**3.1.1 Solution polycondensation**

In this case an organic solvent capable of dissolving the PLA without interfering with the reaction is added, and the mixture is refluxed with removal of the water generated in the polycondensation process, which is beneficial to achieve a high molecular weight. Many procedures yield PLA with a weight-average molecular weight (*Mw*) of over 200,000 by this method (Ohta et al., 1995; Ichikawa et al., 1995). This approach was developed by Carothers and is still used by Mitsui Chemicals. The resultant polymer can be coupled with isocyanates, epoxides or peroxides to produce a range of molecular weights (Lunt et al., 1998). The reaction proceeds smoothly, however solution polymerization suffers from certain disadvantages such as being susceptible to impurities from the solvent and various side reactions including racemization and trans-esterification. It also consumes large volumes of organic solvents, which are potential pollutants to the environment.

Under optimized conditions, Ajioka et al. obtained PLA with *M*w > 300,000 by this method (Ajioka et al., 1995). Characterization data have shown that the glass transition temperatures (*T*g) of PLA and polylactide synthesized by the conventional lactide process are essentially identical (*T*g = 58°C and 59 °C, respectively), but PLA has a lower melting point (*T*m = 163 °C) than polylactide (*T*m = 178°C). The mechanical properties of the two polymers are also very similar.

### **3.1.2 Melt polycondensation**

250 Biomedical Science, Engineering and Technology

tetrahydrofuran (THF), dioxane etc., but insoluble in ethanol, methanol, and aliphatic hydrocarbons

50-60 55-60 Variable

Properties PDLA PLLA PDLLA Solubility All are soluble in benzene, chloroform, acetonitrile,

Crystalline structure Crystalline Hemicrystalline Amorphous Melting temperature (*T*m)/ °C ~180 ~180 Variable

Decomposition temperature/°C ~200 ~200 185-200 Elongation at break/ (%) 20-30 20-30 Variable Breaking strength/ (g/d) 4.0-5.0 5.0-6.0 Variable Half-life in 37°C normal saline 4-6 months 4-6 months 2-3 months

Two main synthetic methods are used to obtain PLA: Direct polycondensation (including solution polycondensation and melt polycondensation), and ring-opening polymerization

Since the LA monomer has both –OH and –COOH groups, necessary for polymerization,

Direct polymerization includes solution and melt polycondensation, depending on whether

In this case an organic solvent capable of dissolving the PLA without interfering with the reaction is added, and the mixture is refluxed with removal of the water generated in the polycondensation process, which is beneficial to achieve a high molecular weight. Many procedures yield PLA with a weight-average molecular weight (*Mw*) of over 200,000 by this method (Ohta et al., 1995; Ichikawa et al., 1995). This approach was developed by Carothers and is still used by Mitsui Chemicals. The resultant polymer can be coupled with isocyanates, epoxides or peroxides to produce a range of molecular weights (Lunt et al.,

C H COO H

Lactic acid Poly(lactic acid)

O C H

CH3

catalyst + (n - 1) H2

*n*

O

Table 1. Selected physical and chemical properties of PLA.

the reaction can take place directly by self-condensation (Scheme 2):

a solvent is used in the reaction to dissolve the PLA or not.

polymerization

Glass transition temperature

**3. Synthesis of PLA** 

**3.1 Direct polymerization** 

H

CH3 n HO COOH

Scheme 2. Direct polymerization.

**3.1.1 Solution polycondensation** 

(*T*g)/ °C

(ROP).

In contrast to solution polycondensation, the melt polycondensation of monomers can proceed without any organic solvent, but only if the temperature of the reaction remains above the *T*m of the polymer (Gao et al., 2002). Moon et al. discovered that high *M*w PLLA [*M*<sup>w</sup> ≥ 100,000] could be produced in this way in a relatively short reaction time (≤ 15 h) (Moon et al., 2000). This method can lower the cost of the synthesis significantly due to the simplified procedure, but major problems still need to be solved before it can be applied industrially because of its sensitivity to reaction conditions (Maharana et al., 2009). Thus Moon et al. worked to develop a melt/solid polycondensation technique using a binary catalyst system (tin dichloride hydrate and *p*-toluenesulfonic acid) (Moon et al., 2001). Simply put, thermal oligocondensates of LA were first subjected to melt polycondensation to obtain a melt polycondensate, which was then subjected to solid state polycondensation at 105°C. As a consequence, the molecular weight of the PLA was as high as 600,000 after a short reaction time under optimized conditions.

In summary, these one-step polymerization processes are relatively economical and easy to control, but they are equilibrium reactions affected by numerous parameters such as the temperature, the reaction time, catalysts, pressure, and so on. These factors can strongly influence the molecular weight of the products obtained. Besides, the water generated in this process can cause high molecular weight PLA to break down at high reaction temperatures. Thus the polymer resulting from these reactions usually has an unsatisfactorily low molecular weight. Attention must be paid to three aspects of the reaction to obtain a high molecular weight, namely controlling the reaction kinetics, removing the water formed, and preventing the degradation of the PLA chains.

### **3.2 Ring-opening polymerization**

Considering the drawbacks of direct polymerization, PLA is typically synthesized by ringopening polymerization (ROP) (Scheme 3), an important and effective method to manufacture high molecular weight PLA. This reaction requires strict purity of the lactide monomer, obtained by dimerization of the lactic acid monomer. PLA is obtained by using a catalyst with the monomer under vacuum or an inert atmosphere. By controlling the residence time and the temperatures in combination with the catalyst type and concentration, it is possible to control the ratio and sequence of *D*- and *L*-lactic acid (LA) units in the final polymer (Lunt et al., 1998). The polymerization mechanism involved can be ionic, coordination, or free-radical, depending on type of catalyst employed (Penczek et al., 2000). Most of the researchers are now exploring new and effective catalysts. Köhn et al. (Köhn et al., 2003) first reported that the ROP of *D,L*-lactide could be catalyzed by bis(trimethyl triazacyclohexane) praseodymium triflate [(Me3TAC)2 Pr(OTf)3] (Cat), while Pr(OTf)3 by itself had a poor catalytic activity. Cat was found to catalyze the polymerization of *D*,*L*-lactide in various solvents (THF, dichloromethane, ethyl acetate, and toluene) without any additional reagents. The optimal polymer yield (95%) and molecular weight (18,000) were obtained after 18 h at 170°C, with a ratio of [LA]:[Cat] = 1000. John et al. (John et al., 2007) produced one of the few reports on lactide polymerization with a Cu-based catalyst. ROP of *L*-lactide catalyzed by {2-[1-(2,6-diethylphenylimino)ethyl]phenoxy}2Cu(II) yielded the highest *M*w (26.3 × 103) with a monomer conversion of 57%. Two other copper complexes, {2-[1-(2,6-dimethylphenylimino)ethyl]phenoxy}2Cu(II) and {2-[1-(2 methylphenylimino)ethyl]phenoxy}2Cu(II), also catalyzed the reaction under solvent-free melt conditions (160°C) but produced polylactides of moderate molecular weights (*M*w = 12.0 × 103 and 15.9 × 103, respectively).

Scheme 3. Ring-opening polymerization.

Numerous studies have examined the influence of different factors such as the concentration and type of catalyst, monomer purity, and temperature on the polymerization of lactide. Special attention has been paid to the catalyst. Currently tin octoate is the most widely used catalyst for the ring-opening polymerization of lactides, but numerous novel efficient metalfree catalytic systems are emerging as valuable alternatives (Jérôme & Lecomte, 2008). The heavy metal-based catalysts are indeed very likely to contaminate the product, which complicates the purification of the PLA obtained and also limits the applications of PLA in the fields of food packaging and biomedicine. Some of the means developed to solve this problem will be addressed later.

### **3.3 New approaches**

The inherent disadvantages of the traditional synthetic methodology have led some researchers to explore solutions such as the development of nontoxic catalysts, unusual polymerization conditions, or other polymerization pathways (Lassalle & Ferreira, 2008).

To solve the potential pollution problems caused by heavy metal catalysts, many nontoxic catalysts derived from magnesium (Wu et al., 2005), calcium (Zhong et al., 2003), zinc (Sarazin et al., 2004), alkali metals (Chisholm et al., 2003), and aluminum (Nomura et al., 2002) have been developed for the ROP of lactides (Deng et al., 2000; Ejfler et al., 2005). For example, Chen et al*.* (Chen et al., 2007) tested a series of β-diketiminate zinc complexes as initiators for the ROP of lactide and they were all highly active, however the *M*w attained was unsatisfactory. It is worth noting that a variety of rare earth derivatives are usually highly reactive, which entitles them to be very promising initiators for the ROP of lactide (Agarwal et al., 2000).

2000). Most of the researchers are now exploring new and effective catalysts. Köhn et al. (Köhn et al., 2003) first reported that the ROP of *D,L*-lactide could be catalyzed by bis(trimethyl triazacyclohexane) praseodymium triflate [(Me3TAC)2 Pr(OTf)3] (Cat), while Pr(OTf)3 by itself had a poor catalytic activity. Cat was found to catalyze the polymerization of *D*,*L*-lactide in various solvents (THF, dichloromethane, ethyl acetate, and toluene) without any additional reagents. The optimal polymer yield (95%) and molecular weight (18,000) were obtained after 18 h at 170°C, with a ratio of [LA]:[Cat] = 1000. John et al. (John et al., 2007) produced one of the few reports on lactide polymerization with a Cu-based catalyst. ROP of *L*-lactide catalyzed by {2-[1-(2,6-diethylphenylimino)ethyl]phenoxy}2Cu(II) yielded the highest *M*w (26.3 × 103) with a monomer conversion of 57%. Two other copper complexes, {2-[1-(2,6-dimethylphenylimino)ethyl]phenoxy}2Cu(II) and {2-[1-(2 methylphenylimino)ethyl]phenoxy}2Cu(II), also catalyzed the reaction under solvent-free melt conditions (160°C) but produced polylactides of moderate molecular weights (*M*w =

HO C COO

Lactic acid Lactide Poly(lactic acid)

Numerous studies have examined the influence of different factors such as the concentration and type of catalyst, monomer purity, and temperature on the polymerization of lactide. Special attention has been paid to the catalyst. Currently tin octoate is the most widely used catalyst for the ring-opening polymerization of lactides, but numerous novel efficient metalfree catalytic systems are emerging as valuable alternatives (Jérôme & Lecomte, 2008). The heavy metal-based catalysts are indeed very likely to contaminate the product, which complicates the purification of the PLA obtained and also limits the applications of PLA in the fields of food packaging and biomedicine. Some of the means developed to solve this

The inherent disadvantages of the traditional synthetic methodology have led some researchers to explore solutions such as the development of nontoxic catalysts, unusual polymerization conditions, or other polymerization pathways (Lassalle & Ferreira, 2008). To solve the potential pollution problems caused by heavy metal catalysts, many nontoxic catalysts derived from magnesium (Wu et al., 2005), calcium (Zhong et al., 2003), zinc (Sarazin et al., 2004), alkali metals (Chisholm et al., 2003), and aluminum (Nomura et al., 2002) have been developed for the ROP of lactides (Deng et al., 2000; Ejfler et al., 2005). For example, Chen et al*.* (Chen et al., 2007) tested a series of β-diketiminate zinc complexes as initiators for the ROP of lactide and they were all highly active, however the *M*w attained was unsatisfactory. It is worth noting that a variety of rare earth derivatives are usually highly reactive, which entitles them to be very promising initiators for the ROP of lactide

cyclization ring-open polymerization

O O

O CH3

O C H

catalyst *n*

CH3

12.0 × 103 and 15.9 × 103, respectively).

Scheme 3. Ring-opening polymerization.

problem will be addressed later.

**3.3 New approaches** 

(Agarwal et al., 2000).


H3C

H

2

CH3

COOH

With respect to unusual reaction conditions, supercritical CO2 (scCO2) technology has attracted much attention because this environmentally friendly, chemically inert, inexpensive, non-toxic, and nonflammable solvent can be substituted for organic solvents (Nalawade et al., 2006). Yoda et al. (Yoda et al., 2004) thus carried out the synthesis of PLLA from an *L*-lactic acid oligomer in scCO2 with dicyclohexyldimethylcarbodiimide (DCC) as an esterification promoter and 4-dimethylaminopyridine (DMAP) as a catalyst. PLLA with a number-average molecular weight *M*n reaching 13,500 was obtained in 95% yield after 24 h at 3500 psi and 80°C. The molecular weight distribution of the products was also narrower than for PLLA prepared by melt–solid phase polymerization under conventional conditions. Not only can scCO2 be used as a medium to synthesize polymers, but it can also serve in the purification and processing of the polymer microparticles obtained (Kang et al., 2008).

The direct polycondensation of lactic acid has been considered to have a promising future due to its low cost; however it is hard to increase the molecule weight due to the difficulty in removing the water from the system under these conditions. One way to solve this problem is a chain-extension method, although the properties of the PLA obtained in this way can be somewhat affected by the procedure. Simply put, hydroxyl- or carboxyl-terminated low molecular weight PLA obtained by direct polymerization can be linked together through a chain extender, which is a bifunctional compound carrying highly reactive functional groups. Many achievements have been reported in this area, hexamethylene diisocyanate (HDI) being the most widely used chain extender for hydroxyl-terminated prepolymers since the work done by Woo and coworkers (Woo et al., 1995). Finding new and satisfactory chain extenders will remain a major goal in the near future, since HDI is toxic and subject to side reaction in this process.

In addition, LA-polymerizing enzymes functioning in replacement of metal catalysts should enable the biosynthesis of PLA, even though it is enormously challenging both in terms of research and industrial implementation. The best solution could be the development of a PLA-producing microorganism, but this has not been reported so far. Taguchi et al. (Taguchi et al. 2008) have nonetheless obtained encouraging results by developing a recombinant *Escherichia coli* strain allowing the synthesis of LA-based polyesters by introducing the gene encoding polyhydroxyalkanoate (PHA) synthase. This is illustrated in Fig. 2. They thus achieved the one-step biosynthesis of a copolymer with 6 mol% of lactate and 94 mol% of 3-hydroxybutyrate units, having a molecular weight of 1.9 × 105. This extremely important result represents a milestone towards the biological synthesis of PLA and confirms that the work is moving in the right direction. At present, the LA fraction in the copolyesters has been enriched up to 96 mol% (Shozui et al., 2011), so the synthesis of homopolymers of LA represents a major goal. To that end, the current microbial cell factory ought to be improved with further evolved LA-polymerizing enzymes (LPE) and metabolic engineering-based optimization (Taguchi, 2010). Matsumura et al. (Matsumura et al., 1997) likewise reported the lipase PC-catalyzed polymerization of cyclic diester-D,L-lactide at a temperature of 80-130°C to yield poly(lactic acid) with molecular masses of up to 12,600. Other novel methods (e.g. metalfree catalysts, non-catalytic systems) are also under development (Zhong et al., 2003; Achmad et al., 2009). The advantages and disadvantages of the PLA synthesis methods mentioned above are summarized in Table 2.

Fig. 2. Mechanism for the bio-synthesis of LA polyester. In the bio-process, the LA monomer is converted into LA-CoA ,which is recognized by the LA-polymerizing enzyme recruited from microbial PHA synthase (Tajima et al., 2009).


Table 2. Comparison of PLA synthesis methods.
