

F. Zhan et al. / Polymer 51 (2010) 3402e3409
3405
the reaction were consumed by the reaction with carboxylic acid
groups. A hyperbranched polymer was finally prepared by poly-
merization of SA and EHO.
3.2. DB, Mn of hyperbranched poly(SA-EHO)
DB is an important parameter for the structural characterization
of a hyperbranched polymer. Theoretically, the DB of a hyper-
branched polymer is 0.5 when one-pot polymerization of an AB2
monomer is employed [28]. However, this value is obtained based
on the assumption of equal reactivity of two B functional groups. In
fact, the reactivity of the residual B functional group becomes lower
after the other one has reacted because of steric effect or some
other reasons. Consequently, the hyperbranched polymers
prepared in experiments have lower DB than the theoretical value,
usually in the range from 0.3 to 0.5. Several methods have been
reported to improve DB, such as higher reactivity of linear vs.
terminal units, polymerization of prefabricated dendron mono-
mers, slow addition techniques, and polymerization in the presence
of core molecules [29,30]. However, most investigations on these
methods were based on theoretical calculation and only a few
hyperbranched polymers with high DB (more than 0.5) were
reported. A. Hult reported the synthesis of hyperbranched aliphatic
polyesters with the DB value of about 0.8 based on 2,2-bis(meth-
ylol)propionic acid and tris(methylol)propane as a core molecule
[31]. Mario Smet and Mitsuru Ueda synthesized hyperbranched
poly(arylene oxindole)s and hyperbranched polythiolketal both
with 100% DB, respectively [32e34].
Fig. 1. 1H NMR spectra of (A) poly(PA-EHO) and (B) poly(CHA-EHO).
methylene adjacent to ether bond, methylene adjacent to hydroxyl
group, and methyl are 2.64, 37.94, 19.42, and 30.00, respectively.
This indicates that more than 93% oxetane groups are consumed in
the polymerization but most of the hydroxyl groups did not react
with carboxylic acid groups. All the facts suggest that a linear or
little branched polymer was obtained when PA or CHA was chosen.
In contrast, when SA, an aliphatic dicarboxylic anhydride, was
selected, the distinct results were obtained. As shown in Fig. 2, the
signal peaks attributed to the oxetane proton at 4.58e4.45 ppm are
obvious and the signals at 3.59e3.42 ppm originated from the
methylene unit adjacent to hydroxyl group are much weaker than
those for poly(PA-EHO) and poly(CHA-EHO). The integrations of the
proton attributed to oxetane ring, methylene adjacent to ether
bond, methylene adjacent to hydroxyl group, and methyl were
obtained to be 12.90, 40.42, 7.13, and 30.00, respectively. This
indicates that about one third oxetane groups exist in the final
product and more than half of the hydroxyl groups formed during
Fortunately, DB of poly(SA-EHO) can be calculated through 1H
NMR spectrum directly when a Bruker 400-MHz NMR spectrom-
eter was used (Fig. 2). It is clear that three groups of the peaks at
around 1.76, 1.49 and 1.42 ppm are attributed to the methylene
proton adjacent the methyl group of three different units. As well-
known, the peak for the terminal unit has the highest chemical shift
because of the oxetane ring next to the methylene group. The peak
for the dendritic unit has a high chemical shift compared with that
of the linear one because of the higher electronegativity of ester
bond group than hydroxyl group. To confirm the above presump-
tion, end-capping of the hydroxyl groups with acetic anhydride was
conducted. The 1H NMR spectrum of end-capped product is also
given as the curve C in Fig. 2. The peak signals at 1.42 and
3.35e3.62 ppm disappeared and a new single peak signal at
2.06 ppm corresponding to the methyl proton attached to the ester
carbonyl appears, suggesting that the above presumption is exactly
true. Surprisingly, the integration of signal attributed to the
dendritic unit was much larger than that corresponding to the
linear unit, which is quite different from usual hyperbranched
polymers reported in literature. The DB value was calculated from
DB ¼ (T þ D)/(T þ L þ D) to be 0.67, and that calculated from
DB ¼ 2D/(2D þ L) was 0.70. These values are higher than both the
usual value of 0.30e0.50 and the theoretical value of 0.5 via one-
pot polymerization of an AB2 monomer. This indicates that the rate
of esterification reaction was higher than that of ring-opening
reaction. As a contrastive experiment, equimolar TMP, instead of
EHO, was polymerized with SA under the same condition. The 1H
NMR spectrum of poly(SA-TMP) is also shown as the curve A in
Fig. 2. In contrast, the integration of signal attributed to the
dendritic unit is much less than that corresponding to the linear
unit and the DB value calculated according to Frey was only 0.42.
The glass transition temperature (Tg) of the obtained hyper-
branched poly(SA-EHO) was determined by DSC to be around
ꢀ11.5 ꢁC. The average molecular weight (Mn) and its poly-
dispersibility index (PDI) were measured with GPC spectroscopy
although this method is limited because of the calibration with
standard linear PSt. The results are listed in Table 1. The Mn of
Fig. 2. 1H NMR spectra of (A) poly(SA-TMP), (B) poly(SA-EHO) and (C) end-capped poly
(SA-EHO).

3406
F. Zhan et al. / Polymer 51 (2010) 3402e3409
Table 1
Br
O
HO
Product yields and characters of hyperbranched polyesters.
O
+
O
OH
OH
Polymer
Yield (%)
DB
Mn (g/mol)
PDI
Tg (ꢁC)
O
O
Poly(PA-EHO)
Poly(CHA-EHO)
Poly(SA-EHO)
CoPoly(SA-EHO)
Poly(SA-EHPO)
82
86
82
79
81
e
e
e
e
O
e
e
e
e
Hyperbranched Polymer
+
O
OH
0.70
0.67
e
1914
2300
2108
2.39
1.60
2.28
ꢀ11.5
ꢀ9.9
14.3
O
Scheme 2. Synthesis routes of EHPO and Poly(SA-EHPO).
hyperbranched poly(SA-EHO) is 1914, perhaps being under-
estimated, and the polydispersity index is 2.39. To obtain a hyper-
branched polymer with higher molecular weight, higher reaction
temperature and longer reaction time were considered. However,
higher molecular weight product could not be obtained and gela-
tion occurred under both conditions. The lower molecular weight
of the polymer is probably due to some intramolecular reactions
and the gelation was probably caused by some side reactions such
as intermolecular etherification between two hydroxyl groups or
a hydroxyl group and an oxetane group.
The 1H NMR spectra of EHPO and poly(SA- EHPO) were shown as
Figs. 3 and 4, respectively. The DB couldn’t be calculated directly
according tothe 1H NMR result, which may be due tothe influence of
benzene ring in the polymer chain. However, the peak attributed to
proton on oxetane ring exits and the signal corresponding to the
methylene adjacent to hydroxyl group is weak. This indicates that
the hyperbranched polymer, poly(SA-EHPO) was obtained. The
average molecular weight and its polydispersity were also charac-
terized with GPC and to be 2108 g/mol and 2.28, respectively. The Tg
of poly(SA-EHPO) was determined to be higher than those of poly
(SA-EHO) and copoly(SA-EHO). The reason is presumed to be the
benzene ring in the main chain of poly(SA-EHPO).
3.3. Copolymerization of SA-EHO with TMP
To avoid gelation and obtain a hyperbranched polymer with
higher molecular weight, the copolymerization approach was
considered. A. Hult developed the copolymerization of 2,2-Bis
(hydroxymethyl)propionic acid (bis-MPA) with a polyol core, and
found that the risk of gelation was greatly reduced [31,35,36].
However, the gelation occurred when SA-EHO as a monomer was
copolymerized with TMP as a three-functional core molecule. The
reaction parameters and characters of a typical soluble copolymer,
which were measured before gelation, are listed in Table 1. The
average molecular weight of the copolymer is 2300, bigger than
that of poly(SA-EHO), but having a narrow polydispersity of 1.6.
In poly(SA-EHPO), the ester bond which makes the cationic
curing activity of oxetane groups decreased is replaced by ether
bond and the deactivation effect will not occur, which enhance the
cationic UV-curing.
3.5. Properties of the UV cured films
The formulations containing different ratios of EP828 to poly
(SA-EHO) or poly(SA-EHPO) were prepared with the addition of
2 wt.% cationic photoinitiator Irgacure 250, and exposed to
a medium pressure mercury lamp (2 KW, Fusion UV systems, USA)
to obtain the cured films. For EP828/poly(SA-EHO) system (Table 2),
a series of EPmPSEn were obtained, where m is the percent content
of EP828, and n is the percent content of poly(SA-EHO). For EP828/
poly(SA-EHPO) system, another series of EPpPSEPq were obtained,
where p is the percent content of EP828, and q is the percent
content of poly(SA-EHPO).
3.4. Cationic curing
The hyperbranched poly(SA-EHO) was investigated as an addi-
tive in cationic UV-curing systems. The samples containing
different ratios of poly(SA-EHO) to EP828 and 2 wt.% cationic
photoinitiator were prepared (Table 2), and exposed to a medium
pressure mercury lamp to obtain the cured films. However, as the
content of poly(SA-EHO) increased, the curing of the formulation
became difficult. When the ratio reached to 30%, the film couldn’t
cure completely even with a long UV exposure time. This is inter-
preted to be due to the influence of the carbonyl group on the ester
bond adjacent to the oxetane ring [37,38]. The carbon atom on the
oxetane ring became more electron-deficient after the oxetane ring
was activated by protonic acid while the oxygen atom on the
carbonyl group was electron-rich. The oxygen atom attacked the
carbon atom, resulting in the formation of a stable six-member
ring. This side reaction played an important role and gave rise to
deactivation of oxetane groups.
The dynamic mechanical thermal analysis was employed to
investigate the dynamic mechanical behavior of the cured films.
To avoid the side reaction, a new hydroxyl oxetane, EHPO, was
synthesized to prepare a novel polymer, poly(SA-EHPO) (Scheme 2).
Table 2
Viscoelastic properties of the cured EP828/Poly(SA-EHO) films.
Sample
Formulation (wt.%)
FWHM (ꢁC)
Ve (mmol/cm3)
EP828
Poly(SA-EHO)
EP100PSE0
EP95PSE5
EP90PSE10
EP85PSE15
EP80PSE20
100
95
90
85
80
0
5
10
15
20
21.5
24.7
23.4
25.8
5.37
4.79
4.56
3.92
Fig. 3. 1H NMR spectrum of EHPO.

F. Zhan et al. / Polymer 51 (2010) 3402e3409
3407
Fig. 6. DMTA curves of cured EP828/Poly(SA-EHPO) films with different Poly(SA-
EHPO) contents.
Moreover, it is well known that the value of tan
viscous modulus to elastic modulus and that the peak value of tan
curve can be used to predict the damping behavior of a material. As
shown in Fig. 5, the peak value of tan increases with an increasing
d is the ratio of
d
Fig. 4. 1H NMR spectrum of poly(SA-EHPO).
d
weight fraction of poly(SA-EHO). This indicates that the cured film
with higher content of poly(SA-EHO) has higher damping property.
The breadth of the glass transition can be defined by the full width
The storage modulus and loss factor (tan
d
) curves of cured EP828/
poly(SA-EHO) films are shown in Fig. 5 and the detailed results are
listed in Table 2. The crosslinking density (Ve) is the molar number
of elastically effective network chains per cubic centimeter of the
cured film. The Tg can be defined as the temperature at which the
at half maximum (FWHM) of the tan d peaks. As listed in Table 2, all
the FWHMs of the peaks are in a temperature range of 23.5 ꢃ 2.5 ꢁC.
The uniformity of the tan
d peak indicates the absence of any
peak of tan
d curve appears. As listed in Table 2, the crosslinking
network heterogeneity, which implies the good miscibility
between poly(SA-EHO) and EP828.
density of cured film decreases with the content of hyperbranched
poly(SA-EHO) increasing. This behavior can be attributed to two
reasons. On the one hand, the relative concentration of oxetane
group in poly(SA-EHO) is lower than that of epoxy group in EP828
resin. On the other hand, the oxetane group of poly(SA-EHO) has
lower cationic UV-curing activity because of the deactivation effect
of the carbonyl group on the ester bond adjacent to the oxetane
ring. The Tg of cured film decreases continuously from 140 ꢁC for
pure EP828 to 104 ꢁC for 20 wt.% poly(SA-EHO) addition. This can
also be ascribed to two factors. Firstly, the crosslinking density
decreases with the content of hyperbranched poly(SA-EHO)
increasing. Secondly, the hyperbranched poly(SA-EHO) chain is
aliphatic while EP828 is an aromatic resin. As a result, the incor-
porated hyperbranched poly(SA-EHO) can reduce the rigidity of the
formed network and make chain motion possible at lower
temperature.
The EP828/poly(SA-EHPO) cured films are also tested via
dynamic mechanical thermal analysis and the results are given in
Fig. 6 and Table 3. It is interesting to note that the crosslinking
density of the cured films first increases and then decreases with
the increasing of poly(SA-EHPO). The increase in the crosslinking
density along with poly(SA-EHPO) addition may be due to the
higher reactivity of the oxetane group in poly(SA-EHPO), compared
with epoxy group in EP828 [13e17]. Due to the small amount of
oxetane group in the formulation and overlapping of peaks in the
range of 940 cmꢀ1 to 1097 cmꢀ1 for oxetane and other ether bonds,
as shown in Fig. 7, the investigation of relative reactivity and the
polymerization kinetics were failed with the real-time FT-IR spec-
troscopic analysis. Moreover, it is found that the further addition of
poly(SA-EHPO) makes the crosslinking density of cured film
decreasing. That can be attributed to two factors. On the one hand,
the relative concentration of reactive group in poly(SA-EHPO) is
much lower than that in EP828 resin. On the other hand, the
hydroxyl group in poly(SA-EHPO) can react with the growing chain
through the chain transfer mechanism, resulting in the decrease in
the formed polymeric chain length, thus the lower crosslinking
density of the cured film [39]. The Tg of cured films decreases with
poly(SA-EHPO) content increasing. That’s mainly due to the less
Table 3
Viscoelastic properties of the cured EP828/Poly(SA-EHPO) films.
Sample
Formulation (wt.%)
FWHM (ꢁC)
Ve (mmol/cm3)
EP828
Poly(SA-EHPO)
EP100PSEP0
EP95PSEP5
EP90PSEP10
EP85PSEP15
EP80PSEP20
EP70PSEP30
100
95
90
85
80
70
0
5
10
15
20
30
29.8
26.4
26.4
22.6
24.7
23.7
4.58
5.06
4.80
4.75
4.36
4.13
Fig. 5. DMTA curves of cured EP828/Poly(SA-EHO) films with different Poly(SA-EHO)
contents.

3408
F. Zhan et al. / Polymer 51 (2010) 3402e3409
4. Conclusion
Two competitive reactions, ring-opening reaction and esterifi-
cation, between carboxylic anhydride and hydroxyl oxetane took
place during the polymerization. In this work, it was found that
a linear or little branched polymer was obtained when cyclo-
hexene-1,2-dicarboxylic anhydride and phthalic anhydride were
selected as monomers. However, when succinic anhydride was
used as a monomer, the hyperbranched polyester, poly(SA-EHO),
was synthesized successfully. The obtained poly(SA-EHO) has
a higher DB of about 0.7 than that reported in literature, calculated
directly according to the 1H NMR result. This indicates that the rate
of esterification was much higher than that of ring-opening reac-
tion. The Tg of poly(SA-EHO) was ꢀ11.5 ꢁC from DSC. And the
average molecular weight was determined to be 1914 g/mol,
perhaps being underestimated, with the polydispersity index of
2.39. The hyperbranched poly(SA-EHPO) with aromatic component
was also prepared. Both hyperbranched polyesters were added to
EP828 resin as additives with different contents, and cured by the
exposure to a UV lamp, respectively.
Fig. 7. FT-IR spectrum of EP50PSEP50 formulation.
The Tg measured by DMTA and tensile strength decrease with
increasing poly(SA-EHO) in the cured EP828/poly(SA-EHO) film.
However, for EP828/poly(SA-EHPO) film, the Tg decreases, but the
tensile strength increases with the addition of a small amount of
poly(SA-EHPO), and then decreases with further addition. The
EP828/poly(SA-EHPO) films have higher tensile strength and lower
elongation at break than EP828/poly(SA-EHO) films.
rigid polymer chain in poly(SA-EHPO). The peak value of tan
d also
increases with the weight fraction of hyperbranched poly(SA-
EHPO) increasing, which implies that the cured film with higher
content of hyperbranched poly(SA-EHO) has higher damping
property. The glass transitions of the cured films keep narrow,
indicating that the hyperbranched poly(SA-EHPO) and EP828 have
good compatibility.
Acknowledgements
The mechanical properties of cured films were measured via
tensile test and the results are listed in Table 4. The tensile strength
and elongation at break are mainly influenced by the crosslinking
density and the rigidity of the network chains. Both the tensile
strength and elongation at break decrease when poly(SA-EHO) was
added. That’s because both the crosslinking densityand the rigidity of
the formed network decrease. However, when a small amount of poly
(SA-EHPO) was added, the increasing of the crosslinking density, as
discussed above, makes the tensile strength increased. As the content
of poly(SA-EHPO) increases continuously, the tensile strength
decreases due to the decrease of aromatic group concentration in the
formulation and thus the rigidity. The elongation at break always
increases for both systems, suggesting that the increased flexibility of
the network is the primary factor in the change.
The authors gratefully acknowledge the financial support of the
National Natural Science Foundation of China (No. 50633010).
References
[1] Yan DY, Gao C. Macromolecules 2000;33:7693e9.
[2] Gao C, Yan DY. Chem Commun 2001;1:107e8.
[3] Forehling P, Brackman J. Macromol Symp 2000;151:581e9.
[4] van Benthem RATM, Meijerink N, Geladé E, de Koster CG, Muscat D,
Froehling PE, et al. Macromolecules 2001;34:3559e66.
[5] Gao C, Yan DY. Prog Polym Sci 2004;29:183e275.
[6] Li XR, Zhan J, Li YS. Macromolecules 2004;37:7584e94.
[7] Li XR, Lu XF, Lin Y, Zhan J, Li YS, Liu ZQ, et al. Macromolecules
2006;39:7889e99.
[8] Wang SJ, Fan XD, Kong J, Lu JR. Polymer 2009;50:3587e94.
[9] Wang SJ, Fan XD, Kong J, Wang X, Liu YY, Zhang GB. J Polym Sci Part A Polym
Chem 2008;46:2708e20.
[10] Sangermano M, Bongiovanni R, Malucelli G, Priola A, Olbrych J, Harden A. J
Polym Sci Part A Polym Chem 2004;42:1415e20.
[11] Malucelli G, Bongiovanni R, Sangermano M, Ronchetti S, Priola A. Polymer
2007;48:7000e7.
[12] Sasaki H, Crivello JV. J Macromol Sci Pure Appl Chem 1992;A29(10):915e30.
[13] Crivello JV, Sasaki H. J Macromol Sci Pure Appl Chem 1993;A30(2&3):189e206.
In particular, the cured EP828/poly(SA-EHPO) films have higher
tensile strength and lower elongation at break than EP828/poly(SA-
EHO) films when the same contents of polymers were added. That’s
attributed to two factors, the reactivity of oxetane group and the
rigidity of the polymer chain.
[14] Sasaki H, Rudzinski JM, Kakuchi T.
J Polym Sci Part A Polym Chem
1995;33:1807e16.
[15] Sangermano M, Bongiovanni R, Malucelli G, Priola A, Thomas RR, Medsker RE,
et al. Polymer 2004;45:2133e9.
[16] Lengvinaite S, Sangermano M, Malucelli G, Priola A, Grigalevicius S,
Grazulevicius JV, et al. Eur Polym J 2007;43:380e7.
Table 4
Tensile strength and elongation at break of the cured EP828/Poly(SA-EHO) and
EP828/Poly(SA-EHPO) films.
[17] Verstegen EJK, Kloosterboer JG, Lub J. J Appl Polym Sci 2005;98:1697e707.
[18] Sangermano M, Gianni AD, Malucelli G, Roncuzzi C, Priola A, Voit B. J Appl
Polym Sci 2005;97:293e9.
[19] Sangermano M, Malucelli G, Bongiovanni R, Priola A, Harden A, Reheberg N.
Polym Eng Sci 2003;43:1460e5.
[20] Sangermano M, Amerio E, Di Gianni A, Priola A, Pospiech D, Voit B. Macromol
Symp 2007;254:9e15.
[21] Zhang DH, Jia DM. J Appl Polym Sci 2006;101:2504e11.
[22] Blaskovich MA, Evindar G, Rose NGW, Wilkinson S, Luo Y, Lajoie GA. J Org
Chem 1998;63:3631e46.
[23] Sato K, Kameyama A, Nishikubo T. Macromolecules 1992;25:1198e9.
[24] Nishikubo T, Kameyama A, Minegishi S. Macromolecules 1994;27:2641e2.
[25] Nishikubo T, Kameyama A, Suzuki A. React Funct Polym 1998;37:19e25.
[26] Nishikubo T, Kudo H, Nakagami T. Polym J 2006;38:145e52.
[27] Kudo H, Nishikubo T. J Polym Sci Part A Polym Chem 2007;45:709e26.
Samples
Tensile strength (MPa)
Elongation at break (%)
EP100
64.9 (1.9)
59.1 (3.2)
52.8 (2.5)
43.3 (3.7)
36.9 (2.3)
3.1 (0.2)
3.7 (0.6)
5.1 (0.2)
7.1 (0.7)
7.9 (0.4)
EP95PSE5
EP90PSE10
EP85PSE15
EP80PSE20
EP95PSEP5
EP90PSEP10
EP85PSEP15
EP80PSEP20
EP70PSEP30
69.7 (3.5)
64.1 (2.7)
58.1 (4.2)
52.7 (3.1)
42.9 (3.9)
3.4 (0.3)
3.8 (0.3)
4.5 (0.6)
5.2 (0.5)
7.2 (0.9)
Standard deviation listed in parenthesis.

F. Zhan et al. / Polymer 51 (2010) 3402e3409
[35] Malmström E, Hult A. Macromolecules 1996;29:1222e8.
3409
[28] Flory PJ. J Am Chem Soc 1952;74:2718e23.
[29] Hölter D, Frey H. Acta Polym 1997;48:298e309.
[30] Cheng KC. Polymer 2003;44:1259e66.
[36] Magnusson H, Malmström E, Hult A. Macromolecules 2000;33:3099e104.
[37] de Ruiter B, El-ghayoury A, Hofmeier H, Schubert US, Manea M. Prog Org Coat
2006;55:154e9.
[38] El-ghayoury A, Boukaftane C, de Ruiter B, van der Linde R. J Polym Sci Part A
Polym Chem 2003;41:469e75.
[31] Malmström E, Johansson M, Hult A. Macromolecules 1995;28:1698e703.
[32] Smet M, Schacht EH, Dehaen W. Angew Chem Int Ed 2002;41:4547e50.
[33] Sinananwanich W, Ueda M. J Polym Sci Part A Polym Chem 2008;46:2689e700.
[34] Fu Y, Van Oosterwijck C, Vandendriessche A, Kowalczuk-Bleja A, Zhang X,
Dworak A, et al. Macromolecules 2008;41:2388e93.
[39] Sangermano M, Messori M, Galleco Martin M, Rizza G, Voit B. Polymer
2009;50:5647e52.