The Effect of Solvent Hydrophilicity on the Enzymatic Ring-Opening Polymerization of L-Lactide by Candida rugosa Lipase

Contradictions have been reported on the effect of organic solvents, especially toluene, on enzymatic ring-opening polymerization (eROP) of L-lactide. Studies have shown that log P, a common measure of hydrophilicity, affects enzyme activity. This study examines the effect of solvents with various log P values on the eROP of L-lactide, performed using Candida rugosa lipase (CRL). N,N-dimethylacetamide (DMA), 1,2-dimethoxybenzene, 1,4-dimethoxybenzene, diphenyl ether, and dodecane were used as the organic solvents. The eROP in ionic liquid 1-butyl-3-methylimidazolium hexafluorophosphate ([BMIM][PF6]) was also conducted to compare its performance with the organic solvents. The results show that [BMIM][PF6]-mediated eROP gave better conversion and molecular weight than the organic solvent-mediated eROP. In this study, the effects of solvents hydrophilicity are discussed, including the possibility of hexafluorophosphate ion ([PF6]−) hydrolysis to occur.

Several things remained unclear on the effect of solvent hydrophilicity on the eROP of L-lactide. A classical study by Laane et al. (1987) shows that, generally, enzymatic activity in organic solvents follows an S-shaped behavior linked to the log P values of the surrounding solvent. Log P, the partition coefficient's logarithm, indicates a compound's preference to dissolve in a hydrophobic or hydrophilic environment. Generally, the partition coefficient is determined by comparing the concentration of the measured compound in octanol (as the hydrophobic media) with its concentration in water. A stronger preference for water would give a lower log P value, which indicates the compound's hydrophilicity. Hydrophobic compounds are those that have higher concentrations in octanol than in water, giving a

Enzymatic Ring-Opening Polymerization (eROP) of L-lactide
For the polymerization, 2.5 g of L-lactide was put in a reaction tube with a predetermined amount of Candida rugosa lipase (CRL). The solvent effect was studied by adding 1.25 mL of it, unless specified otherwise, into the reaction vessel, while the bulk eROP was performed without adding the solvent. The reaction mixture was then purged with Argon at room temperature for 10 min to push out air from the reaction system. The reaction mixture was placed in an oil bath at the desired temperature, where the reaction mixture started to melt (in bulk eROP)/homogenized (for solvent-mediated eROP) and allowed to react according to Figure 1 for 72 h. performed without adding the solvent. The reaction mixture was then purged with Arg at room temperature for 10 min to push out air from the reaction system. The react mixture was placed in an oil bath at the desired temperature, where the reaction mixt started to melt (in bulk eROP)/homogenized (for solvent-mediated eROP) and allowed react according to Figure 1 for 72 h. The reaction was halted by quenching the tube in water at room temperature. Ch roform was added to the reaction mixture to dissolve the PLA formed and the remain lactide. This mixture was centrifuged to separate the CRL. Some samples were taken fr the supernatant for 1 H-NMR analysis, while most were precipitated in ice-cold metha to obtain the solid product. The obtained product was centrifuged and dried in a vacu desiccator at room temperature.

Characterizations
1 H-NMR and 13 C-NMR analysis were performed using Nuclear Magnetic Resonan Spectrometer-Bruker Avance Neo 500 MHz (Bruker Corporation, Billerica, MA, US The samples were dissolved in deuterated chloroform prior to the analysis. The 1 H-NM spectra were used to estimate the monomer conversion according to the method describ by Duchiron et al. [14]. Gel permeation chromatography (GPC) was used to determine the molecular wei and polydispersity of the precipitated samples. The GPC was performed using Shimadzu L 20, with UV-RID detector and LF 804 column (Shimadzu Corporation, Kyoto, Japan). Te hydrofuran (THF) was used as the solvent, and polystyrene was used as the standard.

The Ability of Candida rugosa Lipase (CRL) to Catalyze Enzymatic Ring-Opening Polymerization (eROP) of L-lactide
Compared to lipase B from Candida antarctica (CALB/Novozym-435) and lipase fr Burkholderia cepacia (lipase PS/BCL), Candida rugosa lipase has been considerably less stu ied for the eROP of L-lactide. Bulk eROP of commercial L-lactide at 100 °C was perform in this study to assess CRL activity in catalyzing the polymerization. A lower temperat was not possible since L-lactide only started to melt at around 98 °C, which is in line w the reported thermal properties of L-lactide [28].
The eROP here was performed in a closed batch reaction tube, where the lower p of the tube was submerged in an oil bath while the upper part was in contact with ambi air. With this configuration, crystals formed in the upper part of the tube during the eR process. These crystals seem to be formed from the vapor of the reaction mixture, i.e., lact melt in the case of bulk eROP, that was cooled as it reached the upper part of the reaction tu 1 H-NMR of the crystals from bulk ROP shows that the crystals are indeed lactide. The mas the crystals formed has been considered when determining monomer conversion. 1 H-NMR spectra of the sample reacted without CRL (Figure 2a) indicated that so degree of polymerization could arise by heating the reaction system to 100 °C. Howev the product showed a relatively strong signal at a chemical shift of 4.36-4.41 ppm, a qu tet splitting coming from the -CH-bond of the end chain. It indicated that without CR the products were short-chain oligomers, which was also confirmed by the GPC res The reaction was halted by quenching the tube in water at room temperature. Chloroform was added to the reaction mixture to dissolve the PLA formed and the remaining lactide. This mixture was centrifuged to separate the CRL. Some samples were taken from the supernatant for 1 H-NMR analysis, while most were precipitated in ice-cold methanol to obtain the solid product. The obtained product was centrifuged and dried in a vacuum desiccator at room temperature.

Characterizations
1 H-NMR and 13 C-NMR analysis were performed using Nuclear Magnetic Resonance Spectrometer-Bruker Avance Neo 500 MHz (Bruker Corporation, Billerica, MA, USA). The samples were dissolved in deuterated chloroform prior to the analysis. The 1 H-NMR spectra were used to estimate the monomer conversion according to the method described by Duchiron et al. [14].
Gel permeation chromatography (GPC) was used to determine the molecular weight and polydispersity of the precipitated samples. The GPC was performed using Shimadzu LC-20, with UV-RID detector and LF 804 column (Shimadzu Corporation, Kyoto, Japan). Tetrahydrofuran (THF) was used as the solvent, and polystyrene was used as the standard.

Results and Discussion
3.1. The Ability of Candida rugosa Lipase (CRL) to Catalyze Enzymatic Ring-Opening Polymerization (eROP) of L-lactide Compared to lipase B from Candida antarctica (CALB/Novozym-435) and lipase from Burkholderia cepacia (lipase PS/BCL), Candida rugosa lipase has been considerably less studied for the eROP of L-lactide. Bulk eROP of commercial L-lactide at 100 • C was performed in this study to assess CRL activity in catalyzing the polymerization. A lower temperature was not possible since L-lactide only started to melt at around 98 • C, which is in line with the reported thermal properties of L-lactide [28].
The eROP here was performed in a closed batch reaction tube, where the lower part of the tube was submerged in an oil bath while the upper part was in contact with ambient air. With this configuration, crystals formed in the upper part of the tube during the eROP process. These crystals seem to be formed from the vapor of the reaction mixture, i.e., lactide melt in the case of bulk eROP, that was cooled as it reached the upper part of the reaction tube. 1 H-NMR of the crystals from bulk ROP shows that the crystals are indeed lactide. The mass of the crystals formed has been considered when determining monomer conversion. 1 H-NMR spectra of the sample reacted without CRL ( Figure 2a) indicated that some degree of polymerization could arise by heating the reaction system to 100 • C. However, the product showed a relatively strong signal at a chemical shift of 4.36-4.41 ppm, a quartet splitting coming from the -CH-bond of the end chain. It indicated that without CRL, the products were short-chain oligomers, which was also confirmed by the GPC result (Table 1, Data 1). Conversely, although 100 • C is considered relatively high for enzymatic reaction, this study confirmed that the presence of CRL aided the formation of longer PLA chains. This is evidenced by the higher main-chain to end-chain 1 H-NMR peak intensity ratio, as opposed to without CRL (Figure 2b vs. Figure 2a). The GPC result (Table 1) also supported this finding. The 13 C-NMR spectra of the PLA obtained from bulk eROP with 5% CRL ( Figure 3) also confirms the formation of the polymer. It clearly shows the characteristic peaks of the PLA main chain at 169.62 ppm (C=O), 69.02 ppm (C-H), and 16.64 ppm (CH 3 ), which agrees very well with another study [29].
( Table 1, Data 1). Conversely, although 100 °C is considered relatively high for enzymatic reaction, this study confirmed that the presence of CRL aided the formation of longer PLA chains. This is evidenced by the higher main-chain to end-chain 1 H-NMR peak intensity ratio, as opposed to without CRL (Figure 2b vs. Figure 2a). The GPC result (Table 1) also supported this finding. The 13 C-NMR spectra of the PLA obtained from bulk eROP with 5% CRL (Figure 3) also confirms the formation of the polymer. It clearly shows the characteristic peaks of the PLA main chain at 169.62 ppm (C=O), 69.02 ppm (C-H), and 16.64 ppm (CH3), which agrees very well with another study [29].     Increasing CRL concentration for the bulk eROP resulted in higher conversion and molecular weight. This phenomenon could also be observed visually from the change in viscosity of the reaction mixture. When no CRL was used, there was no visible change in viscosity after 72 h. On the other hand, with 5% CRL, the change in viscosity during the reaction was obvious. The melted lactide was liquid at the beginning of the reaction. It progressively turned more viscous and hardened on the third reaction day. Such solidification was also observed in other studies on the eROP of L-lactide, especially at low temperatures or in solvent-free systems [11,15,16].
The hardening is likely caused by rheological limitation at the temperature used due to the presence of certain amount of polymer within the reaction system. For bulk eROP, such limitation could be reduced by using higher polymerization temperature. However, this was not performed in this study to avoid severe enzyme denaturation. Nonetheless, such hardening also indicated that longer PLA chains were formed when 5% CRL was used than those with lower CRL concentrations, justifying the role of CRL in the polymerization process.
Since the achievable molecular weight is likely to be limited by the rheology of the system, increasing CRL concentration at the same temperature would be misleading. For the same reason, prolonging the reaction time to more than 72 h could not be studied for bulk eROP at 100 °C. Hence, in this study, 5% CRL concentration and 3 days of reaction at 100 °C were taken as the optimum operating condition for the bulk eROP using CRL. The optimum condition obtained would be used as the operating condition for most of the other experiments in this study. Although certain reactions could finish at different reaction times depending on the condition of the mixture, to observe the effect of one factor at a time (OFAT), a similar operating condition would be used for the rest of this study.
The molecular weight achieved from the CRL-catalyzed eROP here was still relatively low compared to most other studies (Table 1). However, this could be attributed to the shorter reaction time and lower reaction temperature used in this study. The optimum condition obtained here also differs from an earlier study on CRL-catalyzed eROP of crude lactide. They used a CRL concentration of 1-10% w/w at a temperature of 70-130 °C. The type of CRL used was the same as in this study. Their optimum operating condition was 2% CRL concentration at a temperature of 90 °C. They used crude lactide (unpurified lactide) synthesized in-house with a purity of 81% [30]. The remaining content is the prepolymer and possibly other compounds from the depolymerization process [31]. Using crude Increasing CRL concentration for the bulk eROP resulted in higher conversion and molecular weight. This phenomenon could also be observed visually from the change in viscosity of the reaction mixture. When no CRL was used, there was no visible change in viscosity after 72 h. On the other hand, with 5% CRL, the change in viscosity during the reaction was obvious. The melted lactide was liquid at the beginning of the reaction. It progressively turned more viscous and hardened on the third reaction day. Such solidification was also observed in other studies on the eROP of L-lactide, especially at low temperatures or in solvent-free systems [11,15,16].
The hardening is likely caused by rheological limitation at the temperature used due to the presence of certain amount of polymer within the reaction system. For bulk eROP, such limitation could be reduced by using higher polymerization temperature. However, this was not performed in this study to avoid severe enzyme denaturation. Nonetheless, such hardening also indicated that longer PLA chains were formed when 5% CRL was used than those with lower CRL concentrations, justifying the role of CRL in the polymerization process.
Since the achievable molecular weight is likely to be limited by the rheology of the system, increasing CRL concentration at the same temperature would be misleading. For the same reason, prolonging the reaction time to more than 72 h could not be studied for bulk eROP at 100 • C. Hence, in this study, 5% CRL concentration and 3 days of reaction at 100 • C were taken as the optimum operating condition for the bulk eROP using CRL. The optimum condition obtained would be used as the operating condition for most of the other experiments in this study. Although certain reactions could finish at different reaction times depending on the condition of the mixture, to observe the effect of one factor at a time (OFAT), a similar operating condition would be used for the rest of this study.
The molecular weight achieved from the CRL-catalyzed eROP here was still relatively low compared to most other studies (Table 1). However, this could be attributed to the shorter reaction time and lower reaction temperature used in this study. The optimum condition obtained here also differs from an earlier study on CRL-catalyzed eROP of crude lactide. They used a CRL concentration of 1-10% w/w at a temperature of 70-130 • C. The type of CRL used was the same as in this study. Their optimum operating condition was 2% CRL concentration at a temperature of 90 • C. They used crude lactide (unpurified lactide) synthesized in-house with a purity of 81% [30]. The remaining content is the prepolymer and possibly other compounds from the depolymerization process [31]. Using crude lactide may explain the possibility of performing the eROP at 90 • C in their study. Whereas for L-lactide with high purity, bulk ROP should only be possible above 95-98 • C.
It may also be interesting to note that for bulk eROP in this study, the reaction system and the CRL only showed slight color change during eROP. This could indicate that the relatively high reaction temperature, compared to typical enzymatic reactions, did not severely damage the CRL. Indeed, an earlier study suggested that the temperature for the thermal denaturation of enzymes increases with decreasing associated water. In other words, the enzyme would denature at a significantly higher temperature in an anhydrous environment versus in an aqueous solution [32]. They showed that CRL denatured at 118 • C in a system with a water activity (a w ) of 0.33, while in water, its denaturation temperature was only 61 • C [32,33].

The Effect of Solvents
Using solvents in the enzymatic reaction can be beneficial as they can reduce the viscosity of the reacting mixture. This could allow the reaction to proceed at a lower temperature, providing a more thermal-friendly environment for the enzyme. An appropriate solvent could also help obtain a higher molecular weight than a solvent-free eROP [15,19]. However, the opposite situation could also occur.
During the eROP, crystals also formed in the upper part of the reaction tubes in the systems whose solvent was diphenyl ether, dodecane, and [BMIM][PF 6 ]. No crystal was formed for the rest of the solvents, i.e., DMA and the dimethoxybenzes. The formed crystals (see Section 3.1) were also dissolved using chloroform and mixed with the reaction product to obtain accurate monomer conversion. The mixing was carried out in the respective reaction tube. An additional 2.5-5 mL of chloroform was sufficient to completely dissolve the formed crystals.
Our results (Figure 4a) show that L-lactide conversion does not follow the S-shaped behavior described by earlier studies [18][19][20]. The highest monomer conversion was achieved when low log P solvents were used. This is partly in line with an earlier study where DMA and [BMIM][PF 6 ] also gave high monomer conversion [15].
The high conversion here might be because of the ability of the low log P solvents to pull the water molecules surrounding the enzymes, and together with agitation, the water molecules become more well distributed in the reaction system ( Figure 5). Considering the proposed mechanism for eROP of lactide and lactones in earlier studies [39,40], such a phenomenon may lead to faster initiation due to the availability of water molecules as initiators near the enzyme-activated monomers. This, coupled with the higher reaction temperature used here (100 • C) compared to typical enzymatic reactions as well as in the eROP of ε-caprolactone in previous studies [19,20], might cause a fast enough initiation rate to consume most of the lactide monomers in [BMIM][PF 6 ] and DMA-mediated eROP. However, in the DMA-mediated one, the molecular weight was low. In a previous study on bulk eROP of ε-caprolactone, the influence of water content on substrate conversion and polymer molecular weight has been elucidated. Higher water content leads to higher conversion. Nevertheless, both too high and too low water content resulted in low polymer molecular weight. The former is due to the possibility of hydrolysis, and the latter case is the consequence of enzyme rigidity [19].  The high conversion here might be because of the ability of the low log P solvents to pull the water molecules surrounding the enzymes, and together with agitation, the water molecules become more well distributed in the reaction system ( Figure 5). Considering the proposed mechanism for eROP of lactide and lactones in earlier studies [39,40], such a phenomenon may lead to faster initiation due to the availability of water molecules as initiators near the enzyme-activated monomers. This, coupled with the higher reaction temperature used here (100 °C) compared to typical enzymatic reactions as well as in the eROP of ε-caprolactone in previous studies [19,20], might cause a fast enough initiation rate to consume most of the lactide monomers in [BMIM][PF6] and DMA-mediated eROP. However, in the DMA-mediated one, the molecular weight was low. In a previous study on bulk eROP of ε-caprolactone, the influence of water content on substrate conversion and polymer molecular weight has been elucidated. Higher water content leads to higher conversion. Nevertheless, both too high and too low water content resulted in low polymer molecular weight. The former is due to the possibility of hydrolysis, and the latter case is the consequence of enzyme rigidity [19].  Figure 5. The influence of hydrophilic/low log P solvent (a) and hydrophobic/high log P solvent on water molecules surrounding the CRL. Water molecules that become better distributed in reaction system and high temperature prompt a faster initiation rate (lactide ring opening) in drophilic solvent (a) than in hydrophobic solvent (b).
The molecular weight of the obtained PLA here generally still follows the comm trend. Hydrophilic solvents tend to induce lower enzymatic activity, thus lower molecu weight, and the opposite happened for hydrophobic solvents (Figure 4b  The molecular weight of the obtained PLA here generally still follows the common trend. Hydrophilic solvents tend to induce lower enzymatic activity, thus lower molecular weight, and the opposite happened for hydrophobic solvents (Figure 4b). Exceptions occur on the eROP of L-lactide in [BMIM][PF 6 ] and diphenyl ether. Indeed, the log P of diphenyl ether falls near the category where its effect on enzymatic activity is unpredictable (log P 2.0-4.0) [18]. The relatively high Mn of the obtained PLA in [BMIM][PF 6 ]-mediated eROP catalyzed by CRL is discussed in the later part of this study (Section 3.4).
The log P of [BMIM] [PF 6 ] mentioned in this study is the log P at saturation [38]. In a common organic solvent, a log p value of −1.03 would normally mean that the solvent is hydrophilic. Other literatures, in fact, reported different (lower) log P values for [BMIM][PF 6 ] [38,41]. However, those studies might not ponder ionic liquids log P's dependence on its initial concentration [38]. In any case, compared to other ionic liquids, [BMIM] [PF 6 ] is considered hydrophobic [23,24]. Although the common measure of properties that have been well accepted for organic solvents is not completely suitable to accurately predict the behavior of ionic liquids [24], many times, enzymes tend to preserve their activity in hydrophobic ionic liquids compared to the hydrophilic ones [23].
At last, it is interesting to note that Figure 4 also indicates that solvents with log P between 2-4 seem to lead to low monomer conversion and molecular weight. Toluene, whose log P is 2.7, has been shown as a poor solvent for the eROP of L-lactide in earlier studies, in which no polymerization of L-lactide could take place or only a very low molecular weight of PLA was formed [9,14,21].

The Effect of Monomer-to-Solvent Ratio
Even though the eROP of L-lactide in dodecane yielded the highest Mn of PLA, the L-lactide melt and dodecane were immiscible. Such immiscibility in solvents with high log P was also observed in other studies [19,20]. This made the reaction mixture behave more like bulk because the substrate and the product are insoluble in the solvent [19]. Due to the immiscibility of the reaction system mediated by dodecane, the effect of the monomer-to-solvent ratio was performed using [BMIM][PF 6 ] as the reaction media.
The above study was done with a monomer-to-solvent ratio of 2:1 (g/mL). Considering that the bulk eROP of L-lactide by CRL (Section 3.1) resulted in higher molecular weight than eROP with solvents in the study above, the monomer-to-solvent ratio was investigated with decreasing solvent content. The amount of L-lactide in each reaction mixture was maintained the same.
A decrease in the amount of [BMIM][PF 6 ] leads to a decrease in L-lactide conversion, except in the solvent-free system (Figure 6a). A similar trend was observed in another study on the eROP of ε-CL in toluene using Novozym-435. The monomer conversion increased with increasing solvent content until a certain ε-CL concentration. This was presumably due to the better partitioning of the polymer product in the system with enough solvent. Thus, the polymer product is not accumulated in the enzyme's active sites, giving better accessibility to the monomer and allowing a faster initiation rate in a system with higher solvent content [42]. However, this does not explain the higher conversion in the solvent-free eROP ('bulk' in Figure 6a). Comparing the bulk and the solvent-mediated eROP, enzymes in the bulk eROP would only collide with the monomer or the formed oligomer/polymer without passing through other solvent molecules. Therefore, the melted L-lactide is thought to be more accessible in solventless systems. The melted lactide might also provide sufficient partitioning for the oligomer/polymer formed in the bulk eROP so as not to saturate the active site of the enzymes, especially in the early stage of the reaction.
Unlike the trend in monomer conversion, the isolated polymer yield and the molecular weight were generally increased with lower [BMIM][PF 6 ] content (Figures 6b and 7). The smaller amount of solvent increases the probability for the enzyme-activated monomer/poly mer chain to meet another opened chain. Thus, the chance of prolonging the polymer chain through the propagation reaction is also increased. the bulk eROP would only collide with the monomer or the formed oligomer/polymer without passing through other solvent molecules. Therefore, the melted L-lactide is thought to be more accessible in solventless systems. The melted lactide might also provide sufficient partitioning for the oligomer/polymer formed in the bulk eROP so as not to saturate the active site of the enzymes, especially in the early stage of the reaction. Unlike the trend in monomer conversion, the isolated polymer yield and the molecular weight were generally increased with lower [BMIM][PF6] content (Figures 6b and 7). The smaller amount of solvent increases the probability for the enzyme-activated monomer/polymer chain to meet another opened chain. Thus, the chance of prolonging the polymer chain through the propagation reaction is also increased.

CRL-Catalyzed eROP in Bulk vs. in [BMIM][PF6]
At the same temperature, catalyst ratio, and reaction time, the CRL-catalyzed L-lactide in [BMIM][PF6] yielded PLA with similar molecular weight as that produce (Table 1, Data 3 vs. Table 2, Data 2). Compared to the case of varying [BMIM][PF above, the propagation is thought to be easier in solvent-free systems due to the a solvent molecules. However, it is undoubted that the viscosity would be higher in bu  6 ] content above, the propagation is thought to be easier in solvent-free systems due to the absence of solvent molecules. However, it is undoubted that the viscosity would be higher in bulk eROP.  6 ] allows the eROP of L-lactide to be performed at a lower temperature compared to bulk ROP ( Table 2, Data 1). However, only oligomeric chains were formed. No precipitation was observed when the product-chloroform mixture of this sample was precipitated in cold methanol, confirming the low molecular weight of the sample. Increasing the temperature to 100 • C significantly improved the monomer conversion and the molecular weight.
Compared to other studies of [BMIM][PF 6 ]-mediated eROP (Table 2), the obtained molecular weight here falls in the medium range. This could potentially be improved if the temperature and reaction time are increased. As can be seen from  6 ], prolonging the reaction time from 3 to 7 days significantly increased the molecular weight of the PLA [11]. The highest Mn reported on ionic liquid-mediated eROP was 54,600 g/mol when [BMIM] [BF 4 ] was used as the solvent [8]. However, another study obtained a molecular weight of only 2700 g/mol with the same solvent, even at a higher eROP temperature [15]. Some other ionic liquids that have been used in the eROP of L-lactide are also shown in Table 2.
The use of CRL in [BMIM][PF 6 ]-mediated ROP resulted in better monomer conversion compared to ROP without CRL ( Table 2, Data 2-3). Nonetheless, the molecular weight and the polydispersity were similar. The question then arose of whether or not the CRL catalyzed the eROP in this ionic liquid media. Before continuing, it is important to note that from the bulk eROP study (Section 3.1), it was clear that CRL could catalyze the eROP in solvent-free systems.
The potential cause of higher conversion in [BMIM][PF 6 ]-mediated ROP with CRL has been described in Section 3.2.1. Moreover, extra water molecules were available from the enzyme for systems containing lipase, in contrast to systems without CRL. Enzymes typically have essential water surrounding their structure, allowing them to maintain activity in an anhydrous environment [46]. Data 2 and 3 from Table 2, as well as Data 1 from Table 1, likely indicate that [BMIM][PF 6 ] could act as a catalyst in the ROP of L-lactide.
[BMIM][PF 6 ] probably took part in the polymerization because the imidazolium cations might induce cationic propagation [17]. However, in another study, no polymerization was observed when lipase was not added to the [BMIM][PF 6 ]-mediated ROP of L-lactide [15]. Hence, it is unclear whether the catalytic effect was given solely by the [BMIM][PF 6 ] or was also due to the CRL. Different perspectives were used and described in the following paragraphs to elaborate on this issue.
Unlike in bulk, a significant color alteration was observed in the presence of [BMIM][PF 6 ]. In a sample without CRL ( Table 2, Data 3), the transparent reaction mixture turned light brown by the end of the reaction. However, the crystals formed at the upper part of the tube were already dark brown within 24 h. As described earlier, crystals were formed in the upper part of the tube for the eROP in bulk and some of the solvents due to the reactor configuration used. The color change became more pronounced in CRL-catalyzed systems mediated by the ionic liquid. However, during the monomer-to-solvent ratio study, it was also noticed that the browning decreased significantly with decreasing [BMIM][PF 6 ] content. Hence, the color alteration phenomena seem to be more caused by the [BMIM][PF 6 ]. It is also important to note that when the CRL-catalyzed eROP in [BMIM] [PF 6 ] was conducted at 80 • C, no browning was observed, both in the reaction mixture and the formed crystals.
Thus, it is suspected that at 100 • C, the ionic liquid might undergo decomposition in this study. Apparently, upon contact with water, hexafluorophosphate anion could hydrolyze; and one of the decomposition products, HF, could denature enzymes [24,47]. Nonetheless, [BMIM][PF 6 ] has been one of the most used ionic liquids in enzymatic reactions. A study on [PF 6 ] − decomposition inferred that [PF 6 ] − hydrolysis occurred particularly at low pH or elevated temperature. While at room temperature, the anion was relatively stable, even after prolonged storage. The study was conducted by contacting [BMIM] [PF 6 ] to the equivolume of water and leaving them to equilibrate [48]. To limit the amount of water, our system was pre-purged with Argon for 10 min before starting the reaction. However, the presence of water might be unavoidable in the eROP system that at 100 • C, the [PF 6 ] − started to hydrolyze. Argon was chosen over nitrogen because Ar is denser than air. Thus, it is expected to remain in proximity to the reacting mixture's surface throughout the reaction. The opened lactide ring during the initiation might also reduce the pH in the system, which could promote the hydrolysis of the [PF 6 ] − . This could also partially explain the similar molecular weight in the [BMIM][PF6]-mediated ROP with and without CRL (  6 ] was put in four closed test tubes along with 10% water, 20% water, 1 mL lactic acid, or a mixture of 1 mL lactic acid and 1 mL water. Each tube consists of a different mixture. These tubes were not pre-purged with Ar. They were placed in an oil bath at 100 • C for several days. In systems with 10% and 20% water, the condensed water was immediately visible on the inner wall of the upper tube, which is in contact with ambient air. No change was observable in these two ionic liquid systems even after 5 days. This could be because the added water was relatively little or because the water immediately evaporated when the system was submerged in the oil bath. No sufficient contact was made between the water and the [PF 6 ] − ions. In the other two systems, lactic acid was added to resemble the pH effect of the opened lactide