Lactide synthesis optimization: investigation of the temperature, catalyst and pressure effects

2.1 Materials

Lactic acid (C₃ H₆ O₃, M_w=90) (90%) was purchased from Merck Millipore (Schwalbach, Germany). Tin (II) octate ([CH₃(CH₂)₃ CH(C₂ H₅)CO₂]₂ Sn), tin(II) 2-ethylhexanoate (95%), tin (II) oxide (SnO) (97%), tin (II) chloride (SnCl₂) (99%), antimony trioxide (Sb₂ O₃) (99%), ethyl acetate, toluene and 18 mol/l of sulfuric acid were all supplied by Sigma Aldrich (Sigma-Aldrich Chemie Gmbh, Munich, Germany) and used as received.

2.2 Experimental design, setup and sample preparation

The reaction conditions used in different experiments are listed in Table 1. Samples were prepared under specific conditions as illustrated in Table 1. For instance, in PC6, a prepolymer sample was prepared using 70 ml of LA in the presence of 0.4 mol% of sulfuric acid as the catalyst, at 170°C for 6 h, in the reaction setup shown in Figure 1A. All reactions were repeated three times to ensure the reproducibility of results with a maximum of ±4% deviation of error. PC reactions were carried out using 70 g of lactic acid in a 100-ml stirred glass reactor with a sealed four-necked cap (Figure 1A). The reactor outlet was connected to a condenser, set at 95°C, to recover back lactic acid vapor to the reaction mixture. To determine the non-extractable content, PC samples were first prepared by producing 10% chloroform solutions. The solutions were then slowly added to the anhydrous alcohol up to eight times the volume, with continuous agitation. The non-extractable content (%) was measured by dividing the sample weight after extraction to its initial weight.

Figure 1

Schematics of the reaction setup used in (A) PC experiments and (B) lactide synthesis experiments.

Lactide synthesis experiments were carried out using 50 g of prepolymer in 75-ml glass reactors connected to a 100-ml two-necked glass balloon via a glass connector. A condenser was installed vertically on top of the 100-ml glass balloon, which was connected to a cold trap and a vacuum pump. All connecting lines were thermally isolated (Figure 1B). Crude lactide was purified, where needed, by means of a solvent recrystallization technique using a 50:50 mixture of ethyl acetate and toluene as solvents.

2.3 Characterizations

In order to measure the amount of material loss during the PC process, samples were weighed before and after each experiment. The crude lactide production rate was evaluated by measuring the lactide weight produced during each experiment. Molecular weight and polydispersity values were measured using a gel permeation chromatography system (Agilent 1100 series, Agilent Technologies) and polystyrene standards. The samples were injected into a column which was suitable for measuring polymers with molecular weights in the range of 1000–10,000 g/mol. The detection was performed using a refractive index detector. Thermal analysis of the samples was carried out using a differential scanning calorimetry (DSC) device (DSC 200 F3, NETZSCH, NETZSCH group) at a heating range of 20–220°C, with a heating rate of 10°C/min. IR spectra of the samples were monitored in the frequency range of 600–4000 cm^-1 using KBr pellets in a Fourier transform infrared spectroscopy device (Equinox 55, Bruker). Gas chromatography (GC) experiments were carried out in an Agilent 6890 series (Agilent Technologies) GC system with a 30-m fused silica capillary column (HP-5ms) at 140°C, using helium as the carrier gas. A 0.1% (v/v) sample solution in dichloromethane was prepared, and 0.2 μl of the solution was injected. Peaks were identified by an Agilent 5973 (Agilent Technologies) network mass selective detector. ¹H NMR sample was prepared by dissolving the purified lactide in CDCl₃. The spectrum was obtained at 400 MHz, 300 K, on a Bruker AVANCE DRX 400 (Bruker).

4.1 Prepolymer synthesis

Reactions in the presence of different catalysts and without a catalyst were performed to fully characterize the PC of LA. The effect of addition of various catalysts at different concentrations on LMW PLA production is discussed here. A pure PLA sample has an exothermic peak of crystallization at 103°C and an endothermic peak with a shoulder at 169°C. As seen in Figure 2A, the T_g, T_c and T_m of this particular sample were at 35°C, 98°C and 122/135°C, respectively. The appearance of the double melting peak could be due to the presence of two distinct crystal structures (α-form and stereocomplex crystals) or to the result of annealing occurring during the DSC scans whereby crystals of low perfection melt have time to recrystallize a few degrees above and to remelt (17).

Table 2I shows that using sulfuric acid, SnCl₂ and Sb₂ O₃ as catalysts resulted in the production of polymers with higher melting points but with different values of enthalpy of fusion. As reported by Hiltunen et al. (9), using sulfuric acid as the catalyst led to the production of a polymer with the highest molecular weight and crystallinity (above 50%), whereas tin-based catalyts produced amorphous LMW PLA. The lower T_m obtained were due to the imperfect crystalline structure of the LMW samples. Polymers with lower optical purity show their crystallization peak and double melting peaks at higher and lower temperatures, respectively. During PC, ester interchange reactions cause racemization, which occurs through the cleavage of an ester group, which can take place at two points: the carbonyl-oxygen bond and the alkyl-oxygen bond. Only the breakage of an alkyl-oxygen bond leads to racemization. It has been reported (9) that Lewis acid catalysts catalyze the breakage of the alkyl-oxygen bond, while strong proton acids, like sulfuric acid, catalyze the carbonyl-oxygen bond. Consequently, LMW PLA samples with lower optical purity produce lactides with higher racemic content.

The prepolymer produced was weighed after each experiment. It is believed that, under these operating conditions, apart from water, the volatilization of other species can be neglected. The relationship between molecular weight and prepolymer weight can be seen in Table 2I. The weight of a product decreased with the increase in polymerization degree. This is because of the evaporation of the water produced as a byproduct during the PC process.

As seen in Table 2I, for a given reaction time and temperature, higher catalyst concentrations resulted in increased polymer polydispersity (PD) and decreased molecular weight. This broadness is due to the enhanced rate of the side reactions at higher catalyst concentrations. Lactic acid, lactide and oligomers in PLA can be extracted through precipitation. When a PLA solution in chloroform is added to excess anhydrous alcohol, most of the PLA, except for some with a low molecular weight, will become a flocculent precipate, while the smaller molecules remain in the solution (18). Table 2II shows the oligomeric mole fractions for all the reactions. In reactions catalyzed by SnCl₂, the mole fractions of monomer and short-chain oligomers decreased with increasing catalyst concentrations, because there were more initiating centers for monomers to participate in the polymerization. Among all catalyts used, sulfuric acid produced the least oligomeric content and the highest non-extractable content.

PLA has a simple molecular structure with no strong hydrogen bonds in its chain, and only weak C-H···O hydrogen bonds, dipole-dipole interaction and van der Waals interactions are expected (19). The weak band at 908 cm^-1 in Figure 2B (4) is the characteristic band of the poly (l-lactide) (PLLA)/poly (d-lactide) (PDLA) stereocomplex (β crystals) that formed because of racemization. In the range of 1500–1000 cm^-1, bands are highly overlapped. Bands at 1460 and 1387 cm^-1 are due to the CH₃ asymmetric and symmetric deformation modes, respectively (20). The band at 1043 cm^-1 is assigned to the stretching vibration of the C-CH₃ group of both amorphous and semicrystalline PLA samples. The crystalline sensitive bands are 1460, 1190, 1120 and 1090 cm^-1. With the transformation of the PLA samples from the amorphous state to the semicrystalline state, the band at 1756 cm^-1 shifts to 1754 cm^-1. Bands at 2999 and 2948 are assigned to the CH₃ and CH₂ stretching modes, respectively. It is well known that the frequency shifts are characteristic of the magnitude of the hydrogen bonds formed, which shift to lower values as a consequence of the weakening of the A-H bond (19). The broad band at 3425–3505 cm^-1 is attributed to the hydroxyl groups of the unreacted monomer (lactic acid).

4.2 Lactide synthesis

There are many mechanisms and factors that influence the depolymerization of lactide. It has been reported that racemization proceeds via the ester-semiacetal tautomerization on the PLA main chain (21). Another mechanism has also been proposed (22), which is a nucleophilic substitution reaction (S_N 2) at an asymmetrical methane carbon that occurs as a back-biting reaction from an active chain end structure: R-COO^-Ca⁺ of PLA.

4.3 Influence of catalyst type

According to Table 3I, using Sb₂ O₃ as the catalyst in the synthesis of lactide from prepolymers resulted in the highest conversion after 5 h, whereas sulfuric acid produced the lowest amount. Therefore, in the specific case of intramolecular transesterification, the best catalysts were Sb₂ O₃, tin (II) octate and SnCl₂, respectively. Tin (II) octate led to the highest overall yield. It has also been reported (23, 24) that aluminum compounds are the least reactive catalysts in intramolecular transesterification. However, neither of these catalysts achieved the yield of lactide obtained by the use of tin-based catalysts.

GC has been used to determine the meso-lactide content in l/d-lactide. Due to the identical physical and chemical properties of d- and l-lactide, they cannot be separated from each other in a conventional capillary GC (25). It has been reported that (26) the optical and chemical yields of lactide are dependent on the coordination activity of the catalysts. Using sulfuric acid and SnCl₂ as catalysts resulted in the production of the purest lactide and to the lowest amount of meso-lactide, respectively. There was also a small trace of lactic acid, dimer and trimer in all the produced lactide samples. The width of the half peak of fusion (WHPF) in a DSC thermogram and the T_m of lactide samples were also measured and used as a reference of purity. l-, d- and meso-Lactides and LA have different melting points, which are 96°C, 96°C, 53°C, and 53°C, respectively. So, samples with a T_m closer to 53°C contained higher amounts of either LA or meso-lactide or both. In general, samples with a lower T_m and a higher WHPF contain more optical or chemical (i.e., LA, dimer and trimer) impurities. Figure 2C and D shows that all chemical and optical impurities have been extracted in the purification step.

4.4 Influence of catalyst concentration

As seen in Table 3II, the lactide production yield increased with the increase in catalyst concentrations up to 0.4 mol%. It was observed that lactide conversion was completed after 5 h using ≥0.4 mol% of SnCl₂ as the catalyst. Any excess amount of catalyst used in the reaction resulted in an enhanced rate of racemization and an increased amount of impurities (26).

As expected, the rate of lactide production increased with the increase in catalyst concentration and temperature (Figure 3A) (15). High product yields were obtained using 0.4, 1 and 2 mol% of the catalyst at 200°C. It was shown that, even at elevated temperatures, lower catalyst concentrations resulted in the decreased rate of lactide production.

4.5 Influence of reaction temperature

Reactions were carried out at temperatures ranging from 195°C to 210°C and from 200°C to 230°C. At higher reaction temperatures, lactide was produced with a higher production rate. According to Figure 3B, reactions in the presence of sulfuric acid and with no catalyst led to low production rates even at 210°C.

During the lactide synthesis experiments, samples were collected after 2 h for analysis. Table 3III summarizes the GC results obtained at different synthesis temperatures. It was observed that the dimer and trimer concentrations decreased with the increase in temperature. LA PC and lactide formation are parallel reactions, and from the experimental data, it can be concluded that increasing the temperature resulted in the suppression of PC and a shift in the back-biting equilibrium reaction toward the ring side in Scheme 2. At temperatures <215°C, conversion to meso-lactide was scarcely found. At a temperature of 230°C, conversions to meso-lactide increased significantly, reaching a value of 25.52 wt.%. Oligomers were found at temperatures below 230°C, and the weight ratio of linear oligomers never exceeded 6 wt.% at any temperatures studied in this work. The measured composition ratio of l/d-lactide at 230°C must be a result of an aggregate of factors that includes not only the conversion to meso-lactide, but also the reverse reactions from oligomers (27, 28). These results indicated that synthesizing lactide in the presence of 0.4 mol% of SnCl₂ at <215°C was a stable reaction with low possibility of racemization. However, a relatively high amount of LA was detected at lower temperatures. LA content could form due to the hydrolysis of lactide in the presence of moisture or as a by-product of the back-biting reactions in Scheme 2 (29) and the remaining unconverted monomer molecules from the PC process. Therefore, a prepolymer with a lower degree of monomer conversion and molecular weight obtained in the PC process leads to the production of a crude lactide with higher LA content.

4.6 Influence of prepolymer molecular weight

Figure 3C shows that the highest lactide production rates were achieved when prepolymers with DP in the range of 15–30 were used. Above this range, the viscosity of the reacting medium was too high for a proper mass transfer. Therefore, it is believed that higher temperatures are needed to distill lactide as the molecular weight and melt viscosity increase (15, 16). Below the mentioned range, the LA content in the prepolymer was too high. At the elevated temperature of lactide synthesis reaction, LA evaporates off the system before lactide does, which results in the production of a crude lactide with higher chemical impurities.

4.7 Influence of reaction pressure

As seen in Figure 3D, at higher pressures, the driving force for lactide evaporation was lower and so was the overall reaction rate. However, GC analysis showed that monomer and LMW oligomer fractions in crude lactide increased for reactions performed at lower pressures. This was because of the ranking of the volatilities, with monomer and dimer being much more volatile than lactide and longer chains.