Poly(phenyllactide): Synthesis, Characterization, and Hydrolytic

The poor solubility of the monomer limited solution polymerizations of ... °C at pH 7.4 show that poly(phenyllactide) degrades at ∼1/5 the rate of ...
0 downloads 0 Views 60KB Size
Biomacromolecules 2001, 2, 658-663

658

Poly(phenyllactide): Synthesis, Characterization, and Hydrolytic Degradation Tara L. Simmons and Gregory L. Baker*,† Department of Chemistry, Center for Fundamental Materials Research, Michigan State University, East Lansing, Michigan 48824 Received November 21, 2000; Revised Manuscript Received May 17, 2001

Poly(phenyllactide) was synthesized via the ring-opening polymerization of phenyllactide, the dimer of phenyllactic acid. Phenyllactide was synthesized by two methods, the solution phase condensation of L-phenyllactic acid and by thermal cracking of low molecular weight phenyllactic acid oligomers. The poor solubility of the monomer limited solution polymerizations of phenyllactide to low yields and low molecular weights, but melt polymerization of phenyllactide with Sn(Oct)2/tert-butylbenzyl alcohol at 180 °C gave high molecular weight polymers in high yields. The resulting polymers were amorphous due to epimerization of ≈10% of the stereocenters during polymerization. Poly(phenyllactide) has a glass transition temperature of 50 °C and degrades to monomer at 320 °C. Experiments run at 55 °C at pH 7.4 show that poly(phenyllactide) degrades at ∼1/5 the rate of rac-polylactide. Introduction Polylactides are an important class of degradable polymers that are derived from renewable resources.1 Long used in medical applications as degradable sutures and implants, polylactides are now being developed for controlled drug delivery2,3 and degradable scaffolds for tissue growth.4,5 Polylactides are also being commercialized as commodity polymers with applications as degradable fibers and packaging materials.6,7 For both classes of applications, it is important to be able to access a broad range of physical properties for polylactides, and manipulation of the stereochemistry, crystallinity, and polymer architecture are tools that are commonly used to gain control over their physical properties. For example, copolymerization of lactide with glycolide,8,9 -caprolactone,10 and other monomers yields random and block copolymers,11-15 which in turn allows for control over polymer degradation rates, cross-linking, and physical properties of the polymers. Largely unexplored are polylactides where the methyl group of lactic acid has been elaborated to give substituted polylactides. Simple changes in the structure of the monomer would provide methods for controlling polymer hydrophobicity, altering the glass transition temperature, and for introducing new chemical functionality to degradable polymers. Ideally, these new monomers should be available from renewable resources. Using phenyllactic acid as an example, Frost has shown that one strategy for obtaining new R-hydroxy acids is to harness the biosynthetic pathways to R-amino acids.16 The normal pathway to phenylalanine leads from glucose to phenylpyruvic acid, which is then converted into the amino acid by a transaminase. Interception of phenyl pyruvate and reduction of the R-carbonyl produces phenyllactic acid from phenylpyruvic acid. Naturally occurring †

E-mail: [email protected].

phenyllactic acid is racemic, although it should be possible to direct the synthesis toward a particular phenyllactic acid enantiomer. Thus, fermentation may prove to be a low-cost source of phenyllactic acid. Phenyllactic acid and its dimer, phenyllactide, are particularly interesting monomers as the aromatic ring should lead to degradable polymers with properties that are different than those now obtained from polylactide. In addition, the aromatic ring is a convenient platform for the introduction of additional functionalities that can be exploited to give more elaborate polymer architectures and provide desirable properties such as the ability to crosslink the polymers. There are few reports of polymers based on phenyllactic acid. All known examples of poly(phenyllactic acid) were formed by the direct polycondensation of the R-hydroxyacid without a catalyst, a method that usually produces low molecular weight materials (Mn < 10 000).17 Little information has been published on the physical properties of these low molecular weight polymers; however, they were shown to be biodegradable.17,18 Copolymerization of phenyllactic acid with lactic acid17,18 has been reported, with a 65/35 poly(lactic acid-co-phenyllactic acid) copolymer having a Tg of 47 °C.17 A practical synthesis of high molecular weight poly(phenyllactic acid) will likely mirror that of polylactide and involve the ring-opening polymerization of phenyllactide, the phenyllactic acid dimer. As described below, the dimer is readily prepared from commercially available phenyllactic acid and can be polymerized to give high molecular weight poly(phenyllactide). Results and Discussion 1. Monomer Synthesis. Because the polycondensation of phenyllactic acid produces low molecular weight material, we elected to prepare poly(phenyllactic acid) by ring-opening

10.1021/bm005639+ CCC: $20.00 © 2001 American Chemical Society Published on Web 07/24/2001

Poly(phenyllactide) Synthesis and Properties

Biomacromolecules, Vol. 2, No. 3, 2001 659

Scheme 1

polymerization of the phenyllactic acid dimer. The easiest method for forming the dimer is acid-catalyzed self-esterification, removing the water by azeotropic distillation as it is formed (Scheme 1). This method gave a mixture of the cyclic dimer and low molecular weight linear oligomers. The reaction was run in dilute solution to favor dimerization over the formation of linear oligomers, but the condensation reaction was slow, taking 1 week to reach 90% conversion in refluxing toluene. Despite its slow rate, an advantage of this method is that there is no racemization of the stereocenters, allowing the formation of enantiopure R,R- and S,S-phenyllactide by simply starting with the corresponding phenyllactic acid enantiomer. Increasing the amount of p-toluenesulfonic acid used to catalyze the reaction increased the rate of conversion of phenyllactic acid to products (dimer plus linear products), however the maximum yield of dimer remained near 40%. Interestingly, catalyzing the reaction with 4 Å molecular sieves in place of p-toluenesulfonic acid gave high conversions but the dimer yield was 70%. Lactide polymerizations are typically run with a monomer concentration of at least 1 M, but phenyllactide’s low solubility forced the polymerizations to be run in dilute solution (0.1 M), decreasing the reaction rate. (Polymerization in THF, a better solvent for the monomer, did not lead to appreciable conversions or molecular weights.) In addition, the polymerizations typically did not reach high conversion but instead approached a limiting value of conversion. Two possibilities can explain this trend: catalyst degradation during polymerization or equilibrium control of the polymerization. The former would result from slow degradation of the catalyst during the course of the polymerization. A degradation process has not been directly observed in phenyllactide polymerizations, but solution polymerizations of ethyl glycolide run under similar conditions with the same Sn(Oct)2/BBA catalyst system led to a solid precipitate which was proposed to be a 2:1 polymer formed from the alkylsubstituted lactic acid and Sn(II).21 A similar precipitate

660

Biomacromolecules, Vol. 2, No. 3, 2001

Simmons and Baker

isolated from lactide polymerizations22 was identified as a cyclic tin-lactic acid complex based on elemental analysis data. Loss of catalyst from polymerizations by precipitation would be consistent with the observed data; however, no precipitate was isolated from phenyllactide polymerizations initiated with Sn(Oct)2/BBA. If catalyst degradation is significant, then the maximum conversion that can be achieved is defined by the relative rates for propagation and catalyst degradation. If the polymerizations are under equilibrium control, then the low conversions to polymer would imply similar rates for propagation and depropagation above room temperature. As was shown for the melt polymerization of alkyl-substituted lactides,21 the solution polymerization data can be fit to kinetic equations that describe both of these cases. For termination due to spontaneous catalyst degradation, the expression for the rate of propagation (eq 1) d[M] ) kobs[M*][M] dt

-

(1)

Figure 2. Kinetics of L-phenyllactide polymerization initiated with Sn(Oct)2/BBA at 90 (filled diamonds), and 100 °C (filled squares). All reactions were carried out at 0.1 M L-phenyllactide in toluene with a monomer:initiator ratio of 100 and a BBA:Sn ratio of 1.

where kobs is the observed rate constant for polymerization and [M*] is the concentration of actively growing chains, is modified to include the loss of propagation sites. Equation 2 assumes spontaneous loss of active sites [M*] ) [M*]0e-kdt

(2)

by a process with a rate constant for decomposition, kd. Substituting eq 2 into eq 1 and integrating gives -ln

( )

kobs [M] ) [M*]0(e-kdt - 1) kd [M]0

(3)

The case of equilibrium control has been studied previously,20 and is described by eq 4 -ln

(

[M] - [M]eq

[M]0 - [M]eq

)

) kobs[I]t

(4)

where [M]eq is the monomer concentration at equilibrium, [M]0 is the initial monomer concentration, kobs is the observed propagation rate constant, and [I] is the initiator concentration. The 100 and 90 °C data of Figure 1 are consistent with equilibrium control of the polymerization. When plotted according to eq 4 (Figure 2), the 100 and 90 °C data are linear with equilibrium monomer concentrations of 0.0275 and 0.0270 M, respectively. The quality of the fits to the 90 and 100 °C data in Figure 2 are good, and the reasonable values for [Meq] point to equilibrium control. The 70 °C data can also be fitted to eq 4, but only with the assumption of an unreasonably large equilibrium monomer concentration. Both the 50 and 70 °C data show substantially slower initial polymerization rates and induction times, indicative of inefficient initiation. Thus, solution polymerizations run at g90 °C appear to be under equilibrium control, but catalyst degradation or poor initiation seems to be responsible for the low conversions seen for polymerizations run at lower temperatures. In addition, the equilibrium monomer concentration should increase with increasing temperature,23 but the

Figure 3. Data for the solution polymerization of phenyllactide initiated by Al(OiPr)3 at 50 °C. The polymerization was run as a 0.1 M L-phenyllactide solution in toluene with a monomer:initiator ratio of 100, assuming each alkoxide initiates one polymer chain.

data for the 50 and 70 °C polymerizations (Figure 1) show the opposite trend. Given the long polymerization times, it is likely that the conversion vs time behavior of Sn(Oct)2/ BBA-initiated phenyllactide polymerizations reflects equilibrium control during the early part of the polymerization with loss of catalyst becoming more important at long times. Additional support for catalyst degradation at long times comes from experiments where additional monomer was added to a polymerization after the polymerization rate had approached zero. Polymerization resumed, but at a lower rate than predicted by the increased monomer concentration. The rates of Al(OiPr)3 initiated phenyllactide polymerizations were faster than those initiated by Sn(Oct)2/BBA; however, the limiting conversions reached in Al(OiPr)3catalyzed polymerizations were generally lower. The kinetic data for Al(OiPr)3-initiated polymerizations also can be fitted to catalyst degradation and equilibrium models. A representative example (Figure 3) shows that the fit of the data to the equilibrium model is reasonable, but as in the case of the Sn system at low temperatures, the [M]eq needed for the fit, 0.06 M, seems too large. In addition, Al(OiPr)3 initiated

Biomacromolecules, Vol. 2, No. 3, 2001 661

Poly(phenyllactide) Synthesis and Properties

polymerizations consistently gave lower molecular weights than predicted by the monomer:initiator ratio, assuming each isopropoxide initiates one polymer chain. The maximum molecular weights were 8000 g/mol, similar to those obtained in Sn(Oct)2/BBA-catalyzed polymerizations. 3. Melt Polymerization. Melt polymerizations of phenyllactide were run at 180 °C, 15 °C above the melting point of the monomer. All melt polymerizations used enantiopure L-phenyllactide. A variety of catalysts were tested for their ability to polymerize phenyllactide, with Sn(Oct)2, Ph4Sn, Bi(Oct)3, and PbO all giving >90% conversion in less than 2 h. With the Sn(Oct)2/BBA system, the molecular weights matched those predicted by the monomer/catalyst ratios, and we were able to synthesize polymers with degrees of polymerization >100. Because melt polymerizations take place at higher temperatures than solution polymerizations and have a higher monomer concentration (no solvent), the reactions are often complete in minutes instead of the hundreds of hours seen for solution polymerizations. The melt reactions resulted in much higher conversion to polymer, with ≈3% of the monomer unchanged. Melt depolymerization of monomer-free poly(phenyllactide) gave the same monomer concentration, proving equilibrium control of the polymerization.20,24-26 There are advantages and disadvantages to melt polymerizations. While the higher temperature allows for faster reactions, and not having solvent reduces the costs associated with the purification and disposal of used solvent, transesterification and epimerization are often prominent side reactions. The extent of epimerization varied with the catalyst used, but typically was less than 10%. Shown in Figure 4 is the change in the degree of polymerization seen at extended polymerization times. Intramolecular transesterification results in the formation of cyclic oligomers, increasing the number of polymer chains, and therefore decreasing the molecular weight. The decrease in molecular weight is consistent with a random chain scission model described by DPt )

1 kSt +

1 DP0

(5)

where DPt is the degree of polymerization, ks is the rate for chain scission, in this case the rate constant for intramolecular transesterification, and DP0 is the initial degree of polymerization. The fit to the model deviates at long times (low degrees of polymerization) because the model does not account for the reversal of the scission process, the polymerization of cyclic oligomers. 4. Polymer Properties. Poly(phenyllactide) is a clear colorless polymer that can be processed into films by solution casting from cyclopentanone or compression molding above 100 °C. DSC scans of poly(phenyllactide) prepared by melt polymerization show a well-defined Tg at 50 °C but no evidence of crystallinity. Polarized optical microscopy confirms that poly(phenyllactide) obtained from melt polymerizations is amorphous. Polylactide produced from Llactide is semicrystalline because of its regular tacticity, but the 5-15% epimerization that occurs during melt polymerization of L-phenyllactide is sufficient to suppress crystal-

Figure 4. Melt polymerization of phenyllactide at longer reaction times. The reaction was carried out at 180 °C with Sn(Oct)2 and BBA. The monomer to catalyst and monomer to cocatalyst rations were 100. The curve is based on the random chain scission model.

lization. Despite its structural similarity to polystyrene, poly(phenyllactide)’s Tg is nearly identical to that of racpolylactide. A Tg of 50 °C is significantly lower than polystyrene (109 °C), and is close to that reported for low molecular weight poly(lactic acid-co-phenyllactic acid).17 The flexible methylene group that links the aromatic ring to the main-chain depresses the Tg relative to polystyrene, and in contrast to data reported for structurally similar poly(hydroxyalkanoate)s, the aromatic ring provides little increase in Tg. For example, poly(3-hydroxy-5-phenylvalerate) has a Tg of 13 °C compared to -15 °C for the unsubstituted poly(hydroxyvalerate).27 Thermogravimetric analysis indicates shows an onset for weight loss at 330 °C. The primary degradation pathway is depolymerization to monomer, which was confirmed by degrading a polymer sample under vacuum and analyzing the decomposition products by 1H NMR. The major product was L-phenyllactide (95%), with 11% recovered as the R,S diastereomer. These results are consistent with the degradation temperature being related to the volatility of the lactide monomer, a correlation seen in alkylsubstituted polylactides.21 5. Polymer Degradation. Hydrolytic degradation of poly(phenyllactide) was carried out at pH 7.4 at 55 °C, the degradation products being phenyllactic acid and low molecular weight oligomers. The profile of the weight loss curve shows degradation behavior typical of polylactides, a period where the sample mass remains nearly constant, followed by a rapid loss in sample weight. The onset for the mass loss occurs when the molecular weight of the degraded polymer falls below the entanglement limit and the polymer loses its mechanical integrity. For substituted polylactides, this occurs when the degree of polymerization falls below 30.28 A number of factors can affect the degradation rate, such as the crystallinity, Tg, and the hydrophobicity of the polymer. Since poly(phenyllactide) is amorphous, it degrades faster than semicrystalline poly(L-lactide), but as shown in Figure 5, at one-fifth the rate of rac-polylactide. Because the degradation experiments were run above the Tg of both polymers, the polymer hydrophobicity should be the dominant factor responsible for the difference in degradation rate.

662

Biomacromolecules, Vol. 2, No. 3, 2001

Simmons and Baker

than the Tg. Thus, once degradation begins, the carboxylic acid end groups of the polymer are concentrated in the center of the sample, where they can catalyze the hydrolysis reaction and increase the degradation rate. Faster degradation in the center of large specimens of polylactide has been reported previously.30 Conclusions

Figure 5. Weight loss during the degradation of polyphenyllactide. The degradation was run in a phosphate buffer solution at a pH of 7.4 and a temperature of 55 °C.

Figure 6. Changes in the degree of polymerization during degradation of poly(phenyllactide) (open rectangles) and poly(rac-lactide) (filled rectangles). Inset: same data plotted to show the relative degradation rates.

Thus, the slower degradation rate for poly(phenyllactide) can be attributed to the aromatic ring in the poly(phenyllactide) structure. The molecular weight curve (Figure 6) also displays an apparent induction time, and then follows the calculated curve based on a random chain scission model. Similar induction times were observed during the degradation of lactic acid/mandelic acid29 and lactic acid/phenyllactic acid copolymers and were attributed to a slower diffusion rate for water in the copolymers. A more plausible explanation is that initially the equilibrium concentration of water in pristine samples of poly(phenyllactide) is low, but hydrolysis renders the polymer increasingly hydrophilic and accelerates the degradation process. Near the end of the experiment, poly(phenyllactide) degraded faster than predicted by the random chain scission model. The samples used in the degradation experiment were initially in the form of small pieces but aggregated into a single mass because the degradation temperature was greater

Solution polymerization of phenyllactide gives low conversions that depend on the initiator (Sn(Oct)2/ROH or Al(OiPr)3) and reaction temperature. For both initiators, the conversion of monomer to polymer saturates due the onset of a monomer-polymer equilibrium and catalyst degradation at long times. Melt polymerizations are the method of choice for obtaining high molecular weight poly(phenyllactide). A variety of catalysts polymerize phenyllactide under melt conditions, with repeatability and control over molecular weight best obtained through the use of an alcohol initiator. At long polymerization times, transesterification reactions broaden the molecular weight distribution and reduce the polymer molecular weight. Poly(phenyllactide) can be processed into clear, colorless films by solution casting or by melt pressing the polymer above the Tg (50 °C). Because of epimerization during polymerization, poly(phenyllactide) has no measurable crystallinity. Poly(phenyllactide) degrades at >300 °C to phenyllactide, which should enable simple recycling of the polymer. The Tg of poly(phenyllactide) is similar to that of polylactide, but the aromatic ring renders the polymer more hydrophobic. Thus, its degradation rate is one-fifth that of polylactide. Substituting phenyllactide for lactide in polymerizations would be a simple route to realize longer lifetimes for polylactide, especially in situations where an amorphous polymer is preferred. In addition, elaboration of the aromatic ring of poly(phenyllactide) may provide materials with higher Tgs and additional chemical functionality. Acknowledgment. We thank the Corn Marketing Board of Michigan, the Center for Fundamental Materials Research, and the Center for Crop and Food Bioprocessing at Michigan State University for financial support of this work. Supporting Information Available. Text giving experimental details of the synthesis of poly(phenyllactide) and protocols for the kinetics experiments (including a table). This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Chiellini, E.; Solaro, R. AdV. Mater. 1996, 8, 305-313. (2) Uhrich, K. E.; Cannizzaro, S. M.; Langer, R. S.; Shakesheff, K. M. Chem. ReV. 1999, 99, 3181-3198. (3) Legrand, P.; Barratt, G.; Mosqueira, V.; Fessi, H.; Devissaguet, J. P. STP Pharma Sci. 1999, 9, 411-418. (4) Agrawal, C. M.; Ray, R. B. J. Biomed. Mater. Res. 2001, 55, 141150. (5) Hutmacher, D. W. Biomaterials 2000, 21, 2529-2543. (6) Jacobsen, S.; Degee, P. H.; Fritz, H. G.; Dubois, P. H.; Jerome, R. Polym. Eng. Sci. 1999, 39, 1311-1319. (7) Drumright, R. E.; Gruber, P. R.; Henton, D. E. AdV. Mater. 2000, 12, 1841-1846. (8) Kasperczyk, J. Polymer 1996, 37, 201-203. (9) Li, Y. X.; Nothnagel, J.; Kissel, T. Polymer 1997, 38, 6197-6206.

Poly(phenyllactide) Synthesis and Properties (10) Grijpma, D. W.; Pennings, A. J. Polym. Bull. 1991, 25, 335-341. (11) Du, Y. J.; Lemstra, P. J.; Nijenhuis, A. J.; Vanaert, H. A. M.; Bastiaansen, C. Macromolecules 1995, 28, 2124-2132. (12) Gotsche, M.; Keul, H.; Hocker, H. Macromol. Chem. Phys. 1995, 196, 3891-3903. (13) Sipos, L.; Zsuga, M.; Deak, G. Macromol. Rapid Commun. 1995, 16, 935-940. (14) Zhang, S.; Hou, Z.; Gonsalves, K. E. J. Polym. Sci., Polym. Chem. 1996, 34, 2737-2742. (15) Wang, Y. B.; Hillmyer, M. A. Macromolecules 2000, 33, 73957403. (16) Snell, K. D.; Draths, K. M.; Frost, J. W. J. Am. Chem. Soc. 1996, 118, 5605-5614. (17) Fukuzaki, H.; Yoshida, M.; Asano, M.; M.; K.; Imasaka, K.; Nagai, T.; Mashimo, T.; Yuasa, H.; Imai, K.; Yamanaka, H. Eur. Polym. J. 1990, 26, 1273-1277. (18) Tabushi, I.; Yamada, H.; Matsuzaki, H.; Furukawa, J. J. Polym. Sci., Polym. Lett. Ed. 1975, 13, 447-450. (19) Nijenhuis, A. J.; Grijpma, D. W.; Pennings, A. J. Macromolecules 1992, 25, 6419-6424.

Biomacromolecules, Vol. 2, No. 3, 2001 663 (20) Kowalski, A.; Duda, A.; Penczek, S. Macromolecules 2000, 33, 7359-7370. (21) Yin, M.; Baker, G. L. Macromolecules 1999, 32, 7711-7718. (22) Kricheldorf, H. R.; Kreiser-Saunders, I.; Stricker, A. Macromolecules 2000, 33, 702-709. (23) Odian, G. Principles of Polymerization; 3rd ed.; Wiley-Interscience: New York, 1991. (24) Duda, A.; Penczek, S. Macromolecules 1990, 23, 1636-1639. (25) Witzke, D. R.; Narayan, R.; Kolstad, J. J. Macromolecules 1997, 30, 7075-7085. (26) Kowalski, A.; Duda, A.; Penczek, S. Macromolecules 2000, 33, 689695. (27) Fritzsche, K.; Lenz, R. W.; Fuller, R. C. Makromol. Chem. 1990, 191, 1957-1965. (28) Yin, M.; Baker, G. L. Unpublished data. (29) Fukuzaki, H.; Aiba, Y. Makromol. Chem. 1989, 190, 2407-2415. (30) Li, S. J. Biomed. Mater. Res. 1999, 48, 342-353.

BM005639+