Poly(butylene terephthalate) modified with ethoxylated bisphenol S with increased glass transition temperature and improved thermal stability

Article abstract of DOI:10.1016/j.polymer.2011.01.018

Novel copolyesters have been prepared by polycondensation and by melt mixing of poly(butylene terephthalate) with an ethoxylated bisphenol S. No side reactions occur during the synthesis of the samples, as proved by NMR analysis. The polyesters were examined by TGA and DSC. The insertion of the bisphenol S (sulfonyldiphenol) group significantly improved the thermal stability of the polymer. The thermal analysis carried out using DSC technique showed that the T_m of the copolymers decreased with increasing co-unit content, differently from T_g, which on the contrary increased, exceeding in some cases 100 °C, and crystallization rate decreased. A polymer containing only terephthalate moieties and ethoxylated bisphenol S has been prepared for the first time.

Full text of DOI:10.1016/j.polymer.2011.01.018

Polymer 52 (2011) 904e911
Contents lists available at ScienceDirect
Polymer
journal homepage: www.elsevier.com/locate/polymer
Poly(butylene terephthalate) modified with ethoxylated bisphenol S with
increased glass transition temperature and improved thermal stability
*
N. Lotti, M. Colonna , M. Fiorini, L. Finelli, C. Berti
Dipartimento di Ingegneria Civile, Ambientale e dei Materiali, Via Terracini 28 40131 Bologna, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 November 2010
Received in revised form
Novel copolyesters have been prepared by polycondensation and by melt mixing of poly(butylene tere-
phthalate) with an ethoxylated bisphenol S. No side reactions occur during the synthesis of the samples, as
proved by NMR analysis. The polyesters were examined by TGA and DSC. The insertion of the bisphenol S
4
January 2011
(
sulfonyldiphenol) group significantly improved the thermal stability of the polymer. The thermal analysis
Accepted 9 January 2011
Available online 15 January 2011
carried out using DSC technique showed that the T
content, differently from T
m
of the copolymers decreased with increasing co-unit
, which on the contrary increased, exceeding in some cases 100 C, and crys-
ꢀ
g
tallization rate decreased. A polymer containing only terephthalate moieties and ethoxylated bisphenol S
has been prepared for the first time.
Keywords:
Terephthalate polyester
Bisphenol S
Ó 2011 Elsevier Ltd. All rights reserved.
Thermal properties
1. Introduction
Bisphenol A cannot be directly incorporated in the polyester
main chain since aromatic diols are not active in esterification
Terephthalate polyesters are widely used as engineering ther-
moplastic materials due to their good chemical resistance, high
thermal and dimensional stability, high strength and rigidity
coupled with good surface hardness and gloss [1]. These polymers
have found the main applications in precision moulding for elec-
trical and electronic devices, domestic and office appliances and
automotive parts. Poly(butylene terephthalate) (PBT) is the most
widely used terephthalate polyester as engineering thermoplastic
due to its high crystallization rate and high crystallinity which
imparts good solvent and hydrolysis resistance. The main drawback
of terephthalate polyesters lies in their low impact resistance
especially at low temperatures. On the contrary, one of the most
important features of bisphenol A polycarbonate (PC) is its high
impact resistance, even at low temperatures. It is reported in the
literature [2,3] that the high impact resistance of PC can be ascribed
to its capability to absorb high impact energy which, in turns, has
been correlated with secondary transitions involving cooperative
reactions with terephthalic acid or dimethyl terephthalate and
therefore has to be modified in order to be inserted in the polyester
backbone during a normal polycondensation process [1].
We have reported [4,5] that the insertion of bisphenol A (BPA)
units inside a terephthalate polyester leads to an increase in glass
transition temperature and in thermal stability of PBT, PET and PPT.
Similar results have been reported by Koning et al. [6] on block
copolymers obtained by solid state polymerisation. The approach we
have followed consists in the derivatization of BPA via ethoxylation
and the functionalized BPA was used directly in melt mixing with
terephthalate polyesters. We studied bis(hydroxyethyl ether) of
bisphenol A (BHEEB), since several easy and low-cost routes for the
synthesis of BHEEB are reported in the literature [7e11], for example
by reaction between ethylene carbonate (ETC) and BPA [7] and by
reaction of PC, ethylene glycol and ETC in presence of sodium
hydroxide. The reaction of BPA with ETC permits the preparation of
ethoxylated bisphenol Awithout the formation of aliphaticealiphatic
ether linkages as usually observed when ethylene oxide is used. The
insertion of ethoxylated BPA units in the polyester backbone has been
performed during the polymerisation or by melt mixing, which in our
opinionisthe mostinteresting methodwhenalowamountofdiolhas
to be incorporated in order to obtain semicrystalline materials. Melt
mixing allows recycledterephthalate polyestersto be used and can be
performed in shorter reaction times compared to the full polymeri-
sation procedure.
motions of bisphenol A (BPA) units that take place at very low
ꢀ
temperature (T
g
¼ ꢁ105 C from dynamic mechanical analysis [2]).
However, PC presents poor solvent resistance. For these reasons
blends of PC with PBT are industrially used for applications where
low temperature impact resistance and solvent resistance are both
required, as for example, in car bumpers.
It has been reported [12] that the insertion of sulfonyldiphenol
units (bisphenol S, BPS) leads to a more consistent improvement in
*
Corresponding author. Tel.: þ390512090331; fax: þ390512090322.
E-mail address: martino.colonna@unibo.it (M. Colonna).
0
032-3861/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymer.2011.01.018

N. Lotti et al. / Polymer 52 (2011) 904e911
905
T
g
of terephthalate copolyesters with respect to the use of BPA
2.4.2. DSC measurements
derivatives. This behavior has been attributed to the higher rigidity
of the BPS unit since in this molecule it is not possible to bend the
CeSeC bonds due to the presence of SeO double bonds. Recently,
the synthesis of copolyesters based on terephthalic acid, ethylene
glycol, cyclohexane dimethanol and bis(hydroxyethyl ether) of
bisphenol S (BHEBS) has been patented by Turner and Sublett from
Eastman Kodak [13]. They reported that the PETG copolymers with
Calorimetric measurements were carried out by means of
a Perkin Elmer DSC7 instrument equipped with a liquid sub
ambient accessory and calibrated with high purity standards
(indium and cyclohexane). With the aim of measuring the glass
transition and the melting temperatures of the polymers under
investigation, the external block temperature control was set at
ꢀ
ꢁ60 C and weighed samples of c.a. 10 mg were encapsulated in
ꢀ
BHEBS present higher T
g
and improvements in environmental
aluminum pans and heated to about 40 C above fusion
ꢀ
stress cracking resistance after lipid exposure. They also claim the
use of butanediol as monomer but always in combination with
another aliphatic glycol. Therefore, it is of interest for the industrial
and scientific community the study of the effect of these units on
other polyesters and the properties of a polymer containing only
terephthalate-BHEBS (BHEBST) units.
temperature at a rate of 20 C/min (first scan), held there for
ꢀ
3 min, and then rapidly quenched to ꢁ10 C. Finally, they were
ꢀ
reheated from ꢁ10 C to a temperature well above the fusion
ꢀ
temperature of the sample at a heating rate of 20 C/min (second
scan). The glass-transition temperature T was taken as the
p
midpoint of the heat capacity increment Dc associated with the
g
The aim of our work was to prepare semicrystalline copolymers
containing bisphenol S moieties. To the best of our knowledge, the
literature does not report a complete and systematic study of the
synthesis and of the properties of terephthalate polyesters con-
taining BHEBS units.
In this paper we report the synthesis of poly(butylene tere-
phthalate) containing BHEBST units and the chemical and thermal
characterization of the copolymers synthesised. The synthesis of
the polyesters have been performed both by melt mixing of PBT
with BHEBS and subsequent polycondensation under vacuum, and
by melt polycondensation starting from monomers. It has to be
emphasized that the synthesis of a polyester containing only
terephthalate and ethoxylated bisphenol S units (PBHEBST) has
been never carried out before.
glass-to-rubber transition. The melting temperature (T
the crystallization temperature (T ) were determined as the peak
value of the endothermal and the exothermal phenomena in the
DSC curve, respectively. The specific heat increment , asso-
ciated with the glass transition of the amorphous phase, was
calculated from the vertical distance between the two extrapo-
lated baselines at the glass transition temperature. The heat of
m
) and
c
Dc
p
m c
fusion (DH ) and the heat of crystallization (DH ) of the crystal
phase were calculated from the areas of the DSC endotherm and
exotherm, respectively. In order to determine the crystallization
rate under non-isothermal conditions, the samples were heated
ꢀ
ꢀ
at 20 C/min to about 40 C above fusion temperature, kept
ꢀ
there for 3 min and then cooled at 10 C/min. The temperature
corresponding to the maximum of the exothermic peak in the
DSC cooling-curve (Tcc) can be correlated to the crystallization
rate.
2
. Experimental
2.1. Materials
2.5. Syntheses
Poly(butylene terephthalate) (PBT, Valox 315, M
w
125000) was
2.5.1. BHEBS synthesis from BPS and ETC
The reaction was carried out in a 500 mL three-neck round-
bottom flask equipped with a nitrogen inlet and a reflux condenser.
ꢀ
a gift of SABIC Innovative Plastics. PBT was dried overnight at 120 C
under vacuum before use. ETC, BPS, DMT, butanediol (BD) and
Titanium butoxide (TBT) (all from Aldrich Chemicals) were high
purity products and were not purified before use.
2 3
BPS (125 g; 0.50 mol), ETC (90.0 g; 1.02 mol) and K CO (0.30 g;
2.83 mmol) were mixed under a stream of nitrogen. The mixture
ꢀ
was heated at 220 C for 2 h following the progress of the reaction
2.2. Gel-permeation chromatography
by thin layer chromatography. The crude product was crushed into
powder and washed three times with a 2M solution of KOH in water
Molecular weight data were obtained by gel permeation chroma-
tography at 30 C using a 1100 Agilent Series system with a UV spec-
trophotometer (at 254 nm wavelength) as detector, equipped with
and twice with water (yield 95%).
H NMR analysis (solvent CDCl ): d (ppm downfield from TMS)
3
7.8 (4H, d), 6.90 (4H, d), 4.10 (4H, t).
ꢀ
1
Agilent PLgel 5
mixture of chloroform/1,1,1,3,3,3-hexafluoro-2-propanol (CHCl
95/5 v/v) was used as eluent with a 0.3 mL/min flow, and sample
concentrations of about 2 mg/mL were applied. A molecular weight
calibration curvewas obtained with several monodisperse polystyrene
standards in the range of molecular weight 2000e200,000 g/mol.
m
MiniMIX-C column (250/4.6 length/i.d., in mm). A
3
/HFIP)
2.5.2. Synthesis of copolyesters by melt mixing of PBT with BHEBS
The melt mixing of PBT and BHEBS was performed in a Bra-
bender Plasticorder PL2000/W50. 50.0 g of PBT were charged in
(
ꢀ
the Brabender mixer heated at 245 C. After the complete mel-
ting of the polymer, BHEBS and 100 ppm of TBT catalyst (as
titanium with respect to the final terephthalate polyester) were
added and the mixing chamber closed with a cover to seal the
mixer.
2.3. NMR spectroscopy
1H NMR and 13C NMR spectra were recorded with a Varian XL-
00 spectrometer (chemical shifts are downfield from tetrame-
The reactions were performed in two stages; in the first one,
carried out at atmospheric pressure for 15 min, the alcoholysis of
the polyester by BHEBS took place. In the second stage, the pres-
sure was carefully reduced down to 2 mbar in 30 min and buta-
nediol was distilled off from the mixing chamber and recovered in
a condenser. The reaction was stopped when no further increase in
the torque signal was observed. The second stage time varied from
30 min to 2 h depending on the BHEEB amount added. The copo-
lyesters obtained and analyzed in this work will be indicated as
PBT/BHEBSTX, where X is the amount of BHEBST co-units exp-
ressed as mol %.
3
thylsilane (TMS)). The solvent was CDCl
3
for monomers and
a mixture of CF COOD/CDCl (20/80, wt.-%) for the polymers.
3
3
2
2
.4. Thermal analysis
.4.1. TGA measurements
Thermogravimetric curves were obtained both in air and under
nitrogen atmosphere using a Perkin Elmer TGA7 apparatus (gas
ꢀ
ꢀ
flow: 50 mL/min) at 10 C/min heating rate up to 900 C.

906
N. Lotti et al. / Polymer 52 (2011) 904e911
O
C
O
S
O
2
_
20°C
CO2
HO
OH + 2
HO CH2 CH2
O
S
O
O CH2 CH2 OH
O
O
O
2
BPS
ETC
BHEBS
Fig. 1. Ethoxylated bisphenol S (BHEBS) synthesis.
2.5.3. Polyester syntheses from BHEBS, dimethyl terephthalate and
3. Results and discussion
butanediol
The synthesis of the copolymers was also performed starting
from the monomers according to the standard two-step procedure
for polyester synthesis [14].
BPS can be easily ethoxylated in very high yields by reaction
with ethylene carbonate (Fig. 1).
The insertion of ethoxylated BPS inside the polyester can be
performed either by melt mixing with the polyester (and subse-
DMT (50.0 g; 257 mmol), BD (29.6 g; 329 mmol), and BHEBS
were introduced in a 250 mL three-neck round-bottom flask
provided with mechanical stirring, a torque meter, a Dean Stark
w
quent polycondensation to reach high M ) (Fig. 2) or starting from
monomers. We have performed mainly the first route but we have
ꢀ
1
13
condenser and a 20 cm reflux column heated at 60 C. TBT
also successfully tested the second method. H NMR and C NMR
analysis showed no significant differences (end-groups, side reac-
tions and degree of randomness) in the chemical composition of
the polymers obtained with the two methods Table 1.
(
100 ppm as titanium with respect to the final polymer) was
added to the reagents and the flask was placed in an oil bath at
ꢀ
ꢀ
1
50 C. The temperature was then increased to 230 C over
3
0 min and kept at that temperature until no significant increase
The BHEBS insertion by reactive blending is quantitative and no
side reactions have been observed. For example, the H NMR
1
in distillate volume was observed (120 min from the catalyst
addition). By 1H NMR analysis, no BHEBS was found in the
distillate. The Dean Stark condenser was removed and the reactor
was connected with a liquid nitrogen cooled trap. The reaction
temperature was then increased to 245 C and dynamic vacuum
was slowly applied. The reaction was conducted at 1 mbar until
no increase in torque signal was observed (150 min from vacuum
spectrum of PBT/BHEBST20 is reported in Fig. 3. The NMR spectrum
shows only the expected peaks of the PBT and of the BHEBST units
according to the assignment of the peaks reported on top of Fig. 3.
No peak due to side reactions (for example vinyl end-groups) has
been found in the NMR spectrum. The amount of BHEBS inserted
has been calculated comparing the peak of the bisphenol S units at
ꢀ
1
application). The chemical structure (by H NMR), molecular
weight (by GPC), and colour of the final polymer do not signifi-
cantly differ from that of the polymer obtained by reactive
blending.
Table 1
ꢀ
PBT/BHEBST copolymers prepared by reactive blending at 245 C.
BHEBS
Feed (mol%)
BHEBS inserted
(mol %)
Second stage
time (h)
M
w
GPC
n
M GPC
(g/mol)
A polymer from DMT and BHEBS was synthesised using the
same procedure described above. In this case a 5 mol% excess of
BHEBS with respect to DMT was used. The reaction was conducted
(g/mol)
5
4.9
10.1
19.9
49.5
2
2
2
3
52660
63740
41370
33590
30350
37500
18960
14570
10
ꢀ
ꢀ
for 2 h at 215 C at atmospheric pressure and for 170 min at 245 C
under dynamic vacuum (1 mbar).
20
50
O
S
O
O
O C
O
m HO CH2 CH2
O
O
CH2 CH2 OH
+
C O CH2
4
n
BHEBS
TBT catalyst
15 min 245°C
O
S
O
O
O
m HO CH2 CH2
O
O
CH2 CH2 O C
C O CH2
4
n/m
O
Dynamic vacuum 245°C
O
S
O
O
O
O
O C
CH2 CH2
O
O
CH2 CH2 O C
C O CH2
C O
4
x
y z
+
BHEBST
BT
x z = m
*
*
y z = m-n
HO CH2 CH2 CH2 CH2 OH
Fig. 2. Insertion of ethoxylated BPS units inside the polyester chain by reactive blending with PBT.

N. Lotti et al. / Polymer 52 (2011) 904e911
907
Fig. 3. 1H NMR of PBT/BHEBST20.
d
d
¼ 7.0 ppm (4H) with that of the terephthalate group at
¼ 8.1 ppm (4H) after dissolution of the polymer in a mixture of
can be seen, PBT/BHEBST5 sample shows a thermal stability similar
to that of pure PBT, whereas in the case of the copolymers con-
taining from 10 to 50 mol % of BHEBST co-units, the thermal
stability appears to depend on composition, being higher as
BHEBST unit content is increased. Furthermore, it can be noted that
for all the samples in Fig. 7a, the weight-loss takes place almost
completely in one-step, except for pure PBHEBST whose TGA curve
shows two weight-loss steps.
3 3
CHCl /CF COOH 80/20 (v/v) and precipitation in methanol.
The distribution of BHEBST units along the polyester chain is
13
random as verified by C NMR (see as an example PBT/BHEBST50
copolymer in Fig. 5). Inthis case fourdifferent triads canbe formed by
reaction between the two diols (butanediol labeled B and BHEBS
labeled S) andtheterephthalate unit (T) (Fig. 4). According to Fig. 5, in
13
C NMR spectrum the triad SeTeS should give rise to a single signal
Lastly, the thermogravimetric curve of PBT is characterized by
a very low char residue (5%), whereas that of pure PBHEEBT shows
the highest char residue among all the samples under investigation
(25%). As far as the copolymers are concerned, the char residue
appears to depend on composition, increasing as the amount of
BHEBST co-units is increased. An analogous dependence on
composition of the thermal stability was observed in air (see
Fig. 7b), even though all the samples are characterized by a weight-
loss of 100% and the thermal degradation process takes place
always in two separate steps, the latter one being more evident as
higher is the amount of BHEBST unit in the polymeric chain. The
improved thermal stability of PBT by introduction of BHEBST units
along the polymeric chain may be explained as due to the steric
hindrance and high stiffness of the co-units, which prevent thermal
degradation reactions, which usually occur in polyesters when
heated [15,16]. An analogous result was obtained previously by us
investigating the effect of the introduction of the 2,2-bis[4-(ethyl-
enoxy)-1,4-phenylene]propane terephthalate co-monomeric units
on the thermal stability of PBT [5].
of the carboxyl groups (a in Fig. 5) and of the aromatic carbon bonded
to the carboxyl groups (b in Fig. 5). Also the triad BeTeB should give
rise to a single signal. On the contrary, both the triad SeTeB and
13
BeTeS should give rise to two different signals in C NMR analysis.
Taking into account that in PBT/BHEBST50 polymer the unit B and S
are present in equal amounts and that the two signals of the triads
SeTeB and BeTeS are equivalent, if the co-monomers are randomly
distributed along the polymer chain, four signals of equal intensity
should be present, since the four triads are equally probable. Indeed,
the two expansions in Fig. 5 show that four signals of the same
intensity are present, indicating that the expected random distribu-
tion of co-monomeric units have been achieved.
The polyester from DMT and BHEBS has been prepared by
a standard two steps melt process using TBT catalyst. The final M
w
1
of the polymer was 30 000 g/mol (measured by GPC). The H NMR
spectrum in Fig. 6 shows only the expected peaks. Again, no side
1
reaction was detected by H NMR analysis. End groups are mainly
methyl ester while a small amount of BHEBS end-groups has been
found. As mentioned in the Introduction, this is the first synthesis
reported in the literature of BHEBS terephthalate polyester.
The copolyesters and the two parent homopolymers were
preliminarily examined using thermogravimetric analysis. Investi-
gation of the thermal stability was carried out both in air and under
nitrogen atmosphere. Typical weight-loss curves as a function of
temperature under nitrogen atmosphere are shown in Fig. 7a. As it
The polyesters were afterwards examined by differential scan-
ning calorimetry: the influence of molecular weight (M
w
) on the
glass transition and melting phenomena of the samples under
investigation can be excluded, having PBT, PBHEBST and
w
PBT/BHEBST polyesters of high and similar M values. It is well
established that the melting behavior of a polymer is affected by its
previous thermal history and therefore, in order to provide the
O
S
O
O
C
O
C
O
C
O
C
CH2 CH2
O
O
CH2 CH2
O
O
CH2 CH2 CH2 CH2
m
O
O
n-m
S
T
B
T
Fig. 4. Repeating units in PBT/BHEBST polymers.

908
N. Lotti et al. / Polymer 52 (2011) 904e911
Fig. 5. Expansion of 13C NMR of PBT/BHEBST50.
same heat treatment to all the samples investigated, prior to
thermal analysis each sample was kept at room temperature for
about 4 weeks. The DSC traces of the so-treated samples are
reported in Fig. 8 and the results obtained in Table 2.
broader, suggesting the presence of a larger distribution of crys-
tallites with different degree of perfection.
The melting point reduction can be examined in order to esti-
ꢀ
mate the equilibrium melting temperature T
m
and the equilibrium
ꢀ
In all cases, a glass transition and a melting endotherm are
evident, with exception of PBHEBST, whose DSC trace is charac-
terized only by an endothermal baseline deviation associated with
the glass transition phenomenon. First of all, it has to be empha-
sized that the phase behavior of the two parent homopolymers is
opposite: PBT is semicrystalline, whereas PBHEBST is completely
amorphous. As it is well known, the crystallization capacity of
a polymer is correlated with several factors, such as the symmetry
and the flexibility of the polymeric chain. In PBHEBST the presence
of the very rigid bisphenol S group causes a drastic reduction of
flexibility, hindering completely the ability of crystallizing of this
polymer.
As far as the melting phenomenon is concerned, the calorimetric
results indicate that an increase in the amount of the co-monomer
BHEBST leads to a reduction in both of the melting temperature and
the heat of fusion (see Fig. 9a), as usually found in random copol-
ymers with only one co-unit taking part in the crystallization
process. Furthermore, in the copolymers the endotherm region is
heat of fusion
DH
m
of crystallizable unit. The Flory’s treatment [17],
commonly used in the past and derived assuming that the fusion
concerns the disappearance of long sequences of crystallizable
units, underestimates the melting point depression of random
copolymers. As a matter of fact, the concentration of long
sequences decreases with increasing co-unit content, and the
m
experimental T values are consequently lower than those postu-
lated by the theory. On the contrary, the equation proposed by Baur
[18] takes into account the effect of sequence length of crystalliz-
able units which can crystallize only when their length corresponds
to the crystal thickness. Baur’s equation is given by [18]:
À
Á
o
o
1=T
m
¼ 1=Tm ꢁ R=
D
H
ðlnx ꢁ 2x ð1 ꢁ x ÞÞ
(1)
m
C
C
C
where Tm is the melting temperature of a random copolymer with
ꢀ
molar fraction xC of crystallizable co-monomer C, Tm is the equi-
librium melting temperature of the homopolymer (in this case PBT)
and R is the gas constant. On the basis of equation (1), the T s were
m
Fig. 6. 1H NMR of PBHEBST polymer.

N. Lotti et al. / Polymer 52 (2011) 904e911
909
ꢀ
Fig. 7. TGA curves of (1) PBT, (2) PBT/BHEBST5, (3) PBT/BHEBST10, (4) PBT/BHEBST20, (5) PBT/BHEBST50, (6) PBHEBST at 10 C/min in: (a) nitrogen; (b) air.
reciprocally plotted against e[lnx
C
ꢁ2x
C
(1ꢁx
C
)] in Fig. 9b and the
equilibrium melting temperature and the heat of fusion for
the completely crystalline PBT were extrapolated. As can be noted,
the plot shows a good linearity and this result can be considered
a further proof of the random nature of the copolymers investigated
aswellasoftheexclusionofthe co-unitsfromthe crystallinelatticeof
ꢀ
ꢀ
PBT.The T
m
and the
DH
m
values estimated on the basis of Baur’s
ꢀ
equationwere foundtobe228 Cand122J/g, ingood agreementwith
the ones reported in the literature for PBT homopolymer [19e21].
It is known that a partially crystalline material usually exhibits
a different glass transition behavior than the analogous completely
amorphous material. In fact, although some conflicting results are
reported in the literature [22], crystallinity usually acts like cross-
g
linking and raises T through its restrictive effect on segmental
motion of amorphous polymer chains. Therefore, in order to study
the influence of the chemical structure on the glass transition of
random copolymers, the phenomenon should be examined in the
total absence of crystallinity. Rapid cooling (quenching) from the
melt is the method commonly used to prevent crystallization and
to obtain polymers in a completely amorphous condition. In this
view, all the samples under investigation were subjected to such
thermal treatment (see the Experimental section for the details).
The calorimetric curves of the polymers so treated are reported in
Fig. 10. In all cases, except for PBHEBST and PBT/BHEBST50, the
calorimetric traces show a glass transition followed by an
exothermal “cold crystallization” peak and a melting endotherm at
higher temperature. As far as PBT and the PBT/BHEBST copolymers
poor in BHEBST units (up to 5 mol %) are concerned, the enthalpy
associated with the crystallization phenomenon is lower than that
of the fusion endotherm, indicating that these samples cannot be
frozen into a completely amorphous state by quenching. Never-
g
theless, a portion of amorphous material, once T is exceeded,
acquires enough mobility to rearrange and crystallize. The DSC
curves of such samples are therefore typical of partially crystalline
Fig. 8. Calorimetric curves of PBT, PBHEBST and their random copolymers (1st scan).
Table 2
Thermal characterization data of PBT/BHEBST random copolymers (enthalpy has not been normalized by composition).
a
ꢀ
b
ꢀ
c
ꢀ
Sample
T
10% ( C)
T
50% ( C)
1st scan
2nd scan
Tcc ( C)
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
m
T
m
( C)
D
H
m
(J/g)
T
g
( C)
D
c
p
(J/g C)
T
c
( C)
D
H
c
(J/g)
T
( C)
m
DH (J/g)
PBT
378
381
384
389
394
411
403
406
410
424
449
470
224
216
207
188
150
e
57
54
45
34
19
e
41
46
50
61
94
0.085
0.226
0.313
0.327
0.331
0.325
47
59
80
119
e
4
27
34
26
e
223
213
204
186
e
56
52
35
26
e
196
183
167
144
e
PBT/BHEBST5
PBT/BHEBST10
PBT/BHEBST20
PBT/BHEBST50
PBHEBST
122
e
e
e
e
e
a
T10% temperature at which the sample loose 10% of its weight.
T50% temperature at which the sample loose 50% of its weight.
Tcc temperature corresponding to the maximum of the exothermic peak in the DSC cooling-curve.
b
c

910
N. Lotti et al. / Polymer 52 (2011) 904e911
2
30
10
90
70
50
30
0.0025
a
b
0
.0024
.0023
.0022
.0021
.0020
2
0
1
0
1
0
1
0
1
0.45
0.60
0.75
0.90
0.0
0.4
-[ln x
0.8
1.2
PBT molar fraction
C
- 2x (1-x )]
C
C
m m
Fig. 9. (a) Melting temperatures (T ) as a function of composition; (b) 1/T -composition plot according to Baur’s equation.
polymers. As a matter of fact, it is well known that PBT cannot be
frozen in a completely amorphous glassy state, due to its high
crystallization rate [23].
As can be seen in Fig. 11 and from the data collected in Table 2,
the glass transition temperature is influenced by the amount of
BHEBST units in the polymer chains. The T values associated with
g
In the case of PBT/BHEBST10 and PBT/BHEBST20, the enthalpy of
crystallization very well compares with the corresponding heat of
fusion, indicating that the polymers were completely amorphous.
As regards the calorimetric curves of pure PBHEBST and copolymer
containing 50 mol % of BHEBST units, only an intense endothermal
baseline deviation associated with the glass transition is observed.
The phase behavior of copolymers was found to depend on
composition: in fact, samples able to crystallize during the scan are
obtained exclusively at high PBT content.
the glass transition are plotted in Fig. 11 as a function of BT unit
content together with those of random PBT/BHEEBT copolymers
previously investigated [5]. As far as the trend of the glass transition
temperature with the composition of PBT/BHEBST copolymers is
g
concerned, one can observe that T values increase as BHEBST unit
content is increased, due to the stiffening effect of the moieties
deriving from bisphenol S in the polymeric chain. Comparing the
data of the two copolymeric systems, it can be evicted that the
increment of T
to the higher rigidity of bisphenol S compared to bisphenol A.
In amorphous random copolymers, T is usually a monotonic fun-
ction of composition [24] and the commonest relationship used to
predictT asafunctionof comonomercontentisthe Foxequation[25]:
g
is more consistent for PBT/BHEBST copolymers, due
g
g
1=T
g
¼ w =T þ w =T
(2)
I
gI
II gII
1
1
1
40
20
00
8
6
4
0
0
0
0
20
40
60
80
100
xBT (%)
Fig. 11. Composition dependence of T
(from ref. 5) random copolymers; theoretical curve of T
the basis of Fox equation.
g
for (A) PBT/BHEBST and (C) PBT/BHEEBT
Fig. 10. Calorimetric curves of PBT, PBHEBST and PBT/BHEBST random copolymers
after melt quenching.
g
vs. composition calculated on

N. Lotti et al. / Polymer 52 (2011) 904e911
911
as obstacles in the regular packing of polymer chains. This behavior
can be explained taking into account that the presence of bisphenol
S groups increases the difficulty of close packing to form a crystal-
line array.
2
1
1
1
00
80
60
40
4. Conclusions
0
10
20
30
40
xBHEEBTS (mol%)
Our research showed that the insertion of bisphenol S units
PBT
inside terephthalate polyesters backbone provides an efficient
route to vary in a proper way the thermal properties. Copolymer
composition turns out to be an important parameter in deter-
mining the characteristics of the samples under investigation.
In particular, such a chemical modification induces:
-
-
a significant increase of the thermal stability, which widens the
processability window;
PBT/BHEBST5
a decrease of the melting temperature, due to a limit to the
lamellar thickness as a result of the presence of BHEBST chain
segments. In fact, in the presence of these ones, the envisaged
mode of chain attachment does affect the driving force term as
well as the barrier, preventing the formation of very thick
crystallites;
PBT/BHEBST10
PBTBHEBST20
-
-
a significant increase of the glass transition temperature, which
in some copolymers exceeds 100 C value;
a decrease of the crystallization rate, due to the rejection of the
stiff BHEBST co-units in the amorphous phase.
ꢀ
8
0
120
160
T (°C)
200
240
Fig. 12. DSCcrystallizationexothermsofPBTandPBT/BHEBSTrandomcopolymerscooled
from the melt at 10 C/min. In the inset: Tcc as a function BHEBST unit content.
References
ꢀ
[1] Pilati F. Polyesters. In: Allen G, editor. Comprehensive polymer science, vol. 5.
Oxford: Pergamon Press; 1989. Chapter 17.
2] Freitag D, Fengler G, Morbitzer L. Angew Chem 1991;30:1598e610.
3] Xiao C, Yee AF. Macromolecules 1992;25:6800e9.
where TgI and TgII are the glass transition temperatures of the pure
homopolymers and w and wII the respective weight fractions. As
I
[
[
showninFig.11, theFoxequationaccuratelyfits the experimentaldata,
[4] Berti C, Colonna M, Fiorini M, Lorenzetti C, Marchese P. Macromol Mat Eng
004;289(1):49e55.
5] Munari A, Finelli L, Lotti N, Berti C, Colonna M, Lorenzetti C. Polymer 2003;44
5):1409e20.
[6] Jansen MAG, Goossens JGP, De Wit G, Bailly C, Koning CE. Macromolecules
005;38(7):2659e64.
2
using for PBHEBST the glass transition temperature experimentally
[
ꢀ
measured by us, and fixing for PBT the value of 37 C reported in the
(
literature [23]. Its applicability can be considered an evidence of the
random nature of the copolymers under investigation.
2
[
[
7] Hait SB, Sivaram S. Polym Intl 1998;47:439e44.
8] Oku A, Tanaka S, Hata S. Polymer 2000;41:6749e53.
Lastly, in order to get preliminary results on the effect of
composition on the crystallization rate of PBT, non-isothermal
experiments were carried out, subjecting the samples to the
thermal treatment described in the Experimental Section. It is
worth remembering that the half-time of primary crystallization in
isothermal experiments correlates with the temperature corre-
sponding to the maximum of the crystallization peak in non-
isothermal experiments (Tcc) [26], being this latter more easily
obtainable. The exothermic crystallization peaks of the samples
under investigation are reported in Fig. 12, with the exception of
pure PBHEBST and PBT/BHEBST50 copolymer, these samples being
not able to crystallize even though cooled from the melt at very low
[9] Hu LC, Oku A, Yamada E. J Appl Polym Sci 1997;63:595e601.
10] Hu LC, Oku A, Yamada E. Polym J 1997;29:708e12.
11] Hu LC, Oku A, Yamada E. Polymer 1998;39:3841e5.
12] Turner RS. J Polym Sci 2004;42:5847e52.
[13] Turner RS, Sublett BJ. USPatent 6120889 2000.
14] Banach TE, Berti C, Colonna M, Fiorini M, Marianucci E, Messori M, et al.
Polymer 2001;42(18):7511e6.
15] Montaudo G, Pugliesi C. In: Grassie N, editor. Developments in polymer
degradation, vol. 7. London: Applied Science; 1987. p. 35.
[16] Garozzo D, Giuffrida M, Montaudo G. Macromolecules 1986;19:1643e9.
[
[
[
[
[
[
[
[
17] Flory PJ. Trans Faraday Soc 1955;51:848e57.
18] Baur H, Baltorowicz M. Makromol Chem 1966;98:297e301.
19] Runt J, Miley DM, Zhang X, Kallagher KP, McFeaters K. Fishbum J Macro-
molecules 1992;25(7):1929e34.
[20] Huo PP, Cebe P. Macromolecules 1993;26(12):3127e30.
21] Righetti MC. Munari A Macromol Chem Phys 1997;198(2):363e78.
22] Boyer RF. Rub Chem Tech 1963;36(5):1303e421.
[23] Bicerano J. Prediction of polymer properties. New York: Marcel Dekker; 1996.
24] Eisenberg A. Physical properties of polymers. Washington: ACS; 1984.
25] Fox G. Bull Am Phys Soc 1956;1:123.
26] Legras R, Dekoninck JM, Vanzieleghem A, Mercier JP, Nield E. Polymer
1986;27(1):109e17.
ꢀ
rate (1 C/min). From the inset, where the Tcc values are reported as
[
[
a function of composition, it can be observed that the temperature
corresponding to the maximum of the exothermal crystallization
peak regularly decrease as the BHEBST unit content is increased.
This trend indicates a decrement of the overall crystallization
rate of PBT, due to the presence of the co-units (BHEBST), which act
[
[
[