Synthesis and characterization of poly(2,5-furan dicarboxylate)s based on a variety of diols

Article abstract of DOI:10.1002/pola.24812

Novel polyesters from renewable resources based on 2,5-dicarboxylic acid and several diols were synthesized and characterized using different polycondensation techniques. The aliphatic diols were sufficiently volatile to allow the use of polytransesterifications, which gave high-molecular weight semicrystalline materials with good thermal stability. In particular, the polyester based on ethylene glycol displayed properties comparable with those of its aromatic counterpart, poly(ethylene terephthalate), namely, the most important industrial polyester. The use of isosorbide gave rise to amorphous polymers with very stiff chains and hence a high glass transition temperature and an enhanced thermal stability. The interfacial polycondensation between the acid dichloride and hydroquinone produced a semicrystalline material with features similar to those of entirely aromatic polyesters, characterized essentially by the absence of melting and poor solubility, both associated with their remarkable chain rigidity. The replacement of hydroquinone with the corresponding benzylic diol was sufficient to provide a more tractable polyester. This study provided ample evidence in favor of the exploitation of furan monomers as renewable alternatives to fossil-based aromatic homologs. Copyright

Full text of DOI:10.1002/pola.24812

Synthesis and Characterization of Poly(2,5-furan dicarboxylate)s
Based on a Variety of Diols
´
MONICA GOMES, ALESSANDRO GANDINI, ARMANDO J. D. SILVESTRE, BRUNO REIS
CICECO and Department of Chemistry, University of Aveiro, 3810-193 Aveiro, Portugal
Received 23 March 2011; accepted 23 May 2011
DOI: 10.1002/pola.24812
Published online 20 June 2011 in Wiley Online Library (wileyonlinelibrary.com).
ABSTRACT: Novel polyesters from renewable resources based
on 2,5-dicarboxylic acid and several diols were synthesized
and characterized using different polycondensation techni-
ques. The aliphatic diols were sufficiently volatile to allow the
use of polytransesterifications, which gave high-molecular
weight semicrystalline materials with good thermal stability.
In particular, the polyester based on ethylene glycol dis-
played properties comparable with those of its aromatic
counterpart, poly(ethylene terephthalate), namely, the most
important industrial polyester. The use of isosorbide gave
rise to amorphous polymers with very stiff chains and hence
a high glass transition temperature and an enhanced thermal
stability. The interfacial polycondensation between the acid
material with features similar to those of entirely aromatic
polyesters, characterized essentially by the absence of melt-
ing and poor solubility, both associated with their remarkable
chain rigidity. The replacement of hydroquinone with the
corresponding benzylic diol was sufficient to provide a more
tractable polyester. This study provided ample evidence in
favor of the exploitation of furan monomers as renewable
C
V
alternatives to fossil-based aromatic homologs.
2011
Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 49:
3759–3768, 2011
KEYWORDS: aliphatic, aromatic and furan diols; crystallinity;
furan polyesters; 2,5-furandicarboxylic acid; polycondensation;
renewable resources; thermal properties
dichloride and hydroquinone produced
a semicrystalline
INTRODUCTION The growing interest in the preparation of
new chemicals and materials based on renewable resources
has naturally led to the development of a correspondingly
original realm of polymers.1 The interest of this strategy is
to provide viable alternatives to the predictable dwindling of
fossil resources associated with their price increase, as well
as to counteract the environmental problems related to fossil
CO2 emissions. Although these actions are more frequently
and more massively directed at novel sources of energy, the
production of chemical commodities from renewable resour-
ces has also become a major issue involving both academic
and industrial research activities.2
years, a massive output of publications3(a,d,e) suggests that
such a process will soon become a reality, motivated by the
practical interest in optimizing its synthesis. This is under-
standable, because HMF is the obvious precursor to 2,5-dis-
ubstituted furan monomers, like 2,5-furandicarboxylic acid
(FDCA) and its dichloride,3 as well as to promising chemicals
such as levulinic acid.4 The key structural feature associated
with these monomers is their close resemblance to aromatic
counterparts and hence the interest in using them to synthe-
size polymers via step-growth mechanisms, namely polyest-
ers and polyamides.
The realm of furan polyesters has been the subject of several
investigations,3 beginning with the pioneering work of Moore
and Kelly5 30 years ago, who studied the synthesis of a num-
ber of linear polyesters prepared from FDCA or its dichlor-
ide. Since then, these types of structures have been virtually
ignored, except for a brief study by Storbeck and Ballauf,6
probably because of the lack of ready availability of FDCA.
Therefore, the interest in furan polyesters shifted to the use
of difuran dicarboxylic monomers prepared by the coupling
of commercial 2-furancarboxylic esters through a condensa-
tion reaction involving aldehydes and, preferably, ketones. A
series of systematic investigations were thus carried out cov-
ering a wide range of diols, as well as of different moieties
In this context, two basic nonpetroleum monomer precursors
are readily accessible from polysaccharides or sugars bear-
ing, respectively, pentose and hexose moieties, namely the
first-generation furan derivatives furfural (F) and hydroxy-
methylfurfural (HMF).2,3 From them, a whole array of furan
monomers can be prepared and polymerized to give materi-
als often comparable to those derived from a whole host of
monomers prepared from fossil resources, that is, the major-
ity of today’s commercial polymers.3 While F has been an
industrial commodity for nearly a century,3 the production of
HMF has been slowed down by difficulties in terms of isolat-
ing it in good yields and purity. However, in the last few
Correspondence to: A. Gandini (E-mail: agandini@ua.pt)
C
V
Journal of Polymer Science Part A: Polymer Chemistry, Vol. 49, 3759–3768 (2011) 2011 Wiley Periodicals, Inc.
SYNTHESIS OF POLY(2,5-FURAN DICARBOXYLATE)S, GOMES ET AL.
3759

JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
bridging the two furan heterocycles,7 which provided a com-
prehensive set of structure–property relationships. Okada
et al.8 extended this domain using diols from renewable
resources.
C2D2Cl4, CD3COCD3, or C3F6OD2. 13C solid-state cross-polar-
ized magic-angle spinning nuclear magnetic resonance (13CP-
MAS [hpdec] NMR) spectra were recorded on a Bru¨cker
Avance 400 spectrometer operating with a magnetic field of
9.4 T, at 9.7 kHz, using glycidine as reference.
The prospect of HMF and its derivatives becoming viable
commodities3(a,d,e) induced us to return to furan polyesters
based on FDCA through a comprehensive investigation of
their synthesis and characterization, including novel struc-
tures and the re-examination of previously reported ones, as
briefly described in two preliminary communications.9,10
Grosshardt et al.11 concurrently tackled some of these
systems.
Elemental analyses of C and H were conducted in duplicate
with a Leco CHNS-932 analyzer. The duplicate elemental
analysis of C, H, O, and F were carried out at the CNRS ‘‘Serv-
ice Central d’Analyse,’’ Vernaison, France.
The thermogravimetric properties were gathered using a Shi-
madzu TGA 50 analyzer equipped with a platinum cell. Sam-
ꢁ
ples were heated at ꢁa constant rate of 10 C/min from room
The present detailed report marks the start of this program
with a first set of furan polyesters and copolyesters based
on FDCA and several diols, prepared by straightforward
routes, using mild conditions. It is important to reiterate that
at the basis of this strategy lies the pressing requirement of
developing novel materials from renewable resources capa-
ble of replacing, or even surpassing, their fossil-based
counterparts.
temperature to 800 C, under a nitrogen flow of 20 mL/min.
DSC thermograms were obtained with a Setaram DSC92 calo-
rimeter using aluminum pans under nitrogen with a heating
ꢁ
ꢁ
rate of 10 C/min in the temperature range of 18–400 C.
The X-ray diffraction (XRD) patterns were recorded with a
Phillips X’pert MPD diffractometer, using Cu Ka radiation.
Monomer Synthesis
The diol comonomers were used as such and the furan
monomers were prepared as described below.
EXPERIMENTAL
Materials
2,5-Furandicarbonyl Chloride 1
FDCA was a generous gift from Dr. Claude Moreau, who pre-
pared it during his pioneering studies related to the optimi-
zation of the preparation of HMF.12 Its purity was deemed
adequate for its subsequent use as a monomer precursor. D-
Isosorbide and isoidide were generous gifts from Roquette
France. Bis(2,5-hydroxymethyl)furan was kindly donated by
TransFurans Chemicals bvba, Geel, Belgium. Ethylene glycol
(99%), propane-1,3-diol (99%), 1,4-di-(hydroxymethyl)-ben-
zene (99%), hydroquinone (99.5%), as well as all other
reagents and solvents were purchased from Aldrich and
used as received. D-Isosorbide, isoidide, and bis(2,5-hydroxy-
methyl)furan were dried under vacuum in the presence of
phosphorus pentoxide and 1,1,2,2-tetrachloroethane (TCE),
which was dried over sodium hydride.
1 g of FDCA, 2 mL of SOCl2, and 20 lL of dimethylforma-
mide (DMF) were introduced into a 25 mL round-bottom
flask fitted with a condenser and a magnetic stirrer, and the
ꢁ
mixture was refluxed at 80 C for 4 h with constant stirring.
The condenser was connected to a washing bottle, filled
with a concentrated sodium hydroxide aqueous solution,
through a glass tube packed with activated silica gel. After
the reaction, the excess of SOCl2 and DMF was removed
under vacuum at room temperature and collected in a trap
cooled with liquid nitrogen. The ensuing monomer was iso-
lated and purified by high-vacuum sublimation (yield: 78%).
White crystalline powder, m.p. 79 ꢁC. FTIR (m/cmꢂ1): 3143
1
(¼¼CH); 1737 (C¼¼O); 1563 (C¼¼C); 974, 824, 712 (CAH). H
NMR (CDCl3, d/ppm): 7.5 (s, H3 and H4). 13C NMR (CDCl3,
d/ppm): 155.9 (C¼¼O); 149.3 (C2 and C5); 123.2 (C3 and
C4).
Techniques
Molecular weights and their distributions were obtained
using a home-made size exclusion chromatographer (SEC)
equipped with a PL-EMD 960 light scattering detector. The
column set consisted of a PL HFIPgel guard followed by two
Bis(hydroxyalkyl)-2,5-furandicarboxylates
The general procedure for the synthesis of these monomers
deals with the classical Fischer esterification. Typically, 1 g
of FDCA (6.41 mmol), a large excess of diol, and 1.8 mL of
concentrated HCl were introduced into a round-bottom flask.
The mixture was heated at 90 ꢁC for 12 h under magnetic
stirring. The excess of diol was then removed under high
vacuum after neutralization with a saturated solution of so-
dium hydroxide in the used diol. The products were isolated
by dissolution in acetone at room temperature, the impur-
ities were filtered, and finally the solvent vacuum was
removed.
2
ꢁ
PL HFIP columns (300 ꢀ 7.5 mm ), kept at 40 C. The HPLC
was set with a flow rate of 1.0 mL/min and a v/v/v mixture
of dichloromethane/chloroform/1,1,1,3,3,3-hexafluoro-2-pro-
panol (HFIP) 70/20/10 was used as the eluent. The sample
solutions with a concentration of about 3 mg/mL were fil-
tered through a PTFE membrane before injection. A molecu-
lar weight calibration curve was obtained with eight polysty-
rene standards in narrow-range of molecular weights
comprised between 790 and 96,000 g/mol.
All FTIR-attenuated total reflection (ATR) spectra were taken
with a Bru¨cker IFS FTIR spectrophotometer equipped with a
single horizontal Golden Gate ATR cell. 1H and 13C NMR
spectra were recorded on a Bru¨cker AMX 300 operating at
300.13 MHz for 1H and 75.47 for 13C spectra in CDCl3,
Bis(hydroxyethyl)-2,5-furandicarboxylate 2. FTIR (m/
cmꢂ1): 3354 (OH); 2952, 2881 (CAH); 1716 (C¼¼O); 1582,
1511 (C¼¼C); 1272 (CAO); 963, 832, 764 (¼¼CH). 1H NMR
(CD3COCD3, d/ppm): 7.4 (s, H3/H4 furan ring); 4.4 (t,
3760
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
CH2AOAC¼¼O); 3.9 (t CH2-OH). 13C NMR (CD3COCD3, d/
ppm): 158.7 (C¼¼O); 147.7 (C2/C5 furan ring); 119.7 (C3/C4
furan ring); 67.6 (COOCH2A); 60.6 (ACH2OH). White crystal-
13C NMR (C2D2Cl4, d/ppm): 157.07 and 156.85 (C¼¼O);
146.13 and 146.00 (C2/C5 furan ring); 119.41 and 119.18
(C3/C4 furan ring); 73.06 (C7); 85.74 (C8); 78.80 (C9);
74.95 (C10); 80.95 (C11); 70.76 (C12).
ꢁ
line powder, m.p. 91 C.
Bis(3-hydroxypropyl)-2,5-furandicarboxylate 3. FTIR (m/
cmꢂ1): 3228 (OH); 3119 (¼¼CH); 2930, 2873 (CAH); 1725
(C¼¼O); 1573, 1505 (C¼¼C); 1279 (CAO); 960, 855, 771
(¼¼CH). 1H NMR (CDCl3, d/ppm): 7.3 (s, H3/H4 furan ring);
4.5 (t, CH2AOAC¼¼O); 3.8 (t, ACH2AOH); 2.0 (q,
ACH2ACH2ACH2A). 13C NMR, (CDCl3, d/ppm): 158.2 (C¼¼O);
146.6 (C2/C5 furan ring); 118.5 (C3/C4 furan ring); 62.6
(CH2AOAC¼¼O); 58.9 (ACH2OH); 33.9 (ACH2ACH2ACH2A).
Interfacial Polycondensation
The reactions were carried out at room temperature using a
biphase system stirred mechanically at 900 rpm, consisting
of (i) a 0.19 M aqueous solution of NaOH containing hydro-
quinone (5.18 mmol) and tetrabutylammonium bromide
(72.4 mg) as the phase-transfer agent and (ii) dichlorome-
thane containing the 2,5-furandicarbonyl chloride (5.18
mmol). A solid precipitate appeared almost immediately and
the stirring was maintained further for an hour. The medium
was then acidified to pH 2, and the polyester was isolated
by filtration, washed with water, ethanol and acetone, and
vacuum dried.
ꢁ
White crystalline powder, m.p. 79 C.
Polymerization Procedures
Three different techniques were adopted taking into account
the specific monomers involved. Solution and interfacial
polycondensations were the procedures chosen for the syn-
thesis of polyesters based on 1 and nonvolatile diols,
whereas monomers 2 and 3 were submitted to a polytran-
sesterification approach for preparing both polyesters and
copolyesters. The detailed procedures are provided below.
Poly(1,4-phenylene-2,5-furandicarboxylate). FTIR (m/cmꢂ1):
3122 (¼¼CH); 1739 (C¼¼O); 1573, 1572 (C¼¼C). 13C MAS
(hpdec) NMR, d/ppm: 160.6 (C¼¼O); 147.4 (C2/C5 furan ring
and C7/C10 aromatic ring); 122.3 (C3/C4 furan ring and
C8/C9/C11/C12 aromatic ring).
Poly(2,5-furandimethylene 2,5-furandicarboxylate). FTIR
(m/cmꢂ1): 1716 (C¼¼O); 1580 (C¼¼C); 1267 (CAO); 960, 821,
Solution Polycondensation
The procedure was adapted from Storbeck and Ballauff6 and
Gharbi et al.7e The reactions were carried out at low temper-
ature under nitrogen with magnetic stirring and TCE as sol-
vent. The diol monomer (2.59 mmol) was first dissolved in 1
mL of TCE and, after complete dissolution, 1.7 mL of pyri-
1
764 (¼¼CH). H NMR (C3F6DOD, d/ppm): 7.3 (s, H3/H4 furan
ring); 6.5 (s, H3/H4 furan ring), and 5.3 (s, COOCH2); 13C
NMR (C3F6DOD, d/ppm) 160.5 (C¼¼O), 147.1 (C2 and C5),
0
0
0
0
127.0 (C3 and C4 ), 123.7 (C2 and C5 ), 120.0 (C3 and C4),
and 68.0 (COOCH2).
ꢁ
dine was added. This mixture was cooled to about 0 C with
an ice bath and an equal molar amount of 2,5-furandicar-
bonyl chloride (2.59 mmol), dissolved in 1.5 mL of TCE, was
added. The reaction medium was kept under constant stir-
ring at room temperature while its viscosity increased pro-
gressively before the actual precipitation of the polymer,
which was isolated by filtration, washed repeatedly with
methanol, and dried.
Polytransesterification
The homopolymerization reactions were carried out in bulk
using 1 g of the corresponding bis(hydroxyalkyl)-2,5-furandi-
carboxylate mixed with 1% (w/w) of the catalyst Sb2O3 in a
round-bottom flask equipped with a magnetic stirrer. The
flask was connected to a high-vacuum line through a trap
cooledꢁ with liquid nitrogen. The mixture was heated rapidly
ꢁ
Poly(1,4-phenylbismethylene-2,5-furandicarboxylate). FTIR
(m/cmꢂ1): 3124 (¼¼CH); 2880 (CAH); 1718 (C¼¼O); 1578
(C¼¼C); 1267 (CAO); 976, 826, 761 (¼¼CH). 1H NMR
(C3F6DOD, d/ppm): 7.4 (s, H9/H10/H12/H13 aromatic ring);
7.3 (s, H3/H4 furan ring); 5.4 (s, CH2AOAC¼¼O). 13C NMR,
(C3F6DOD, d/ppm): 160.7 (C¼¼O); 147.1 (C2/C5 furan ring);
135.8 (C9/C10/C12/C13 aromatic ring); 128.9 (C8/C11 aro-
matic ring); 120.4 (C3/C4 furan ring); 68.3 (CH2AOAC¼¼O).
to 70 C and then progressively, in steps of about 10 C at a
time, to 240–250 ꢁC, under constant stirring. The tempera-
ture increases were applied when a stagnancy occurred in
the release of the glycol. The reaction was stopped when the
product turned into a solid mass by letting the system
return to room temperature. The ensuing polyester was dis-
solved in trifluoroacetic acid (TFA) or in HFIP and then pre-
cipitated into an excess of methanol, filtered, and vacuum
dried.
Polyester from 1 and Isosorbide 6. FTIR (m/cmꢂ1): 1713
(C¼¼O); 1579 (C¼¼C); 1267 (CAO); 967, 810, 761 (¼¼CH). 1H
NMR (C2D2Cl4, d/ppm): 6.61–6.56 (H3/H4 furan ring); 3.40–
3.29 (H7/H12); 4.76–4.71(H8/H11); 3.98 (H9); 4.38 (H10).
13C NMR (C2D2Cl4, d/ppm): 157.07 and 157.02 (C¼¼O);
146.13 and 145.99 (C2/C5 furan ring); 119.41 and 119.17
(C3/C4 furan ring); 73.05 (C7); 85.74 (C8); 78.84 (C9);
74.96 (C10); 80.96 (C11); 70.73 (C12).
Poly(ethylene 2,5-furandicarboxylate). FTIR (m/cmꢂ1):
3123 (¼¼CH); 1716 (C¼¼O); 1578 (C¼¼C); 1264 (CAO); 960,
834, 761 (¼¼CH). 1H NMR (C3F6DOD, d/ppm): 7.4 (s, H3/H4
furan ring); 4.8 (s, ACH2ACH2A). 13C NMR (C3F6DOD, d/
ppm): 161.0 (C¼¼O); 147.1 (C2/C5 furan ring); 121.1 (C3/C4
furan ring); 64.7 (ACH2ACH2A).
Poly(3-propylene 2,5-furandicarboxylate). FTIR (m/cmꢂ1):
3129 (¼¼CH); 2968 and 2906 (CAH); 1715 (C¼¼O); 1576
(C¼¼C); 1267 (CAO); 966, 825, 763 (¼¼CH). 1H NMR
Polyester from 1 and Isoidide 7. FTIR (m/cmꢂ1): 1719
(C¼¼O); 1579 (C¼¼C); 1267 (CAO); 976, 826, 761 (¼¼CH). 1H
NMR (C2D2Cl4, d/ppm): 6.61–6.56 (H3/H4 furan ring); 3.40–
3.32 (H7/H12); 4.76–4.71(H8/H11); 3.97 (H9); 4.38 (H10).
(C2D2Cl4,
d/ppm):
7.2
(s,
H3/H4);
4.5
(t,
ACH2ACH2ACH2A); 2.3 (q, ACH2ACH2ACH2A). 13C NMR,
SYNTHESIS OF POLY(2,5-FURAN DICARBOXYLATE)S, GOMES ET AL.
3761

JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
SCHEME 1 The synthesis of monomers 1–3.
(C2D2Cl4, d/ppm): 158.3 (C¼¼O); 146.9 (C2/C5); 119.3 (C3/
C4); 62.6 (ACH2ACH2ACH2); 28.4 (ACH2ACH2ACH2A).
SCHEME 2 The diols used in this study.
The copolymerization reactions involved monomers 2 and 3
in a molar feed of 1:3 and were conducted following the
same protocol used for the homopolymerizations detailed
above.
to that of their terephthalic polyester counterparts. With the
purpose of preparing progressively stiffer polymer chains,
we used D-isosorbide 6, isoidide 7, bis(2,5-hydroxymethyl)-
furan 8, bis-(1,4-hydroxymethyl)benzene 9, and hydroqui-
none 10. Additionally, diols 6, 7, and 8 were selected for the
preparation of polyesters fully derived from renewable
resources, considering moreover that isosorbide is a compet-
itively priced chemical commodity. The latter considerations
also apply, in principle, to 4 and 5, because they can be pre-
pared from glycerol, an abundant and cheap commodity pro-
duced in the preparation of biodiesel.
Poly(ethylene 2,5-furandicarboxylate-ran-3-propylene-2,
5-furanodicarboxylate). 1H NMR (C2D2Cl4, d/ppm): 7.2 (s,
H3/H4 furan rings); 4.7 (s, ACH2ACH2A 4.5 (t,
ACH2ACH2ACH2A); 2.3 (m, ACH2ACH2ACH2A 13C NMR
(C2D2Cl4, d/ppm): 158.2 (C¼¼O); 147.0 (C2/C5 furan rings);
119.5 (C3/C4 furan rings); 63.5 (ACH2ACH2A); 62.7
(ACH2ACH2ACH2); 28.4 (ACH2ACH2ACH2A).
Polymer Synthesis and Characterization
RESULTS AND DISCUSSION
The polycondensation processes adopted here were based
on well-known mechanisms and procedures, adapted to the
present context and to the different types of monomers and
comonomers. For the reactions conducted in solution, a pro-
ton trap insured the removal of HCl, whereas in the case of
interfacial procedures, the presence of NaOH in the aqueous
phase produced the corresponding neutralization. The poly-
transesterification approach was limited to those diols pos-
sessing a sufficient volatility to be vacuum removed, that is,
the two aliphatic diols 4 and 5.
Monomer Synthesis and Characterization
The purity of the FDCA we used was deemed satisfactory for
1
its use as a precursor in this study, as confirmed by TLC, H
and 13C NMR spectroscopy, as well as elemental analysis.
The synthetic pathways leading to the furan monomers 1–3
from it are sketched in Scheme 1, and the data relevant to
their structure are summarized in Experimental section.
The FTIR spectra of these monomers were in tune with their
expected structures, through the absence (for 1) or the char-
acteristic shift and narrowing of the OH band (from carbox-
ylic to primary alcohol for 2 and 3), respectively. The C¼¼O
bands of the carboxylic moieties shifted to the characteristi-
cally higher wavenumbers, compared with that of FDCA. Fur-
ther corroboration was acquired from both 1H NMR and 13C
NMR spectra, in which the multiplicity, integration, and
assignments were in tune with each monomer structuꢁre.
Interestingly, the melting temperatures of 79, 91 and, 78 C,
respectively, for 1, 2, and 3 were similar to those reported
for their terephthalic counterparts, namely, 79–83, 109–110,
and 77–79 ꢁC, respectively, indicating that the 2,5-disubsti-
tuted furan moiety introduces a cohesive energy similar to
that of the 1,4-phenylene homolog.
The polyester from (1) and isosorbide 6 (PDASF), polyester
from 1 and isoidide 7 (PDAIF), and PHMBF were prepared
from monomer 1 and diols 6, 7, and 9 using a conventional
solution polycondensation protocol at low temperature, as
shown in Scheme 3.
The interfacial polycondensation between 1 and 10 (see
Scheme 4) was carried out at room temperature using a
sodium hydroxide water solution/dichloromethane system
and tetrabutylammonium bromide as the phase transfer cat-
alyst. These specific conditions were selected on the basis of
The diol comonomers 4–10 (Scheme 2) were used as
received, given their high purity, and were chosen for their
specific structural peculiarities. The absence of detectable
impurities in these commercial products was confirmed by
1
TLC, FTIR, and both H and 13C NMR spectroscopy. Ethylene
glycol 4 and propane-1,3-diol 5 were aliphatic structures
used to achieve materials with a degree of flexibility similar
SCHEME 3 Solution polyesterification.
3762
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
SCHEME 4 Interfacial polycondensation for the synthesis of
PHQF.
SCHEME 5 Synthesis of PEF and PPF via polytransesterification.
a previously reported systematic study,7(d) which showed
that they gave the best results for the interfacial polyconden-
sation of furan monomers.
The number-average degree of polymerization (DPn) of the
polyesters prepared by polytransesterification was also
assessed by end-group determination. Samples from different
syntheses were dissolved in TFA and treated with an excess
of pentafluorobenzoyl chloride to esterify both terminal
groups, as this polymerization technique generated macro-
molecules bearing a primary OH function at each chain end.
After precipitation in methanol and washing several times
with it, the modified polymers were submitted to elemental
analysis in terms of C, H, O, and F and their DPn was calcu-
lated, assuming complete end-group modification. These val-
ues turned out to be higher than those obtained from the
SEC tracings, either because the fluorination had not been
complete or, more probably, because of the partial insolubil-
ity of the polyesters in the eluent medium used for the SEC
analyses (We did not use pure HFIP because of its high
price!), which excluded the higher molecular weights.
Another source of discrepancy could have been the differ-
ence in hydrodynamic volume of the polyesters compared
with that of the polystyrene standards, which would also
account for the fact that some polyesters gave polydispersity
index (PDI) values lower than two. Cyclic products were not
This synthetic approach was also selected for the prepara-
tion of the fully furan-based polyester, poly(2,5-furandi-
methylene 2,5-furandicarboxylate) (PBHMF; Fig. 1), because
of the low thermal stability of bis-(2,5-hydroxymethyl)furan.
Poly(ethylene 2,5-furandicarboxylate) (PEF) was prepared by
the polytransesterification of monomer 2. After an investiga-
tion aimed at optimizing the process, the best results, in
terms of molecular weights for this polyester, were obtained
with the synthetic approach described in Experimental sec-
tion (Scheme 5), which was also applied to monomer 3 to
prepare the homolog polyester, poly(3-propylene 2,5-furandi-
carboxylate) (PPF).
The co-polyester PEF-ran-PPF was also synthesized in these
conditions, with a threefold excess of monomer 3 in the
molar feed ratio. The composition of the copolymer was
1
determined by H NMR spectroscopy, using the relative inte-
gration of the ACH2CH2A equivalent protons of the ethylene
glycol units and the ACH2ACH2ACH2 protons of the propyl-
ene glycol units. The respective values indicated 24% of eth-
ylene units and 76% of propylene moieties, in good agree-
ment with the monomer feed compositions, as one would
indeed expect from very similar structures, both bearing ter-
minal primary OH functions.
identified and this aspect requires
investigation.
a more thorough
The lower molecular weight of PDAIF, compared with that of
PDASF, was attributed to a difference in the purity of the
corresponding diols, with D-isosorbide being certified as a
thoroughly purified compound.
The results of the optimized synthesis, in terms of yields of
the isolated polymers and their higher molecular weights,
are shown in Table 1. For instance, in the case of PEF, the
transesterification of monomer 2 gave much better results
when compared with those obtained from its synthesis via
the solution polycondensation from monomer 1 and ethylene
glycol 4, as already anticipated in our previous communica-
tions.9,10 Therefore, the transesterification procedure appears
to be particularly suited for the synthesis of furan polyesters
using aliphatic diols.
The solubility of the polyesters was examined using different
potential solvents. Common organic solvents such as
dichloromethane, chloroform, and tetrahydrofuran did not
dissolve poly(1,4-phenylbismethylene-2,5-furandicarboxylate)
(PHMBF), PEF, PPF and their copolymers, and indeed, the
only two compounds that induced total dissolution at room
temperature were trifluoroacetic acid and HFIP, with hot
TCE as the third candidate. Mixtures of some common sol-
vents containing some 10% of HFIP were successfully used
as solvents for low polyester concentrations, as mentioned
above in the context of the SEC analyses. PDASF was soluble
in more manageable solvents like chloroform, whereas
poly(1,4-phenylene-2,5-furandicarboxylate) (PHQF) was, as
expected, the most intractable of all, with acceptable solubil-
ity only in hot m-cresol.
Regarding the furan polyesters obtained with nonvolatile
diols, solution and interfacial polycondensations were shown
to be very good alternatives for the preparation of a wider
range of materials from renewable resources with a large
spectrum of applications.
The FTIR spectra of all the polyesters reported in Table 2
were found to be consistent with the corresponding macro-
molecular structures. The characteristic bands of the ester
carbonyl group appeared in the range of 1713–1739 cmꢂ1
,
FIGURE 1 The fully furan-based polyester.
depending on the nature of the group directly attached to it,
SYNTHESIS OF POLY(2,5-FURAN DICARBOXYLATE)S, GOMES ET AL.
3763

JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
TABLE 1 Conditions and Results Related to the Optimized Polyesterifications Carried Out in This Study
Polymer
Monomers
Polycondensation Method
Yield (%)a
Mn (g/mol)b
Mw (g/mol)b
PDIb
Mn (g/mol)c
PEF
2
Transesterification
Transesterification
Solution
79
76
23
91
80
60
77
75
58
c
22,400
21,600
2,000
13,750
5,670
3,880
–
44,500
27,600
2,500
23,670
7,270
5,410
–
1.99
1.28
1.25
1.72
1.28
1.39
–
ꢃ45,500
ꢃ49,000
PPF
3
PEF
1 þ 4
1 þ 6
1 þ 7
1 þ 8
1 þ 10
1 þ 9
2 þ 3
PDASF
PDAIF
PBHMF
PHQF
PHMBF
PEF-ran-PPF
a
Solution
Solution
Interfacial
Interfacial
Solution
22,100
14,100
47,500
32,000
2.15
2.27
Transesterification
ꢃ48,000
Related to the amount of polymer recovered after reprecipitation.
Determined by end-group perfluorination.
b
SEC analysis, PDI ¼ Mw/Mn.
1
and the CAO peak appeared around 1270 cmꢂ1. The typical
and characteristic bands of 2,5-disubstituted furan rings
were also present as given in the table. No significant
absorption in the OH stretching region was detected, sug-
gesting that the polymers had reached a plausibly high mo-
lecular weight, thus confirming the SEC data. As an example
of these features, Figure 2 displays the typical FTIR spec-
trum of the furan–aromatic polyester PHQF.
As a typical example of these spectra, Figure 3 shows the H
version of the PEF-ran-PPF copolyester, together with the
chemical shift assignments. A particular feature, observed in
both the 1H and 13C NMR spectra of this copolymer, was a
slight split in all furan-related resonances, including that of
the carbonyl carbons, pointing to a different chemical envi-
ronment, as could indeed be anticipated for a random distri-
bution of the two monomer units in its macromolecules.
1
Similar splits in the chemical shifts were also observed in
the 1H and 13C NMR spectra of the PDASF polyester,6 which
was fully derived from renewable resources, namely, in the
H3 and H4 of the furan ring (6.61–6.56 ppm), in the C2/C5
and C3/C4 carbons of the furan ring (146.13; 145.99 and
119.41; 119.17 ppm, respectively), as well as in the carbonyl
carbon (157.07; 157.02 ppm). In this case, the splits were a
consequence of the nonequivalence of the two OH groups of
isosorbide.
The polyesters were also unambiguously characterized by H
and 13C NMR spectroscopy. The data related to the spectra
corresponding to the polyesters prepared by polytransesteri-
fication are given in Table 3. In all instances, the spectra
were found to be consistent with the expected structure in
terms of both chemical shifts and relative integrations. The
spectra of PEF and PPF were very similar in terms of the
resonance peaks associated to the furan ring and the methyl-
ene moieties directly attached to the ester group. The only
difference arose from the extra middle CH2 group in the PPF
structure. Their integration ratio (1:2) in the PEF and PPF
(1:2:1) spectra reflected the expected behavior, together with
the corresponding peak multiplicities.
The other polymer prepared by solution polycondensation,
PHMBF, also gave NMR spectra, which corroborated its
expected structure. The 1H NMR spectrum of its solution in
C3F6DOD gave resonances peaks at 7.3 and 7.4 ppm,
TABLE 2 FTIR Data of the Polyesters
Frequency (cmꢂ1
)
Assignment
PEF
PPF
PDASF
PDAIF
PBHMF
PHQF
PHMBF
PEF-ran-PPF
¼CH Fu
3123
2972
1716
1578
1264
1015
3119
3124
3111
3127
2973
1716
1580
1267
1024
3121
–
3109
2925
1713
1579
1267
1014
3121
CAH (CH2)
C¼O (ester)
C¼C Fu
2968, 2906
1715
2979, 2880
1718
2931, 2870
1717
2965, 2903
1720
1739
1572
1269
1576
1577
1573
1578
CAO (ester)
1267
1267
1266
1269
Fu breathing
1024
1017
1019
Superposition
with aromatic
bands
1017
2,5-Disubstituted Fu
Fu: furan ring.
960, 834,
761
966, 825,
763
976, 826,
761
976, 822,
762
960, 821,
764
967, 810,
761
967, 826,
764
3764
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
phere was consistently good with an expected higher resist-
ance for the PHQF. These features are entirely similar to
those associated with the well-known polyesters in which ar-
omatic rings are present instead of furan heterocycles, thus
indicating that this structural substitution does not affect the
thermal stability of the ensuing polymers. This is particularly
relevant in the case of PEF and its aromatic counterpart
poly(ethylene terephthalate) (PET), namely, the most impor-
tant polyester on the market.
The DSC thermograms, one of which was published in our
previous communication,9,10 showed that the polyesters
exhibited glass transition temperatures (Tgs) ranging from
50 to 180 ꢁC, depending on the chain stiffness induced by
the structure of the diol used.
FIGURE 2 FTIR spectrum of PHQF.
The Tg, as well the Tc, of both PEF and PPF only appeared
on the second DSC scan trace after quenching the melted
polymers. This observation indicates a very high aptitude to
crystallization and, in fact, both polyesters displayed a high
degree of crystallinity when examined as pristine materials.
The differences in their Tg, Tc, and Tm are clearly related to
the increase in macromolecular flexibility associated with the
additional methylene group in PPF, which results in a sub-
stantial reduction in all these thermal transitions. Obviously,
their copolymer lost all aptitude to crystallize given its ran-
dom nature, which was confirmed by the fact that its Tg was
intermediate between those of the corresponding homopoly-
mers (Table 5).
respectively, for the furan H3/H4 and aromatic protons, with
the methylene protons between the two rings resonating at
5.4 ppm. In the same solvent, the 13C NMR spectrum showed
the corresponding assignments of the carbon from the car-
bonyl moieties at 160.7 ppm, the C2/C5 furan carbons at
147.1 ppm and C3/C4 at 120.4 ppm. The peaks assigned to
the aromatic ring carbons appeared at 128.9 ppm (unsubsti-
tuted carbons) and 135.7 ppm (p-substituted carbons) and
those of the methylene carbon at 66.4 ppm.
As stated earlier, the furanic–aromatic polyester PHQF was
insoluble in all common solvents and consequently its struc-
tural characterization was conducted by solid-state 13CP-MAS
NMR. This spectrum, shown in Figure 4, gave three intense
peaks at 160.0, 147.4, and 122.3 ppm, assigned, respectively,
to (i) C¼¼O, (ii) furan C2/C5 and p-substituted aromatic car-
bons, and (iii) C3/C4 furan and unsubstituted aromatic
carbons.
PDASF displayed a very high Tg, arising from the particularly
rigid isosorbide structure. As expected, this polymer did not
display any crystallinity because of its irregular structure, as
discussed above. Similar results were obtained for PDAIF,
with a slightly lower Tg, attributed to its lower molecular
weight (see Table 1), but in this case some crystallinity was
expected, given its more regular structure. This feature was,
however, not detected by X-ray analysis and indeed by the
The average values of the elemental analyses of all of the
polyesters are given in Table 4. The slightly lower carbon
percentages, together with the correspondingly higher hydro-
gen contents, suggest the presence of residual moisture, a
feature which became more pronounced with the isosorbide-
based polyester.
ꢁ
absence of a melting feature in its DSC tracing up to 275 C,
where its thermal decomposition begun. The low molecular
weight of this polyester could have been the cause of its fail-
ure to crystallize, which indicates the need of further work
aimed at improving its synthesis.
Table 5 collects the results related to the thermal properties
of the polyesters. The thermal stability in a nitrogen atmos-
TABLE 3 1H and 13C Chemical Shifts (d, ppm) from TMS in Different Solvents of the
Polyesters Synthesized by Polytransesterification
1H
13C
PEF
PPF
PEF-ran-PPF
PEF
PPF
PEF-ran-PPF
Assignment
C3F6DOD
C2D2Cl4
C3F6DOD
C3F6DOD
C2D2Cl4
C3F6DOD
2/5 Fu
–
–
–
147.1
121.1
161.0
64.7
–
146.9
119.3
158.3
–
147.0
119.5
158.2
63.5
3/4 Fu
7.4 (s)
7.2 (s)
–
7.2 (s)
–
2/5-CO
–
ACH2CH2A
ACH2CH2CH2A
ACH2CH2CH2A
4.8 (s)
–
–
4.7 (s)
4.5 (t)
2.3 (m)
4.5 (t)
2.3 (q)
62.6
28.4
62.7
–
28.4
Multiplicity—s: singlet; t: triplet; q: quintet; m: multiplet; Fu: furan ring.
SYNTHESIS OF POLY(2,5-FURAN DICARBOXYLATE)S, GOMES ET AL.
3765

JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
FIGURE 3 1H NMR spectrum of PEF-ran-PPF in TCEd2.
The Tg of PHMBF was not much higher than that of PEF and
its lack of pristine or induced crystallinity, confirmed by the
X-ray diffractogram, is difficult to rationalize in view of its
regular structure. A typical DSC thermogram of this polymer
is shown in Figure 5 and its TGA tracing in Figure 6. PHQF
did not display any thermal transition up to 400 ꢁC, despite
its high crystallinity (see below), a feature which is typical of
very highly rigid macromolecular structures like those of ar-
omatic polyesters, hence the replacement of a benzene ring
by a furan heterocycle did not reduce this stiffness in any
detectable way. In other words, the cohesive energy of these
polymers places their melting temperature above that of
their thermal degradation.
sharp signals at 2h ¼ 16.8, 24.8, and 29.5ꢁ (Fig. 7), suggest-
ing a high degree of macromolecular order.
CONCLUSIONS
The first family of furan polyesters described in this study
showed structures, molecular weights, and properties which
make them thoroughly viable as macromolecular materials,
TABLE 4 Elemental Analysis Data
%C
%H
PEF
Calculated
Found
PPF
The presence of crystallinity on the pristine polymers was
also assessed by XRD. PEF diffractograms showed three
sharp signals, respectively, at 2h ¼ 16.0, 20.1, and 27.8ꢁ.
PPF showed a similar pattern with 2h ¼ 11.0, 22.5, and
27.8ꢁ. The other polyesters did not display any crystallinity
pattern, with the exception to PHQF, which gave three rather
52.76
51.88
3.32
3.60
Calculated
Found
55.10
54.79
4.08
3.88
PDASF
Calculated
Found
54.14
52.97
3.79
3.80
PHMBF
Calculated
Found
65.12
65.18
3.90
4.03
PHQF
Calculated
Found
62.62
61.90
2.63
2.72
PEF-ran-PPF
Calculateda
Found
54.50
54.54
3.87
4.02
FIGURE 4 Solid-state 13CP-MAS NMR spectrum of PHQF.
a
Calculated on the basis of the copolymer composition.
3766
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
TABLE 5 Onset of Thermal Decomposition (Tdi), Maximum
Decomposition Temperature (Td), Glass Transition
Temperature (Tg), Crystallization Temperature (Tc), and Melting
Temperature (Tm) of the Polyesters
a
a
b
b
b
Polymer
Tdi (ꢁC) Td (ꢁC) Tg (ꢁC) Tc (ꢁC) Tm (ꢁC)
PEF
300
295
350
275
205
280
300
398
390
450
396
345
490
390
398
80
165
127
–
215
174
–
PPF
50
PDASF
PDAIF
PBHMF
PHQF
PHMBF
180
140
–
–
FIGURE 7 X-ray diffractogram of the pristine PHQF.
nd
87
80
nd
–
nd
–
PEF-ran-PPF 300
165
215
glycol can be prepared from glycerol. Work is in progress, on
the one hand, to extend the characterizations of these poly-
esters to mechanical and other properties, including mois-
ture uptake, and, on the other hand, to synthesize other orig-
inal furan polyesters.
a
TGA measurements.
DSC measurements (nd ¼ not detected below 400 ꢁC).
b
The authors thank the Portuguese Fundaça˜o para a Cieˆncia e
a Tecnologia (FCT) for a research grant (PTDC/QUI-QUI/
101058/2008) sponsoring this study.
REFERENCES AND NOTES
1 (a) Monomers, Polymers and Composites from Renewable
Resources; Belgacem, M. N.; Gandini, A. (Eds.); Elsevier: Ames-
terdam, 2008; (b) Gandini, A. Macromolecules, 2008, 41, 9491;
(c) Gandini, A. Green Chem 2011, 13, 1061.
2 (a) Werpy, T.; Petersen, G. Top Value Added Chemicals from
Biomass. Results of Screening for Potential Candidates from
Sugars and Synthesis Gas; Pacific Northwest National Labora-
tory, National Renewable Energy Laboratory for the Office of
Biomass Program (US Dept. of Energy): Washington, 2004; Vol.
I; (b) Petersen, G.; Bozell, J. Green Chem 2010, 12, 539.
FIGURE 5 Second DSC scan of PHMBF after a rapid cooling of
200 ꢁC/min.
3 (a) Gandini, A. Polym Chem 2010, 8, 245; (b) Belgacem, M.
N.; Gandini, A. In Monomers, Polymers and Composites from
Renewable Resources; Belgacem, M. N.; Gandini, A.; Eds;
Elsevier: Amesterdam, 2008; Chapter 6; (c) Gandini, A.; Belga-
cem, N. M. Prog Polym Sci 1997, 22, 1203; (d) Gandini, A. In
Green Polymerization Methods: Renewable Starting Materials,
Catalysis and Waste Reduction; Mathers, R. T.; Meier, M. A.,
Eds.; Wiley-VCH: Weinheim, 2011; p. 29; (e) Gandini, A. In
Films and Coatings from Renewable Resources; Plackett, D.,
Ed.; Wiley-VCH: Weinheim, 2011; p. 180.
quite comparable with aromatic–aliphatic homologs derived
from fossil resources. The most interesting of these compari-
sons is of course that relating PEF to PET, because of the
enormous relevance of the latter polymer. All the features
tested here indicate that PEF could successfully replace PET
and thus constitute a valuable contribution to polymers from
renewable resources, considering moreover that ethylene
4 Boisen, A.; Christensen, T. B.; Fu, W.; Gorbanev, Y. Y.; Han-
sen, T. S.; Jensen, J. S.; Klitgaard, S. K.; Pedersen, S.; Riisager,
˚
A.; Stahlberg, T.; Woodley, J. M. Chem Eng Res Des 2009, 87,
1318.
5 (a) Kelly, J. E. Ph.D. Thesis, Rensselear Polytechnic, NY,
1975; (b) Moore, J. A.; Kelly, J. E. Macromolecules 1978, 11,
568; (c) Moore, J. A.; Kelly, J. E. Polymer 1979, 20, 627.
6 Storbeck, R.; Ballauff, M. Polymer 1993, 34, 5003.
7 (a) Khrouf, A.; Boufi, S.; El Gharbi, R.; Belgacem, N. M.; Gan-
dini, A. Polym Bull 1996, 37, 589; (b) Khrouf, A.; Abid, M.;
Boufi, S.; El Gharbi, R.; Gandini, A. Macromol Chem Phys
FIGURE 6 TGA of PHMBF in a nitrogen flow.
SYNTHESIS OF POLY(2,5-FURAN DICARBOXYLATE)S, GOMES ET AL.
3767

JOURNAL OF POLYMER SCIENCE PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
1998, 199, 2755; (c) Khrouf, A.; Boufi, S.; El Gharbi, R.; Gan-
dini, A. Polym Int 1999, 48, 649; (d) Chaabouni, A.; Gharbi,
9 Gandini, A.; Silvestre, A.; Neto, C. P.; Sousa, A. F.; Gomes,
M. J Polym Sci Part A: Polym Chem 2008, 47, 295.
S.; Abid, M.; Boufi, S.; El Gharbi, R.; Gandini, A.
Chim Tunisie 1999, 4, 547; (e) Gharbi, S.; Andreolety, J.-P.;
Gandini, A. Eur Polym 2000, 36, 463; (f) Abid, M.;
J Soc
10 Gandini, A.; Coelho, D.; Gomes, M.; Reis, B.; Silvestre, A.
J Mater Chem 2009, 19, 8656.
J
11 Grosshardt, O.; Tubke, B.; Kowollik, K.; Fehrenbacher, U.;
¨
Kamoun, W.; El Gharbi, R.; Fradet, A. Macromol Mater Eng
2008, 293, 39.
Dingenouts, N.; Wilhelm, M. Chem Ing Technik 2009, 81,
1829.
8 (a) Okada, M.; Tachikawa, K.; Aoi, K. J Polym Sci Part A
Polym Chem 1997, 35, 2729; (b) Okada, M.; Tachikawa, K; Aoi,
K. J Appl Polym Sci 1999, 74, 3342.
12 Moreau, C.; Belgacem, M. N.; Gandini, A. Topics Catal 2004,
27, 11.
3768
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA