Partially renewable copolyesters prepared from acetalized d -glucitol by solid-state modification of poly(butylene terephthalate)

Article abstract of DOI:10.1002/pola.26985

The backbone of poly(butylene terephthalate) (PBT) was modified with 2,4:3,5-di-O-methylene-D-glucitol (Glux) using solid-state modification (SSM). The obtained copolyesters proved to have a non-random overall chemical microstructure. The thermal properties of these semicrystalline, block-like, Glux-based materials were extraordinary, showing higher melting points, and glass transition temperatures compared with other sugar-based copolyesters prepared by SSM. These remarkable thermal properties were a direct result of the inherently rigid structure of Glux and the relatively slow randomization of the block-like chemical microstructure of the Glux-based copolyesters in the melt. SSM proved to be a versatile tool for preparing partially biobased copolyesters with superior thermal properties. Copyright

Full text of DOI:10.1002/pola.26985

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
Partially Renewable Copolyesters Prepared from Acetalized D-Glucitol by
Solid-State Modification of Poly(butylene terephthalate)
1
2
2
1
ꢀ
Erik Gubbels, Cristina Lavilla, Antxon Martınez de Ilarduya, Bart A. J. Noordover,
Cor E. Koning,1,3 Sebastian Munoz-Guerra
1Department of Chemical Engineering and Chemistry, Laboratory of Polymer Materials, Eindhoven University of Technology,
2
ꢀ
~
Den Dolech 2, P.O. Box 513, 5600 MB Eindhoven, The Netherlands
2
ꢀ
Departament d’Enginyeria Quımica, Universitat Polite`cnica de Catalunya, ETSEIB, Diagonal 647, 08028 Barcelona, Spain
3DSM Coating Resins, Ceintuurbaan 5, Zwolle, The Netherlands
Correspondence to: B. A. J. Noordover (E-mail: B.A.J.Noordover@tue.nl)
Received 12 August 2013; accepted 15 October 2013; published online 6 November 2013
DOI: 10.1002/pola.26985
ABSTRACT: The backbone of poly(butylene terephthalate) (PBT)
was modified with 2,4:3,5-di-O-methylene-D-glucitol (Glux)
using solid-state modification (SSM). The obtained copolyest-
ers proved to have a non-random overall chemical microstruc-
ture. The thermal properties of these semicrystalline, block-
like, Glux-based materials were extraordinary, showing higher
melting points, and glass transition temperatures compared
with other sugar-based copolyesters prepared by SSM. These
remarkable thermal properties were a direct result of the inher-
ently rigid structure of Glux and the relatively slow randomiza-
tion of the block-like chemical microstructure of the Glux-
based copolyesters in the melt. SSM proved to be a versatile
tool for preparing partially biobased copolyesters with superior
C
V
thermal properties.
2013 Wiley Periodicals, Inc. J. Polym.
Sci., Part A: Polym. Chem. 2014, 52, 164–177
KEYWORDS: bicyclic acetalized alditol; carbohydrate-based poly-
esters; polybutylene terephthalate; remarkable thermal proper-
ties;
solid-state
modification;
polyesters;
solid-state
polymerization, thermal properties; renewable resources
polycondensates containing IS monomer residues have been
prepared by solid-state modification (SSM) of an existing
semicrystalline polyester. However, due to its volatility and
relatively low reactivity the direct incorporation of IS into
the polyester backbone using this technique was not possi-
ble, which made a macrodiol approach necessary.18
INTRODUCTION The development of biobased polymers is
nowadays attracting a great deal of interest, due to the high
price and future depletion of fossil fuel stocks.1–3 Various
groups of biobased monomers can be used in this respect,
but one class has received special attention, namely, the car-
bohydrates. Carbohydrates are one of the most abundant
classes of renewable resources; they are readily available
and are found in a rich variety of chemical structures with
great stereochemical diversity. Nowadays, there is an increas-
ing research drive to obtain bifunctional monomers from
carbohydrate resources for step-growth polymerization,
which is clearly reflected in the recently published litera-
ture.4–7 Especially bifunctional, sugar-based monomers with
a bicyclic structure are of interest, due to their ability to add
stiffness to the polymer backbone, resulting in enhanced Tg
values.8–10 Many examples of polyesters containing 1,4:3,6-
dianhydro-D-glucitol (isosorbide, IS), a bicyclic diol prepared
from the widely available D-glucose, have been described in
the literature.11–14 Even though the two secondary hydroxyl
groups of IS display limited reactivity under conventional
polycondensation conditions,15–17 high molecular weight
SSM19 is based on the same principles as the commonly used
solid-state polycondensation (SSP), which is usually employed
to increase the molecular weight of semicrystalline polyesters
and polyamides by transesterification and transamidation
reactions, respectively.20 Both techniques are carried out at a
ꢀ
reaction temperature just (20–30 C) below the melting tem-
perature of the crystalline phase of the polymer, and they
take advantage of the molecular mobility in the amorphous
phase of the polymer above Tg. During SSM, only chain seg-
ments present in the mobile amorphous fraction are modified
with the (macro)monomer. Mobility restrictions prevent the
crystalline phase as well as the rigid amorphous fraction
(RAF) from taking part in transesterification and transamida-
tion reactions occurring during the SSM process.18,19,21–23
Additional Supporting Information may be found in the online version of this article.
C
V
2013 Wiley Periodicals, Inc.
164
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
Poly(butylene terephthalate) (PBT) is a well-known semi-
crystalline aromatic polyester, which is used for a wide vari-
ety of engineering plastic applications and is often applied in
injection molding processes.24 However, its relatively low
glass transition temperature (Tg 5 45 ꢀC) precludes the use
of PBT in higher temperature applications, such as contain-
ers for pressurized gaseous beverages. Most of the
approaches aimed at increasing the Tg of PBT, such as reac-
tive blending with more rigid polycondensates and melt
copolymerization with rigid comonomers, are successful in
increasing the glass transition temperature of PBT but at the
same time severely compromise the melting and crystalliza-
tion behavior.25–28 However, with SSM it is possible to exclu-
sively modify the properties related to the amorphous phase
(i.e. glass transition temperature) while to a large extent
retaining the crystallization behavior of the parent polyester.
In particular, attention has been given to the modification of
PBT with sugar-derived bicyclic diols, since these lead to
copolyesters, partially based on renewable resources, exhibit-
ing remarkable thermal properties. Besides the incorporation
by transesterification of a macrodiol derived from IS,18
recently the diols 2,3:4,5-di-O-methylene-galactitol (Galx)
and 2,4:3,5-di-O-methylene-D-mannitol (Manx) have been
incorporated into PBT by SSM.29 Unlike IS, which is a cyclic
diether obtained by the dehydration of D-glucose, these two
symmetric compounds are cyclic diacetals derived from D-
galactose and D-mannose, respectively.
tope Laboratories. Solvents used for purification and charac-
terization were purchased from Biosolve (Valkenswaard, the
Netherlands) and they all were of either technical or high-
purity grade. All chemicals were used as received unless
stated otherwise.
General Methods
1H-NMR and 13C-NMR spectra were recorded on a Bruker
AMX-300 spectrometer at 25.0 ꢀC operating at 300.1 and
75.5 MHz, respectively. Samples were dissolved in a mixture
of deuterated chloroform and TFA-d1 (9:1), and spectra were
internally referenced to tetramethylsilane. About 10 and 50
mg of sample dissolved in 1 mL of the solvent mixture were
1
used for H-NMR and 13C-NMR, respectively. Sixty-four scans
1
were acquired for H and 1,000–10,000 for 13C with 32 and
64-K data points as well as relaxation delays of 1 and 2 s,
respectively. 2D 1H-1H homonuclear (COSY) and 13C-1H het-
eronuclear shift correlation (HETCOR) spectra were respec-
tively recorded by means of the cosy and hxco pulse
sequences implemented in the Bruker NMR instrument pack-
age. For conformational studiꢀes, 1H-NMR spectra were
recorded within the 60 to 260 C temperature range in the
NMR spectrometer equipped with a variable-temperature
ꢀ
unit. Temperatures were selected at 20 C intervals. For each
temperature, the sample was held at that temperature for 10
min to reach thermal equilibrium. 2D NOE spectra (NOESY)
were recorded at a fixed temperature (T 5 298.1 K) with a
standard pulse sequence (noesytp) with a mixing time of 2 s
over a sweep width of 1397 Hz using 2048 data points in
the t2 dimension and 256 increments in the t1 dimension.
The repetition delay was 2 s, and 256 scans were collected
for each t1 increment. 1H-NMR simulated spectra were
obtained with SpinWorks 3.1.8 program (Kirk Marat, Univer-
sity of Manitoba). Size-exclusion chromatography (SEC) was
performed on a system equipped with a Waters 1515 Iso-
cratic HPLC pump, a Waters 2414 refractive index detector
working at 40 ꢀC, a Waters 2707 autosampler, and a PSS
PFG guard column followed by a 2PFG-linear-XL (7 lm, 8300
mm) columns in series at 40 ꢀC. HFIP with potassium tri-
fluoroacetate (3 g L21) was used as eluent at a flow rate of
0.8 mL min21. The molecular weights were calculated
against poly(methyl methacrylate) standards (Polymer Labo-
ratories, Mp 5 580 Da to Mp 5 7.1 3ꢁ106 Da). Differential
scanning calorimetry (DSC) was performed using a DSC
Q100 from TA-Instruments. DSC data were obtained from 4
to 6 mg samples at heating/cooling rates of 10 ꢀC min21
under a nitrogen flow. Indium was used as standard for tem-
perature and enthalpy calibration. The glass transition tem-
peratures were determined from the third heating run at a
heating rate of 20 ꢀC min21. Thermogravimetric analyses
(TGA) for polymers were performed under a nitrogen flow
In this work, the bicyclic acetalized diol 2,4:3,5-di-O-methyl-
ene-D-glucitol, abbreviated as Glux, is incorporated into PBT
by transesterification in the solid-state. Glux is a D-glucose-
derived, nonsymmetric bicyclic diol such as IS, but affords
the advantage of being more reactive because of its two pri-
mary hydroxyl groups. Given the structural similarities
between Glux and its diastereoisomer Manx, a conforma-
tional analysis will be performed to elucidate the differences
between both sugar-based monomers in this regard. The
synthesis, chemical microstructure, and thermal properties
of the new Glux-based copolyesters will be studied and dis-
cussed in detail. We will show that Glux allows a much eas-
ier synthesis of D-glucose-derived copolyesters compared
with IS and that these partially biomass-based copolyesters
display outstanding properties due to their unique chemical
microstructure.
EXPERIMENTAL
Materials
2,4:3,5-Di-O-methylene-D-glucitol (Glux) was prepared by
acetalization of 1,5-D-gluconolactone with paraformaldehyde,
followed by esterification with methanol and subsequent
reduction as described elsewere.30 PBT was a kind gift from
DSM (Geleen, the Netherlands). The reagents benzoyl chlo-
ride (991%) and dimethyl terephthalate (DMT, 991%), and
the transesterification catalyst dibutyl tin oxide (DBTO, 98%)
were purchased from Sigma-Aldrich. Deuterated chloroform
(CDCl3, 99.8% atom D) and deuterated trifluoroacetic acid
(TFA-d1, 99% atom D) were obtained from Cambridge Iso-
at a heating rate of 10 C min21, within a temperature range
ꢀ
of 25–600 ꢀC, using a TA Q500 equipment. Sample weights
of about 5 mg were used in these experiments. TGA for
monomers were performed isothermally at 160 ꢀC for 6 h
under a nitrogen flow, and the volatiles were analyzed by
1H-NMR. Sample weights of about 15 mg were used in these
experiments. X-ray diffraction (XRD) measurements were
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
165

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
performed on a Rigaku Geigerflex Bragg-Brentano Powder
Diffractometer using Cu radiation, wavelength 1.54056 Å, at
40 kV and 30 mA. The scans were performed with 0.02ꢀ
steps in 2-theta and a dwell time of 3–6 s in the 2-theta
range from 10ꢀ until 35ꢀ. The analyses were performed on
the crude copolyester samples.
of 1,4-butanediol and Glux, respectively, in the resulting
copolyester.
Synthesis of Glux-Derived Model Compound and
Reactivity Studies
Synthesis of 1,6-di-O-benzoyl-2,4:3,5-di-O-methylene-D-
glucitol
A mixture of 1.7 g of benzoyl chloride (12.1 mmol) and 1.0 g
(4.9 mmol) of 2,4:3,5-di-O-methylene-D-glucitol in dry pyri-
dine (20 mL) was refluxed at 80 C for 30 min. After cooling
down, the benzoic acid was filtered. Na2CO3 was added to
the remaining solution until 1,6-di-O-benzoyl-2,4:3,5-di-O-
methylene-D-glucitol precipitated, which was filtered and
washed with water and diethyl ether.
The Solution Preparation of Physical PBT/Glux Mixtures
Physical mixtures of PBT and 2,4:3,5-di-O-methylene-D-gluci-
tol (Glux) were prepared from solution, using a common sol-
vent approach. Prior to the preparation of these mixtures,
the PBT was dissolved in a 3:1 (v/v) mixture of chloroform
and 1,1,1,3,3,3-hexafluorisopropanol (HFIP) and precipitated
in an excess of ethanol in order to remove any titanium-
based catalysts residues. Predetermined amounts of the pre-
ꢀ
cipitated PBT (Mn 5 23.3 kg mol21 and Mw 5 47.1 kg mol21
,
Reaction of 2,4:3,5-di-O-methylene-D-glucitol with TFA-d1
2,4:3,5-Di-O-methylene-D-glucitol (20 mg) was placed in an
NMR tube containing 1 mL of TFA-d1. The reaction was car-
determined by SEC), Glux and transesterification catalyst
DBTO (0.36% molar with respect to PBT) were dissolved in
HFIP. As an example, the preparation of the physical mixture
for the PB84Glux16T copolyester (83.3/16.7, 1,4-butanediol/
Glux molar feed ratio) is described; 1.3642 g PBT (6.20
mmol), 0.2552 g Glux (1.24 mmol), and 0.0056 g DBTO
(0.022 mmol) were dissolved in 10 mL HFIP at 55 ꢀC. After
the complete dissolution of all compounds, the HFIP was dis-
tilled off. As soon as the material started to precipitate, a
vacuum of 190–300 mbar was applied to facilitate a more
efficient removal of the HFIP. Finally, the obtained physical
mixture was dried under vacuum for 24 h, cooled in liquid
nitrogen, and subsequently ground into powder using an IKA
A11 Basic Analytical mill. ꢀThis powder was subsequently
dried under vacuum at 20 C for a period of 24 h. Different
molar ratios of PBT and Glux were used to obtain copolyest-
ers with varying compositions.
ried out at 25 C and was followed by 13C-NMR.
ꢀ
Reaction of 2,4:3,5-di-O-methylene-D-glucitol with DMT
The reaction was performed in a three-necked, cylindrical-
bottom flask equipped with a mechanical stirrer, a nitrogen
inlet, and a distillation outlet. A 2:1 molar excess of 2,4:3,5-
di-O-methylene-D-glucitol with respect to DMT was used, and
DBTO (0.36% molar respect to DMT) was the catalyst of
choice. The apparatus was purged with nitrogen at room
temperature to remove traces of air. The transesterification
reaction was carried out under a low nitrogen flow at 160
ꢀC, and was followed by 13C-NMR.
RESULTS AND DISCUSSION
Conformational Study of 2,4:3,5-Di-O-methylene-D-
glucitol Monomer
SSM of PBT
To determine the preferred conformation of the 2,4:3,5-di-O-
methylene-D-glucitol (Glux) unit, and to study the possible
differences of the Glux monomer with respect to its diaster-
eoisomer 2,4:3,5-di-O-methylene-D-mannitol (Manx) in terms
of reactivity, a conformational analysis of 1,6-di-O-benzoyl-
2,4:3,5-di-O-methylene-D-glucitol was undertaken. This die-
ster is thought to be an adequate model compound for the
PBT/Glux-based copolyesters. In principle, fused six-
membered bicyclic acetals could assume two conformations
containing a cis junction, one being referred to as “O-inside”
(I) and another being referred to as “H-inside” (II),31,32
which are depicted in Scheme 1 for the Glux unit.
The SSM of PBT was performed in a reactor comprising a
glass tube (inner diameter 5 2.4 cm) with a sintered glass
plate at the bottom. A heat exchange glass coil (inner diame-
ter 5 0.5 mm) surrounded the reactor and entered the inner
glass tube at the bottom just below the glass plate. The
nitrogen gas was heated by passing through this coil prior to
entering the reactor, which was immersed in an oil bath kept
at 160–175 ꢀC. The nitrogen flow was controlled by a flow-
meter. Typically, 0.4 g of a powdered PBT/Glux mixture was
placed on the sintered glass plate. The powder was fixed by
addition of glass pearls (diameter 5 2 mm) on top of the
powder, and the reactor was purged with a nitrogen flow of
0.5 L min21 during 30 min. Reactions were left to proceed
at 160 ꢀC for 6 h and at 170 ꢀC for 15 h under a 0.5 L
min21 nitrogen flow, and at 175 ꢀC for 24 h under a 1 L
min21 nitrogen flow. The molecular weights were followed
as function of reaction time using SEC. After completion of
the reaction, the reactor was gradually cooled down to room
temperature by removing the heat source from the reactor
and by continuing the purging of the reactor with nitrogen,
and the obtained polymer was dried under vacuum. The
Glux copolyesters obtained via SSM are abbreviated as
PBxGluxyT, where x and y are the mole percentages (mol %)
To assess the relative preference of Glux for one of these
two conformations, a variable temperature proton nuclear
magnetic resonance (1H-NMR) spectroscopic study was car-
ꢀ
ried out in the 260 to 60 C range. In addition to this study,
nuclear Overhauser effect (2D NOESY) spectrum was
recorded at room temperature. When both conformers I and
II are present, two signals for each proton should appear in
1
the H-NMR spectra at low temperatures. Since, under those
experimental conditions, no interconversion is expected to
take place, both conformations would be observed if present.
However, only single proton signals are observed along the
166
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
SCHEME 1 Two possible conformations of the Glux unit. I: “O-inside”; II: “H-inside.”
whole temperature range and the only detected change is an
upfield shift of the signals with heating (Fig. 1).
results are in good agreement with the angles determined
for conformer I, but differ significantly from those deter-
mined for conformer II. In the case that conformer II was
the preferred one, characterized by U 5 180ꢀ, the J45 value
would be around 9 Hz.
It can be concluded that no conformational interconversion
occurred and that only one conformer was present along the
whole range of temperatures. The analysis of the coupling
constants obtained from the 1H-NMR spectra indicates con-
former I to be the preferred conformation for Glux. This was
already suggested by Mills32 and Bourne et al.,33 based on
the steric hindrance reasons between the protons and the R1
group. In fact, the spin-spin coupling constants for this sys-
tem (see Supporting Information Fig. S1) provided by NMR
spin simulation (SpinWorks 3) are:
The identification of conformer I being the preferred of the
Glux residues is supported by data obtained from 2D NOESY
experiments (Fig. 2). As expected for conformation I, strong
4
2
3
5
NOE peaks were observed for H with H, H, H, and one of
the acetal methylene protons. However, in conformation II
4H would only be in the immediate vicinity of 3H, thus this
4
NOE peak would be the only one detected for this H, which
is not the case. The valuable conclusion, which can be
derived from this conformational study is that Glux is
J
J
J
2351:5 Hz
3451:4 Hz
4551:2 Hz
Applying these coupling constants to the Karplus equation
derived for alditols,34,35 the torsional angles x, w, and U
could be estimated. The x, w, and U values resulting from
this estimation were 69ꢀ, 70ꢀ, and 72ꢀ, respectively. These
1
1H-NMR spectra of 1,6-di-O-benzoyl-2,4:3,5-di-O-
FIGURE 2 NOESY spectrum of 1,6-di-O-benzoyl-2,4:3,5-di-O-
methylene-D-glucitol in CDCl3 at 25 ꢀC with assigned NOE
peaks. R1, R6: CH2OCOPh. [Color figure can be viewed in the
online issue, which is available at wileyonlinelibrary.com.]
FIGURE
methylene-D-glucitol in CDCl3 at the indicated temperatures,
including the assignment of the signals corresponding to the
indications in Scheme 1.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
167

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
SCHEME 2 Preferred conformation of the Glux and Manx units.7
present in conformation I (“O-inside,” Scheme 1) and that
the two R-groups (e.g., methylol group in the Glux monomer)
have different orientations in this conformation, namely, one
axial- (R6) and one equatorial-oriented (R1) functional group.
this case. These two model experiments proved that the
equatorially oriented hydroxyl groups are more reactive in
(trans)esterification reactions. Since Manx has two equatori-
ally oriented hydroxyl groups, it is expected that the incorpo-
ration of Glux into the PBT backbone will take more time
compared with Manx.
For comparison, the preferred conformation of Manx7 is
depicted in Scheme 2. The Manx monomer has a c2 symme-
try and, in contrast to Glux, its preferred conformation is the
“H-inside” with two equatorially oriented hydroxyl groups.
Polymer Synthesis
A series of poly(butylene terephthalate) (PBT)/2,4:3,5-di-O-
methylene-D-glucitol (Glux)-based copolyesters with varying
Glux content was successfully prepared using SSM
(Scheme 3). Typically, SSM is performed at 20–30 ꢀC below
the melting point of the semicrystalline polymer to allow
for sufficient molecular mobility and condensate diffu-
sion,18,22,23,38 the condensate being 1,4-butanediol which is
replaced in the polymer backbone by the less volatile Glux
units by transesterification reactions. However, to minimize
the volatilization of the Glux monomer, lower reaction tem-
peratures were used in this study. DBTO was the catalyst of
choice, because of its high activity in SSM reactions con-
ducted at relatively low temperatures.29 The initial reaction
conditions (160 C, 0.5 L min N2) allowed for the gradual
incorporation of the sugar-based comonomer into the back-
bone of PBT. Once a significant part of the Glux monomer
was incorporated into the polyester chain by alcoholysis, the
reaction temperature and nitrogen flow were augmented
(170–175 ꢀC, 1 L min21 N2) to obtain copolyesters of suffi-
ciently high molecular weight by postcondensation.
This difference in the orientation of the functional groups
could lead to a difference in their reactivity. Reactivity differ-
ences between the two reactive groups of several Glux deriv-
atives have been reported in the literature. One chlorine
atom in 1,6-dichloro-1,6-dideoxy-2,4:3,5-di-O-methylene-D-
glucitol was more reactive than the other in the replacement
by iodine or in elimination by treatment with metallic
sodium.36 Also one bromine atom in 1,6-dibromo-1,6-
dideoxy-2,4:3,5-di-O-benzylidene-D-glucitol was more readily
replaced by iodine than the other.37 To study if during a pol-
ycondensation reaction
a reactivity difference could be
detected, Glux was esterified using TFA-d1 or transesterified
with DMT, and the reactions were followed by 13C-NMR.
Illustrative spectra are presented in Supporting Information
Figures S2 and S3, which show a clear difference in the reac-
tivity of the two hydroxyl groups of Glux in such reactions.
In the esterification with TFA-d1 at 25 ꢀC, the equatorial
hydroxyl group (R1) reacted more readily than the axial (R6),
the reactivity ratio (axial:equatorial) being 1:1.7. The reac-
21
ꢀ
ꢀ
tion of Glux with DMT at 160 C showed a similar result, the
equatorial hydroxyl group was transesterified more readily,
although the reactivity ratio (axial:equatorial) was 1:1.2 in
The molecular weight and polydispersity were followed as a
function of the SSM reaction time (tssm) using size exclusion
SCHEME 3 A schematic representation of the SSM of PBT with 2,4:3,5-di-O-methylene-D-glucitol.
168
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
-
weight increase, the D increased because of the chain recom-
-
bination taking place. Eventually, the increase in D leveled
off at a value of 2.4, which is higher than the expected value
of 2 for polycondensation reactions. This increase is most
probably due to the deviation from the Flory distribution.
Interestingly, the molecular weight build-up took a signifi-
cant longer time compared with the chain scission reaction.
Consequently, the total reaction time for the preparation of
Glux-based copolyesters is significantly longer compared
with the reaction times necessary to incorporate similar
amounts of 2,3:4,5-di-O-methylene-galactitol (Galx) or
2,4:3,5-di-O-methylene-D-mannitol (Manx) under comparable
SSM reaction conditions.29 The peculiar differences in reac-
tion times can be explained by the differences in the molecu-
lar orientation and the melting point of the comonomers.
Schematic representations of the chemical structures of these
sugar-based comonomers, including IS for comparison, are
depicted in Scheme 4.
The nonfused bicyclic structure of Galx is more flexible than
the fused bicyclic structure of Glux, which could facilitate the
higher reaction rate exhibited by Galx. Moreover, Manx is
symmetric with two equatorially oriented methylol groups.
As proven by the conformational study of Glux, this mono-
mer has one axial and one equatorial oriented methylol
group. Additionally, it was shown that the equatorially ori-
ented methylol groups have a higher reactivity toward
(trans)esterification reactions. Due to their higher reactivity,
it is likely that predominantly the equatorially oriented
hydroxyl groups would take part in the chain scission reac-
tions (i.e., alcoholysis), leading to low molecular weight
copolyester chains end-capped with mostly axially oriented
methylol groups of Glux. Therefore, it is expected that the
chain recombination reactions in the case of Glux progress at
a lower rate than for Manx. Besides the difference in orienta-
tion and reactivity of the functional groups, the higher melt-
ing point of the Glux monomer should be considered (Tm,
Glux 5 192–194 ꢀC; Tm, Manx 5 139–140 ꢀC). Before the SSM
FIGURE 3 (a) SEC chromatograms of PB76Glux24T and (b) the
evolution of Mn (•) and Ð (~) as a function of tssm. [Color fig-
ure can be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
chromatography (SEC). Figure 3 depicts the number average
molecular weight trend as a function of tssm as commonly
observed when SSM is used to modify polyester. The pres-
ence of the unreacted Glux monomer [marked as * in Fig.
3(a)] can be clearly observed in the SEC chromatogram of
the initial PBT/Glux mixture (tssm5 0 h), whereas the broad
peak at lower elution times corresponds to the unmodified
PBT. Initially, the molecular weight of the polymer decreased,
indicating that chain scission of the PBT chains by alcoholy-
sis was the dominating process. After 1 h, the molecular
weight reached the lowest value and the strongly reduced
size of the Glux monomer peak indicates that a significant
amount of the Glux monomer had already reacted with the
PBT backbone. Subsequently, the molecular weight of the
PBT/Glux-based copolyesters increased steadily over time,
showing that chain recombination with elimination of 1,4-
butanediol condensation product became the dominating
process. After 45 h of reaction time, the copolyesters
attained Mn values of 17–20 kg mol21 and Mw values of 41–
45 kg mol21 and the reaction was stopped. Such molecular
weight values were desired to allow for a comparison of
their thermal and crystallization properties with other PBT/
sugar-based copolyesters prepared by SSM in our previous
study.29 The D decreased initially because of the formation
-
SCHEME 4 Chemical structure of 2,4:3,5-di-O-methylene-D-glu-
citol (Glux), IS, 2,4:3,5-di-O-methylene-D-mannitol (Manx) and
2,3:4,5-di-O-methylene-galactitol (Galx).
of many low molecular weight species, due to the significant
extent of chain scission, which takes place at these high
comonomer contents. Concomitantly with the molecular
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
169

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
reaction, some of the Glux molecules are present as crystals
within the amorphous phase of the physical mixture (Sup-
porting Information Fig. S4), due to the limited miscibility of
this monomer in the amorphous phase of PBT. When the
SSM reaction starts, the non-crystallized Glux molecules are
directly available for the alcoholysis reaction, but the crystal-
lized Glux first has to dissolve in the amorphous phase. The
latter process is thought to cause the presence of unreacted
Glux in the amorphous phase of PBT for a longer period of
time compared with the Manx case. Consequently, the chain
scission would take place for a significantly longer time,
which could lead to longer reaction times. This is not
expected for Manx or Galx because these monomers melt
below the SSM reaction temperature. Therefore, all the
comonomer is available for incorporation from the beginning
of the SSM reaction. Interestingly, no significant differences
are observed in the time frame within which the chain scis-
sion reaction is dominating when the Manx and Glux reac-
tions are compared. Moreover, the molecular weight
decrease in both reactions is similar. Only the subsequent
recovery of the molecular weight is much slower for Glux.
This indicates that, although both the differences in orienta-
tion and physical behavior may play a role, the reactivity dif-
ference between the two differently oriented functional
groups of Glux has the larger influence. Even though long
reaction times were required, the incorporation of Glux into
the backbone of PBT was successful under the applied condi-
tions, as proven by NMR spectroscopy.
porting information (Supporting Information Figs. S5 and
S6). The resonance of the aromatic protons at 8.2 ppm (sig-
nal e, Fig. 4) is clearly resolved. The signals belonging to the
1,4-butanediol residues are assigned at 4.5 ppm (signal a,
Fig. 4) and 2.0 ppm (signal b, Fig. 4), respectively. The
resonances corresponding to the molecular structure of Glux
are observed in the interval between 3.5 and 5.5 ppm. The
signals relating to the acetal protons (signals m and l) are
clearly visible, indicating that the cyclic structure was
retained during the SSM process.
The carbon nuclear magnetic resonance (13C-NMR) spectrum
of PB76Glux24T and the graphical representation of the corre-
sponding structure are presented in Figure 5, with all the
resonances assigned to the magnetically different carbon
atoms. The eight resonances arising from the Glux moieties
along the backbone of the copolyester are well resolved in
the 60–100 ppm interval. Resonances from the 1,4-butane-
diol residues in the backbone of the copolyester are also
clearly visible (signals a and b, Fig. 5). The resonances
assigned to the terephthalate ring (signals d and e, Fig. 5)
and carbonyl residue (C@O, Fig. 5) are situated above 120
ppm. Interestingly, the resonance of the substituted aromatic
carbon atom has split up into eight resolved signals (inset
Fig. 5). This splitting is caused by the sensitivity of this car-
bon atom to the chemical environment and it will be used
for the determination of the chemical microstructure of the
copolyester backbone, as described below.
The proton nuclear magnetic (1H-NMR) resonance spectrum
of PB76Glux24T and the representation of the copolyester
backbone are presented in Figure 4. All the observed
resonances are assigned to the various magnetically different
protons along the PBT/Glux-based copolyester backbone.
The assignment was corroborated by 2D 1H-1H correlation
(COSY) and 13C-1H heteronuclear shift correlation (HETCOR)
NMR spectroscopy. These spectra can be found in the sup-
The composition of the copolyesters was determined by the
integration of the 1H-NMR signals arising from the 1,4-buta-
nediol and Glux units. The final compositions of the
PBxGluxyT copolyesters proved to be in good agreement with
the initial feed composition (Table 1). Hence, using a rela-
tively low SSM reaction temperature, even though this tem-
perature was below the melting point of the Glux crystals,
was an appropriate approach to obtain PBT/Glux-based
FIGURE 4 1H-NMR spectrum of the PB76Glux24T copolyester recorded in CDCl3/TFA-d1 at room temperature.
170
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
FIGURE 5 13C-NMR spectrum of the PB76Glux24T copolyester
recorded in CDCl3/TFA-d1 at room temperature. The inset is the
zoomed-in region between 133 and 135 ppm.
materials where a significant part of the comonomer units
are incorporated into the copolyester backbone. Similar
results were obtained for the SSM of PBT with other sugar
diols, namely, Galx and Manx.29 However, not all attempts to
modify PBT with a sugar-based diol were successful. Sablong
et al.18 showed that when SSM was used to modify PBT with
IS (IS, Tm 5 60263 ꢀC), no IS was incorporated into the
backbone of the polyester under the applied reaction condi-
tions. Only when a macrodiol approach was used, implying
that the SSM of PBT was performed with a non-volatile pen-
tamer diol containing an IS residue, it became possible to
incorporate IS into PBT by SSM. Two reasons for the suc-
cessful incorporation of Glux can be invoked, namely, its
lower volatility compared with IS and the higher reactivity
exhibited by the primary hydroxyl groups of Glux. It is well
known that primary hydroxyl groups show a higher reactiv-
ity compared with secondary hydroxyl groups when used in
a polycondensation reaction. To study the relative volatilities
of IS and Glux, thermogravimetric (TGA) analyses were per-
formed isothermally at 160 ꢀC. The weight loss of the two
monomers was studied as function of time. A significant dif-
ference between the behavior of IS and Glux was found; it
was observed that IS completely volatilized within 5.5 h,
whereas the Glux sample only lost 5% of its initial weight in
that timeframe (Supporting Information Fig. S7). To prove
that no degradation had taken place and the weight loss was
1
only caused by volatilization, a H-NMR spectrum of the sub-
limed material was performed, which showed only the pres-
ence of the respective monomers. It is probable that both
these factors, viz. reactivity and volatility, play a role deter-
mining the difference in the incorporation behavior of the D-
glucose-derived bicyclic diols Glux and IS.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
171

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
Chemical Microstructure
were found, revealing that SSM produced aregic PBT/Glux-
based copolyesters, with Glux units showing no preferred
orientation. The functional groups of Glux are incorporated
into the backbone of the copolyester irrespective of their ori-
entation, of which the reaction sequence was already dis-
cussed above.
The chemical microstructure of the PBT/Glux-based copo-
lyesters was studied by quantitative 13C-NMR spectroscopy,
making use of the sensitivity of the substituted aromatic
carbon atoms (indicated by dots in Fig. 6) to sequence
effects at the dyad level. This approach has been previously
used to examine the chemical microstructure of several
copolymers.7,21,29,39,40
Based on the dyad integrals results, the average sequence
length (n) and the overall randomness (R) were determined
using the following equations:41,42
An overview of the 13C-NMR resonances in the region of
133–135 ppm obtained for the PBxGluxyT copolyesters is
presented in Figure 6. Since Glux lacks symmetry (Scheme
4), the obtained splitting is much more complex than for
Manx.7,29 Nine different dyads, corresponding to the possible
diol sequences of 1,4-butanediol (B) and Glux (G), are possi-
ble along the PBxGluxyT copolyester backbone. Evidently, the
most intense resonance originates from the dyad assigned to
the BB sequence. The GG sequences show four peaks, due to
the two possible orientations of the Glux residues. For the
same reason, four additional peaks will also result from the
BG/GB sequences. However, because one GG and one BG/GB
resonance overlap, only eight signals can be deduced in the
133–135 ppm interval of the 13C-NMR spectra (Fig. 6). By
integrating the various dyad resonances, their respective
molar fractions (N) were determined. Interestingly, when the
molar fractions of the individual GG or BG/GB resonances
were compared with each other no significant differences
nB5ðNBB 1 0:5 ꢁ ðNBG 1NGB ÞÞ=ð0:5 ꢁ ðNBG 1NGB ÞÞ
nG5ðNGG 1 0:5 ꢁ ðNBG 1NGB ÞÞ=ð0:5 ꢁ ðNBG 1NGB ÞÞ
R5ð1=nBÞ 1 ð1=nGÞ
R-values around unity are indicative of an essentially random
chemical microstructure. When the R-value decreases toward
zero or increases toward two, the chemical microstructure
becomes blocky or alternating, respectively. From Table 1, it
can be seen that the R-values obtained from these copolyest-
ers deviated significantly from unity. These copolyesters
possess a block-like, non-random overall chemical micro-
structure, as reported for other copolyesters obtained by
SSM.29,38 Moreover, there seems to be a relation between the
comonomer content and the obtained overall randomness,
since the R-value decreases with increasing comonomer con-
tent. At this point, it needs to be stated that at the lowest
comonomer contents it was challenging to obtain an accurate
integral value from the dyad signals.
During SSM, large PBT sequences integrated in the crystal-
line phase are unaltered due to mobility restrictions but
they are included in the calculation of the overall chemical
microstructure, since this is determined from solution 13C-
NMR data. It is desirable in this study to evaluate the chemi-
cal microstructure of the amorphous phase only, excluding
the contribution of the crystalline phase. To this end, Jansen
et al. have developed a method to correct the R-values for
the presence of the crystalline phase, yielding the random-
ness of the amorphous phase (Ra).21 Using the enthalpy of
melting (discussed below) of the copolyesters determined by
DSC, the amorphous content of the samples was estimated,
and the integrals obtained from 13C-NMR spectra were cor-
rected accordingly. The Ra values as a function of the Glux
content are presented in Table 1, showing that the chemical
microstructure of the amorphous fraction is practically ran-
dom (Ra 5 0.8–0.9). Moreover, the relation between the Ra
and the composition of the copolyesters (decreasing R-value
with increasing comonomer content) has become less signifi-
cant. Furthermore, it is very well possible that if these
results would be corrected for the presence of the RAF,
which according to Jansen et al. is not participating in the
transesterification reactions, the chemical microstructure of
the mobile amorphous phase would prove to be fully ran-
dom, as was observed in a previous study.21 The main con-
clusion is that the PBxGluxyT copolyesters prepared by SSM
show a block-like, non-random overall chemical microstruc-
ture, caused by PBT homogeneous sequences which make up
FIGURE 6 13C-NMR signals used for the microstructure analy-
sis of PBxGluxyT copolyesters with a graphical representation
of the generic dyad structure. [Color figure can be viewed in
the online issue, which is available at wileyonlinelibrary.com.]
172
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
TABLE 2 Thermal properties of PBxGluxyT copolyesters.
DSC
Coolinga
TGA
First heatinga
Second heatinga
Copolyester T5% (ꢀC) Td (ꢀC) Wd (%) Tg (ꢀC) Tm (ꢀC) DHm (J g21) Tc (ꢀC) DHc (J g21) Tc (ꢀC) DHc (J g21) Tm (ꢀC) DHm (J g21
)
b
c
e
PBT
348
385
400
399
388
400
390
0
3
2
1
9
2
45
225
211
209
208
210
211
60.5f
66.7
62.4
59.6
53.6
52.2
193
179
173
169
163
143
52.0
43.4
42.2
38.9
32.1
23.6
–
–
224
210
206
205
201
194
49.0
39.7
35.1
32.9
28.9
28.1
PB91Glux9T 360
PB89Glux11T 355
PB84Glux16T 349
PB79Glux21T 361
PB76Glux24T 350
66
68
74
85
88
–
–
–
–
–
–
–
–
132
4.1
a
d
e
Melting (Tm) and crystallization (Tc) temperatures, and melting (DHm
)
Remaining weight at 600 ꢀC.
Glass transition temperature taken as the inflection point of the heat-
and crystallization (DHc) enthalpies measured by DSC at heating/cooling
rates of 10 ꢀC min21
.
ing DSC traces recorded at 20 ꢀC min21 during the third heating run.
f
b
Temperature at which 5% weight loss was observed.
Temperature at the maximum degradation rate.
After applying a similar thermal treatment which was used for the
c
copolyesters the DHm increased to 74.4 J g21
.
the crystalline phase of the copolyester during the SSM.
However, the chemical microstructure of the overall amor-
phous phase is practically random.
diffraction (WAXD) was used. The WAXD profiles are
depicted in Figure 7 and the most important Bragg spacings
present therein are listed in Table 3.
When all the WAXD patterns in Figure 7 are studied as a
function of Glux content, it is clear that these patterns are
essentially similar to PBT regarding both spacings and rela-
tive intensities. Clearly, (co-)crystallization of GluxT sequen-
ces does not occur, a result, which was already reported for
other Glux-based semi-aromatic copolyesters.45,46 The crys-
tallinity was estimated from these WAXD diffractograms by
estimating the crystalline area and divide this value by the
total area under the pattern (Table 3). A small decrease of
the crystallinity with increasing Glux content can be noted.
Thus, the crystalline phase is not completely retained during
the SSM process, something which was observed before in
Thermal Properties
The thermal properties of the PBT/Glux-based copolyesters
prepared by SSM have been systematically studied by TGA
and DSC. TGA was used to evaluate the influence of the Glux
residues on the thermal stability of the copolyesters (Sup-
porting Information Fig. S8). The thermal properties derived
from the TGA thermograms, viz. the onset decomposition
temperature (T5%) and the temperature of maximum degra-
dation rate (Td) are listed in Table 2. The degradation takes
ꢀ
place in one step and the residual weight at 600 C is in the
order of 0–9 wt %. The thermal stability of the materials
increases after the incorporation of Glux when these are
compared with the parent PBT. The ability of sugar-based
bicyclic diols to increase the thermal stability of
terephthalate-based polyesters has been reported before in
the literature.7,29,43,44
studies performed by Jansen et al.23 and Lavilla et al.29
A
DSC was used to evaluate the thermal properties of the
series of PBT/Glux-based copolyesters. The obtained results
(Table 2) are discussed and also compared with other PBT/
sugar-based copolyesters prepared by SSM.18,29 The chemical
structures of the sugar-based monomers involved in the
comparative study are depicted in Scheme 4.
All the PBxGluxyT copolyesters are semicrystalline materials
ꢀ
with melting temperatures in the range of 210 C, without a
clear dependence on the composition of the copolyester.
These melting points are related to the PBT crystalline
phase, which did not participate in the SSM reactions
because of mobility restrictions. The copolyesters show
strong melting endotherms as indicated by the high enthalpy
of melting (DHm). To determine the crystallinity of the
PBxGluxyT samples, to elucidate if co-crystallization takes
place and to complement the DSC data, wide-angle X-ray
FIGURE 7 WAXD profiles of the PBxGluxyT copolyesters with
˚
the indication of the dhkl in A.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
173

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
TABLE 3 Powder XRD Data of PBxGluxyT Copolyesters
da (A)
Xc
b
˚
Copolyester
PB91Glux9T
5.6 s
5.6 s
5.6 s
5.6 s
5.6 s
5.1 s
5.1 s
5.1 s
5.1 s
5.1 s
4.3 m
4.3 m
4.3 m
4.3 m
4.3 m
3.8 s
3.8 s
3.8 s
3.8 s
3.8 s
3.6 s
3.6 s
3.6 s
3.6 s
3.6 s
3.1 w
3.1 w
3.1 w
3.1 w
3.1 w
2.9 w
2.9 w
2.9 w
2.9 w
2.9 w
0.65
0.61
0.59
0.53
0.49
PB89Glux11
PB84Glux16
PB79Glux21
PB76Glux24
T
T
T
T
a
b
Bragg spacings measured in powder diffraction patterns for samples
coming directly from synthesis. Intensities visually estimated as follows:
m, medium; s, strong; w, weak.
Crystallinity index calculated as the quotient between crystalline area
and total area. Crystalline and amorphous areas in the XRD pattern
were quantified using PeakFit v4.12 software.
combination of two physical processes (viz., annealing and
swelling of the amorphous phase) are at the basis of this
particular behavior. During SSM, annealing and crystal per-
fectioning take place simultaneously along with the Glux
incorporation process. Annealing converts the RAF into
lamellar PBT crystals and this leads to an increase in the
overall crystallinity of the copolyesters. Furthermore, it is
known that a small quantity of comonomer can lead to an
increase in crystallinity due to the plasticizing effect of the
comonomer on the parent polymer. At the same time, swel-
ling of the amorphous phase with the comonomer can take
place. Because of this swelling, the RAF becomes mobile and
accessible for the modification reaction. Once all the RAF has
become mobile, it is possible that the monomer starts to dis-
solve the PBT crystals too.23 Hence, the swelling of the amor-
phous phase and the annealing are two competitive
processes, which occur simultaneously during the SSM reac-
tion. Nevertheless, it is clear that it is possible to prepare
PBT/Glux-based copolyesters without significantly compro-
mising the degree of crystallinity.
predetermined times and after this treatment their chemical
microstructure and thermal properties were studied (Fig. 8)
using 13C-NMR and DSC, respectively.
The melting temperature decreases as a function of the time
during which the sample was kept in the melt. This decrease
of the Tm was accompanied by an increase of the R-value
from 0.63 to 0.83, showing that some randomization of the
chemical microstructure occurred. This randomization pro-
cess is probably catalyzed by the excellent transesterification
catalyst DBTO, which was added to the reaction mixtures to
facilitate the inꢀcorporation process. Interestingly, even after
30 min at 260 C no fully random Glux-based copolyester is
obtained.
The rate of randomization of the Glux-based copolyesters
was compared with randomization results obtained for other
copolyesters prepared by SSM, containing similar amounts of
sugar-based diols.29 An overview is presented in Figure 9. It
is observed that the Glux-based samples randomized rela-
tively slowly compared with the Galx- and Manx-based copo-
lyesters. As was explained above, differences in the
orientation of the two functional groups of Glux lead to a dif-
ference in their reactivity, which causes the need for
extended reaction times. This argument most probably also
DSC was also used to study the crystallization behavior of
the PBT/Glux-based copolyesters (an example is presented
in Supporting Information Fig. S9). From the data presented
in Table 2, it can be noted that the crystallization tempera-
ture (Tc) and the crystallization enthalpy (DHc) both
decrease with increasing Glux content. The amount of non-
crystallizable Glux segments is increasing, which makes crys-
tallization from the melt more difficult. The PB76Glux24
T
copolyester also exhibits cold crystallization. This behavior
was observed before in copolyesters, which were prepared
by SSM from different sugar-based diols.29
As mentioned above, these materials are semicrystalline and
can recrystallize from the melt. However, it needs to be
stated that the melting temperature (Tm) and the melting
enthalpy (DHm) are not fully retained after the first heating
run (Table 2). A reason for this difference between the first
and second heating run is the randomization of the chemical
microstructure, an entropically driven process, which takes
place by transesterification when the sample is kept in the
melt for a period of time. Moreover, the presence of modified
chain segments could severely hinder the crystallization
from the melt. To study the randomization phenomenon,
FIGURE 8 DSC traces of the first heating run of PB79Glux21
T
copolyester, which was kept in the melt for different predeter-
mined times, with indication of their degree of randomness R
and their melting temperatures. [Color figure can be viewed in
the online issue, which is available at wileyonlinelibrary.com.]
ꢀ
samples of PB79Glux21T were kept in the melt at 260 C for
174
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
FIGURE 10 Overview of the melting behavior during the sec-
ond heating runs of the PBT/sugar-based copolyesters pre-
pared by SSM. Data obtained from this work (Glux) and
literature (IS, Galx, and Manx).18,29 The open triangle in the
upper left corner of the figure represents the melting tempera-
ture of the homopolyester PBT. [Color figure can be viewed in
the online issue, which is available at wileyonlinelibrary.com.]
FIGURE 9 Overview of the R-values as function of time in the
melt at 260 ꢀC for the sugar-based copolyesters PB79Glux21T,
PB82Manx18T, and PB81Galx19T. [Color figure can be viewed in
the online issue, which is available at wileyonlinelibrary.com.]
applies to the randomization reaction, where the axially ori-
ented OH-groups have a lower reactivity, yielding a lower
randomization rate compared with Manx. Moreover, from
Figure 9, it can be concluded that the Galx-based copolyest-
ers randomize at the highest rate, which can be explained by
its more flexible structure. Galx was also shown to be incor-
porated within the shortest timeframe at similar reaction
mining the thermal properties of the materials. Glux shows a
relatively slow randomization and therefore longer crystalliz-
able PBT segments are retained during the time in the melt.
Unfortunately, no data was reported in ref. 18 for IS-based
materials. Galx- and Manx-based copolyesters randomize
faster (Fig. 9) and therefore yield much shorter crystallizable
segments, thus a lower melting temperature in the second
heating compared with Glux-based materials. The important
conclusions, which can be drawn are that the SSM Glux-
based copolyesters are able to retain most of their thermal
features, and that this is due to the relatively slow random-
ization of the chemical microstructure attained in these
materials upon melting.
temperatures.29 These results clearly show
a relation
between the rate of incorporation and the rate of randomiza-
tion; both depend on the orientation of the methylol groups
and the flexibility of the monomer structure.
Thermal properties of the second heating run are worth
studying and comparing with other PBT/sugar-based copo-
lyesters prepared by SSM, since these all have similar ther-
mal histories. As mentioned above, melting temperatures
and enthalpies of PBT/Glux-based copolyesters were not
fully retained after the first heating run, which was previ-
ously observed for other PBT/sugar-based copolyesters pre-
pared by SSM.18,29 In accordance with those results, the
randomization of the chemical microstructure leads to a
change in the thermal behavior between the first and second
heating run. When the Tm values obtained from the second
heating runs for PBT/Glux-based copolymers are compared
with other PBT/sugar-based copolyesters prepared by SSM,
noticeable differences are observed and an overview is given
in Figure 10.
The Tg data obtained for the copolyesters during the third
heating run is presented in Table 2. The Tg increases asꢀ a
ꢀ
function of the Glux content from 45 C in PBT up to 88 C
in PB76Glux24T. The increase of the Tg values was largely
expected because of the stiffer nature of Glux compared with
1,4-butanediol. To compare the effect of the various sugar-
based diols on the Tg values of their respective copolyesters
obtained by SSM, an overview is given in Figure 11.
The sugar-based comonomers IS, Galx and Manx all lead to
increased glass transition temperatures of PBT. However, the
increase of the Tg value is significantly larger for Glux-based
copolyesters compared with IS, Galx, and Manx. Based on
data reported in the literature for PGluxT and PIST homopo-
lyesters (i.e., prepared from the sugar-based diol and DMT),
such a significant difference would in principle not be
expected. The Tg of a low molecular weight PGluxT homopo-
lyester (Mn 5 3.0–5.5 kg mol21) has been reported to be in
the order of 150–154 ꢀC.45,46 Moreover, from a Fox-Flory
data representation of Tg vs. 1/Mn for several PIST sam-
ples,10 the Tg for PIST homopolyesters having comparable
molecular weights (Mn 5 3.0–5.5 kg mol21) was estimated to
lie in the range 150–180 ꢀC. Therefore, a comparable or
The copolyester series prepared from the sugar-based como-
nomers IS, Galx, and Manx show a decrease of the Tm values
with increasing sugar content, as is observed for Glux. How-
ever, noteworthy differences between the melting tempera-
tures of the various copolyesters are not observed at the
lowest comonomer contents. The differences are much more
significant at comonomer contents exceeding 15 mol %. The
copolyesters, which contain Glux or IS show higher melting
points compared with the materials with similar amounts of
Galx and Manx. At higher comonomer contents, the differ-
ence in randomization rate could play a key role in deter-
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
175

JOURNAL OF
POLYMER SCIENCE
ARTICLE
WWW.POLYMERCHEMISTRY.ORG
CONCLUSIONS
Using SSM, the D-glucose-derived diol 2,4:3,5-di-O-methylene-
D-glucitol (Glux) was successfully introduced into the back-
bone of PBT by transesterification. The conformation of this
diol was determined and it was shown that just one confor-
mation is assumed by Glux, which was significantly different
from the conformation of the structurally similar Manx. A rel-
atively low reaction temperature and nitrogen flow were used
to synthesize the copolyesters without significant volatiliza-
tion of the sugar-based comonomer. The final composition of
the copolyesters, which was determined by 1H-NMR, was
close to the initial feed ratio. The overall chemical microstruc-
ture of the resulting polymers was proven to be nonrandom.
Long PBT sequences, forming part of the crystalline phase,
remained unmodified during the SSM reaction, whereas the
chemical microstructure of the amorphous phase was practi-
cally random. The Glux-based copolyesters randomized in the
melt, but relatively slowly compared with other sugar-based
copolyesters prepared by SSM. This was attributed to the
lower reactivity of the axial hydroxyl group of Glux. The semi-
crystalline Glux-based copolyesters showed higher melting
points and higher glass transition temperatures compared
with materials prepared from IS, Manx, and Galx by the same
technique. Clearly, the conformation of the diol residue had a
significant influence on the incorporation behavior as well as
on the thermal properties of the copolyesters.
FIGURE 11 Comparison of the Tg values of PBT/sugar-based
copolyesters prepared using SSM. Data obtained from this
work (Glux) and literature (IS, Galx, and Manx).18,29 The open
triangle represents the homopolyester PBT. [Color figure can
be viewed in the online issue, which is available at
wileyonlinelibrary.com.]
slightly higher Tg was expected for the IS-based copolyesters.
Hence, another factor is playing a crucial role in this devia-
tion. The crystallinity of the IS-based samples is much lower
compared with the samples based on Glux. Hence, the amor-
phous phase of the Glux-based copolyesters is much richer
in rigid comonomer segments, which in turn leads to an
increase of the Tg values of Glux-based materials. Moreover,
it is possible the Glux allows a closer packing of the chains
in the amorphous phase, and therefore the Tg value
increases. The observation that the Glux-based SSM copo-
lyesters show higher Tg values compared with their Galx-
based counterparts was expected. Since the two 1,3-dioxo-
lane rings of Galx are not fused, its molecular structure is
more flexible than the fused bicyclic 1,3-dioxane structure of
Glux. At first glance, the structure of Manx seems to be very
comparable with Glux, since both are composed of two fused
1,3-dioxane rings. However, the PManxT homopolyester
shows a Tg value of 137 ꢀC,7 which is significantly lower
than that of PGluxT, despite PManxT having a higher molecu-
lar weight (Mn 5 12.9 kg mol21). The conformational study
conducted on Glux and Manx7 (Scheme 2) proved that these
two compounds have very different conformations. It is clear
that the conformations of these compounds lead to a signifi-
cant difference in the Tg values obtained for the copolyesters
prepared from these diastereoisomers. Additionally, at higher
comonomer contents the crystallinity of the Glux-based sam-
ples is significantly higher, due to the slower randomization
of the chemical microstructure (Fig. 9). Therefore, the rela-
tive amount of comonomer present in the amorphous phase
from the Glux-based samples is higher, leading to an addi-
tional increase of the Tg value.
Hence, by incorporating Glux into the backbone of PBT by
SSM, partially renewable copolyesters with outstanding ther-
mal properties were prepared. Moreover, in contrast to IS,
which required the synthesis of a macrodiol derivative prior
to SSM, this D-glucose-based monomer could be incorporated
directly into the backbone of the polyester.
ACKNOWLEDGMENTS
The authors thank the Bio-Based Performance Materials (BPM)
Program for financial support. Financial support for this work
was also provided by MINECO (Spain) MAT2009–14053-C02-
02 and MAT2012–38044-CO3-03 and by AGAUR (Catalonia)
with grant 2009SGR1469. The authors are also indebted to
MEC (Spain) for the FPU grant awarded to C. Lavilla. The
ꢀ
authors thank to L. Sanchez and C. Japu for their contribution to
the synthesis of 2,4:3,5-di-O-methylene-D-glucitol and 1,6-di-O-
benzoyl-2,4:3,5-di-O-methylene-D-glucitol.
REFERENCES AND NOTES
1 R. Wool, S. Sun, In Biobased Polymers and Composites; Aca-
demic Press: New York, 2005.
2 C. K. S. Pillai, Design. Monom. Polym. 2010, 13, 87–121.
3 G. Q. Chen, M. K. Patel, Chem. Rev. 2012, 112, 2082–2099.
ꢀ
ꢀ
4 J. A. Galbis, M. G. Garcıa-Martın, In Monomers, Polymers
and Composites from Renewable Resources; M. N. Belgacem,
A. Gandini, Eds.; Elselvier: Oxford, 2008; pp 89–114.
It can be concluded therefore that the D-glucose derived diol
Glux, unlike IS, can be directly incorporated into PBT, yield-
ing partially renewable materials with remarkable thermal
properties, which are superior to other PBT/sugar-based
copolyesters prepared by SSM.
ꢀ
ꢀ
5 J. A. Galbis, M. G. Garcıa-Martın, Chem. Curr. Top. 2010, 295,
147–176.
6 A. Gandini, Green Chem. 2011, 13, 1061–1083.
176
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177

JOURNAL OF
POLYMER SCIENCE
WWW.POLYMERCHEMISTRY.ORG
ARTICLE
ꢀ
ꢀ
7 C. Lavilla, A. Martınez de Ilarduya, A. Alla, M.G. Garcıa-
26 Z. Denchev, M. Sarkissova, Macromol. Chem. Phys. 1996,
197, 2869–2887.
ꢀ
~
Martın, J.A. Galbis, S. Munoz-Guerra, Macromolecules 2012,
45, 8257–8266.
27 D. R. Kelsey, B. M. Scardino, J. S. Grebowicz, H. H. Chuah,
Macromolecules 2000, 33, 5810–5818.
€
8 J. Thiem, H. Luders, Polym. Bull. 1984, 11, 365–369.
9 D. Braun, M. Bergmann, J. Prakt. Chem. 1992, 334, 298–310.
28 N. Lotti, M. Colonna, M. Fiorini, L. Finelli, C. Berti, Polymer
2011, 52, 904–911.
10 F. Fenouillot, A. Rosseau, G. Colomines, R. Saint-Loup, J. P.
Pascault, Prog. Polym. Sci. 2010, 35, 578–622.
ꢀ
29 C. Lavilla, E. Gubbels, A. Martınez de Ilarduya, B. A. J.
~
Noordover, C. E. Koning, S. Munoz-Guerra, Macromolecules
2013, 46, 4335–4345.
11 R. Storbeck, M. Ballauff, J. Appl. Polym. Sci. 1996, 59, 1199–
1202.
ꢀ
~
30 R. Marın, S. Munoz-Guerra, J. Appl. Polym. Sci. 2009, 114,
ꢀ
~
12 R. Quintana, A. Martınez de Ilarduya, A. Alla, S. Munoz-
Guerra, J. Polym. Sci. Part A: Polym. Chem. 2011, 49, 2252–
2260.
3723–3736.
31 D. W. Conder, The structures and conformations of some
cyclic O-benzylidene acetals of hexitols. PhD Thesis, University
of British Columbia, 1961.
13 B. A. J. Noordover, V. G. van Staalduinen, R. Duchateau, C.
E. Koning, R. A. T. M. van Benthem, M. Mak, A. Heise, A. E.
Frissen, J. van Haveren, Biomacromolecules 2006, 7, 3406–
3416.
32 J. A. Mills, Adv. Carbohydr. Chem. 1955, 10, 1–53.
33 E. J. Bourne, J. Burdon, J. C. Taltow, J. Chem. Soc. 1958,
14 B. A. J Noordover, R. Duchateau, R. A. T. M van Benthem,
W. Ming, C. E. Koning, Biomacromolecules 2007, 8, 3860–3870.
1274–1279.
34 F. Franks, J. Dadok, S. Ying, R. L. Kay, J. R. Grigera, J.
Chem. Soc. Faraday Trans. 1991, 87, 579–585.
15 H. R. Kricheldorf, G. Behnken, M. Sell, J. Macromol. Sci.
Pure Appl. Chem. 2007, 44, 679–684.
35 O. Jardetzky, G. C. K. Roberts, Molecular Biology; Academic
Press: New York, 1981; pp 201.
16 M. Okada, Y. Okada, A. Tao, K. Aoi, J. Appl. Polym. Sci.
1996, 62, 2257–2265.
36 J. F. Wiggins, D. J. C. Wood, J. Chem. Soc. 1951, 1180–
17 M. Okada, K. Tsunoda, K. Tachikawa, K. Aoi, J. Appl. Polym.
Sci. 2000, 77, 338–346.
1184.
37 W. G. Overend, R. Montgomery, L. F. Wiggins, J. Chem.
Soc. 1948, 2201–2203.
18 R. Sablong, R. Duchateau, C.E. Koning, G. de Wit, D. van Es, R.
Koelewijn, J. Van Haveren, Biomacromolecules 2008, 9, 3090–
3097.
38 E. Gubbels, L. Jasinska-Walc, D. Hermida Merino, H.
Goossens, C. Koning, Macromolecules 2013, 46, 3975–3984.
19 M. A. G. Jansen, J. G. P. Goossens, G. de Wit, C. Bailly, C.
E. Koning, Anal. Chim. Acta 2006, 557, 19–30.
39 R. A. Newmark, J. Polym. Sci. 1980, 18, 559–563.
40 H. R. Kricheldorf, Makromol. Chem. 1978, 179, 2133–2143.
20 C. D. Papaspyrides, S. N. Vouyiouka, In Fundamentals of
Solid State Polymerization; C. D. Papaspyrides, S. N.
Vouyiouka, Eds.; Wiley: Hoboken, 2009; pp 2–158.
41 Y. Yamadera, M. Murano, J. Polym. Sci. Part A: Polym.
Chem. 1967, 5, 2259–2268.
42 J. C. Randall, In Polymer Sequence Determination; Aca-
demic press: New York, 1977; pp 71.
21 M. A. G. Jansen, J. G. P. Goossens, G. de Wit, C. Bailly, C.
Schick,C.E. Koning, Macromolecules 2005, 38, 10658–10666.
ꢀ
43 (a) C. Lavilla, A. Alla, A. Martınez de Ilarduya, E. Benito, M.
22 M. A. G. Jansen, L. H. Wu, J. G. P. Goossens, G. de Wit, C.
Bailly, C. E. Koning, J. Polym. Sci. Part A: Polym. Chem. 2007, 45,
882–899.
ꢀ
ꢀ
~
G. Garcıa-Martın, J. A. Galbis, S. Munoz-Guerra, Polymer 2012,
ꢀ
53, 3432-3445; (b) C. Lavilla, A. Martınez de Ilarduya, A. Alla, S.
~
Munoz-Guerra, Polym. Chem. 2013, 4, 282–289.
23 M. A. G. Jansen, L. H. Wu, J. G. P. Goossens, G. de Wit, C.
Bailly, C. E. Koning, G. Portale, J. Polym. Sci. Part A: Polym.
Chem. 2008, 46, 1203–1217.
~
44 C. Lavilla, S. Munoz-Guerra, Green Chem., 2013, 15, 144–
151.
ꢀ
ꢀ
45 C. Japu, A. Alla, A. Martınez de Ilarduya, M. G. Garcıa-
Martın, E. Benito, J. A. Galbis, S. Munoz-Guerra, Polym. Chem
24 R. R. Gallucci, B. R. Patel, In Poly(Butylene Terephthalate),
in Modern Polyesters, Chemistry and Technology of Polyesters
and Copolyesters. J. Scheirs, T. E. Long, Eds.; Wiley: Chiches-
ter, 2004; pp. 293–321.
ꢀ
~
2012, 3, 2092–2101.
ꢀ
ꢀ
46 C. Japu, A. Martınez de Ilarduya, A. Alla, M. G. Garcıa-
ꢀ
~
Martın, J. A. Galbis, S. Munoz-Guerra, Polym. Chem 2013, 4,
25 J. M. R. C. A. Santos, J. T. Guthrie, J. Mater. Chem. 2006,
16, 237–245.
3524–3536.
WWW.MATERIALSVIEWS.COM
JOURNAL OF POLYMER SCIENCE, PART A: POLYMER CHEMISTRY 2014, 52, 164–177
177