Poly(lactic acid). Группа авторов

is that the reaction from oligomer to lactide is an equilibrium reaction. To pull the reaction toward the right, lactide must be withdrawn from the system. In reaction engineering terms, this means that the chemical kinetics of the reaction cannot be understood without consideration of the method and efficiency of lactide removal. In terms of know‐how described in patents, this means that reported lactide production rates depend to a large extent on the geometry of the equipment in which lactide synthesis is performed and that provides for removal of lactide vapor from the reaction zone.

In modern chemical technology, one of the goals is to fully understand a given system, capture the knowledge in models to describe experimental work, and ultimately use these models to design, optimize, and debottleneck large‐scale equipment. For the present system, this means that one must develop process know‐how on chemical kinetics and thermodynamics of lactide and HL oligomers, and on physical phenomena related to equipment design.

      These aspects will be relevant for both the prepolymerization and the synthesis of lactide, as these chemical systems are highly similar. In practice, however, lactide synthesis is more complex as chemistry, recovery and type of equipment are intertwined, and the viscous nature of reaction mixtures requires special attention.

      With these aspects in mind, the information on the lactide synthesis that can be found in the literature is summarized below.

       1.3.2.1 Prepolymerization

      A general procedure for batch prepolymerization is described in a patent by O'Brien et al. [75]. Typically, vacuum pressures of 70–250 mbar and temperatures up to 190°C are used to dewater lactic acid to a prepolymer with an average degree of polymerization (DP) of around 10 in a batch process time of 6 h. For lab‐scale equipment, it was also found that thin film and rotating flask vacuum equipment showed faster reaction times than a stirred tank, indicating the importance of mass transfer of water in the already viscous prepolymer.

      Continuous prepolymerization has also been described in a number of patents, for example, in stirred tanks in series or in evaporator‐type equipment [68, 76, 77]. Usually patents describe prepolymers with a DP of 7–20 as feed to the lactide synthesis. Using modern HPLC methods, it has been shown that in oligomeric systems up to DP 10, an equilibrium is present with constant equilibrium constants between the oligomers [6, 72].

       1.3.2.2 Lactide Synthesis During Prepolymerization

      Because the composition of a mixture comprising lactic acid oligomers and lactide is governed by chemical equilibria, a prepolymerization exhibits relatively high concentrations of lactide (HL₂–H₂O–L₂ equilibrium) around DP 2. Sinclair et al. distilled these fractions to recover lactide, but the crude lactide was quite impure, which may prevent economical processing [73]. In hindsight, the patent describes trials to optimize Pelouze's original lactide synthesis without catalyst [71].

       1.3.2.3 Basic Research on Batch Lactide Synthesis and the Catalysts Used

      Noda and Okuyama reported on the batch synthesis of lactide from DP 15 prepolymer with various catalysts at 4–5 mbar and 190–245°C [74]. In a batch synthesis with 50 g of oligomer in a stirred flask, the evolution rate of crude lactide is rather constant and then starts to decline and the conversion levels off at 80–90%. The tin catalyst performed best compared with other catalysts and showed the lowest levels of racemization. Tin octoate (stannous 2‐ethylhexanoate) is a liquid catalyst that can be handled easily, is food grade, and is widely available.

      Thinking in terms of mechanisms, the equilibrium concentration of lactide in an oligomer mixture is 5% or less, and it will boil off at low vacuum [6, 68]. The catalyst increases the rate of lactide formation by facilitating lactide formation by backbiting from hydroxyl chain ends of oligomers [4, 74]. In a batch experiment, the rate is initially constant, but during synthesis esterification also occurs, and the DP of the polyester rises concomitantly. The melt viscosity of the reaction mixture increases accordingly and at the end of a batch process, mixing the highly viscous residue becomes very difficult, which limits the extent to which the residue can be depleted of lactide.

      In engineering terms, this means that mass transfer of lactide from the liquid to the gas phase decreases as viscosity increases. The balance between lactide production and lactide removal plays a role in all experiments that one might want to investigate on lab scale. For example, catalyst concentrations of 0.05–0.2 wt% tin(II) octoate are mentioned in the literature, but traditional experiments to verify the order of the reaction for the catalyst are difficult because of the influence of mass transfer limitations.

       1.3.2.4 Continuous Synthesis

      In 1992, Gruber et al. [68] described a continuous lactide synthesis in which prepolymer is fed continuously to a reactor, crude lactide is evaporated under vacuum, and residue is removed. Typical operating conditions for the reactor were residence time around 1 h, vacuum pressure 4 mbar, temperature 213°C, and catalyst amount 0.05 wt% tin(II) octoate on feed. The conversion per pass was around 70%, and the overall yield was increased by recycling the residue to the lactic acid section of the process, where the oligomers are hydrolyzed again.

      Especially in the patent literature, several different reactor types are described for continuous lactide synthesis:

       Stirred tank reactor with different stirrer types [76]. On a bench scale, the reactor is jacketed for heating.

       Stirred reactor with a distillation section on top of the reactor to fractionate the product [50].

       Thin film evaporator with a typical conversion of 80% on pilot scale [70].

       Horizontal wiped film evaporator. In a patent by Kamikawa et al. [77], the use of horizontal wiped film is described. In the horizontal mode, the residence time of the reaction mixture can be controlled and a conical form is used in which wipers transport the viscous residue.

       Distillation column. In a patent by O'Brien et al. [75], a distillation column with perforated plates and optional use of packing material and heating on the stage are described. In an experiment with a single tray, a DP 10 feed was fed to the top, and N2 was used to strip the lactide from the liquid. At different residence times, the conversion on the tray could be as high as 93% at 210–215°C. In other patents, the use of N2 gas as a stripping agent is mentioned, but it is to be expected that in large‐scale equipment the processing of large amounts of inert gases will be less economical compared with the use of vacuum systems.

      Reviewing the literature provides a list of process aspects that need consideration in the design of a solventless synthesis operated with vacuum equipment.

       Temperature. Intrinsic reaction rates increase with temperature. At higher temperature also, the vapor pressure of lactide above the reaction mixtures increases. The reaction rate of racemization will also increase with temperature. In Witzke's Ph.D. study, information on activation energies can be found [6].

       Pressure. Pressures of 10 mbar or less are used. At higher pressures, the driving force for lactide evaporation will be lower, and the overall reaction rate will be lower. Low pressures will require detailed considerations of equipment size, vacuum systems, condensers, and so on.

       Feed DP. The feed DP has two effects. First, a low DP feed will contain more monomer lactic acid that boils at a lower temperature than lactide, and this will contaminate the crude lactide distilled off from the reactor. Also, monomer lactic acid can be released from DP 3 with the catalyst, leading to more acidity in the crude lactide. Second, it is to be expected that at a higher feed DP the residue in the reactor will have a higher DP and viscosity with consequences for equipment design. The influence of prepolymer DP on the meso‐lactide level formed during lactide synthesis was discussed by Gruber et al. [69]. Increasing feed DP clearly resulted in a decrease in the lactic acid concentration in the crude lactide. A drawback is that the meso‐lactide concentration also increased