Versatile copolymerization of glycolide and rac-lactide by dimethyl(salicylaldiminato)aluminum compounds

Article abstract of DOI:10.1021/ma402174y

The dimethylaluminum compounds Al(CH₃)₂[(O-2-{(C ₆F₅)N=CH}-4,6-R₂C₆H₂)], R = H (1) or cumyl (2), were synthesized and tested as initiators in the homo- and copolymerization of ra

Full text of DOI:10.1021/ma402174y

Article
pubs.acs.org/Macromolecules
Versatile Copolymerization of Glycolide and rac-Lactide by
Dimethyl(salicylaldiminato)aluminum Compounds
Angelo Meduri,†,§ Tiziana Fuoco,‡,§ Marina Lamberti,‡ Claudio Pellecchia,‡ and Daniela Pappalardo†,*
†Dipartimento di Scienze e Tecnologie, Universita
̀
del Sannio, I-82100, via dei Mulini 59/A, Benevento, Italy
‡Dipartimento di Chimica e Biologia, Universita
̀
di Salerno, I-84084, via Giovanni Paolo II 132, Fisciano, Salerno, Italy
S
* Supporting Information
ABSTRACT: The dimethylaluminum compounds Al(CH3)2[(O-
2-{(C6F5)NCH}-4,6-R2C6H2)], R = H (1) or cumyl (2), were
synthesized and tested as initiators in the homo- and
copolymerization of rac-lactide and glycolide. These complexes
resulted active for the production of PLGA copolymers with
variable microstructure. All the copolymers were fully charac-
terized by NMR, GPC, and DSC analysis. The copolymerization
reactions were performed in bulk and in solution, by varying
comonomers ratio, monomer/catalyst feed ratio, temperature,
reaction time, and solvent. Interestingly, by changing the reaction conditions, copolymers from random, to blocky, to diblock
were obtained, demonstrating the effectiveness and versatility of such systems in modulating the copolymers microstructure and
the related thermal properties.
consequence, the properties of the copolymers widely vary
from batch to batch. This problem, obviously, involves all
polymeric materials, but it is crucial for those used in
biomedical field, where the in vivo applications require an
absolute control on the polymer microstructure and monomers
sequences.6
This aspect was elegantly addressed by Meyer et al. with the
preparation of poly(lactic-co-glycolic acid) repeating sequence
copolymers.7,8 In their work, the strategy for producing these
repeating sequence copolymers involved the assembly by
condensation polymerization of preformed segmers comprising
high degree of sequence and stereocontrol. The work of Meyer
allowed a really extensive, systematic and thorough inves-
tigation of PLGA microstructure. Remarkably, it was
demonstrated that an alternating PLGA exhibited a dramatic
different hydrolysis behavior in comparison with a random one.
Truly alternated poly(glycolide-alt-lactide) copolymers were
also obtained by polymerization of the 3-methyl-1,4-dioxan-2,5-
dione, synthesized ad hoc, using stannous octanoate9 or
bimetallic alkoxide of Al and Zn as initiators.10
INTRODUCTION
■
Over the past decades aliphatic poly(esters), such as poly-
(glycolide) (PGA), poly(lactide) (PLA), and their copolymers
(PLGAs), have found rapidly increasing research interest.
Because of their intrinsic biodegradability and bioassimilability,
they are already used in a large range of applications from
packaging and textile fibers to more sophisticated drug delivery
systems, absorbable surgical fibers and stem cell scaffoldings.1
The final applications of PLGAs are obviously dictated by
their properties, such as degradation rate and thermal and
mechanical behaviors, which can be tuned by carefully
modifying the polymer chain parameters such as molecular
weight, molecular-weight dispersity, monomers ratio and
sequence, and polymer chain-ends. In this regard, the ring-
opening polymerization (ROP) of commercially available
lactide (LA) and glycolide (GA) represents the most efficient
method to produce these polymers.2 The industrial and the
most commonly used catalyst for the preparation of such
polyesters is stannous octanoate. The first systematic studies on
the preparation of PLGAs copolymers by this initiator revealed
that the synthesized copolymers did not show a truly random
monomer distribution.3,4 Because of the higher reactivity of GA
in comparison to LA, the copolymers initially formed were
richer in GA than the monomer feed mixture. The blocky
structure, however, was randomized to various extent by
transesterification reactions. In other cases (i.e., for bulk
Although both the above-mentioned approaches allowed a
precision synthesis of PLGAs, they are less efficient than ROP.
Therefore, the search for ROP initiators that allow the
preparation of PLGAs in a controlled and reproducible fashion
still represents a challenge.
In fact, in addition to stannous octanoate, several catalysts
and initiators have been tested in GA/LA copolymerization.
Early studies include the testing of commercially available
polymerization at 150 °C), the detailed H and 13C NMR
1
investigations of the polymer microstructure revealed the
formation of shorter blocks, with average block lengths close to
2.5 However, one drawback of the use of the stannous
octanoate is the poor control of the polymerization and the
scarce reproducibility of the polymerization results. As a
Received: October 22, 2013
Revised: December 17, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
in CDCl3 at 25 °C on a Bruker Avance 300 spectrometer (1H, 300.13
MHz; 13C, 75.47 MHz). 2D DOSY PGSE NMR spectrum of the block
copolymer was performed on the Bruker Avance 400 spectrometer in
DMSO-d6 at 80 °C without spinning; parameters δ (1100 μs) and Δ
(0.1 s) were kept constant during the experiments, whereas G was
varied from 2 to 95% in 16 steps. The resonances are reported in ppm
(δ) and coupling constants in Hz (J), and they are referenced to the
residual solvent peak versus Si(CH3)4: C6D6 at δ 7.16 (1H) and δ
128.1 (13C), DMSO-d6 at δ 2.50 (1H) and δ 39.5 (13C), CDCl3 at δ
7.26 (1H) and δ 77.0 (13C); in the case of 19F, resonances were
automatically referenced versus CF3C6H5 by the software. All spectra
recording and data processing were performed on Bruker TopSpin
v2.1 software.
Molecular weights (Mn and Mw) and molecular-weight dispersities
(Mw/Mn) were measured by gel permeation chromatography (GPC).
The measurements were performed at 30 °C on a Waters 1525 binary
system equipped with a Waters 2414 Refractive Index (RI) detector
and a Waters 2487 Dual λ Absorbtion (UV, λabs = 220 nm) detector.
In the case of the analyses performed using THF as eluent (1.0 mL
min−1) a system of four Styragel HR columns (7.8 × 300 mm; range
103 − 106 Å) was employed. In the case of the analyses performed
using CHCl3/1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) 99/1 as eluent
(1.0 mL min−1) a system of two Styragel HR columns (7.8 × 300 mm;
range 103−104 Å) was employed. Narrow polystyrene standards were
used as reference and Waters Breeze v3.30 software for data
processing.
chlorides, alkoxides, oxides or sulfides of main groups and
transition metals (Sn, Al, Zr, Ti, Pd, Cd, and Zn).11 In this
study only tin-based initiators were claimed to produce
“random” copolymers, however the average blocks sequences
were not reported. Cationic copolymerization in the presence
of organic acids and salts (i.e., methyl triflate) clearly produced
non random macromolecules, with average blocks sequences
higher than 2.12 Afterward, homoleptic metal-complexes of Li,13
Mg,13 Al,5 Zn,5,14 Ca,15 Zr,16 Fe,17 and Bi18 have been also
tested and produced multiblock non random copolymers.
Subsequently, the preparation of poly(ethylene glycol)-block-
PLGAs was obtained by an elaborate semibatch polymerization
strategy requiring the stepwise addition of the more reactive
glycolide to a solution containing the lactide monomer, the
poly(ethylene glycol) macroinitiator and an organocatalyst.19
It is apparent, then, that in the literature there is still a
paucity of ROP catalysts capable of producing truly random
PLGAs. In the framework of our interest in the ROP of cyclic
esters promoted by various well-defined organometallic
catalysts,20 we recently reported dimethyl(salicylaldiminato)
aluminum compounds able to efficiently catalyze the living ROP
of L- and D,L-lactide and ε-caprolactone.20b Interestingly, in the
presence of these initiators random and block copolymers of ε-
caprolactone and lactide were prepared, with a controlled chain
growth, in the absence of transesterification reaction. Moreover,
the simple formulation and the easy activation represent
advantageous features with respect to most ROP aluminum
catalysts based on more complex polydentate ligand systems.
Although analogous aluminum phenoxy-imine complexes have
been largely studied for the ROP of a variety of cyclic esters,21
this class of catalysts has not been reported for the
homopolymerization of glycolide and for PLGAs production.
With these premises, in this work we have tested these
compounds as precatalysts in the ROP of glycolide and rac-
lactide, and studied the feasibility of random and block
copolymerization in different experimental conditions.
Glass transition temperatures (Tg), melting points (Tm), and
enthalpy of fusion (ΔHm) of the (co)polymers were measured by
differential scanning calorimetry (DSC) using a DSC 2920 TA
Instruments in nitrogen flow with a heating and cooling rate of 10 °C
min−1 in the range of −20 to +260 °C. The data were processed with
TA Universal Analysis v2.3 software and are reported for the second
heating cycle.
Synthesis of Al(CH3)2[O-2-{(C6F5)NCH}C6H4] (1). To a toluene
solution (15 mL) of Lig1 (1.5 g, 5.1 mmol) were added 10 mL of a n-
hexane solution 0.56 M of Al(CH3)3 (5.5 mmol) dropwise via cannula
at 0 °C. The reaction mixture was magnetically stirred for 1 h at 0 °C,
then for 1.5 h at room temperature. After this time, the solvent was
removed, the solid was washed with n-pentane and dried in vacuo.
1
Yield: 1.610 g (92%). H NMR (400 MHz, C6D6, 25 °C): δ 7.18 (s,
1H; NCH), 7.04 (td, J = 8.6, 1.4 Hz, 1H; ArH), 6.89 (dd, J = 8.6,
1.1 Hz, 1H; ArH), 6.63 (dd, J = 7.8, 1.4 Hz, 1H; ArH), 6.40 (td, J =
7.8, 1.1 Hz, 1H; ArH), −0.28 (bs, 6H; Al−CH3).
EXPERIMENTAL SECTION
■
General Procedures. Moisture and air sensitive materials were
manipulated under nitrogen using Schlenk techniques or a MBraun
Labmaster glovebox. Before use, glassware was dried overnight in an
oven at 120 °C and solvents were refluxed over a drying agent
(indicated below) and distilled under nitrogen: toluene, xylenes, and
methanol (Sigma-Aldrich) over Na; THF (Delchimica), n-pentane and
n-hexane (Sigma-Aldrich) over Na/benzophenone. Monomers
(Sigma-Aldrich) were purified prior to use: L-lactide and rac-lactide
were dried in vacuo with P2O5 for 72 h, and afterward stored at −30
°C in glovebox; glycolide was recrystallized from tetrahydrofuran
(THF).
13C NMR (101 MHz, C6D6, 25 °C): δ 176.5 (C(H)N), 167.3
(ArC−O), 143.0 (ArC−F), 140.3 (ArC−H), 139.4 (ArC−F), 136.9
(ArC−F), 136.2 (ArC−H), 123.5 (ArC−H), 121.4 (ArC−N), 118.8
(ArC−C(H)N), 118.2 (ArC−H), −9.2 (Al−CH3).
19F NMR (376 MHz, C6D6, 25 °C): δ −148.44 (d, J = 18.8 Hz, 2F;
o-F), −154.07 (t, J = 22.5 Hz, 1F; p-F), −160.50 (td, J = 22.5, 5.4 Hz,
2F; m-F).
Synthesis of Al(CH3 )2 [(O-2-{(C6 F5 )NCH}-4,6-{C-
(CH3)2C6H4}2C6H2)] (2). To a toluene (25 mL) solution of Lig2
(0.97 g, 1.86 mmol) were added 3 mL of a toluene solution 0.68 M of
Al(CH3)3 (2.05 M) dropwise via cannula at 0 °C. The reaction
mixture was magnetically stirred for 1 h at 0 °C, then for 2 h at room
temperature. After this time, the solvent was removed and the solid
dried in vacuo. Yield: 0.918 g (83%).
Ligands Lig122 and Lig223 were synthesized according to literature
methodologies.
Deuterated solvents were stored and used in agreement with the
recommendations by the producer (Eurisotop: dimethyl sulfoxide d6
(DMSO-d6) and CDCl3; Aldrich: C6D6); C6D6 was dried over
molecular sieves before use.
All other reagents and solvents were commercially available and
used without further purification for synthetic, spectroscopic and
catalytic purposes.
Instruments and Measurements. Elemental analyses were
recorded on a Thermo Finningan Flash EA 1112 series C, H, N, S
analyzer in the microanalytical laboratory of the institute.
NMR spectra of complexes were performed in C6D6 at 25 °C on a
Bruker Avance 400 spectrometer (1H, 400.13 MHz; 13C, 100.61 MHz;
19F, 376.50 MHz), using NMR tubes equipped with J. Young valves.
NMR spectra of polymers were performed in DMSO-d6 at 100 °C and
1H NMR (400 MHz, C6D6, 25 °C): δ 7.73 (d, J = 2.5 Hz, 1H;
ArH), 7.29−7.13 (m, 8H; ArH cumyl), 7.12−7.06 (m, 2H; p-ArH
cumyl), 7.03 (s, 1H; NCH), 6.77 (d, J = 2.5 Hz, 1H; ArH), 1.63 (s,
6H; CH3), 1.62 (s, 6H; CH3), −0.57 (s, 6H; Al−CH3).
13C NMR (101 MHz, C6D6, 25 °C): δ 176.7 (C(H)N), 164.2
(ArC−O), 150.3 (ArC−C), 150.0 (ArC−C), 143.0 (ArC−F), 142.5
(ArC−C), 140.5 (ArC−F),139.9 (ArC−C), 139.3 (ArC−F), 136.7 (ArC−
H), 131.5 (ArC−H), 128.7 (ArC−H), 127.1 (ArC−H), 126.5 (ArC−H),
125.8 (ArC−H), 125.6 (ArC−H), 121.6 (ArC−N), 118.5 (ArC-C(H)
N), 42.7(C(CH3)2), 42.4 (C(CH3)2), 30.9 (C(CH3)2), 29.0 (C-
(CH3)2), −10.3 (Al-CH3).
B
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
1
19F NMR (376 MHz, C6D6, 25 °C): δ −148.28 (dd, J = 18.2, 5.6
Hz, 2F; o-F), −154.76 (t, J = 22.3 Hz, 1F; p-F), −160.87 (td, J = 22.3,
5.6 Hz, 2F; m-F).
Poly(glycolide-co-rac-lactide). H NMR (300 MHz, DMSO-d6, 100
°C): δ 5.34−5.14 (m, 1H; CH(CH3)C(O)O), 4.98−4.71 (m, 2H;
CH2C(O)O), 1.57−1.44 (m, 3H; CH(CH3)C(O)O). 13C NMR (75
MHz, DMSO-d6): δ 168.4, 168.3, 168.2, 168.15, 168.1 (CH(CH3)
C(O)O), 165.8, 165.7 (CH2C(O)O),68.5, 68.3 (CH(CH3)C(O)O),
60.3, 60.2 (CH2C(O)O), 15.8, 15.7 (CH(CH3)C(O)O).
Homopolymerization in Bulk. In a typical homopolymerization
run, a vial (20 mL) was charged sequentially with monomer (2.50
mmol), precatalyst (25 μmol), and MeOH (25 μmol; 0.25 mL of a 0.1
M toluene solution). The vial was put into an oil bath, preheated and
thermostated at 140 °C, and was magnetically stirred. After 75 min,
the vial was allowed to cool at room temperature. Product purification
was obtained by dissolving the reaction mixture in CH2Cl2, followed
by a dropwise addition of this solution to rapidly stirring methanol.
The precipitated polymer was recovered by filtration, washed with
methanol, and dried at 60 °C in a vacuum oven overnight.
Poly(glycolide) = 1H NMR (300 MHz, DMSO-d6, 100 °C): 4.87 (s,
2H; CH2C(O)O), 4.13 (s, 2H; CH2OH), 3.72 (s, 3H; OCH3).
Synthesis of Poly(glycolide-block-rac-lactide). The Schlenk
tube (10 mL) was charged sequentially with rac-lactide (1.25 mmol),
precatalyst (25 μmol; 5 mM in xylenes), xylenes, and MeOH (25
μmol; 0.25 mL of a 0.1 M toluene solution). The Schlenk tube was put
into an oil bath, thermostated at 130 °C. After 4.5 h, glycolide (0.39
mmol) was added as a solid to the reaction mixture. The reaction was
quenched after 10 min by addition of 2 mL of wet CH2Cl2. The
mixture was then added to methanol (20 mL). The precipitated
polymer was recovered by filtration, washed with methanol and dried
1
1
at 60 °C overnight in a vacuum oven. The Mn,NMR evaluated by H
Poly(rac-lactide) = H NMR (300 MHz, DMSO-d6, 100 °C): δ
NMR was 3.7 KDa.
5.25−5.16 (m, 1H; CH(CH3)C(O)O), 4.23 (m, 1H; CH(CH3)OH),
3.70 (s, 3H; CH3O), 1.53−1.45 (m, 3H; CH(CH3)C(O)O), 1.32 (d, J
= 7.0 Hz, 3H; CH(CH3)OH).
1H NMR (300 MHz, DMSO-d6, 100 °C): δ 5.27−5.14 (m, 1H;
CH(CH3)C(O)O), 4.87 (s, 2H; CH2C(O)O), 1.54−1.44 (m, 3H;
CH(CH3)C(O)O). 13C NMR (75 MHz, DMSO-d6): δ 168.3, 168.15,
168.1 (CH(CH3)C(O)O), 165.8 (CH2C(O)O), 68.5, 68.3 (CH-
(CH3)C(O)O), 60.3 (CH2C(O)O), 15.8, 15.7 (CH(CH3)C(O)O).
13C NMR (75 MHz, DMSO-d6): δ 168.3, 168.15, 168.1 (CH(CH3)
C(O)O), 68.5, 68.3 (CH(CH3)C(O)O), 15.8, 15.7 (CH(CH3)C(O)-
O).
Copolymerization in Bulk. In a typical copolymerization run, a
vial (20 mL) was charged sequentially with monomers (total amount =
2.50 mmol, if not stated otherwise), precatalyst (25 μmol) and MeOH
(25 μmol; 0.25 mL of a 0.1 M toluene solution). The vial was put into
an oil bath, preheated, and thermostated at 140 °C, and was
magnetically stirred. The polymerization work-up was performed as
above.
RESULTS AND DISCUSSION
■
Synthesis and Characterization of the Salicylaldimi-
nato Aluminum Complexes. The complexes 1 and 2 were
synthesized in toluene by the alkane elimination reaction
between the corresponding proligand and Al(CH3)3, as
previously described for analogous compounds.24 The
phenoxy-imine proligands have been synthesized following
previously published procedures.22,23
The phenoxy-imine compounds coordinate to the aluminum
atom as monoanionic ligands, yielding the dimethyl compounds
1 and 2 (Scheme 1) and one equivalent of methane. Complexes
Poly(glycolide-co-rac-lactide) = 1H NMR (300 MHz, DMSO-d6,
100 °C): δ 5.34−5.14 (m, 1H; CH(CH3)C(O)O), 4.98−4.71 (m, 2H;
CH2C(O)O), 1.57−1.44 (m, 3H; CH(CH3)C(O)O). 13C NMR (75
MHz, DMSO): δ 168.4, 168.3, 168.2, 168.15, 168.1 (CH(CH3)
C(O)O), 165.8, 165.7 (CH2C(O)O), 68.5, 68.3 (CH(CH3)C(O)O),
60.3, 60.2 (CH2C(O)O), 15.8, 15.7 (CH(CH3)C(O)O).
1H NMR (300 MHz, CDCl3, 25 °C): δ 5.31−5.11 (m, 1H;
CH(CH3)C(O)O), 4.92−4.57 (m, 2H; CH2C(O)O), 1.65−1.52 (m,
3H; CH(CH3)C(O)O). 13C NMR (75 MHz, DMSO): δ 169.6, 169.4
169.3, 169.2 (CH(CH3)C(O)O), 166.4, 166.74 (CH2C(O)O), 69.3,
69.2,69.0 (CH(CH3)C(O)O), 60.9, 60.8, 60.7 (CH2C(O)O), 16.7,
16.6 (CH(CH3)C(O)O).
Scheme 1. Synthetic Route for Complexes 1 and 2
Synthesis of Low Molecular Weight Poly(glycolide-co-rac-
lactide). The copolymers were prepared as above, but 0.50 mmol of
glycolide and 0.50 mmol of rac-lactide were used.
1H NMR (300 MHz, DMSO-d6, 100 °C): δ 5.34−5.14 (m, 1H;
CH(CH3)C(O)O), 4.98−4.71 (m, 2H; CH2C(O)O), 4.23 (m, 1H;
CH(CH3)OH), 4.29−4.18 (m, 1H; CH(CH3)OH), 4.13 (s, 2H;
CH2OH), 4.09 (m, 2H; CH2OH), 3.72 (s, 3H; OCH3), 3.70 (s, 3H;
CH3O), 1.57−1.44 (m, 3H; CH(CH3)C(O)O), 1.32 (d, J = 7.0 Hz,
3H; CH(CH3)OH). 13C NMR (75 MHz, DMSO-d6): δ 168.4, 168.3,
168.2, 168.15, 168.1 (CH(CH3)C(O)O), 165.8, 165.7 (CH2C(O)O),
68.5, 68.3 (CH(CH3)C(O)O), 60.3, 60.2 (CH2C(O)O), 59.1
(CH2OH), 15.8, 15.7 (CH(CH3)C(O)O).
1 and 2 were subsequently recovered, by evaporation of the
solvent in vacuo, as yellow powders in good yields (1, 92%; 2,
83%). They were fully characterized by multinuclear NMR
spectroscopy.
1H NMR (300 MHz, CDCl3, 25 °C): δ 5.34−5.10 (m, 1H;
CH(CH3)C(O)O), 4.95−4.55 (m, 2H; CH2C(O)O), 4.46−4.34 (m,
1H; CH(CH3)OH), 4.30 (s, 2H; CH2OH), 4.28−4.23 (m, 2H;
CH2OH), 3.72 (s, 3H; OCH3), 3.70 (s, 3H; OCH3), 1.65−1.48 (m,
3H; CH(CH3)C(O)O), 1.32 (d, J = 7.0 Hz, 3H; CH(CH3)OH).
Copolymerization in Solution. In a typical polymerization run, a
Schlenk tube (10 mL) was charged sequentially with monomer(s)
(total = 5.00 mmol), precatalyst (25 μmol; 5 mM in the solvent), the
solvent and MeOH (25 μmol; 0.25 mL of a 0.1 M toluene solution).
The Schlenk tube was put into an oil bath, preheated and
thermostated at the desired temperature, and was magnetically stirred.
After the established time, the mixture was cooled to room
temperature. Product purification was attained by dropwise addition
of the reaction mixture, dissolved in CH2Cl2, to rapidly stirring
methanol. The precipitated polymers were recovered by filtration,
washed with methanol and dried at 60 °C overnight in a vacuum oven.
1
The H and 13C NMR spectra of the obtained products
indicated the formation of the desired complexes 1-2 bearing
1
one salicylaldiminato ligand and two methyl groups. In the H
NMR spectra sharp singlets at −0.28 ppm and −0.57 ppm,
respectively for complexes 1 and 2, were observed for the
methyl protons of the Al(CH3)2. The pattern of the protons of
the salicylaldiminato ligands was unequivocally recognized in
each spectrum and showed significant shifts with respect to the
signals of the protons of free proligands (see Supporting
Information). Accordingly, the 19F NMR spectra showed three
signals for the ortho, meta, and para-fluorine atoms on the
aromatic ring bound to the nitrogen. 13C NMR characterization
was coherent with these data showing, in particular, signals at
C
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
−9.2 and −10.3 ppm, respectively for complexes 1 and 2, for
the methyl carbons on the aluminum (see Supporting
Information).
Table 2. Molecular Weights and Dispersities of the Homo-
and Copolymer Samples Obtained in Bulk
a
b
c
d
e
run
catalyst
FGA
Mn,th (kDa)
Mn,NMR (kDa)
Mw/Mn
Homo- and Copolymerization of Glycolide and rac-
Lactide in Bulk. Complexes 1 and 2 were tested in the ring-
opening copolymerization of rac-lactide and glycolide carrying
out the reaction under several experimental conditions. The
homo- and copolymerizations of glycolide and rac-lactide were
first performed in bulk at 140 °C in the presence of catalysts 1
or 2 and one equivalent of methanol. The obtained polymer
samples were characterized by 1H and 13C NMR spectroscopy,
GPC and DSC analysis. The main results are summarized in
Tables 1 and 2.
f
g
2
1
2
1
1
1
2
2
2
2
2
2
2
1
2
1
2
0
11.0
13.9
12.2
12.7
8.9
8.9
13.9
9.2
1.6
f
g
4
0
1.5
5
70
51
30
81
72
59
53
41
34
22
50
55
53
53
2.4
2.2
1.9
1.3
1.1
1.4
1.4
1.5
1.3
1.8
6
9.4
7
15.4
10.7
9.5
8
11.7
10.3
12.0
12.2
10.7
12.2
12.2
5.2
9
10
11
12
13
14
8.3
8.5
7.6
7.3
Table 1. Homo- and Copolymerization of Glycolide and rac-
8.1
a
Lactide in Bulk
h
g
15
4.3
2.0
h
g
16
5.2
5.5
1.4
yield
(%)
b
c
d
d
e
e
i
run catalyst f GA
FGA
LGG
LLL
TLGL
TGLG
17
34.1
32.7
27.2
19.0
1.6
2.0
i
18
1
1
1
2
2
1
1
1
2
2
2
2
2
2
2
100
>99
76
>99
92
78
75
64
89
83
92
77
81
89
74
100
−
100
−
−
−
−
3.55
1.67
1.17
6.13
3.44
2.29
2.05
1.40
1.14
1.18
−
−
−
−
1.52
1.61
2.72
1.44
1.34
1.59
1.82
2.01
2.21
3.05
−
−
−
−
1.59
1.19
0.71
4.39
1.85
1.34
1.07
0.90
0.94
0.74
−
−
−
−
0.01
0.07
0.11
0.10
0.08
0.10
0.15
0.12
0.15
0.32
a
2
0
100
0
Polymerization conditions: precatalyst = 25 μmol; MeOH = 25 μmol
(0.25 mL of a 0.1 M toluene solution); T = 140 °C; t = 75 min; mol
ratio of monomer(s) to precatalyst in the feed = 100. FGA, content of
3
b
4
−
c
glycolide in the copolymer (mol %). Theoretical molecular weight.
5
70
50
30
80
70
60
50
40
30
20
70
51
30
81
72
59
53
41
34
22
d
e
Molecular weight determined by 1H NMR. Molecular weights
6
dispersities determined by gel permeation chromatography (GPC) vs
polystyrene standards, elution solvent mixture: chloroform/HFIP 99/
1. Poly(rac-LA). Elution solvent: tetrahydrofuran (THF). The mol
ratio of monomer(s) to precatalyst in the feed = 40. The mol ratio of
7
8
f
g
h
9
i
10
11
12
13
14
monomer(s) to precatalyst in the feed = 300.
a
Polymerization conditions: precatalyst = 25 μmol; MeOH = 25 μmol
(0.25 mL of a 0.1 M toluene solution); T = 140 °C; t = 75 min; mol
b
ratio of monomer(s) to precatalyst in the feed = 100. Molar
c
percentage of glycolide in the feed. FGA, molar percentage of glycolide
1
in the copolymer, as determined by H NMR (DMSO-d6, 100 °C).
d
Average length of glycolidyl (GG) and lactydyl (LL) blocks in the
e
copolymer; calculated from 13C NMR (DMSO-d6, 100 °C). Yield of
the second mode of transesterification (%) of glycolidyl (LGL) and
lactydyl (GLG) sequences; calculated from H NMR (DMSO-d6, 100
1
°C).
For both the catalysts, after 75 min of reaction, full
conversion in the homopolymerization of glycolide was
assessed; almost complete conversion of rac-lactide was reached
in the same time (Table 1, runs 1−4).
The copolymerizations were performed systematically
varying the comonomers ratio and the monomer/catalyst
feed ratio, and almost complete monomer conversion was
reached in 75 min with both the catalysts. As shown in Table 1,
1
the composition of the copolymers, evaluated by the H NMR
spectrum, parallels the feed ratio, as it would be expected for a
copolymer at full conversion.
Because the chemical shifts of the carbonyl carbons are
highly sensitive to their surroundings,25 a detailed micro-
structure characterization of the copolymer chain of the
samples was achieved through inspection of the 13C NMR
spectra.
Figure 1. 13C NMR (75 MHz, DMSO-d6, 100 °C) spectra in the
carbonyl region of polymers obtained with complex 1: (i) poly(rac-
lactide) (Table 1, run 2); (ii) poly(glycolide-co-rac-lactide), FGA = 30
(Table 1, run 7); (iii) poly(glycolide-co-rac-lactide), FGA = 51 (Table 1,
run 6); (iv) poly(glycolide-co-rac-lactide), FGA = 70 (Table 1, run 5).
The carbonyl regions of the 13C NMR spectra (DMSO-d6,
100 °C) of the copolymer samples prepared with catalyst 1
with different monomers feed (Table 1, runs 5−7) are shown in
Figure 1. For comparison, the 13C NMR spectrum (DMSO-d6,
100 °C) of a poly(rac-lactide) prepared in the same conditions
(Table 1, run 2) is also shown (Figure 1i). Providing that L and
D
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
G represent respectively a lactyl −CH(CH3)C(O)O− and a
glycolyl −CH2−C(O)O− unit, two resonances attributable to
the hetero- and homosequences centered on the carbonyl of
the glycolyl (GGLL at δ 165.7 ppm and GGGG at δ 165.8
ppm) were observed, accordingly with the literature.26 At lower
field, in the region centered on the carbonyl of the lactyl group,
five resonances were observed. As previously reported, the
resonance at δ 168.4 ppm was attributed to the heterosequence
LLGG, while the resonance at δ 168.2 ppm was attributed to
the GLG sequence.26 The latter sequence cannot be formed by
ring-opening of lactide and glycolide during the chain growth,
but it derives from the transesterification of the second mode,
during which the lactidyl and glycolidyl units undergo bond
cleavage. Indeed, the GLG sequence could be generated by a
transesterification reaction involving the attack of an active
glycolidyl chain end −GGAl* on a preformed LLGG sequence
(Scheme 2a).
Figure 2. Plot of average length of glycolidyl (GG) blocks vs
copolymer composition ratio (G/L) for the copolymers obtained with
complexes 1 and 2 (Table 1, runs 5−7 and 8−14).
Signals at 4.83 ppm were attributed, according to the literature,
to the presence of LGL sequences.5,28 These sequences are
formed by transesterification reaction of glycolidyl segments by
active lactidyl chain end (Scheme 2b). The amount of
transesterification sequences LGL and GLG (see above) have
been evaluated by using the coefficients of the second mode of
transesterification, TLGL and TGLG, as previously reported.5,16
According to the definitions, the TLGL and TGLG values are close
to 1 when the contribution of glycolyl and lactyl units in the
chain are close to Bernoullian statistics, while they are higher
than 1 when longer alternated sequences are present in the
chains. The TLGL values increase by increasing the amount of
glycolide in the feed, and values higher than 1 are calculated for
the copolymers obtained when the molar percentage of
glycolide in the feed is higher than 50%.
For both the catalysts, the TLGL values are higher than the
TGLG ones of 1 order of magnitude, thus indicating that the
transesterification reaction involving the attack of active lactidyl
chain end on preformed glycolidyl segments is preferred
(Scheme 2b). Perusal of the literature showed only a similar
precedent in the copolymers obtained in the presence of
butyllitium, that also contained higher amount of TLGL with
respect to the TGLG ones.13 This behavior is definitely in
contrast with previous results obtained with the classical
Sn(Oct)2 catalyst, and with Zr(acac)216 or Fe based catalysts,17
where the TGLG values were higher than the TLGL ones.
This feature can be tentatively explained taking into account
that in the homopolymerization of rac-lactide by this class of
aluminum catalysts, transesterification reactions were absent.20b
It is therefore confirmed that the tendency of these complexes
to break the lactidyl unit into two lactyl fragments is low.
Overall, the two initiators showed roughly analogous
behavior in the polymerization performed in bulk. An accurate
analysis of the copolymerization results, however, showed that
transesterifications of the second mode were slightly higher for
catalyst 2, bearing a bulky cumyl groups as ortho-phenoxy
substituents. Probably the steric hindrance of this group could
have an influence on the relative rate of chain propagation and
transesterification reaction.
Scheme 2. Transesterification Processes Occurring During
the Copolymerization
The remaining three resonances (at δ 168.3, 168.15, 168.1
ppm) are attributable to the different stereochemical
combination of the LLLL homosequence.27 It is worth to
note that in the homopolymerization of rac-lactide, in the
absence of a stereoselective catalyst/initiator, 11 hexads or 5
tetrads of stereochemical sequences, resulting from the addition
of D- and L-lactide molecules, are expected. However, due to
either insufficient resolution or overlapping of the chemical
shifts in DMSO-d6 solvent, only 3 resonances are observed. The
same resonances appeared in the 13C NMR spectrum (DMSO-
d6) of the poly(rac-lactide) prepared with the same catalyst
(Figure 1i). For this sample the methine region of the
27
1
homonuclear decoupled H NMR spectrum in CDCl3 (see
Figure S5) showed a slightly isotactically enriched micro-
structure (Pm = 0.64), as previously observed for the poly(rac-
lactide) obtained in toluene solution with similar salicylaldimi-
nato aluminum catalysts.20b
The average lengths of glycolidyl and lactidyl blocks (LGG
and LLL) were also calculated from the 13C NMR spectra, by
using previously reported equations.12 The so-calculated
lengths were confirmed by using as control the monomers
12
1
composition ratio (G/L) evaluated by H NMR.
The average block lengths linearly depend on the copolymer
composition ratio (Figures 2 and Supporting Information) and
the monomers feed. As a result, the copolymers microstructure
could be easily tuned by adjusting the feed.
Interestingly, with catalyst 2, in the case of a 50 to 50
monomer feed composition (Table 1, run 11), the LGG and LLL
values were close to a value of 2, as expected for a random
copolymer.
The final microstructure of copolymer chain should
reasonably result from the reactivity of comonomers as well
as transesterification processes taking place together with the
main copolymerization reaction. In particular, the main
transesterification process operating in this system is involving
the attack of active lactidyl chain end on preformed glycolidyl
segments.
More information on the copolymers microstructures can be
1
derived from the H NMR spectra in the methylene region.
E
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
End Groups Analysis by NMR. In order to get more
information on the mechanism involved in these copolymeriza-
tion reactions an accurate end group analysis was carried out by
1H NMR spectroscopy in DMSO-d6 at 100 °C. For this
purpose, low molecular weight copolymer samples were
prepared by conversion of 20 equiv of each monomer. The
assignment of the different end groups was made by
comparison with the spectra of the homopolymer samples
(Figure 3, parts i and ii) and literature data.9 Although PGA
preference is more significant with catalyst 2 (Figure 3iv)
suggesting a stronger discrimination in favor of the less
hindered monomer by the most encumbered complex.
The most abundant hydroxyl end groups, generated by
hydrolysis of the growing chain, were the HOCH2C(O)-
OCH2− (HOGG−) (4.13 ppm) and the HOCH2C(O)OCH-
(CH3)− (HOGL−) (4.09 ppm) groups (Figure 3, parts iii and
iv). In particular, the latter may only derive from trans-
esterification reactions generating the LGL sequence (Scheme
2b). As a matter of fact, this kind of transesterifications was the
most abundant for the explored aluminum catalysts (see
above). Accordingly, signals due to the hydroxyl end groups
bound to a lactyl unit (HOLG−) (4.20 ppm), which can be
only generated through the less abundant transesterification
reactions depicted in Scheme 2a, always showed a negligible
intensity.
Finally, the low intensity observed for the HOLL− (4.23
ppm) groups may be rationalized taking into account that the
Al−lactidyl active centers, from which these end groups can be
generated, are the most involved in the transesterification
reactions.
The whole picture suggests that a coordination−insertion
mechanism, proceeding through acyl−oxygen cleavage of both
the monomers, should be operative in these systems. The
occurrence of the different transesterification reactions with the
relative frequencies detailed above well explains the relative
ratio of the observed end groups.
Determination of the Molecular Weight of the
Samples. The molecular weights of the obtained polymers
were evaluated by gel permeation chromatography (GPC) and
by NMR, being known the end group signals (see above).
Representative results are reported in Table 2.
The molecular weight of the poly(rac-lactide)s, evaluated by
GPC vs polystyrene standards, using THF as elution solvent,
corrected by a factor of 0.58,29 resulted Mn,GPC = 11.4 kDa for
run 2 and 12.2 kDa for run 4. Monomodal molecular weight
distributions were observed. A good agreement between the
molecular weights evaluated by NMR, Mn,NMR, and the
theoretical molecular weights, Mn,th, calculated by the
monomer/catalyst feed ratio was observed. On the contrary,
the assessment of the molecular weights for the PGA
homopolymers (see runs 1 and 3, Table 1) was not possible
by either GPC or NMR analysis, since the polymers are
insoluble in almost all solvents.25
Figure 3. 1H NMR (300 MHz, DMSO-d6, 100 °C) spectra: (i)
poly(glycolide) obtained with complex 1 (Table 1, run 1); (ii)
poly(rac-lactide) obtained with complex 1 (Table 1, run 2); (iii)
poly(glycolide-co-rac-lactide) obtained with complex 1 (Table 2, run
15); (iv) poly(glycolide-co-rac-lactide) obtained with complex 2
(Table 2, run 16).
The determination of the molecular weights of the
poly(glycolide-co-rac-lactide) samples was performed by GPC
in a chloroform/HFIP 99/1 solvents mixture. Low molecular
weight samples, prepared by a lower monomer/inititator feed
ratio, dissolved even in THF, therefore in these cases the GPC
analysis was performed by using THF as eluent. As previously
highlighted in the literature,7 the radius of gyration Rg of the
poly(glycolide-co-rac-lactide) samples is extremely sequence
and solvent dependent, thus the values obtained by GPC
should be regarded with special care. However, the GPC
analysis performed on all the samples disclosed monomodal
molecular weight distributions with variable molecular-weight
dispersities (1.1−2.4). In detail, catalyst 1 generally produced
polymers having narrower dispersities than those obtained with
catalyst 2; this behavior should be related to the presence of the
bulkier cumyl substituent on the phenolate ring, which should
hamper the transesterification reactions.
homopolymer is scarcely soluble in common solvent,
olygomers of PGA were found to be soluble in DMSO-d6 at
100 °C, thus allowing the analysis of end groups.
Easily recognizable were the singlets due to the terminal
alkoxide −OCH3 group. Similar signals in the homopolymer
spectra (Figure 3, parts i and ii) allowed us to distinguish the
−CH2C(O)OCH3 (G−OCH3) and the −CH(CH3)C(O)-
OCH3 (L−OCH3) end groups in the copolymers. The
presence of both signals indicated that the first step of these
copolymerization reactions can be the insertion of either the
glycolide unit or the lactide unit into the Al−OCH3 bond.
Although the partial overlapping of these signals did not permit
an exact estimation of their relative abundance, it is possible to
claim that, for both complexes, the preferred first step is the
insertion of the glycolide monomer into the Al−OCHH3 bond.
This is in agreement with the higher reactivity of this monomer
with respect to that of lactide. Moreover, the observed
Interestingly, the molecular weights evaluated by NMR are in
reasonable agreement with the theoretical molecular weights,
F
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
Mn,th, obviously indicating that the molecular weight could be
easily tuned by adjusting the monomer/catalyst feed ratio. As a
matter of fact, lower molecular weight samples (Mn ≈ 5 kDa;
Table 1, runs 15−16) and higher molecular weight samples
(Mn ≈ 20−30 kDa; Table 1, runs 17−18) have been obtained
by adjusting the monomers feed.
Thermal Characterization of Poly(glycolide-co-rac-
lactide). Thermal analysis of the copolymers was carried out
by means of differential scanning calorimetry (DSC), from −20
to +260 °C. The glass transition temperature, Tg, and the
melting temperature, Tm, are given in Table 3.
previously reported cases of poly(glycolide-co-rac-lactide) with
a content of glycolide of 80 mol % or higher.3,16,17
The DSC thermograms recorded during the second scan for
all the samples displayed a unique glass transition temperature
with values intermediate between those of the pure
homopolymers and changing as a function of the composition.
The experimental Tg values linearly increase by decreasing the
glycolide content in the copolymer (Figures 4 and Supporting
Information), which in turn reflects the feed composition. It is
noteworthy to highlight that it is possible to control the
microstructure and the thermal properties, by choosing the
appropriate monomers feed.
Copolymerization of Lactide and Glycolide in
Solution. Copolymerization of glycolide and lactide were
also performed in solution. In order to elucidate the influence
of the reaction conditions on yields, molecular weights and
composition of the copolyesters, the following experimental
parameters were systematically varied: (i) nature of the catalyst,
(ii) nature of the solvent, (iii) temperature, and (iv) reaction
time. In all the cases equimolar amounts of the two monomers
were used.
Table 3. Homo- and Copolymerization of Glycolide and rac-
Lactide in Bulk: Thermal Properties
a
b
c
d
d
d
run
3
catalyst
f GA
FGA
Tg (°C)
Tm (°C)
ΔHm (J g−1
)
2
2
1
1
1
2
2
2
2
100
0
100
0
n.o.
48.3
43.5
47.2
49.2
41.2
40.7
45.8
51.4
222.6
n.o.
83.8
n.o.
n.o.
n.o.
n.o.
50.3
n.o.
n.o.
n.o.
e
4
5
70
50
30
80
60
40
20
70
51
30
81
59
41
22
n.o.
6
n.o.
7
n.o.
1
8
201.3
n.o.
The obtained polymeric samples were characterized by H
and 13C NMR, GPC, and DSC analysis. The main results are
summarized in Tables 4 and 5.
10
12
14
n.o.
n.o.
Carrying out the polymerization experiments in toluene at 90
°C, only poly(glycolide) was obtained in agreement with the
higher reactivity of glycolide. Polymerization tests performed at
90 °C with catalyst 1 in two solvents of different polarity,
namely chlorobenzene and xylenes, afforded copolymers with
higher incorporation of glycolide and very long glycolide
sequences (Table 4, runs 20 and 22), thus confirming the
higher reactivity of glycolide. The boiling points of these
solvents allowed to carry out the polymerization runs at higher
temperatures (see Table 4). In these cases, polymeric samples
with a glycolide content ranging between 49 and 66% were
obtained showing that, in these experimental conditions,
comparable incorporation of both monomers is obtained.
The randomness of these copolymers was assessed by the
calculation of the block lengths of rac-lactide and glycolide
within the copolymer chain, while the occurrence of the second
mode of transesterification was quantified by analysis of the
signals due to the GLG and LGL groups in the NMR spectra of
the copolymers.5,16
a
Polymerization conditions: precatalyst = 25 μmol; MeOH = 25 μmol
(0.25 mL of a 0.1 M toluene solution); T = 140 °C; t = 75 min; mol
ratio of monomer(s) to precatalyst in the feed = 100. f GA, molar
b
c
percentage of glycolide in the feed. FGA. content of glycolide in the
copolymer (mol %), as determined by 1H NMR (DMSO-d6, 100 °C).
d
e
Values reported for the second heating cycle. FLA = 100. n.o. = not
observed.
In Figure 4 are shown the thermograms of the poly-
(glycolide) and of poly(glycolide-co-rac-lactide) samples
For copolymers produced by complex 1 in the two different
solvents, the average lactide and glycolide block lengths were
higher than 2 (Table 4, runs 21, 23), indicating a nonrandom
copolymer chain. DSC analysis of the copolymers obtained by
complex 1 showed two Tg values (Table 5, runs 21, 23; Figures
S11−S12). Melting endotherm of glycolide blocks crystalline
phase was observed for the copolymer obtained in run 21,
whereas the sample obtained in run 23 should have shorter
glycolide blocks, in agreement with the presence of the GLG
sequences. The whole picture is compatible with the formation
of copolymer samples with a blocky structure showing a first
sequence comprising glycolidyl blocks separated by short lactyl
and lactidyl groups and a second part of the chain with a
complementary distribution of the two monomers.
Copolymers obtained by complex 2 at higher temperature
showed lower LGG and LLL values, and higher second mode of
transesterification values (Table 4, runs 25−26). In particular,
average lactide and glycolide block lengths for the copolymer
obtained in run 26 was close to 2, the value expected for a
completely random copolymer. Accordingly, these samples
Figure 4. DSC thermograms (run II) of poly(glycolide-co-rac-lactide)
obtained with complex 2: (a) FGA = 100 (Table 3, run 3); (b) FGA = 81
(Table 3, run 8); (c) FGA = 59 (Table 3, run 10); (d) FGA = 41 (Table
3, run 12); (e) FGA = 22 (Table 3, run 14).
obtained with catalyst 2. The thermogram of the poly-
(glycolide) displays a melting peak at 226 °C, with endotherm
of fusion of 83.8 J·g−1 (Figure 4a). The Tg of this
poly(glycolide) sample was not observed with our analytical
settings in agreement with the thermal behavior of a medium
molecular weight poly(glycolide) (Mw = 10−20 kDa).30
All the copolymers were amorphous, apart from the sample
prepared with 80 mol % of glycolide (Table 3, run 8); in this
case a crystallization exotherm and a melting peak can be seen
in the thermogram. This observation is in agreement with
G
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
a
Table 4. Copolymerization of rac-Lactide and Glycolide Promoted by Complexes 1 and 2 in Solution
b
c
c
d
d
run
catalyst
solvent
toluene
T (°C)
yield (%)
FGA
LGG
LLL
TLGL
TGLG
19
20
21
22
23
24
25
26
1
1
1
1
1
2
2
2
90
90
45
68
66
43
66
41
67
79
99
72
54
90
59
89
66
49
n.d.
n.d.
n.o.
n.o.
0.05
n.o.
0.94
0.13
n.o.
0.08
0.09
chlorobenzene
chlorobenzene
xylenes
5.91
2.94
7.23
2.96
n.d.
2.32
2.50
0.80
2.06
n.d.
0.28
0.63
n.o.
120
90
xylenes
130
90
0.73
n.o.
toluene
chlorobenzene
xylenes
120
130
2.75
1.94
1.32
2.02
1.39
0.95
a
Polymerization conditions: precatalyst = 25 μmol; MeOH = 25 μmol (0.25 mL of a 0.1 M toluene solution); solvent = 5 mL; glycolide = 2.50
b
1
mmol, rac-lactide = 2.50 mmol, t = 180 min. FGA, content of glycolide in copolymer (mol %), as determined by H NMR (DMSO-d6, 100 °C).
c
d
Average sequences length of glicolidyl (GG) and lactidyl (LL) blocks in the copolymer; as calculated by 13C NMR (DMSO-d6, 100 °C). Yield of
the second mode of transesterification (%) of glycolidyl (LGL) and lactydyl (GLG) sequences; calculated from 1H NMR (DMSO-d6, 100 °C). n.d. =
not determined; n.o. = not observed.
Table 5. Molecular Characterization and Thermal Analysis
of PLGAs Prepared with Complexes 1 and 2 in Solution
Block Copolymerization. The synthesis of block copoly-
mers was attempted by using catalyst 1 in xylenes at 130 °C.
The block copolymer was obtained by sequential addition of
the two monomers, polymerizing first the rac-lactide. After 4.5
h an aliquot was withdrawn from the reaction mixture to assess
the molecular weight of the poly(lactide) block by NMR (2.5
kDa). The addition of the glycolide monomer to the mixture
yielded the product in 10 min, and the precipitated polymer
was analyzed by NMR.
a
M
Mn,NMR
c
Tg
T
ΔH
n,thb
d
e
m e
m e
run
20
(kDa)
(kDa)
Mw/Mn,GPC
(°C)
(°C)
(J g−1
36.7
)
21.4
22.3
1.3
41.3;
51.1
199.0
23
23.6
22.8
1.3
42.0;
52.7
n.o.
n.o.
25
26
21.0
26.0
14.9
21.4
1.3
42.7
45.9
187.7
n.o.
6.9
The 13C NMR analysis showed the exclusive presence of the
carbonyl signals attributed to the homosequences LLLL and
GGGG. Signals due to transesterification processes were
n.d.
n.o.
a
General conditions: precatalyst = 25 μmol; MeOH = 25 μmol (0.25
mL of a 0.1 M toluene solution); solvent = 5 mL; glycolide, 2.50
b
1
mmol, rac-lactide = 2.50 mmol; t = 180 min. Theoretical molecular
negligible. The H NMR analysis of the copolymer confirmed
c
1
weight. Molecular weight determined by H NMR (DMSO-d6, 100
the reflection of the feed in the copolymer composition (see
Supporting Information). The lengths of the glycolidyl and
lactidyl blocks were determined by evaluation of the integrals of
the main signals, and were found to be as follows: LGG = 15;
LLL = 31. The 1H NMR analysis, moreover, showed the
exclusive presence of end groups LL−OCH3 (at 3.70 ppm),
derived from the insertion step of the rac-lactide monomer into
the Al−OCH3 bond, and the HOGG− end group (at 4.13
ppm), generated by hydrolysis of the growing poly(glycolide)
block.
d
°C). Determined by gel permeation chromatography (GPC) vs
polystyrene standards, elution solvent mixture: chloroform/HFIP 99/
1. Values reported for the second heating cycle. n.d. = not
e
determined; n.o. = not observed.
displayed unique glass transition temperature (Table 5, runs
25−26; see also Supporting Information), as observed in the
copolymerizations performed in bulk (see above).
To get more insights on the copolymerization behavior of
catalyst 2, the effect of the polymerization time was studied. A
Formation of the poly(rac-lactide-block-glycolide) copolymer
was definitely proved by DOSY NMR experiment (Figure 5).
polymerization run, performed in the same conditions (f GA
=
50) than run 26 in Table 4, was quenched at low conversion (t
= 0.5 h). Comparison of the two products showed that in the
beginning of the polymerization the LGG (2.95) are higher than
LLL (0.89), thus indicating that glycolide is polymerized first
(FGA = 70). At higher conversion the LGG and LLL values were
close to 2, as expected for a random copolymer. Thus, the
transesterification reactions, taking place during the polymer-
izations, are mainly responsible of the random structure.
1
End group analysis performed by H NMR spectroscopy on
the obtained copolymers showed polymer chains end-capped
with a methyl ester and a hydroxyl group.
The molecular weights of selected samples were evaluated by
GPC in the chloroform/HFIP 99/1 solvents mixture. As
discussed above, the GPC analysis is not reliable for the
assessment of the real molecular weight of the glycolide/lactide
copolymers. Nevertheless, the GPC results indicated fine
dispersities (Mw/Mn = 1.3), narrower than those of the
copolymers prepared in bulk (Mw/Mn = 1.1−2.4; see Table 2).
Molecular weights calculated from 1H NMR spectra are in good
agreement with the theoretical values for the samples obtained
with catalyst 1, while they are lower for the samples obtained
with catalyst 2.
Figure 5. 2D DOSY NMR (400 MHz, DMSO-d6, 80 °C) of the block
copolymer obtained with compound 1. Signals at 2.50 and 3.06 ppm
are relative to the deuterated solvent (DMSO-d6) and adventitious
water, respectively.
H
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
This experiment, indeed, providing diffusion coefficients of
molecules related to hydrodynamic radius and molecular
weight, is becoming a very powerful tool in investigating
polymer properties.31 In our case, the DOSY spectrum of the
sample obtained by the block copolymerization reaction
showed that the multiplets of the poly(rac-lactide) block
(centered at 5.20 and 1.49 ppm) and the singlet of the
poly(glycolide) block (at 4.87 ppm) lied at the same diffusion
coefficient, and therefore belonged to the same polymeric
chains.
The molecular weight estimated by NMR was found to be
close to the theoretical one (Mn,NMR = 3.7 kDa vs Mn,th = 3.1
kDa). DSC analysis (see Supporting Information) evidenced the
presence of only one Tg at 45.4 °C, attributable to the rac-
lactide block, while no Tg was observed for the homoglycolide
block, as observed above for the poly(glycolide) (see Figure 4a;
Table 3, run 3).
Thus, the sequential addition of the two monomers leads to
the achievement of a poly(rac-lactide-block-glycolide) copoly-
mer, which represents an experimental evidence of the
tendentially living behavior of the polymerization promoted
by this class of initiators.
It was found, moreover, that, in order to obtain the block
copolymer, glycolide had to be added to living PLA chains. The
opposite sequence of monomers addition led mainly to
poly(glycolide) and a low amount of the block copolymer.
The importance of the order of the monomer addition in the
block copolymerization was previously underlined in the
literature.10,32
In all the cases, GPC analysis disclosed monomodal
molecular weight distribution with narrow dispersities. A
reasonable agreement between the theoretical molecular
weights and the experimental ones evaluated by NMR analysis
was observed. As a result a controlled polymerization behavior
can be recognized in the polymerization, and the copolymers
molecular weight could be adjusted by regulating the
monomers/initiator feed ratio.
We can envisage that these results should be of interest in
applications where modulated thermal, physical and degrada-
tion properties of PGA/PLA based materials are required.
ASSOCIATED CONTENT
* Supporting Information
■
S
NMR spectra of complexes 1 and 2, NMR spectra and DSC
thermograms of selected polymers, plot of LLL vs L/G, data of
copolymerization of L-LA and GA. This material is available free
of charge via the Internet at http://pubs.acs.org/.
AUTHOR INFORMATION
Corresponding Author
*(D.P.) Telephone: +39 0824323632. E-mail: pappalardo@
unisannio.it.
■
Author Contributions
§These authors equally contributed to the paper.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
CONCLUSIONS
■
The authors thank Dr. Patrizia Oliva for NMR technical
assistance, Dr. Ilaria D’Auria for GPC technical assistance and
Miss Alessandra Carbone for some synthetic preparations. This
work was supported by the Italian Ministry of University and
Research (PRIN 2010-2011: Nanostructured polymeric
materials with tailored molecular and crystalline structures,
for advanced technologies and for the environment).
The copolymerization of glycolide and rac-lactide produces
biocompatible and biodegradable materials, which have long
been of interest for biomedical application. In this regards the
search for efficient ROP initiators for the synthesis of PLGA
copolymers having controlled composition and microstrure is a
very stimulating field.
We have shown that salicylaldiminato aluminum compounds
are efficient initiators in the homo- and copolymerization of
rac-lactide and glycolide. A highly versatile behavior has been
recognized, and PLGAs having different microstructures, from
random to blocky to multiblock, have been obtained as the
polymerization conditions have been changed.
Copolymerization in bulk produced random copolymers,
whose average block lengths linearly increase with the
monomer feed ratio. The copolymers were amorphous, and
their Tg could be nicely modulated by the feed.
DEDICATION
■
Dedicated to Prof. Adolfo Zambelli, on the occasion of his 80th
birthday
REFERENCES
■
(1) (a) Chiellini, E.; Solaro, R. Adv. Mater. 1996, 8, 1375−1381.
(b) Albertsson, A.-C.; Varma, I. K. Biomacromolecules 2003, 4, 1466−
1486. (c) Dechy-Cabaret, O.; Martin-Vaca, B.; Bourissou, D. Chem.
Rev. 2004, 104, 6147−6176.
The copolymer microstructure reasonably should result from
transesterification processes taking place together with the main
copolymerization reaction. Interestingly, the values of the
coefficients of the second mode of transesterification TLGL were
higher than the TGLG ones of 1 order of magnitude, thus
indicating that the transterification reaction involving the attack
of active lactidyl chain end on preformed glycolidyl segments
was preferred. On the contrary, as previously observed, the
tendency of these complexes to break the lactidyl unit into two
lactyl fragments was low. Such a behavior is in contrast with
previously reported ROP initiators for the synthesis of PLGAs.
On the contrary, copolymerization performed in several
solvents afforded mainly blocky copolymers, with sequence
blocks lengths higher than 2. Finally, the sequential addition of
the two monomers afforded diblock copolymers, thus
indicating a certain living character of the polymerization.
(2) (a) Handbook of Ring-Opening Polymerization; Dubois, P.,
Coulembier, O., Raquez, J. M., Eds.; Wiley-VCH Verlagh GmbH &
Co.: KGaA: Weinheim, Germany, 2009. (b) O’Keefe, B. J.; Hillmyer,
M. A.; Tolman, W. B. J. Chem. Soc., Dalton Trans. 2001, 2215−2224.
(c) Thomas, C. M. Chem. Soc. Rev. 2010, 39, 165−173.
(3) Gilding, D. K.; Reed, A. M. Polymer 1979, 20, 1459−1464.
(4) Grijpma, D. W.; Nijenhuis, A. J.; Pennings, A. J. Polymer 1990, 31,
2201−2206.
(5) Kasperczyk, J. Polymer 1996, 37, 201−203.
(6) Vert, M.; Schwach, G.; Engel, R.; Coudane, J. J. Controlled Release
1998, 53, 85−92.
(7) Stayshich, R. M.; Meyer, T. Y. J. Am. Chem. Soc. 2010, 132,
10920−10934.
(8) Li, J.; Stayshich, R. M.; Meyer, T. Y. J. Am. Chem. Soc. 2011, 133,
6910−6913.
(9) Dong, C.-M.; Qiu, K.-Y.; Gu, Z.-W.; Feng, X.-D. J. Polym. Sci.,
Part A: Polym. Chem. 2000, 38, 4179−4184.
I
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX

Macromolecules
Article
(10) Dong, C.-M.; Qiu, K.-Y.; Gu, Z.-W.; Feng, X.-D. J. Polym. Sci.,
Part A: Polym. Chem. 2001, 39, 357−367.
Soc. 2011, 133, 13828−13831. (c) Li, W.; Chung, H.; Daeffler, C.;
Johnson, J. A.; Grubbs, R. H. Macromolecules 2012, 45, 9595−9603.
́ ̂
(32) Barakat, I.; Dubois, P.; Grandfils, C.; Jerome, R. J. Polym. Sci.,
(11) Kricheldorf, H. R.; Jonte,
́
J. M.; Berl, M. Makromol. Chem. 1985,
Part A: Polym. Chem. 2001, 39, 294−306.
Suppl. 12, 25−38.
(12) Kricheldorf, H. R.; Kreiser, I. Makromol. Chem. 1987, 188,
1861−1873.
́
(13) Dobrzynski, P.; Kasperczyk, J.; Jelonek, K.; Ryba, M.; Walski,
M.; Bero, M. J. Biomed. Mater. Res., Part A 2006, 79, 865−873.
(14) Kreiser-Saunders, I.; Kricheldorf, H. R. Macromol. Chem. Phys.
1998, 199, 1081−1087.
́
(15) Dobrzynski, P.; Karperczyk, J.; Bero, M. Macromolecules 1999,
32, 4735−4737.
(16) Dobrzynski, P.; Kasperczyk, J.; Janeczek, H.; Bero, M.
Macromolecules 2001, 34, 5090−5098.
(17) Dobrzynski, P.; Karperczyk, J.; Janeczek, H.; Bero, M. Polymer
2002, 43, 2595−2601.
(18) Kricheldorf, H. R.; Behnken, G. J. Macromol. Sci, Part A: Pure
Appl. Chem. 2007, 44, 795−800.
(19) Qian, H.; Wohl, A. R.; Crow, J. T.; Macosko, C. W.; Hoye, T. R.
Macromolecules 2011, 44, 7132−7140.
(20) (a) Pappalardo, D.; Annunziata, L.; Pellecchia, C.; Biesemans,
M.; Willem, R. Macromolecules 2007, 40, 1886−1890. (b) Pappalardo,
D.; Annunziata, L.; Pellecchia, C. Macromolecules 2009, 42, 6056−
6062. (c) Gorrasi, G.; Vertuccio, L.; Annunziata, L.; Pellecchia, C.;
Pappalardo, D. React. Funct. Polym. 2010, 70, 151−158. (d) Li, G.;
Lamberti, M.; Mazzeo, M.; Pappalardo, D.; Roviello, G.; Pellecchia, C.
Organometallics 2012, 31, 1180−1188. (e) Lamberti, M.; D’Auria, I.;
Mazzeo, M.; Milione, S.; Bertolasi, V.; Pappalardo, D. Organometallics
2012, 31, 5551−5560. (f) Li, G.; Lamberti, M.; Pappalardo, D.;
Pellecchia, C. Macromolecules 2012, 45, 8614−8620.
(21) (a) Nomura, N.; Aoyama, T.; Ishii, R.; Kondo, T. Macro-
molecules 2005, 5363−5366. (b) Iwasa, N.; Liu, J.; Nomura, K. Catal.
Commun. 2008, 9, 1148−1152. (c) Liu, J.; Iwasa, N.; Nomura, K.
Dalton Trans. 2008, 30, 3978−3988. (d) Iwasa, N.; Fujiki, M.;
Nomura, K. J. Mol. Catal. A: Chem. 2008, 292, 67−75. (e) Normand,
M.; Dorcet, V.; Kirillov, E.; Carpentier, J.-F. Organometallics 2013, 32,
1694−1709.
(22) Song, D.-P.; Li, Y.-G.; Lu, R.; Hu, N.-H.; Li, Y.-S. Appl.
Organomet. Chem. 2008, 22, 333−340.
(23) Axenov, K. V.; Klinga, M.; Lehtonen, O.; Koskela, H. T.;
Leskela, M.; Repo, T. Organometallics 2007, 26, 1444−1460.
̈
(24) Pappalardo, D.; Tedesco, C.; Pellecchia, C. Eur. J. Inorg. Chem.
2002, 621−628.
(25) Kricheldorf, H. R.; Mang, T.; Jonte, J. M. Macromolecules 1984,
17, 2173−2181.
(26) Dobrzynski, P. J. Polym. Sci., Part A: Polym. Chem. 2002, 40,
3129−3143.
(27) Thakur, K. A. M.; Kean, R. T.; Hall, E. S.; Kolstad, J. J.;
Lindgren, T. A.; Doscotch, M. A.; Siepmann, J. I.; Munson, E. J.
Macromolecules 1997, 30, 2422−2428.
(28) The signal due to LGL sequence is reported at 4.83 ppm for
GA/L-LA copolymers. When using rac-LA instead of L-LA, signals
coming from different stereosequences could overlap with the named
signal. To address this issue, a copolymerization run was carried out in
the same conditions of run 26, but by using L-LA instead of rac-LA.
The TLGL value calculated for this sample resulted in good agreement
with the those obtained for sample 26 (see Table S1 in the Supporting
Information), suggesting that the quantification of the TLGL value is
reliable.
(29) (a) Barak, I.; Dubois, P.; Jerome, R.; Teyssie, P. J. Polym. Sci.,
Part A: Polym. Chem. 1993, 31, 505. (b) Baran, J.; Duda, A.; Kowalski,
A.; Szymanski, R.; Penczek, S. Macromol. Rapid Commun. 1997, 18,
325−333.
(30) Cohn, D.; Younes, H.; Marom, G. Polymer 1987, 28, 2018−
2022.
(31) (a) Chen, A.; Wu, D.; Johnson, C. S. J. Am. Chem. Soc. 1995,
117, 7965−7970. (b) Hansell, C. F.; Espeel, P.; Stamenovic,
́
M. M.;
Barker, I. A.; Dove, A. P.; Du Prez, F. E.; O’Reilly, R. K. J. Am. Chem.
J
dx.doi.org/10.1021/ma402174y | Macromolecules XXXX, XXX, XXX−XXX