INTRODUCTION
Homo-, co- and terpolyesters of D, L-lactic a cid, glycolic acid and ε-caprolactone
have received much attention in the search for biodegradable polymers for potential
use in biomedical and packaging applications. Each of the three homopolyesters:
poly (DL-lactic acid), poly (glycolic acid) and poly (ε-caprolactone) is
biodegradable via a simple, non-enzymatic hydrolysis mechanism (Srisa-ard
et al., 2001). The hydrolysis products are non-toxic. However, the
properties of these homopolyesters invariably only partially match the property
requirements of the application such as biodegradation rate and mechanical properties.
The co- and terpolyesters of these monomers have all been reported either as
random or block co- and terpolymers (Florczak et al.,
2007; Yaoming et al., 2007). Random copolymers with different compositions
and types of monomer units give a range of materials with different mechanical
and biodegradative properties (Saha and Tsuji, 2006).
Usually, terpolyesters of these monomers were synthesized by ring-opening polymerization.
But this technique involves too many steps and is not cost-effective (Yaoming
et al., 2007). Direct polycondensation for preparing polyesters is
a faster and cheaper method than the ring-opening polymerization. Many works
have been reported on polycondensation of homopolyesters of L-lactic and glycolic
acids (Takahashi et al., 2000; Moon
et al., 2001; Chen et al., 2006; Takasu
et al., 2006; Achmad et al., 2009).
The synthesis of copolyesters by polycondensation has been scarcely published
(Yaoming et al., 2007; Hirata
and Kimura, 2008).
In this study, we report the direct polycondensation of terpolyesters of D,
L-lactic acid, glycolic acid and ε-caprolactone. The optimum polymerization
time was determined. The structures and thermal properties of terpolyesters
with varying compositions have been investigated and their structure-property
relationships are discussed.
MATERIALS AND METHODS
This research was conducted on October 2008-August 2009 at Mahasarakham University, Mahasarakham, Thailand.
Excess water in DL-lactic acid (DLLA) aqueous solution (90 wt.%, Fluka, Switzerland)
was distilled in the absence of any catalyst under reduced pressure at 150°C
before use (Chen et al., 2006). The ε-caprolactone
(CL, 99%, Acros Organics, USA) was purified via drying with CaH2
followed by distillation under reduced pressure before being stored over molecular
sieves in a refrigerator. Glycolic acid (GA, 99%, Acros Organics, USA) and tin
(II) chloride dehydrate (SnCl2.2H2O, 98%, Carlo Erba,
USA) were used without further purification. Other reagents were of analytical
grade.
Synthesis of Terpolyesters
Poly(D,L-lactic acid-co-glycolic acid-co-ε-caprolactone)
terpolyesters (PDLLGACL) was synthesized by polycondensation in a round-bottom
flask equipped with a magnetic stirrer, which had been dried in an oven at 150°C
for 24 h before use. Appropriate amounts of monomers and SnCl2.2H2O
were charged into the reactor. Table 1 shows DLLA:GA:CL feed
ratios of 100: 0: 0, 90: 10: 0 and 90: 5: 5 by weight for the synthesis of PDLLA,
PDLLGA and PDLLGACL, respectively. The polymerization procedure was described
by Yaoming et al. (2007) with some modification.
The round-bottom flask was immersed in an oil bath maintained at 160°C under
reduced pressure for 48 h. The SnCl2.2H2O concentration
was kept constant at 0.5 wt. %.
| Table 1: |
Characteristics of the terpolyesters |
 |
aCalculated from monomer feed ratios, bCalculated
from 1H-NMR spectra, cObtained from GPC curves, =
No.-average molecular weight, MWD: Molecular weight distribution, dCalculated
from Fox equation, eMid-point of DSC glass transition |
|
| Scheme 1: |
Polymerization reaction of PDLLGACL |
The DLLA, GA and CL monomers were synthesized to obtain terpolymer using SnCl2.2H2O
as a catalyst at 160°C, as shown in Scheme 1. The as-polymerized
terpolyesters were purified by dissolving in chloroform before being precipitated
in cold n-hexane. Finally, they were dried to constant weight in a vacuum oven
at room temperature before characterization.
Characterization of Terpolyesters
The intrinsic viscosity (η) of terpolyesters were determined from flow-time
measurements on a diluted series of solutions in chloroform (CHCl3)
as solvent at 30°C using viscometer. Terpolymer compositions and microstructures
of the polyesters were characterized by 1H-NMR using a Bruker Avance
DPX 300 1H-NMR Spectrometer. Spectra were obtained from copolymer
solutions in deuterated chloroform using tetramethysilane as internal reference.
Number-average molecular weights,
and
molecular weight distributions, MWD, were determined by Gel Permeation Chromatography
(GPC) using a Waters 717 plus Autosampler GPC equipped with an Ultrastyragel®
column operating at 40°C and employing universal calibration. Tetrahydrofuran
was used as the solvent at a flow rate of 1.0 mL min-1. Thermal analysis
was carried out by means of Differential Scanning Calorimetry (DSC) using a
Perkin-Elmer DSC Pyris Diamond. For DSC analysis, terpolymer samples weighing
5-10 mg were heated at 10°C min-1 under a helium atmosphere in
order to observe their glass transition temperatures (Tg) from their
second heating scans. For the second heating scans, the terpolyesters were first
heated to 100°C before fast cooling (quenching) according to the DSC instruments
own default cooling mode before the second run.
RESULTS
Figure 1 shows percentage yields with different polymerization times of polyesters. The percentage yields increased with polymerization time increased until 48 h. It can be seen that the percentage yields of PDLLA and PDLLGA at the same polymerization time are similar. When CL monomer was copolymerized, the percentage yield slightly decreased. The influence of polymerization time on the intrinsic viscosity, as shown in Fig. 2 indicated that the intrinsic viscosity increased as the increasing of polymerization time. When the polymerization time was increased to 72 h, the intrinsic viscosity was decreased. The results suggested that the appropriate polymerization time at 160°C was 48 h for all polyesters.
The compositions of co-and terpolyesters were determined from their 1H-NMR
spectra by using the peak areas corresponding to the DLL methine protons at
δ = 5.0-5.3 ppm, the G methylene protons at δ = 4.5-4.9 ppm and the
CL ε-methylene protons at δ = 3.9-4.2 ppm. The 1H-NMR
spectra of the polyesters are shown in Fig. 3a-c
and the calculated compositions of DLLA/GA/CL are given in Table
1.
|
| Fig. 1: |
Relationship between percentage yields and polymerization
time |
|
| Fig. 2: |
Relationship between intrinsic viscosity and polymerization
time |
|
| Fig. 3: |
1H-NMR spectra of (a) PDLLA, (b) PDLLGA and (c)
PDLLGACL (peak assignments as shown) |
As would be expected, the terpolymer compositions are similar to the monomer
feed ratios.
Molecular weight characterization was carried out by means of GPC. The molecular
weight characteristics are also reported in Table 1. It was
found that the
of
PDLLGA and PDLLGACL were lower than the PDLLA. All polyesters gave similar unimodal
GPC molecular weight distributions.
The chain microstructures of polyesters are reflected in the fine structures
of the 1H-NMR spectra. The appearance of multiple resonances for
the same proton can be attributed to the presence of different monomer sequences
and therefore slightly different chemical environments in the co- and terpolymer
chains.
|
| Fig. 4: |
Expanded carbonyl regions of 13C-NMR spectra of
(a) PDLLGA and (b) PDLLGACL |
The CH2 protons in the G units and α-CH2 and ε-CH2
protons in the CL units are seen to be particularly sensitive to this. The band
at 4.8 ppm corresponding to the CH2 protons of the G units shows
different resonance lines indicating various sequences of lactyl (or lactic
acid) units and glycolyl (or glycolic acid) units (Baimark
et al., 2007). The bands at 2.4 and 4.1 ppm corresponding to the
α-CH2 and ε-CH2 protons in the CL units,
respectively are split into two quite distinct triplets adjacent to one another.
Monomer sequencing was characterized from the 13C-NMR spectra, specially
from the expanded carbonyl carbon (C = O) region from δ = 166-173 ppm,
as shown in Fig. 4 for PDLLGA and PDLLGACL. The various peaks
can be assigned to various carbonyl carbons of the middle units of triad sequences,
as also labelled in Fig. 4a and b (L = lactyl
unit, G = glycolyl unit and C = ε-caprolactone unit). The expanded carbonyl
region of 13C-NMR spectrum of PDLLA showed only LLL triad peak. The
various triad peaks, excepted LLL and GGG peaks in Fig. 4
indicated the random monomer sequencing of the co-and terpolyesters (Baimark
and Molloy, 2005; Baimark et al., 2007).
Thermal analysis of the terpolyesters was carried out by means of DSC. The DSC curves of the terpolyesters each exhibited a single Tg over the range of 20-40°C, as shown in Fig. 5 and also summarized in Table 1. The Tgs of the PDLLA, PDLLGA and PDLLGACL are 38, 27 and 22°C, respectively. The Tg of PDLLA decreased by incorporating glycolide and ε-caprolactone units into the polyester chain.
|
| Fig. 5: |
DSC thermograms of (a) PDLLA, (b) PDLLGA and (c) PDLLGACL |
DISCUSSION
The tin (II) chloride dehydrate has been used as a catalyst for polycondensation of poly(ε-caprolactone-co-lactic acid) (Yaoming et al., 2007). It is reported that the tin (II) chloride dehydrate with concentration of 0.5% by weight showed good dispersibility and dissolubility during polymerization and gave high polyester molecular weight. From Figure 1, the percentage yields of all polyesters did not increase after polymerization time of 48 h. It suggested that the highest %conversion was obtained at 48 h. The percentage yield was nearly 100% indicating that the synthesis reactions proceeded to near-quantitative conversion. Meanwhile, the intrinsic viscosities of all polyesters were decreased when polymerization time was higher than 48 h. This may be interpreted in term of degradative side reactions catalyzed by tin (II) chloride dehydrate taking place. Then, the polyesters prepared using polymerization time at 48 h were chosen for characterization of molecular weight, chain microstructure and thermal property.
The
of
all polyesters obtained from GPC analysis were directly related to their intrinsic
viscosities, as shown in Table 1. The molecular weights of
PDLLA and PDLLGA were slightly higher than PDLLGACL. This may be due to the
percentage yield of PDLLGACL was lower than those of PDLLA and PDLLGA (Fig.
1).
The CH2 peaks of G and CL units in 1H-NMR spectra can be used to determine comonomer sequencing (Baimark et al., 2007). The multiple peak split of CH2 protons of the G units (Fig. 3b) attributed to random sequencing of DLL and G units. While the multiple peak split of α-CH2 and ε-CH2 protons of the CL units (Fig. 3c) designed as random sequencing of DLL and CL units. The randomization of monomer sequencing of co- and terpolyesters can be confirmed from expanded carbonyl region of 13C-NMR spectra, as shown in Fig. 4. The L, G and C in the peak labeling are haft-lactide, haft-glycolide and ε-caprolactone, respectively. These triad sequencing occurred by tranesterification reaction during polycondensation. The both 1H- and 13C-NMR results supported the random character of the co- and terpolyesters.
For DSC analysis, the Tg of polyesters are in order of PDLLA>PDLLGA>PDLLGACL because the Tg of homopolyester of PDLLA>PDLLGA>PCL. Each experimentally observed Tg of the polyesters is comparable with the weight-averaged value calculated from the Fox Eq. 1 for a random terpolymer.
where, wDLLA, wGA and wCL are the respective
weight fractions of the DLLA, GA and CL units, as calculated from the corresponding
mole fractions from 1H-NMR. TgPDLLA (338 K), TgPGA
(308 K) and TgPCL (213 K) are the respective Tg (K) values
of the PDLLA, PGA and PCL homopolymers, as obtained from the reference literature
(Baimark et al., 2007).
The calculated Tg values of the terpolyesters from the Fox Equation
are summarized in Table 1. It is found that the Tgs
from the DSC curves are generally lower than the calculated Tg values
from the Fox Equation. This may be due to these polyesters are low molecular
weight polymers. The results suggested that the Tg of polymer also
directly related to the molecular weight of
polymer.
The Tg of poly(L-lactide-co-glycolide-co-ε-caprolactone) terpolyester
obtained from DSC curve (35°C) was closer to the Tg calculated
from the Fox equation (38°C) because of its was approximately 105
g mol-1 (Srisa-ard et al., 2001).
Finally, it should be noted that the polycondensation method is a simple and
low-cost method for synthesizing terpolyesters compared with ring-opening polymerization.
It is possible in large-scale production for use as biodegradable polyesters
in drug delivery applications (Zhao et al., 2004).
CONCLUSION
The PDLLGACL with different chemical compositions were successfully synthesized by direct polycondensation under reduced pressure at 160°C using tin (II) chloride dehydrate as the catalyst. The polymerization time at 48 h gave the highest molecular weight polyesters prepared by this method. The tin (II) chloride dehydrate shows superior to the conversional systems in regard to randomization of the monomers in resultant co- and terpolyesters. The random character of polyesters can be analyzed from their 1H- and 13C-NMR spectra. The Tgs of polyesters depended on their compositions and molecular weights. It is expected that these low molecular weight terpolyesters will find potential applications in biomedical and pharmaceutical fields.
ACKNOWLEDGMENTS
This study was supported by the National Research of Council of Thailand, the Research Development and Support Unit, Mahasarakham University (fiscal year 2009) and the Center of Excellence for Innovation in Chemistry (PERCH-CIC), Commission on Higher Education, Ministry of Education, Thailand.