3.1. Polyester Synthesis in the Miniemulsion
The influence of three different lipases, in free (CRL, BCL, PBL) and immobilized (CRLI, BCLI, and PBLI) forms, on the polyester synthesis from different diacids and diols in a miniemulsion was evaluated.
Initially, an equimolar monomer concentration of 0.5 M of octanedioic acid and 1,8-octanediol was tested for the POS synthesis with free lipases (CRL, BCL, PBL) (
Figure 2,
Table 2). The free lipases present different catalytic efficiencies, mainly dependent on the initial pH and temperature.
Enzyme activity depends on the pH of the reaction media, which in the case of a miniemulsion system is determined by the continuous (water) phase [
40]. A comparison of the polymer molecular weight at pH 3.3 (pH of the original miniemulsion) and miniemulsions with the corrected pH value was made (
Figure 2A and
Table 2). A higher POS molecular weight and activity were obtained for CRL at pH 4 (2750 g mol
−1) and for BCL (3250 g mol
−1) and PBL (6950 g mol
−1) at pH 5 at 25 °C (
Figure 2B and
Table 2). POS formation catalyzed with BCL at pH 3.3 was not observed (
Figure 2A).
The effect of temperature on lipase activity was evaluated in the range of 25–45 °C at optimum pH previously observed for each enzyme (CRL at pH 4; BCL and PBL at pH 5). The CRL and BCL activities decreased significantly at temperatures higher than 30 °C, obtaining the highest molecular weight of 2300 and 3400 g mol
−1, respectively, at 25 °C (
Figure 2B). These results were due to the loss of the enzyme secondary structure with a temperature increase and a parallel decrease in activity in the polycondensation media [
49]. However, no variation in the molecular weight (2600 g mol
−1) was observed by a temperature change at pH 3.3 for PBL (
Table 2). An inhibition effect of temperature on PBL at 45 °C (
Mw of 3550 g mol
−1) in comparison to 25 °C (
Mw of 6950 g mol
−1) was observed for a pH value of 5.0 (
Table 2).
Due to the lower catalytic efficiency of CRL and BLC than PBL in the POS synthesis at 0.5 M monomer concentration, the free enzymes (CRL, BCL, and PBL) were tested for different equimolar monomer (1,8-octanediol and octanedioic acid) concentrations in the range from 0.05 to 0.6 M at 25 °C and 48 h of polycondensation. The maximum POS molecular weight of 3550 g mol−1 for CRL was observed at a concentration of 0.1 M. In the case of BCL and PBL, the maximum POS molecular weight of 3850 and 6950 g mol−1, respectively, was detected at a concentration of 0.5 M. Based on these data, an increase in the monomer concentration above 0.1 M showed a high inhibition effect on CRL. Besides the maximum polymer weight observed for the same monomer concentration (0.5 M) for BCL and PBL, PBL presented higher monomer selectivity with regard to the higher molecular weight obtained.
The respective immobilized enzyme preparations (CRLI, BCLI, and PBLI) were also tested for the same initial polycondensation conditions in the miniemulsion. The three immobilized lipase preparations showed different behaviors in comparison to the free enzymes. The POS synthesis using CLRI and BCLI did not occur, while PBLI catalytic activity in POS synthesis was more efficient than the free enzyme (PBL) (
Table 2). The POS molecular weight obtained by PBLI at pH 3.3 (2700 g mol
−1 for 25 °C and 3450 g mol
−1 for 45 °C) was slightly higher than the POS molecular weight obtained by PBL at the same pH (3.3) (2600 g mol
−1 for both temperatures). POS synthesis at pH 5 led to significant molecular weight differences, especially at 45 °C (i.e., 3550 and 7600 g mol
−1 for PBL and PBLI, respectively) (
Table 2).
The lack of polycondensation with immobilized enzymes, CRLI and BCLI, can be due to the size of the miniemulsion droplets that are unable to penetrate inside the support pores where the immobilized lipases (CRLI and BCLI) are localized. The other possibility is that the lipases immobilized in CRLI and BCLI biocatalysts could not successfully perform the interface activation mechanism characteristic of these specific lipases. This hypothesis could be the more plausive explanation as PBLI presented excellent results (7600 g mol
−1, 45 °C), and PBL is an atypical lipase not characterized by the interface activation mechanism, according to several authors [
50]. Additionally, the high catalytic efficiency of PBLI to catalyze polyester synthesis in a miniemulsion may also be related to the specific support–enzyme interactions already observed for a similar biocatalyst named Novozym 435 and the intrinsic stability characteristics of PBL [
51].
The initial pH correction from pH 3.3 to pH 5.0 in the miniemulsion for free and immobilized enzyme preparations (PBL and PBLI) at 25 °C increased the polymer molecular weight 2.7-fold. However, with the same pH correction at 45 °C, the polymer molecular weight increased only 1.4- and 2.2-fold for free (PBL) and immobilized (PBLI) enzyme preparations, respectively (
Table 2). These results confirm higher activity and probably stability of the immobilized enzyme preparation (PBLI) for polycondensation at 45 °C in miniemulsion systems compared to the free enzyme (PBL).
Barrère et al. (2003) [
11], by direct dehydration polycondensation of dodecanediol with dodecanedioic acid or
tert-butyl-3-adipic acid at 70 °C for 100 h in an emulsion system containing 16 wt % of
p-dodecylbenzene sulphonic acid (DBSA), obtained a conversion yield of 85% and a molecular weight of 3480 or 1100 g mol
−1, respectively. Takasu (2006) [
12], using a similar emulsion system with the DBSA catalyst, obtained 10100 g mol
−1 and excellent yield (99%) at higher polymerization temperature (85 °C) for 48 h. However, in this last work, the direct dehydration polycondensation occurred concomitantly with chain extension using hexamethylene diisocyanate as the chain extender. In our work, it was possible to run the polycondensation at a lower temperature (25 or 45 °C) and also easily separate the biocatalyst (PBLI) in comparison to a soluble catalyst, like DBSA [
11,
12].
The effect of the acid and alcohol carbon chain length on polyester synthesis by PBLI in a miniemulsion was assessed for hexanedionic (C6) and octanedioic (C8) diacids, while the diol carbon chain length increased from C6 to C10 atoms for both diacids.
The polycondensation was carried out at equimolar monomer concentrations (0.5 M) in the miniemulsion, at initial pH 5.0, 250 rpm, and 45 °C, for 48 h using 8 mg mL
−1 of PBLI (
Figure 3).
The diacid carbon chain length showed an evident influence on the molecular weight of the polymer, except for the C
6 dialcohol. By increasing the carbon chain length of dialcohol (C
6 to C
10) in polycondensation with hexanedioic acid (C
6), the polymer molecular weight gradually increased from 1900 g mol
−1 (C
6) to 3300 g mol
−1 (C
10). Furthermore, high conversion (>98%) for monomers with the longer carbon chain length of dialcohol (C
10) with hexanedioic acid (C
6) was also observed (
Figure 3). Enzyme stabilization with the dicarboxylic acids in opposition to the destabilization induced by the dialcohols in miniemulsion systems was previously observed [
49].
For octanedioic acid, with an increasing carbon chain length of dialcohol from C
6 to C
8, a significant increase in polymer molecular weight from 2100 to 7800 g mol
−1 was observed. However, using the dialcohol C
10, the polymer molecular weight decreased by 45% (from 7800 to 4300 g mol
−1 for C
8 and C
10, respectively) (
Figure 3). The results demonstrated that PBLI mostly shows higher specificity for longer carbon chains of dicarboxylic acids and dialcohols under the experimental conditions tested (
Figure 3). PBLI showed higher substrate specificity for the C
8 carbon chain length of the dicarboxylic acid and dialcohol in polyester synthesis in the miniemulsion.
Candida antarctica lipase B (CALB) has been reported to present substrate specificity [
50,
51] and high selectivity for substrates with a longer alcohol carbon chain [
52].
Duwensee et al. (2010) [
16] performed polycondensation of sebacic acid and 1,4-butanediol catalyzed in a biphasic system comprising citric acid/phosphate buffer and an organic solvent diisopropyl ether (DIPE) or
tert-butyl methyl ether (MTBE) using CALB in free and immobilized forms. At the end of the polycondensation at 50 °C, for substrate concentrations of 0.2 M, at pH 6, and after 48 h of polycondensation, molecular weights of 1890 or 2520 g mol
−1, respectively, were detected.
Due to the high specificity shown by PBLI for dicarboxylic acids and dialcohols with a carbon chain length of C8, these monomers were chosen in this work to analyze the effect of reactor conditions on polyester synthesis.
High conversion values (>98%) were achieved for the dicarboxylic acid C
8 (octanedioic acid) with an increasing chain length of dialcohols (C
6, C
8, and C
10) (
Figure 3). This high conversion was confirmed by the only traces of monomers detected by SEC analysis illustrated in
Figure 4. The monomers’ (octanedioic acid and 1,8-octanediol) peak presents a residence time of 22.6 min corresponding to an area of 3421201 at time zero (
Figure 4A), and no significant peak was detected by SEC after 48 h of polycondensation (
Figure 4B), confirming that conversion of about 100% occurred.
3.2. Effect of Reactor Conditions on POS Synthesis in the Miniemulsion
When considering the properties of polymers prepared by emulsion polymerization, the reactor conditions have a critical impact. The viscosity of the monomer/polymer solution increases fast, with monomer conversion influencing colloidal stability and monomer–monomer and enzyme–monomer interactions, so efficient mixing is essential to obtaining good polymer properties. The polymerization media must be stirred continuously during the polycondensation using a suitable agitator design and speed [
37,
45].
The effects of the miniemulsion preparation and composition and stirring conditions were evaluated on the POS synthesis at equimolar concentrations (0.5 M) of 1,8-octanediol and octanedioic acid for 48 h, catalyzed by the PBLI biocatalyst (8 mg mL
−1) (
Table 3). The polymer molecular weight and conversion (%) obtained under these conditions are illustrated in
Table 3 and
Figure 5.
An increase in the direct magnetic stirring to 500 rpm for initial pH 3.3 and pH 5.0 (conditions A and B,
Table 3 and
Figure 5) at 45 °C led to a polymer molecular weight of 3500 and 7800 g mol
−1, respectively, and under both pH conditions, with conversion higher than 90% (
Table 3). The pH influence on the polymer molecular weight was already observed at 250 rpm (
Table 2). The increase in direct magnetic stirring from 250 rpm to 500 rpm in the miniemulsion showed a relatively low polymer molecular weight increase at 45 °C (7.7%, conditions B and C,
Table 3 and
Figure 5). Additionally, at 500 rpm, formation of agglomerates occurred, containing probably monomers, the polymer, and biocatalyst, glued to the magnetic stirrer and the reactor wall at the end of the polymerization. This singularity was not observed in the experiment carried out at direct magnetic stirring at 250 rpm.
For this reason, the polycondensation was also tested in an orbital shaker at 250 rpm (condition D,
Table 3 and
Figure 5). In the orbital shaker, the PBLI beads were almost retained in the bottom of the reactor with low dispersion and homogenization, and additionally, there was no agglomerate formation. The use of orbital stirring led to a 60% polymer molecular weight decrease compared to direct magnetic stirring at 250 rpm. Due to high external mass transfer resistance, less contact occurred between the biocatalyst and the polymer already formed, affecting the polymer growth (conditions C and D,
Table 3 and
Figure 5). Despite this, there was contact between the biocatalyst and the monomers as the conversion was 99%.
As an increase in magnetic stirring to 500 rpm showed a low influence on polymer molecular weight and conversion (%), further experiments were carried out at 250 rpm.
The viscosity dependence of temperature is an important property of polymerization systems, so the temperature of 65 °C was tested to analyze its influence on the POS molecular weight (condition F,
Table 3 and
Figure 5) compared to 25 °C and 45 °C (conditions E, C,
Table 3 and
Figure 5). At 25 and 45 °C, the polymer molecular weight change was less than 0.5%, but a critical drop in the polymer molecular weight from 7200 to 4300 g mol
−1 was detected by increasing the temperature from 45 to 65 °C for 48 h of polycondensation. Simultaneously, the conversion dropped from 98 ± 0.2% to 91 ± 0.4% (conditions C and F,
Table 3 and
Figure 5). The lower POS molecular weight and conversion degree observed at 65 °C could be due to the thermal inhibition of the enzyme activity of the PBLI biocatalyst.
Energy consumption is an important parameter for industrial application, so the polycondensation was tested for the emulsion system without the sonication step and the use of pure water as a polymerization medium (conditions G and H,
Table 3 and
Figure 5). Since miniemulsion droplets are like nanoreactors, their size and stability can have a significant impact on the polycondensation performance in miniemulsions. A POS molecular weight of 7600 g mol
−1 was achieved in the miniemulsion, considerably higher than 3500 and 3650 g mol
−1 obtained for the emulsion without sonication and pure water, respectively, although in all the polycondensation media tested, the conversion was higher than 90%.
By applying ultrasonication on the emulsion formation, the droplet size decreased, increasing the droplet interfacial area available for enzyme action on monomers or the polymer already formed and accumulated inside miniemulsion droplets [
11,
36]. This phenomenon is confirmed by the lower molecular weights obtained in pure water and the emulsion without ultrasonication.
On the other hand, the POS synthesis in water also proves that PBLI can perform polycondensation without the emulsion or organic solvent indispensable for promoting the interface activation mechanism of a typical lipase [
50]. This result was surprising, in particular for pure water, as there was a considerable amount of monomers that are insoluble in water and remain initially in solid form (precipitate). The same precipitation behavior in water is expected for POS synthesized due to its high hydrophobicity and low water solubility. The polycondensation results obtained in pure water suggest that there is a small fraction of monomers and polyester soluble and in equilibrium with respective precipitates in pure water. The monomers and polyester gradually dissolved and diffused into the active site of the immobilized enzyme (PBLI), where they were converted initially to small-chain-length polyesters but increasing with polymerization time. Thus, it is possible to classify this as polycondensation from solid to solid via biocatalysis in water [
53].
The best performance regarding the POS molecular weight and substrate conversion was accomplished at 45 °C with magnetic stirring at 250 rpm in the miniemulsion.
3.3. Effect of the Reaction Operation Mode: Batch vs. Fed-Batch for Different Polymerization Media
The enzymatic POS synthesis was carried out in the previous subsection in batch operation mode with the initial concentration of 0.5 M of each monomer (1,8-octanediol and octanedioic acid). Polycondensation in batch operation mode allows flexibility and adaptability to obtain different polyesters [
54]. However, the inhibition effect of the high substrate concentrations on lipase activity in the miniemulsion, as well as in an organic solvent, was previously observed for the esterification [
38,
39,
40,
55]. To overcome the inhibition effect of high substrate concentration on enzyme activity, the reactions can be carried out in fed-batch operation mode, which means the stepwise addition of substrates [
55].
The initial concentration of the monomers was fixed at 0.1 M (equimolar) for the implementation of the fed-batch operation mode for the synthesis of POS. Then, by consecutive stepwise addition of 0.1 M/h of each monomer, after 5 h, 0.5 M of each monomer was attained in the polymerization media. Simultaneously, the polycondensation in batch operation mode started with an initial monomer concentration of 0.5 M. Both polycondensation operation modes (batch vs. fed-batch) were evaluated in three different polymerization media: miniemulsion, water, and an organic solvent (cyclohexane:THF 5:1
v/
v) (
Table 4). The initial pH value of the miniemulsion and water was corrected to 5.0 before adding the biocatalyst (PBLI).
High conversion and rapid polyester formation with molecular weights between 2600 and 2700 g mol
−1 occurred in the first hour of polycondensation for all polycondensation media in batch operation mode. After 24 h of polycondensation, the molecular weight in the miniemulsion and water was very similar at 3700 g mol
−1 and 3450 g mol
−1, respectively. The polymer molecular weight at 24 h was significantly higher in the organic solvent than the other two polymerization media (5900 g mol
−1). However, after 48 h, the polymer molecular weight in the organic solvent did not change from 24 h, while there was an increase of 52% and 50% in the miniemulsion and water, reaching 7800 and 6900 g mol
−1, respectively, after 48 h of polycondensation. Furthermore, a lower conversion (86.7%) was observed in the organic solvent (
Table 4). These results clearly show that the biocatalyst reacts at different rates as a function of the monomer or polymer molecular weight already formed. In this case, PBLI showed a higher activity for monomers and short-chain polymer molecules. The enzyme activity slowed down due to the high sterical hindrance for longer-chain polymers accumulated in the three polycondensation media.
The highest conversion (98.3 ± 0.7%) and a molecular weight of 7800 g mol−1 after 48 h of polycondensation were achieved in the miniemulsion in batch operation mode.
The POS synthesis in water at pH 5.0 showed a polyester molecular weight of 6900 g mol
−1, with high conversion (up to 94.3 ± 1.3%) after 48 h. A significant amount of monomers and polymer continued in solid precipitates due to their very low solubility in water but presenting a similar performance as that obtained in the miniemulsion. These results for the polycondensation of suberic acid and 1,8-octanediol at equimolar monomer concentrations (0.5 M) in water at an initial pH of 5.0 are superior to those obtained in the previous
Section 3.2 with initial pH 3.3. These results are also superior to those obtained by several authors with the same monomer and 0.1 M substrate concentrations, with molecular weights and conversions of 1200 g mol
−1 and 12%, and 1600 g mol
−1 and 3%, respectively [
32,
56].
In comparison, the polycondensation in the organic solvent presented in this work the lowest values of polymer molecular weight and conversion after 48 h (5900 g mol
−1 and 86.7 ± 1.8%, respectively). For example, Azim 2006 [
24] observed that the polyester precipitation in diphenyl ether during the polycondensation overcame increasing temperatures from 80 °C to 95 °C, keeping the polyester soluble and proceeding with the enzymatic polyester synthesis in a monophasic polymerization medium.
The lower polycondensation efficiency in the organic solvent may be due to the low initial water content that negatively affects enzyme activity. Another possibility could be associated with the sequestration of water molecules from the enzyme microenvironment layer when using polar organic solvents like THF. The organic solvents are also usually associated with enzyme unfolding or relaxing of the right enzyme 3D structure and, consequently, lower enzyme activity [
15].
The polymer molecular weights obtained in fed-batch operation mode were smaller than the ones assayed in batch operation mode for all polycondensation media after 48 h (
Table 4). These results indicate that a high monomer concentration (0.5 M) may not have significant enzyme inhibition of the PBLI biocatalyst in POS synthesis in the miniemulsion and water but may negatively affect enzyme activity during polycondensation in the organic solvent (
Table 4). This enzyme activity inhibition in the organic solvent is confirmed, as for the lowest monomer concentration (0.1 M) in the first hour of polycondensation in fed-batch operation mode, the polymer molecular weight reached a higher value (4500 g mol
−1) in comparison to batch operation mode (2700 g mol
−1) for the same time of polymerization.
Additionally, high POS solubility in the cyclohexane:THF (5:1
v/
v) solvent could enhance the initial polymer molecular weight. The conversion difference achieved after 1 h of polycondensation of 74.4 ± 1.4% and 53 ± 2.3% for fed-batch and batch operation modes, respectively, also confirm the enzyme inhibition in the organic solvent for a high monomer concentration (0.5 M) (
Table 4). Other authors have already observed a similar effect where the fed-batch operation mode protected the enzyme preparation against the inhibitory effect of a high substrate concentration in an organic solvent [
47,
55].
The fed-batch operation mode in the miniemulsion and water showed rapid and efficient formation of oligomers or small-chain polyesters after 1 h due to the high conversion (100 ± 0.5%) observed (
Table 4).
Despite the better polycondensation performance of PBLI in fed-batch operation mode in the organic solvent after 1 h of reaction, equal molecular weights were achieved after 48 h of polycondensation (5900 g/mol−1) in batch operation mode at 24 and 48 h. This result suggests that there is another factor affecting the PBLI activity negatively in the organic solvent.
These results showed that regarding the polymer molecular weight and conversion, the batch operation mode at equimolar monomer concentrations (0.5 M) is more appropriate for the synthesis of POS in the miniemulsion and water. In contrast, the fed-batch operation mode should be implemented for polycondensation in an organic solvent.
3.4. Effect of Stirring Type on POS Synthesis
Regardless of polymerization, the media must be homogeneously stirred for the full period of the polycondensation using an appropriate agitator design. Radial flow impellers are frequently used in emulsion polymerization at the laboratory scale [
36].
The POS synthesis was tested in a small reactor containing an impeller (RI) constituted by a magnetic stirrer rod with a two-blade propeller and compared with the previously described glass vial reactor with a magnetic stirrer (RMS) to analyze the effect of stirring type on the polycondensation conditions. The polymerization working volume in the RI increases 5-fold (50 mL) compared to the RMS (10 mL). The polycondensation was performed in batch operation mode at equimolar monomer concentrations (0.5 M) and using PBLI (8 mg mL
−1) at 45 °C in the miniemulsion, water, and the organic solvent (cyclohexane:THF 5:1
v/
v). The initial pH was corrected to 5.0 in the aqueous polycondensation media (miniemulsion and water) before adding the biocatalyst (
Table 5).
The stirring rate of 250 rpm of the impeller was insufficient for efficient biocatalyst dispersion in the reactor with an impeller (RI), so the stirring rate of the impeller was enhanced to 500 rpm.
Despite the increase in the impeller stirring rate (500 rpm), the homogenization of polycondensation media after 1 h was inefficient inside the reactor with an impeller (RI) due to increased viscosity with polyester formation. Despite the higher solubility of the monomers and polyester in the organic solvent, the formation of a like-gummy was very viscous, which made the homogenization of the polycondensation media difficult inside the RI. Consequently, the enzyme–polymer–monomer contact was limited after 1 h of the POS synthesis in the RI. Additionally, stirring inside the RI during the POS synthesis started to be irregular between 1 and 8 h in the three polycondensation media. The impeller showed difficulty in moving and stopped several times, and for this reason, the polymerization was interrupted after 8 h. For comparison, the polycondensation in the RMS was also finished at 8 h.
The conversion after 1 h of polycondensation in the miniemulsion was higher (92 ± 0.9%) in the RMS in comparison to the RI (45 ± 2.1%). Yet, the same POS molecular weight was observed in both types of reactors (
Table 5). The same tendency was observed for the polycondensation in water. In the organic solvent, a lower conversion for the RMS (53.9 ± 2.0%) compared with the RI (63.1 ± 1.8%) was observed. The polymer molecular weight stayed equal for both types of reactors.
The conversion in the miniemulsion (92.3 ± 0.9%) was higher than in water (72.6 ± 1.4%) and the organic solvent (53.9 ± 2.0%) in the first hour of polycondensation in the RMS, but in the RI, after 1 h, the conversion was lower (45.3 ± 2.1%) in the miniemulsion than in water (56.8 ± 2.0%) and the organic solvent (63.1 ± 1.8%) (
Table 5). After 8 h of polycondensation in water, the polymer molecular weight was 3460 and 2690 g mol
−1 in the RMS and the RI, respectively (
Table 5).
The conversion in the RI (78.8 ± 1.3%) after 8 h of polymerization in the organic solvent was lower, around 8%, compared to the conversion in the RMS and approximately 16% and 20% lower compared to the conversion in the RI for the miniemulsion and water, respectively. The polymer molecular weight obtained after 8 h in the organic solvent was 4900 g mol−1 (RMS) and 2690 g mol−1 (RI).
Even though the conversion increased slightly between 1 and 8 h for POS synthesis in the RI, the polymer molecular weight stabilized to about 2690 g mol
−1 for all polycondensation media tested (
Table 5). These molecular weights are similar to those achieved in
Section 3.3 with direct magnetic stirring (250 rpm) in glass vials (RMS) in the miniemulsion and water (2600 g mol
−1 and 2700 g mol
−1, respectively) after 1 h polycondensation. However, they are inferior than the POS molecular weights obtained in these polymerization media after 48 h (5900 and 7800 g mol
−1, respectively) (
Table 4). These results suggest poor performance of the polycondensation reaction in the RI for 8 h due to high viscosity and deficient homogenization of the biocatalyst independent of the polycondensation media tested.
The geometry, diameter reactor/impeller ratio, type of impeller, and viscosity of the polycondensation media should be optimized due to the complexity of these parameters and their effect on mass transfer resistance, mainly when using a biocatalyst constituted by an enzyme immobilized on a support surface. Meyer (2003) [
57] confirmed that “scale-up procedures are not trivial, and special attention has to be paid to this viscosity increase during polymerization, decreasing the heat- and mass-transfer rates” [
58]. Therefore, different reactor/impeller geometry and type of impeller (e.g., marine propeller) are fundamental to be studied and optimized, considering these performance limitation factors on POS synthesis in this work.
3.5. Reutilization of the PBLI Biocatalyst during POS Synthesis in the Reactor with an Impeller (RI)
Despite the lower performance of the POS synthesis observed in the reactor with an impeller (RI), the RI was chosen to evaluate the reutilization of the PBLI biocatalyst as the stainless steel network can retain the biocatalyst totally at the end of each reutilization cycle. This way, the biocatalyst loss was avoided if the filtration was performed outside the reactor, and it could contribute to a significant experimental error, especially when working with a small amount of biocatalyst.
Based on the assumption that an increase in polymerization temperature decreases the viscosity, the polycondensation temperature was tested now at 60 °C in the small reactor with an impeller (RI) for up to 2 h of POS synthesis since no significant variation of the polymer molecular weight and conversion was observed for a more extended period (
Table 5).
Additionally, the PBLI reutilization (8 mg mL
−1) through successive cycles of the POS synthesis was tested. The reuse of this biocatalyst was performed and compared in the miniemulsion, water, and the organic solvent system (cyclohexane:THF 5:1
v/
v) in batch operation mode at an equimolar substrate concentration (0.5 M) at 500 rpm and 60 °C and initial pH 5.0 in the miniemulsion and water (
Figure 6). The immobilized enzyme preparation (PBLI) was reused for seven successive cycles. The polycondensation time for the three media was 2 h in the first five cycles and 24 h and 48 h for the sixth and seventh cycles, respectively. This way, the immobilized enzyme preparation (PBLI) operational stability was evaluated for a more extended period at 60 °C. The efficiency of enzyme reutilization was expressed through the dicarboxylic acid conversion for the three polymerization media (
Figure 6).
Despite the high viscosity in all polycondensation media, the impeller inside the small reactors showed adequate homogenization under these polymerization conditions (60 °C and 500 rpm).
Figure 6 shows that PBLI could be reused efficiently in the miniemulsion and water at least seven cycles corresponding to 72 h at 60 °C without significant polymer molecular weight variation (around 2700 g mol
−1) and conversion higher than 90% (
Figure 6). The miniemulsion comprised 81% of water, promoted high enzyme stability, and minimized the monomers’ enzyme inhibitory effect observed in the organic solvent. These results also showed that the enzyme was not desorbed from the support surface inclusively at 60 °C despite several incubation cycles of this biocatalyst with Triton X-100 as a surfactant used in the miniemulsion formulation.
The biocatalyst stability in water was also high, and as the monomers are in solid state (precipitated) during the time course of polycondensation, the inhibitory effect was also insignificant (
Figure 6).
In contrast, the loss of the catalytic capacity of the immobilized enzyme preparation (PBLI) in the organic solvent (cyclohexane:THF 5:1
v/
v) was evident. After the second cycle of PBLI reuse in this organic solvent, only about 35 ± 0.2% of conversion was observed after polycondensation for 2 h. A total lack of conversion and no polymer formation occurred after the sixth cycle (
Figure 6). Other authors have confirmed this lack of stability of biocatalysts in an organic solvent when reused in consecutive batch cycles of reaction. For example, Lerin (2011) [
59], using Novozym 435, showed lower conversion (around 10%) after the sixth cycle for 2-ethylhexyl palmitate synthesis and 50% after the eighth cycle of reuse for ascorbyl palmitate synthesis in
tert-butanol. For 1-glyceryl benzoate production using 2-propanol as the organic solvent, 20% conversion was observed after the tenth cycle of reuse [
57].
The low capacity of reutilization of the PBLI biocatalyst during POS synthesis in the organic solvent can be due to the enzyme desorption from the support surface in contact with the mixture of cyclohexane and THF (5:1
v/
v) at 60 °C as this lipase is likely physisorbed onto the support mainly through hydrophobic interactions [
59]. However, this solvent is primarily composed of cyclohexane with a high log
P (~3.4), and during filtration, a precipitate that could reveal the PBL desorption was not observed. Additionally, the
1H NMR spectrum (
Figure 7) did not reveal any peak associated with protein trace contaminating the synthesized polyester in the organic solvent (cyclohexane:THF 5:1
v/
v) (
Section 3.6.1).
The low water content in the cyclohexane:THF (5:1 v/v) solvent, high temperature (60 °C), and high monomer concentration (0.5 M) may decrease enzyme activity and stability after the second cycle of polycondensation in the organic solvent.