Aggregation of polylactide with carboxyl groups at one chain end in the presence of metal cations

A Schlenk tube was charged with l- or d-lactide (2 g, 13.9 mmol) and glycolic acid (9.5 mg, 0.125 mmol), closed with a rubber septum, degassed on the vacuum line and filled with nitrogen. The tube was placed in an oil bath and heated to 105 °C to melt the monomer. Then triflic acid (30 μL of 10 % solution in dichloroethane) was added with a syringe through the rubber septum. The polymerization was conducted at ~105 °C in nitrogen atmosphere with stirring (magnetic stirrer) for about 3 h (after about 2.5 h solidification of polymer was observed, the polymerization vessel was heated for the next 0.5 h). Solid polymer was crushed, washed three times with distilled water to remove catalyst and dried for 5 h at about 40 °C in vacuum. The polymer was analyzed by ¹H NMR and MALDI TOF.

l- or d-Lactide (2 g, 13.9 mmol) was placed in a Schlenk tube which was degassed on vacuum line. The tube was filled with nitrogen, next dichloromethane was added with a syringe through the rubber septum followed by the addition of propargyl alcohol (17 μL, 16.5 mg, 0.29 mmol) and triflic acid (8 μL). Polymerization was conducted during about 18 h at room temperature with stirring (magnetic stirrer), then solvent was evaporated, solid polymer was crushed and washed three times with water. The polymer was analyzed by ¹H NMR and MALDI TOF. Mn = (¹H NMR) ~3,600.

0.4 g of propargyl-PLA (Mn ~3,600, ~0.1 mmol of alkyne groups) together with 0.60 g of mercaptosuccinic acid (4 mmol) and ~3 mg of DMPA (~0.012 mmol) were dissolved in 10 mL of THF in a round-bottom flask which was closed with a rubber septum. The flask was degassed on vacuum line (using a needle) and filled with nitrogen. This cycle was repeated three times. Next the reaction mixture was irradiated with UV lamp (365 nm) for 110 min at room temperature with stirring. Then polymer was precipitated into cold diethyl ether and the precipitate (after decantation of the solvent) was washed twice with Et₂O. The polymer was dried on vacuum line.

To compare the behavior of solutions of polymers with different numbers of –COOH groups in the presence of calcium cations, 10 wt% solutions of PL-LA-(COOH)₁ and PL-LA-(COOH)_2.8 in 1,2-dichloropropane were prepared and twofold excess of CaO with respect to –COOH groups was added. Thus, 0.1 g of PLLA-(COOH)₁ (~0.027 mmol of –COOH groups) was dissolved in 1 mL of 1,2-dichloropropane and was mixed with 0.003 g of CaO (0.054 mmol). The suspension was vigorously stirred (magnetic stirrer) for about 15 min until the solution became almost clear. Then ~0.3 mL of solution was transferred to the measuring cell of viscosimeter with a pipette, the cell was placed in the instrument and the viscosity was measured periodically without removing the solution from the measuring cell.

10 wt% solutions of PLLA-(COOH)₁ and PDLA-(COOH)₁ were prepared by dissolving 0.05 g of each polymer (~0.013 mmol of –COOH groups) in 0.5 mL of 1,2-dichloropropane. Solutions were mixed together with 0.003 g of CaO (0.054 mmol, twofold molar excess with respect to the total number of –COOH groups). The suspension was vigorously stirred (magnetic stirrer) for about 15 min and then was left for about 2 h. After that time two phases were well separated––the precipitate (fraction 1) at the bottom of the flask and the solution above (fraction 2). After decantation, the precipitate was dried in vacuum and analyzed by SEM.

In the attempt of vitamin D₃ encapsulation during PLA stereocomplexation, the same procedure was applied, vitamin D₃ (0.0157 g, 16 wt% with respect to PLAs) was added together with CaO during mixing of PLLA and PDLA solutions. Fraction 1 (precipitate) (0.063 g) and Fraction 2 (solution) (0.096 g) were separated. Both fractions were analyzed by ¹H NMR. Fraction 2 contained vitamin D₃ with traces of PLA not involved in stereocomplex. Fraction 1 contained PLA and vitamin D₃ (¹H NMR spectrum of precipitate and SEM picture are presented in Fig. 13).

¹H NMR spectra were recorded in CDCl₃ using a Bruker DRX500 instrument operating at 500 MHz.

Size-exclusion chromatography (SEC) was performed using an Agilent Pump 1100 Series with Agilent G1379A Degasser and a set of two PL-Gel 5 μ mixed-C columns. Wyatt Optilab Rex interferometric refractometer and multi-angle laser light scattering (MALLS) Dawn Eos laser photometer (Wyatt Technology Corp., Santa Barbara, CA) were used as detectors. Dichloromethane was used as an eluent at a flow rate of 0.8 mL min⁻¹ at room temperature. The system was calibrated according to polystyrene standards.

Matrix-assisted laser desorption/ionization time-of-flight (MALDI TOF) measurements were performed with the Voyager Elite (PerSeptive Biosystems, Framingham, MA) time-of-flight instrument equipped with a pulsed N₂ laser (337 nm) and time-delayed extraction source. The accelerating voltage of 20 kV was used. Dithranol was used as a matrix, CF₃COOK as a cationating agent and THF as a solvent.

Viscosity measurements were performed at 25 °C using Brookfield DV-II + viscosimeter with S52 spindle and shear rate (SR) of 10 s⁻¹ (which corresponds to 5 RPM spindle rotation).

Scanning electron microscopy (SEM) images were taken using Jeol JSM-5500LV apparatus working in the secondary electron mode with an accelerating voltage of 10 kV.

Differential scanning calorimetry (DSC) analysis was performed under nitrogen at a heating and cooling rate of 10 °C/min on DSC 2920 Modulated TA Instrument. Both temperature and heat flow were calibrated with indium.

Wide angle X-ray scattering (WAXS) analysis was performed by measuring X-ray intensity versus 2θ (from 10° to 40°) in the transmission mode-coupled θ/2θ. A wide-angle computer controlled goniometer coupled to a sealed-tube source of filtered Cu Kα radiation operating at 30 kV and 50 mA (Philips PW3830) was used. The split system of the diffractometer was adjusted to measure the integral intensity of a given diffraction peak.

Cationic polymerization in which hydroxyacids were used as initiators was successfully used by us earlier for the synthesis of poly(ε-caprolactone) containing different numbers of –COOH groups at one chain end [3]. We tried to apply the same method to the synthesis of functional PLAs. Thus, poly(l-lactide) (PLLA) and poly(d-lactide) (PDLA) with one carboxyl group at the chain end were obtained by cationic bulk polymerization performed in the presence of glycolic acid as an initiator as it is shown in Fig. 1.

Cationic polymerization, in contrast to anionic or coordination polymerization, proceeds smoothly in the presence of free carboxyl groups. Bulk polymerization of lactide was performed at 105 °C, i.e., slightly above the monomer melting point (T _m of lactide = 96 °C [17]).

In GPC chromatograms monomodal molecular weight distribution was observed but the determination of true M _n values from GPC analysis is doubtful because of the lack of appropriate standards (in some publications correction factors are used but such procedure may be questioned for relatively low molecular weight polymers) [18, 19]. Therefore, for the determination of M _n we relied mainly on ¹H NMR analysis which allowed the calculation of molecular weights more reliably by comparison of the intensities of signals corresponding to terminal methine group with those corresponding to methine groups from polymer backbone. On the basis of ¹H NMR spectra number average molecular weights M _n were determined as equal to ~3,600 for both PLLA-COOH and for PDLA-COOH.

MALDI TOF analysis (see Fig. 2) showed, however, that only part of macromolecules was initiated with glycolic acid and about half of the whole macromolecules population was initiated with water. Water was probably formed at elevated temperature as a by-product of condensation of hydroxyacid or its higher analogs. Fortunately macromolecules initiated with water have also –COOH and –OH end groups, thus, from the point of view of end-group structure, both series of macromolecules are almost the same, which is important for our studies.

First approach to the synthesis of polylactide with several carboxyl groups at one chain end was based on the same strategy as the synthesis of PLA with one –COOH group. Thus, cationic polymerization was performed in the presence of hydroxy acids with two (malic acid) or three (citric acid) carboxyl groups. Such method was successfully used earlier for polymerization of ε-caprolactone [3]. First experiments concerning bulk or solution cationic lactide polymerization showed, however, that it is not possible to obtain fully functionalized polymers by this approach. MALDI TOF analysis of obtained products indicated that significant fraction (over 80 %) of macromolecules was initiated with water. Water was present in the polymerization system as a by-product of the condensation reaction or dehydration product of β-hydroxyacids. Dehydration is competitive to polymerization and was fast enough in comparison with the initiation of polymerization by hydroxyacid to generate a new much more reactive initiating molecule, i.e., water (dehydration was not competitive in the case of the much more reactive monomer––ε-caprolactone). The initiation of polymerization with water resulted in macromolecules containing one –COOH group at the chain end so this method could not lead to PLAs with two or three –COOH groups at the chain end. In this situation another strategy was applied. The PLA functionalization was achieved in two steps. First step was the synthesis of PLA with unsaturated end group to which, in the next step, a thiol containing –COOH groups was attached. We have recently described the results of such functionalization and compared several combinations of PLAs, thiols containing –COOH groups and reaction conditions [20]. The application of thiol-yne coupling by which two thiol molecules could be attached to one PLA chain seemed to be most promising. Thus, functionalized PLA with more than one carboxyl group at the chain end was obtained by the synthesis of PLA with propargyl end group with subsequent thiol-yne coupling with mercaptosuccinic acid. Figure 3 presents the applied strategy.

Efficiencies of both stages were checked by ¹H NMR and MALDI TOF analysis. As it was found, in the first step PLA with almost all (over 95 %) macromolecules functionalized with alkyne end groups was obtained. The efficiency of functionalization by thiol-yne coupling was not so good (see Fig. 4) which was in accordance with our earlier observation [20].

In the MALDI spectrum of the product of coupling reaction (Fig. 4) four main series of signals are visible. The assignment of signals is complicated because of the overlapping of signals of macromolecules with different structures. As it can be seen in Fig. 4a, all signals in the groups 1 and 2 are wide, as they contain several signals. Signals 1A and 2A are narrower. The MALDI spectrum shown in Fig. 4b registered in reflector mode displays a splitting of main signals which confirms the presence of different signals under the wide signal. Thus, we concluded that Signal 1 around m/z ~ 3,414 corresponds to functionalized macromolecules: M ₁ = 21 × 144.13 + 56 + 150 + 150 + 39 = 3,421.7 and M ₂ = 22 × 144.13 + 56 + 150 + 39 = 3,415.9 and unfunctionalized macromolecule: M ₃ = 23 × 144.13 + 56 + 39 = 3,410.0. Signal 1A around m/z ~ 3,373 corresponds to macromolecules cationated with proton (instead of potassium): M ₄ = 22 × 144.13 + 56 + 150 + 1 = 3,377.9 and M ₅ = 23 × 144.13 + 56 + 1 = 3,372.0. Signal 2 around m/z ~ 3,342 corresponds to macromolecules which underwent transesterification: M ₆ = 41 × 72.07 + 56 + 150 + 150 + 39 = 3,349.9 and M ₇ = 43 × 72.07 + 56 + 150 + 39 = 3,344.0 and possibly also unfunctionalized macromolecules: M ₈ = 45 × 72.07 + 56 + 39 = 3,338.2. Signal 2A around m/z ~ 3,301 corresponds to transesterified macromolecules cationated with proton: M ₉ = 43 × 72.07 + 56 + 150 + 1 = 3,306.0 and unfunctionalized ones: M ₁₀ = 45 × 72.07 + 56 + 1 = 3,300.2.

Summarizing the analysis of MALDI TOF spectra recorded for the product of thiol-yne coupling between propargyl-PDLA and MSA (mercaptosuccinic acid), the majority of macromolecules were functionalized with two thiol molecules but unfunctionalized ones and those functionalized with one thiol molecule were also present.

Thus, MALDI TOF analysis cannot give unequivocal information about the average functionalization efficiency. In this situation, we rather relied on ¹H NMR analysis. On the basis of ¹H NMR spectrum presented in Fig. 5 the functionalization degree of propargyl-PDLA with MSA was estimated as ~70 % (similar value was found in the case of propargyl-PLLA functionalization), with an assumption that two thiol molecules can be attached to one alkyne group. Calculations were performed by comparison of the intensity of methylene and methine groups in the vicinity of sulfur present in attached thiol group with the intensity of methine end-group –CH(CH₃)–OH or by comparison of the intensity of methylene from residual unreacted propargyl group with the intensity of –CH(CH₃)–OH group.

Thus, we concluded that although the applied method did not yield a uniform product of desired structure, majority of macromolecules indeed contained several carboxyl groups at the chain end. Such products were used for aggregation studies.