3.4.2.2 Melt Crystallization
The differences in the melting points for D‐ or L‐lactide and the meso‐lactide can be used for separating the different lactides from each other and from other impurities. Crystallization of lactide from the melt has been described as a three‐step industrial process including crystallization–sweating–melting [102]. The process can also be combined with other lactide manufacturing techniques to improve the lactide quality. Another process is described as an integrated process for the manufacture of purified lactide, where the final step requires that the concentrated lactide is subjected to melt crystallization to separate lactide fractions [103].
3.4.2.3 Separation in the Gas Phase
The differences in the boiling point of the different lactic acid species can be utilized in the purification of lactide. The most volatile compounds are water and LA; meso‐ and racemic pure lactides are less volatile, and lactic acid oligomers are often in the liquid phase. One purification method employs a gas stream for purification of an impure cyclic ester by passing a gaseous inert substance through the impure cyclic ester in a molten state and removing the gas stream, whereby the purified cyclic ester is recovered from the gas stream [104]. There are several references on lactide distillation processes applied on an industrial scale, and even if the processes are similar, they all have different detailed technical solutions. Some basic differences between the processes described are in many cases found in how the outlet for the purified lactide is arranged. In one method, the liquefied gaseous impurities are separated from the solidified lactide, after which the liquefied impurities are returned to the lactide synthesis step [105]. Another process involves producing pure lactide using a final step of purifying meso‐lactide from L‐lactide and/or D‐lactide by distillation to give one meso‐enriched purified lactide stream and one meso‐depleted purified lactide stream [106]. A further and improved process uses a partial condensation of the rectified lactide, whereby the low boiling point gaseous fraction remains as a vapor and is discarded. The lactide fraction is condensed and passed in the liquid phase to a distillation column [107].
3.4.2.4 Separation by Cooling
In this method, a four‐step process is applied. After polycondensation and depolymerization, the depolymerization product is led by a carrier gas, which preferably is nitrogen, through a cooling zone (−78–10°C) into a cold cyclone for separation of unreacted PLA from the mixture of lactide and depolymerization impurities. In the last step, the mixture of lactide and impurities are mixed with water (v/v ratio of 1/0.5–5) to separate lactide in a separation vessel to remove a liquid containing water and impurities formed at the top and to recover only lactide crystals from the bottom. The mixing may be performed at a temperature between 5 and 30°C. The unreacted LA separated can be recycled, and the process may be a continuous cyclic process. The method is advantageous in that the lactide can be obtained at a high yield by a simple method, compared to the above‐mentioned conventional production methods [108].
3.4.3 Ring‐Opening Polymerization
ROP of L‐lactide is generally the most preferred route for preparing high‐molecular‐weight polylactide due to the possibility of an accurate control of the chemistry and thus varying the properties of the resulting polymers in a more controlled manner. This makes ROP well suited for a large‐scale process. Polymerizations of lactide have successfully been carried out by using melt polymerization, bulk polymerization, solution polymerization, and suspension polymerization techniques. Each of these methods has its own advantages and disadvantages, but melt polymerization is generally considered the most simple and reproducible method and will be discussed later in detail [109].
3.4.3.1 Reactor Design
The simplest type of reactor system is a reaction vessel with an agitator. The number of vessels can vary depending on the desired polymerization conditions [110]. A combination of this type of reactor and a static mixer has also been developed for a continuous polymerization process for preparing polyesters from glycolide, lactide, or CL. The column type of plug‐flow reactor is preferably equipped with agitation blades in order to ensure appropriate mixing [111]. A similar concept is described in another patent, but the static mixer here can optionally be linked to an extruder as the final process step [112]. A static mixer for continuous ROP of lactide is described in another US patent. The mixer is equipped with mixing elements designed to enable mixing in both axial and crosswise directions [113]. ROP can also be performed by reactive extrusion, provided that the residence time and catalyst efficiency match [114]. A more recent approach for ROP of lactide is described where the lactide is fed into a first polymerization step (e.g., a stirred vessel or loop‐type bubble column) where pre‐polymerization takes place. The pre‐polymerized product is transferred to a second polymerization reactor where the polymerization step is conducted in a tubular reactor equipped with non‐mixing baffles [115]. Another continuous ROP process for cyclic ester monomers has been developed to operate at temperatures between 100 and 240°C by continuously providing cyclic ester monomer and polymerization catalyst to a mixing reactor to form a pre‐polymerized reaction mixture, which is transferred to a plug flow reactor where the reaction mixture is polymerized to a degree of polymerization of at least 90% whereafter the polymer is continuously removed from the plug flow reactor [116].
3.4.3.2 Catalyst Systems
A vast number of catalysts have been utilized in the ROP of lactide, of which the most studied are the carboxylates and alkoxides of Sn [117–126] and Al [127–133]. Of these, stannous 2‐ethylhexanoate (tin octanoate) is the most intensively studied. The polymerization mechanism is believed to involve a preinitiation step, in which stannous 2‐ethylhexanoate is converted to a stannous alkoxide by reaction with a hydroxyl‐bearing compound. Then, the polymerization proceeds on the tin–oxygen bond of the alkoxide ligand, whereas the carboxylate itself is inactive in the polymerization [120]. Reviews with emphasis on Sn‐ and Al‐catalyzed ring‐opening polymerization have been published by Stridsberg et al. [134] and Slomkowski et al. [135]. Some kinetic studies were also included in the reviews. However, the highly active catalysts based on, for example, tin compounds are toxic [136], and efficient catalysts showing less toxicity based on Ca [137–143], Fe [95,144–154], Mg [155–158], and Zn [158–165] have, therefore, been developed for lactide and lactone polymerization. Many of these, however, tend to cause racemization of PLA, especially when polymerizing at high temperatures. In addition to the aforementioned metals, Kricheldorf et al. [166] used other salts prepared from cations and anions belonging to the human metabolism in the ROP. Zinc lactate was found to be the most efficient of the tested catalysts with regard to reactivity and obtaining high‐molecular‐weight PLA. More recently, however, a potassium‐based catalyst been reported to be more efficient in the ROP of polylactide to high molecular weight [167]. Other catalyst/initiator systems of low‐toxicity metals for ROP have been discussed in a study by Okada [168].
Catalysts have been developed for the stereoselective ROP of lactides. In the early publications, semicrystalline PLAs were prepared from both meso‐lactide (yielding syndiotactic PLA) and racemic lactide (yielding stereoblock isotactic PLA) using