Synthesis of Glycerol-Based Biopolyesters as Toughness Enhancers for Polylactic Acid Bioplastic through Reactive Extrusion.

The synthesis of polyesters based on glycerol, succinic acid [poly(glycerol succinate), PGS] and/or maleic anhydride [poly(glycerol succinate-co-maleate), PGSMA] was investigated aiming to produce a green product suitable for toughening of polylactic acid (PLA) using melt blending technologies. The molar ratio of reactants and the synthesis temperature were screened to find optimum synthesis conditions leading to the highest toughness enhancement of PLA. It was found that a molar ratio of reactants of 1:1 glycerol/succinic acid increases the effectiveness of PGS as a toughening agent for PLA, which correlates with the achievement of a higher molecular weight on the synthesis of PGS. The introduction of maleic anhydride as a comonomer for the synthesis of the partial replacement of succinic acid was advantageous for making PGS suitable for reactive extrusion (REX) mediated by free radical initiators. The tensile toughness of the REX PLA/PGSMA blends was improved by 392% compared with that of neat PLA, which was caused by the simultaneous cross-linking of PGSMA within the PLA matrix, and the in situ formation of PLA-g-PGSMA graft copolymers acting as interfacial compatibilizers. Two-dimensional nuclear magnetic resonance and Fourier transform infrared analysis confirmed the formation of PLA-g-PGSMA species on REX experiments. This in turn caused a decrease in the diameter of the PGS particles dispersed within the PLA matrix from >10 μm to approximately 2 μm as observed using scanning electron microscopy. A further increase of 1600% in the toughness of the blends was achieved by lowering the synthesis temperature of PGSMA from 180 to 150 °C. The optimum synthesis conditions for PGSMA leading to the highest increase in the toughness of 80/20 PLA/PGSMA blends were found to be 1:0.5:0.5 mol glycerol/succinic acid/maleic anhydride synthesized at a temperature of 150 °C for 5 h.

Introduction

Polylactic acid (PLA), or polylactide, is a semicrystalline compostable polyester produced industrially from renewable resources that are fermented to form lactic acid and subsequently polymerized to yield high-molecular-weight products.[1] This polymer presents high tensile strength and modulus but inherent brittleness manifested as a low elongation at break (∼3%) and an impact resistance that limits its adoption for flexible plastic parts. Overcoming the inherent brittleness of PLA to expand its application field has been the subject of intensive research over the past decade, with several review articles published on the ways of turning PLA from a stiff and brittle material to a ductile one.[2−4] Successful toughening of PLA has been achieved by melt blending with several commercial and noncommercial polymers including synthetic polyesters, natural and epoxidized rubbers, thermoplastic elastomers, acrylic copolymers, and so forth.[5] Although the successful modification of PLA to yield a tough blend has been achieved, a limiting factor for the introduction of such blends to commercial applications is usually the high cost of the secondary polymer used for the toughening of PLA. An ideal toughening agent for PLA should be synthesized using inexpensive synthesis routes and preferentially from bio-based monomers produced industrially at a low cost.

Glycerol and succinic acid are both bio-based monomers that gather the aforementioned requisites. Glycerol (1,2,3-propanetriol) is the major coproduct of biodiesel industries with a production capacity that has steadily grown in the last decade, reaching over 2 million tonnes per year.[6] Because of this fact, new applications for glycerol are desired to expand the biorefinery concept in the biodiesel sector.[7] This inexpensive molecule is a trifunctional alcohol suitable for performing polycondensation reactions in the presence of dicarboxylic acids yielding hyperbranched polyesters. Among the dicarboxylic acids produced industrially, succinic acid (1,4-butanedioic acid) is currently one of the few dicarboxylic acids produced in a commercial scale from biomass resources through fermentation technologies. The sustainable production of bio-based succinic acid from biomass resources represents a major breakthrough toward the establishment of a bio-based economy. The usage of bio-based molecules such as biodiesel-derived glycerol and fermentation-derived succinic acid for the synthesis of products of commercial interest is in line with the principles of green chemistry.[8] Hence, the development of a glycerol- and succinic acid-based toughening agent for PLA can open the possibility for creating new toughened bio-based blends of PLA with reduced cost and increased sustainability compared with blends prepared from PLA and petroleum-based polymeric additives. Biodegradable polyesters can be synthesized using glycerol and succinic acid in bulk polycondensations in the absence of a solvent and a catalyst.[9−11] These polyesters can be called “green” polyesters because they are synthesized using renewable resources under environmentally favorable reaction conditions and produce no toxic wastes as outlined in green chemistry principles.[12,13] Synthesizing a hyperbranched polyester using glycerol and succinic acid and tuning these polyester properties for an effective toughening of PLA appears as an advantageous strategy toward the creation of sustainable materials with an improved cost/performance balance.

In earlier literature, two examples of PLA toughened by melt blending with glycerol polyesters can be found. The first reported attempt of developing a PLA/glycerol polyester tough material belongs to Gu and co-workers.[14] In this pioneering work, the authors synthesized a polyester using glycerol and sebacic acid as the starting materials and used this polyester for melt blending with PLA. Upon the addition of 15 and 30 wt % glycerol polyester, the elongation at break of PLA was increased from 7 to 155% and 143%, respectively. A second example of the usage of glycerol-based polyesters for the modification of PLA has been recently reported by Coativy and co-workers.[15] In this study, the starting monomers for synthesizing polyesters were glycerol, sebacic acid, and stearic acid, and the reaction was performed using microwave heating. The synthesized polyester material was blended with PLA at different mass ratios, and the optimum mechanical performance was achieved in the blend ratio of 90/10 PLA/glycerol polyester. In this case, the elongation at break of the system was increased from 2.3% for pure PLA to 93% in the 90/10 PLA/glycerol polyester blend. It is noteworthy to mention that in both of these publications, the diacid of choice was not succinic acid and that the reaction conditions were fixed for synthesizing the polyester. To the best of our knowledge, no study has been conducted on the behavior of the PLA/glycerol polyester blend system using a glycerol polyester synthesized using succinic acid as the acid monomer. Moreover, there is no available literature correlating the synthesis conditions of the glycerol polyester with the mechanical performance of the PLA/glycerol polyester blends.

The bulk polycondensation of glycerol and dicarboxylic acids can yield a wide range of products depending on the reaction conditions used. This system leads to elastomeric gel polyesters if enough time is given for the reaction to achieve gelation.[16,17] If the reaction is stopped before the gelation occurs, the polyesters synthesized are recovered as highly viscous liquids with a low molecular weight and a high polydispersity index.[17,18] The monomer ratio and the synthesis temperature can affect the molecular weight and architecture of the products. This in turn can affect the important properties of these glycerol-based polyesters, such as their solubility properties and the mechanical behavior.[9,19] For the development of a polyester material that could be used as a toughness enhancer for PLA using melt blending technology, it is of interest to investigate how the synthesis conditions for the glycerol-based polyester can affect the mechanical behavior of the blend system. Therefore, the objective of the present research work is to investigate the correlations between the reaction conditions used for the synthesis of glycerol-based polyesters and the mechanical behavior of PLA/glycerol polyester blends, aiming to select synthesis conditions that ensure a successful toughening of PLA.

Results and Discussion

Synthesis of Glycerol-Based Polyesters

Table presents the polyester products synthesized in this study. When glycerol (gly) and succinic acid (succ) were used as monomers for the synthesis, the resultant product was poly(glycerol succinate) (PGS). When maleic anhydride (mah) was used in addition to glycerol and succinic acid as monomers, the product was termed poly(glycerol succinate-co-maleate) (PGSMA). The reaction parameters were chosen to produce a wide range of products with different molecular features. First, a gel material was synthesized (PGS gel), which was obtained as an elastomeric material insoluble in common solvents such as tetrahydrofuran (THF) and acetone because of extensive cross-linking.[17] Subsequently, liquid polyesters were synthesized by stopping the reaction before the occurrence of cross-linking. The molar ratio of the reactants was changed around the stoichiometric balance of the reactants (1:1 gly/succ) to produce polyesters with different terminal groups (hydroxyl or carboxyl) according to polymerization theory.[20] Three different liquid products synthesized from glycerol and succinic acid were produced (PGS1, PGS2, and PGS3) with increasing molar ratio of reactants in the range of 0.6:1 to 1.2:1. Finally, polyesters were synthesized using maleic anhydride as a comonomer in addition to glycerol and succinic acid. These polyesters (PGSMA1 and PGSMA2) possess unsaturations on its polymer backbone as a result of the usage of maleic anhydride as the comonomer.[21]

Table 1

PGS Formulations Used

sample ID	monomers	molar ratio	temp (°C)	reaction time (h)	yield (g PGS/g reactants)	Mw (Da)	Mn (Da)	Mw/Mn
PGS gel	gly/succ	0.6:1	180	1.57	0.78	ND	ND	ND
PGS1	gly/succ	0.6:1	180	1.06	0.74	3119	938	3.3
PGS2	gly/succ	0.9:1	180	2.93	0.75	2569	1011	2.5
PGS3	gly/succ	1.2:1	180	5.27	0.79	1234	754	1.6
PGSMA1	gly/succ/mah	1:0.5:0.5	180	2.5	0.89	4379	1119	3.9
PGSMA2	gly/succ/mah	1:0.5:0.5	150	5	0.86	4442	1138	3.9

The Fourier transform infrared (FTIR) spectra of the synthesized polyesters are shown in Figure . A broad peak appears at 3422 cm–1, which corresponds to OH functional groups in the polyester backbone. As the molar ratio of glycerol to succinic acid is increased in the PGS products, the peak at 3422 cm–1 becomes more visible suggesting that the polyesters are hydroxyl-terminated. The peaks at 1716 and 1153 cm–1 correspond to the C=O and C–O–C groups found commonly in polyesters.[22] When maleic anhydride is incorporated to the polyester formulations (PGSMA), a new peak appears at 1644 cm–1 corresponding to the C=C unsaturation in the polymer backbone. In fact, previous researchers observed peaks corresponding to C=C bonds at 1637 cm–1 in methacrylate derivatives.[23] The C=C double bonds in the PGSMA backbone could be used as a site for cross-linking of the polymers in a reactive extrusion (REX) mode using free radical initiators.

Figure 1

Infrared spectra of the polyesters synthesized.

Effect of Addition of Gel Versus Liquid PGS on the Mechanical Behavior of PLA/PGS Blends

As a first approach for the toughening of PLA by PGS, the blending of PLA and gel PGS was performed using a weight ratio of 80/20 PLA/PGS. The gel PGS was dispersed in particles of approximately 6–15 μm in diameter (Figure S1), and the tensile strength and modulus were decreased by 40 and 26%, respectively (Table S1), which was expected from the addition of amorphous PGS to the semicrystalline PLA. Regarding tensile elongation at break and notched Izod impact results, which are indicative of toughness of the system, the elongation at break raised from 4% for pure PLA to 11% for the 80/20 PLA/gel PGS blend, whereas the impact resistance remained unchanged. The limited increase in the ductility of the system was due to the poor dispersion of cross-linked PGS onto the PLA matrix and a weak interfacial adhesion of the blend components. In fact, the cross-linked nature of gel PGS impedes the dispersion of this material onto smaller diameter rubbery particles within the PLA matrix, which has been proven to be one of the key points for the effective toughening of PLA.[24−27]

Aiming to improve the dispersion of PGS within the PLA matrix during the extrusion process, we synthesized a series of liquid PGS by stopping the reaction before the gelation occurrence, as suggested by previous researchers.[18] PGS products synthesized using three different molar ratios of reactants were produced and blended with PLA at a mass ratio of 90/10 PLA/PGS. The mechanical behavior of the system is presented in Table S2. Interestingly, the tensile behavior of PLA/PGS blends varies significantly according to the molar ratio of monomers used for the synthesis of PGS (Figure ). This phenomenon is a result of differences in molecular interactions between blend components because of changes in the molecular weight and the branching architecture of PGS, which arise by changing the molar ratio of the reactants.[9] The elongation at break of the PLA/PGS blends revealed relevant information about the source of the differences observed in the mechanical behavior. In fact, the elongation at break of the system was maximum when the molar ratio of the reactants used for the synthesis of PGS was closer to the stoichiometric balance of the monomers (PGS2, 0.9:1 mol glycerol/mol succinic acid) (Figure a). Gel permeation chromatography traces of the PGS synthesized showed that these hyperbranched condensation products are composed of different molecular weight species ranging from a few hundred to several thousand daltons (Figure b). Elongation at break in a PLA blend system is closely related to the ability of the secondary polymer to establish molecular bonding interactions and entanglements with the PLA matrix. The ability of entanglement of both polymers in the blend is in turn related to their molecular weights and molecular conformations. In particular, for linear PLA, the chain molecular weight required for entanglement occurrence has been reported earlier as 9000 g/mol, whereas the distance between entanglements was of 4000 g/mol.[28] Additionally, it has been reported that for branched PLA, the ability for entanglement is higher due to the higher volume occupied by the branched structures in comparison with that for linear ones.[29] In the hyperbranched PGS synthesized in our study, there are oligomers with a molecular weight superior to 4000 g/mol, particularly, in the products PGS1 and PGS2, whereas in the PGS3 product most of the oligomers have a molecular weight below this value. Accordingly, the elongation at break of the system was higher for blends of PLA with PGS1 and PGS2. This suggests that the molecular weight of the PGS synthesized plays a crucial role in the ability of these oligomers to interact and entangle effectively with PLA to raise the elongation at break of the blend. Because of the hyperbranched nature of PGS polyesters, it is reasonable to expect that these dendrimer-like molecules can entangle with PLA chains if they reach a molecular weight close to the molecular weight between entanglements reported for PLA. In fact, the maximum elongation at break for the PLA/PGS blend was obtained when the number-average molecular weight of PGS was maximum (Table ) and the presence of oligomers of a lower molecular weight (Mn < 1000 g/mol) was minimum (Figure b), indicating that a higher degree of polymerization was achieved.

Figure 2

(a) Mechanical properties of 90/10 PLA/PGS blends and (b) gel permeation chromatography traces of the PGS synthesized. Calibration curve employed is shown above with molecular weight values.

Interestingly, the impact strength of the PLA/PGS blends increases in average with the increasing molar ratio of monomers (Figure a). The effect of the glass transition temperature of the PGS phase on the blend offers an explanation for this mechanical behavior. In fact, the glass transition temperatures (Tg) of the synthesized PGS determined using differential scanning calorimetry (DSC) were −15.2, −14.8, and −29.9 °C for PGS1, PGS2, and PGS3, respectively. This could be attributed to the lower molecular weight of the PGS3 synthesized compared with that of PGS1 and PGS2 (Figure b). A lower Tg of the included phase is recognized as one of the factors leading to a higher impact strength in thermoplastic-elastomer blends because the capability of the elastomer for absorbing the impact energy increases when used at room temperature.[30]

Scanning electron microscopy (SEM) observations of the PLA/PGS blends were performed at the fracture site on notched Izod impact tested specimens (Figure ). The observations seem to be in agreement with the trend observed in the mechanical properties, where the elongation at break was maximized when using PGS2. In this PLA/PGS blend (Figure b), the fracture site surface shows roughness and there is evidence of plastic deformation of the material upon stress, which is commonly seen in ductile materials (Figure b, inset). On the other hand, blends of PLA/PGS synthesized at different molar ratios of reactants displayed a smoother fracture surface indicating a higher brittleness of the material and there was no evidence of formation of plastic fibrils as seen in the PLA/PGS2 case. Similar morphological observations have been provided by previous studies on PLA and hyperbranched polymer blends.[31] These results confirm that while using a stoichiometric balance of monomers for the synthesis of PGS, a higher toughness of the PLA/PGS blend can be achieved. This phenomenon apparently is caused by reaching a higher Mn of the PGS synthesized, which enables molecular entanglements with PLA to occur.

Figure 3

SEM pictures at the fracture site of PLA/PGS notched Izod impact tested samples. PLA blends with (a) PGS1, (b) PGS2, and (c) PGS3.

Reactive Blends of PLA and PGS

In PLA blend systems, previous researchers showed that REX involving the simultaneous cross-linking of the secondary polymer as well as the graft copolymer formation of PLA and the secondary polymer could efficiently increase the ductility of PLA.[27,32] In PGS polyesters, a way to turn them into reactive molecules suitable for REX with PLA is the incorporation of C=C double bonds in their backbones by using maleic anhydride as a comonomer for the PGS synthesis, as demonstrated by previous researchers.[21] In the present work, liquid glycerol-based polyesters were synthesized using a combination of 1:0.5:0.5 mol of glycerol/succinic acid/ maleic anhydride as monomers (PGSMA1 entry, Table ). Although maleic anhydride is currently industrially obtained from petroleum-based resources that decrease the bio-based content in glycerol polyesters, there is scientific interest in the production of bio-based maleic acid and anhydride and thus these molecules could be obtained from bio-based resources in the future.[33] The PGSMA synthesized showed similar molecular weight and polydispersity as PGS synthesized using only succinic acid as the COOH-bearing monomer at an equal molar ratio of glycerol/diacid monomers (Table , entries PGS2 and PGSMA1), but possesses reactive C=C bonds in its backbone as shown in the FTIR spectra (Figure ).

Upon REX of PLA and unsaturated polymers (polymers bearing C=C bonds in their backbones) using peroxides as free radical initiators, there are two simultaneous and competitive mechanisms taking place, as proposed by previous researchers and depicted in Figure .[23,27,32,34] The first phenomenon corresponds to the cross-linking of the secondary polymer within itself by an addition mediated by free radical initiation (Figure , right). The second phenomenon is the formation of graft copolymers of PLA and the secondary polymer created upon hydrogen abstraction in the PLA chain and the addition of secondary polymer free radicals (Figure , left). A third reaction could take place during the REX with the free radical initiator corresponding to the PLA–PLA self-cross-linking or branching after the formation of PLA macroradicals (not shown in Figure ). Although this reaction is possible, previous researchers demonstrated that because the reactivity of C=C bearing molecules is higher than the reactivity of tertiary hydrogen in PLA chains, this reaction is not favored at the concentration of the free radical initiator used in the present study (0.2 phr).[27,32] The simultaneous occurrence of these phenomena ultimately results in the formation of a tough PLA blend system with a secondary cross-linked phase within and with enhanced interfacial adhesion mediated by the graft copolymers that are located primarily at the interface. In our case, these mechanisms could allow for an enhanced toughening of PLA by the creation of secondary cross-linked structures of PGSMA embedded onto the PLA matrix and compatibilized by in situ-formed graft copolymers of PLA-g-PGSMA. The concentration of the free radical initiator used in the REX process can affect the mechanical behavior of the system because of an enhanced formation of PLA graft copolymers at higher initiator concentrations.[27] In the present study, this parameter was fixed to 0.2 phr based on previous literature, where optimum mechanical behavior was achieved at similar initiator loadings.[27] Nevertheless, a more detailed analysis of the initiator loading effect in the present system will be provided in a future work.

Figure 4

Possible routes of reaction in the REX of PLA and PGSMA mediated by free radical initiators. (I) Grafting of PGSMA onto the PLA backbone and/or (II) PGSMA cross-linking.

The mechanical behavior and the percentage of crystallinity of the PLA reactive blends with PGS and PGSMA are shown in Figure and Table S3, where reactive blends compounded in the presence of a free radical initiator are denoted as RPLA. For a proper analysis of REX effects, the PLA matrix was blended both with saturated glycerol polyesters containing no C=C double bonds in their backbones (PGS2) and with unsaturated glycerol polyesters containing reactive C=C bonds in their backbones (PGSMA1). Also, both blend systems were processed with and without the addition of a free radical initiator to properly analyze the different effects contributing to the mechanical behavior of the system. Upon addition of glycerol polyesters (either PGS2 or PGSMA1) to PLA, the tensile strength and the modulus decreased by approximately 40 and 30%, respectively as observed in the previous section by the addition of the soft liquid PGS to the semicrystalline PLA (Table S3). Figure displays the results of elongation at break of the reactive and unreactive blends of 80/20 wt% PLA/PGS or 80/20 wt% PLA/PGSMA, where the most illustrative results of mechanical behavior could be observed. First, it was seen that the addition of the free radical initiator to PLA itself did not significantly change its strain at break (Figure , entry RPLA) and thus this effect is not considered in the analysis of the mechanical behavior of the reactive blend samples. Not surprisingly, in the saturated glycerol polyester case (PLA/PGS2 blend), the ductility of the material that manifested as elongation at break remained unchanged when comparing the nonreactive extrusion and REX cases because of the absence of C=C double bonds in the PGS2 backbone. On the other hand, when unsaturated polyester (PGSMA1) was blended with PLA, the mechanical behavior of the system showed clear differences between the nonreactive extrusion and REXs. The elongation at break of the system was dramatically increased more than 50% when comparing the nonreactive extrusion and REX cases. The increased elongation at break of the reactive PLA/PGSMA1 blend can be attributed to the simultaneous formation of PGSMA cross-linked particles within the PLA matrix and to the formation of PLA-g-PGSMA copolymers, as depicted in Figure . Similar behavior has been observed in reactive PLA blends and has been attributed to an effective stress transfer to the secondary soft phase within PLA mediated by the presence of graft copolymers formed in situ during the extrusion process.[27,32,35] The role of crystallinity of the PLA phase on the blend samples was also investigated by calculating the percentage of crystallinity of the PLA phase based on DSC analysis of the first melting cycle of the samples (Figure S2 and Table S4). The addition of PGS or PGSMA to PLA causes an increase in the crystallinity percentage of the PLA phase of nearly 10% at the most, according to DSC results (Figure ). The increase in the crystallinity of the PLA phase after the addition of PGS or PGSMA could be attributed to PGS droplets acting as heterogeneous nucleating sites, which could enhance the ability of PLA for initiating crystallization. In fact, the cold crystallization temperature of the PLA/PGS and PLA/PGSMA blends was lower than that of neat PLA in all cases, indicating that PLA can initiate crystallization at a lower temperature in the melting cycle as a result of the nucleation sites provided by PGS particles. This effect has been observed previously in PLA blends with amorphous branched additives and attributed to heterogeneous nucleation mechanisms and enhanced mobility of PLA chains because of the presence of the secondary polymer.[32,36] Regarding the effects of crystallinity of the PLA phase on the mechanical behavior of the blend samples, the increase in crystallinity caused by the addition of PGS does not seem to correlate with the elongation at break of the system. In fact, previous researchers have demonstrated that increases in crystallinity of the PLA phase in the range of 10% such as those reported in the present work have a negligible effect on the mechanical behavior of PLA blends.[32] These findings suggest that the enhanced elongation at break obtained in the REX of PLA and PGSMA is a result of an efficient stress transfer from PLA to the PGSMA phase rather than a change of nature of the PLA phase itself from brittle to ductile such as in plasticization strategies. In fact, as pointed out by previous researchers in PLA REX strategies involving the cross-linking of the secondary polymer within PLA, the tensile toughening achieved is mainly due to a stress concentration effect on the rubbery particles created in the extrusion process.[26,35] These soft particles act as stress concentrators inducing multiple crazing on the material. The increase in the crazing volume causes a decrease in the actual stress in each individual craze, which allows for a stable void growing.[35] Upon further tensile load, particle debonding and elongation is achieved, preventing early failure. Thus, the amount of energy dissipated in multiple crazes is higher than in a single craze, which explains the increased tensile toughness observed in the PLA blends studied.

Figure 5

Elongation at break and the percentage of crystallinity of reactive and unreactive blends of PLA and glycerol polyesters [80/20 wt% PLA/PGS or 80/20 wt% PLA/PGSMA].

Another parameter used commonly to evaluate the toughness of materials is impact resistance. This parameter is related to the energy absorption capability until failure of a material under a sudden load, unlike in the tensile testing where the load is usually applied at much lower strain rates.[30] When comparing the PLA toughened materials from the present study with previously reported PLA reactive blends, we could observe that in numerous cases, the impact resistance is also remarkably enhanced in the final material (>500 J/m).[24,25,37−39] In our case, the impact resistance remained statistically unchanged with respect to PLA even in the blends displaying a high elongation at break (24 ± 8.5 vs 22.9 ± 6.3 J/m for RPLA/PGSMA1 and RPLA, respectively). A similar observation was made by Zhang et al. in reactive blends of PLA and unsaturated thermoplastic polyurethane achieving a high elongation at break but unchanged impact resistance.[27] These authors have attributed the low impact resistance of the materials to the thickness of the PLA ligament between dispersed rubbery phases (interparticle distance). In fact, several researchers have emphasized the existence of a critical ligament thickness of PLA of around 1 μm as a necessary condition for achieving impact resistance in PLA.[25,26,40] Our future research will be directed toward achieving a higher impact resistance of the PLA-based materials developed by increasing the amount of PGS content maintaining a small PGS particle size aiming to decrease the PLA ligament thickness in the system.

Further evidence supporting the findings of increased ductility on the reactive blends of PLA and PGSMA was provided by SEM observations of the fracture site on notched Izod impact tested specimens (Figure ). In these photographs, it was observed that when saturated glycerol polyesters (PGS2) were blended with PLA both in nonreactive and reactive modes, the PGS material was distributed within the PLA matrix as droplets of varying sizes, where some of the particles showed diameters smaller than 1 μm and others displayed larger diameters in the 5–15 μm range. This suggests that when the blend is compounded, the PGS material is distributed as small submicronic droplets on the molten PLA matrix by the action of shear in the extrusion process but upon cooling, the PGS particles coalesce due to low compatibility of the phases forming larger PGS droplets within the solid PLA matrix. In the case of unsaturated glycerol polyester (PGSMA1), a clear difference can be observed between the nonreactive and reactive PLA blends (Figure c,d). In the nonreactive case, PGSMA droplets of large diameters greater than 10 μm can be observed indicating that spreading and coalescence of PGSMA also occur upon extrusion of the materials. In the reactive case, the PGSMA droplets showed a significantly lower diameter in the range of 1–3 μm and less and in fact, no PGSMA particles with a large diameter (∼10 μm) were observed as in previous blends. This interesting observation indicates that there was a decrease in surface tension between both components, leading to the decrease of coalescence of the PGSMA phase, which in turn resulted in smaller PGSMA particles uniformly distributed onto the PLA matrix. The decreased surface tension between the blend components can be attributed to the formation of PLA-g-PGSMA copolymers that would be mainly located at the interface of PLA and PGSMA domains acting as a compatibilizer, which prevents the occurrence of coalescence of PGSMA droplets (Figure ). Similar decrease in the diameter of the secondary polymer phase has been reported by previous researchers upon the in situ formation of PLA graft copolymers in REX studies.[34,41]

Figure 6

SEM pictures at the fracture site in notched Izod impact tested 80/20 wt% PLA/PGS or 80/20 wt% PLA/PGSMA blend samples. (a) PLA/PGS2 nonreactive blend, (b) RPLA/PGS2 reactive blend, (c) PLA/PGSMA1 nonreactive blend, and (d) RPLA/PGSMA1 reactive blend.

Direct evidence about the formation of graft copolymers PLA-g-PGSMA was provided by two-dimensional nuclear magnetic resonance (NMR) analysis of both nonreactive and reactive PLA/PGSMA1 blends (Figure ). Heteronuclear multiple bond correlation (HMBC) NMR has proven to be useful in the identification of PLA graft copolymers formed in situ during REX.[41] In this study, the HMBC maps of the PLA/PGSMA nonreactive and reactive blends clearly showed the formation of PLA-g-PGSMA copolymers. In fact, new cross peaks correlating 1H and 13C at 1.52 and 71.6 ppm and at 1.52 and 172.8 ppm could be observed in the blend prepared in the presence of a free radical initiator (H1C2′ and H1C3′ in Figure ). These new correlations can be assigned to the new chemical shifts of the carboxylic and methine groups in PLA as a result of the grafting of PGSMA to the PLA backbone. The formation of graft copolymers of PLA and unsaturated polymers through REX in the presence of free radical initiators has also been reported by previous researchers under similar processing conditions.[23,27,32,34]

Figure 7

HMBC spectra of 80/20 PLA/PGSMA blends prepared with and without initiators.

To further demonstrate the formation of graft copolymers of PLA and PGSMA, the blends were fractionated selectively in solvents. Results obtained for the fractionation are presented in Table S5. It is worth mentioning that only the reactive blend of RPLA/PGSMA1 processed in the presence of a free radical initiator showed a gel fraction significantly higher than zero (6.9 wt %), which indicates that the proposed mechanism of reaction depicted in Figure is correct. That is, a portion of PGSMA cross-links and forms a gel fraction within the PLA matrix, whereas another portion is grafted onto the PLA backbone. Moreover, the absence of a gel fraction on the PLA samples processed in the presence of a free radical initiator (RPLA) confirms that the PLA–PLA cross-linking reactions are not favored and the dominant reactions are the ones depicted in Figure (PGSMA cross-linking and PLA-g-PGSMA copolymer formation), as suggested by previous researchers.[27,32] The exact amount of PGSMA grafting onto PLA is very difficult to determine precisely using NMR because these polyesters are not uniform in molecular weight. Nevertheless, an estimation of the grafting percentage was obtained from the fractionation in the solvents of the blends. In the unreactive PLA/PGSMA blends, 18.1 wt % of the initial load of PGSMA was recovered as soluble PGSMA in acetone. This indicates that most of the PGSMA remains unreacted after the extrusion process. In the reactive blend of PLA and PGSMA, a total of 16.6 wt % PGSMA was recovered as a gel and an extracted fraction. The difference between the PGSMA recovery in the unreactive and reactive cases is thought to correspond to the percentage of PGSMA converted onto PLA-g-PGSMA copolymers (1.5 wt %). Previous researchers have reported similar values of 0.6 wt % of conversion of secondary polymer to PLA graft copolymers in PLA reactive binary blend studies.[41] After drying of the samples dissolved in THF to evaporate the residual solvent, both the soluble and gel fractions were scanned using infrared spectroscopy. Figure a shows the collected spectra in the region of carboxylic and ester functionalities (1700–1750 cm–1) for the RPLA/PGSMA1 blend. The full wavenumber range spectra can be seen in Figure S3. RPLA showed a distinctive peak at 1747 cm–1 corresponding to both ester groups and carboxylic acid terminal groups in the molecule. PGSMA, on the other hand, shows a peak at 1716 cm–1 also corresponding to terminal carboxylic groups and ester bonds and a peak at 1644 cm–1 corresponding to C=C double bonds within its backbone. The gel fraction recovered from the blend shows absence of the peak at 1644 cm–1 indicating consumption of the C=C double bonds via cross-linking or grafting, as depicted in Figure . This observation is in agreement with previous researchers who have observed the disappearance of C=C double bonds at 1637 cm–1 in free radical-mediated REXs of PLA with methacrylate derivatives.[23] Additionally, the major carboxylic peak in the gel fraction appears at 1728 cm–1 where a shoulder in the region of 1747 cm–1 can be observed. In fact, this peak was deconvoluted into two peaks at 1726 and 1758 cm–1 (Figure b) corresponding to PGSMA and PLA carboxylic functionalities, which indicates that this fraction is composed of both PLA and PGSMA molecules. Considering that the gel fraction of PLA/PGS blends was significant only for the RPLA/PGSMA blend and that the gel recovered from this sample was washed several times with fresh solvent after dissolution, this fraction is free of any un-cross-linked PLA and PGSMA. Therefore, the presence of both PLA and PGSMA ester peaks in the gel fraction suggests that some PLA molecules are grafted onto PGSMA as the free radical-mediated reactions take place. Upon cross-linking of PGSMA, these PLA segments are retained in the gel fraction of the blend, which causes the double peak observed at 1728 cm–1. Previous researchers have found similar observations on the infrared spectroscopy analysis of PLA blends with unsaturated polymers performed in REX mediated by free radical initiators.[27] This result is in agreement with the results obtained in the 2D NMR analysis.

Figure 8

FTIR spectra of the fractionated RPLA/PGSMA1 blend. (a) Zoomed-in carboxylic acid-ester region of neat polymers and blend fractions and (b) deconvolution of the carboxylic peak on the gel fraction extracted.

Effect of the Synthesis Temperature

The effect of the synthesis temperature on the mechanical properties of PLA/PGS was investigated by synthesizing another PGSMA formulation at 150 °C as opposed to the first PGSMA synthesized at 180 °C (Table , entries PGSMA1 and PGSMA2). Subsequently, the reactive blending of PLA and PGSMA2 was performed under identical conditions as those for the blend of PLA and PGSMA1 (180 °C, 2 min extrusion, 100 rpm, 0.2 phr free radical initiator). Tensile testing results were similar for both blends regarding tensile strength and modulus (Table S6). Elongation at break data presented a very significant difference as shown in the representative curve of stress versus strain for both blend systems (Figure ). In fact, the elongation at break of the system was remarkably enhanced from 58.6% in RPLA/PGSMA1 to 150% in the blend of PLA and PGSMA synthesized at a lower temperature (PGSMA2 synthesized at 150 °C). Consequently, the toughness of the blend (measured as area under the stress–strain curve) was increased fourfold from 10.2 to 41.6 MJ/m3 comparing both blend systems and increased 16-fold comparing PLA and RPLA/PGSMA2 blends. On the basis of molecular weight determinations (Table ), the difference in the molecular weight between both PGSMA1 and PGSMA2 synthesized seems to be not large enough for causing such a dramatic change in the mechanical behavior (Mn equal to 1119 and 1138 for PGSMA1 and PGSMA2, respectively). The increased toughness of RPLA/PGSMA2 is likely to be caused by changes in the architecture of the PGSMA molecule because of the differences in the synthesis temperature. In fact, several researchers have demonstrated that in the glycerol-based polyester synthesis, the temperature of the synthesis can significantly affect the topology of the hyperbranched molecules. Li et al. showed that in glycerol and sebacic acid polycondensations, the reactivity of the secondary hydroxyl group on glycerol is lower than the reactivity of primary hydroxyl groups.[42] Another study dealing with the polymerization of glycerol and sebacic acid showed that a decrease in the reaction temperature from 140 to 120 °C decreased the conversion of carboxylic acid groups at gelation from 0.999 to 0.994.[19] These findings indicate that because of the lower reactivity of secondary hydroxyl groups on glycerol monomers, an increase in the reaction temperature will likely cause an increase in the esterification of secondary hydroxyl groups and thus lead to more branched products, as suggested by previous researchers.[17] Thus, in our study, it is reasonable to expect products with longer branches and a lower degree of branching when decreasing the synthesis temperature from 180 to 150 °C. In polymeric systems, the ability of the polymer chains to entangle with each other is inversely proportional to their branching degrees.[43] The increased toughness of the RPLA/PGSMA2 reactive blend could be attributed to a higher level of entanglement between both blend components because of a lower branching degree and longer linear segments produced by decreasing the esterification on secondary hydroxyl groups when the synthesis temperature is decreased.

Figure 9

Stress vs strain curves indicating fracture toughness (UT) and appearance of (A) neat PLA, (B) 80/20 RPLA/PGSMA1, and (C) 80/20 RPLA/PGSMA2.

The toughened PLA/PGSMA materials developed in the present study via REX are comparable in terms of tensile toughness with PLA/poly(glycerol sebacate) (PGSeb) blends reported previously (155% elongation at break in the 85/15 PLA/PGSeb blend)[14] and superior to PLA/poly(glycerol sebacate-co-stearate) (PGSebSt) (90% elongation at break in the 90/10 PLA/PGSebSt blend).[15] Interestingly, these previous PLA blends were developed by conventional unreactive extrusion as opposed to the REX strategy illustrated in this study. In unreactive blends presented in this study with a similar addition of PGS (80/20 PLA/PGS), the elongation at break achieved was much lower (11.7%) than that in previous PLA/PGSeb unreactive blends. PGSeb presents a higher hydrophobic character compared to PGS because of the longer aliphatic chain on sebacic acid as compared with that of succinic acid. An increase in aliphatic chains in the diacid comonomer in the glycerol polyester could increase the compatibility with PLA as predicted by the solubility parameter theory.[44] This could contribute in achieving a better interface between the blend components mimicking the effect of the graft copolymer formation observed in reactive blends in the present study and enhancing the stress transfer between blend components to achieve a high elongation at break.

Conclusions

The synthesis of glycerol-based polyesters (PGS) oriented to the application of these biomacromolecules as toughness enhancers for PLA has been investigated. The effect of the molar ratio of reactants, the monomer type, and the temperature of synthesis on PGS physicochemical properties and its impact on the suitability of PGS as a toughening agent for PLA has been established. It was found that the utilization of liquid polyesters synthesized at a stoichiometric ratio of monomers yielded tougher PLA blends because of achieving a higher molecular weight of the synthesized PGS. The usage of maleic anhydride as a comonomer for the synthesis of PGS can turn these polyesters into reactive species, which is advantageous for the compatibilization of the PLA/PGS blends producing a significant improvement in their toughness. Upon REX in the presence of free radical initiators, an in situ compatibilization effect was realized involving simultaneous cross-linking of PGS within the PLA matrix and the formation of PLA-g-PGSMA copolymers, which act as interfacial compatibilizers in the system increasing the toughness of the system by 392% (from 2.6 to 10.2 MJ/m3). Furthermore, a decrease in the temperature of the synthesis of PGS from 180 to 150 °C led to a higher improvement in the toughness of the blends of 1600% (from 2.6 to 41.6 MJ/m3), which was attributed to a decrease in the branching degree of the synthesized polyesters. For an effective toughening of PLA by reactive melt blending with glycerol-based polyesters, the optimum synthesis conditions found were 1:0.5:0.5 mol glycerol/succinic acid/maleic anhydride synthesized at a temperature of 150 °C for 5 h.

Experimental Section

Materials

Technical glycerol (gly) was obtained from a local distributor (BIOX Corporation, Canada), specified as purified glycerol from biodiesel production. The average glycerol content of this source was 95 wt %.[16] Pure succinic acid (succ) (99+ wt %, KIC chemicals, UK) and maleic anhydride (mah) (99 wt %, Sigma Aldrich, Canada) were purchased and used as received. THF, acetone (99.8 wt %, Fisher Scientific, Canada), and deuterated THF ([D8] THF, ≥99.5 atom % D, contains 0.03% (v/v) TMS, Sigma Aldrich, Canada) were used as received. Polylactic acid (Ingeo 3251D), a product of Nature Works, as white pellets with a melting point of 155–170 °C and a melt flow index of 30–40 g/10 min (190 °C, 2.16 kg) according to the supplier was used in this study. 2,5-Bis(tert-butyl-peroxy)-2,5-dimethylhexane (Luperox 101) with a technical grade of 90% was used as a free radical initiator for REXs (Sigma Aldrich, Canada).

Synthesis of Glycerol-Based Polyesters

Polyesters were synthesized using a one-pot polycondensation procedure without the addition of a solvent or a catalyst to the system. For the reactions, a mixture of 200 g of monomers (glycerol, succinic acid, and maleic anhydride) was put together in a 1 L glass reactor equipped with a stirrer and a Dean–Stark apparatus to collect the water formed during the condensation reaction. The temperature was fixed at a certain value, and the mixture was agitated at a fixed speed (250 rpm). When glycerol and succinic acid were used as monomers for the synthesis, the product was called PGS. When maleic anhydride was added as a third monomer to the formulations, the synthesis product was termed PGSMA. For the synthesis of PGS gel materials, the reaction was continued until the material changed from a viscous liquid to an insoluble gel because of extensive cross-linking. For the synthesis of PGS liquid materials, the time for gelation was recorded on an initial screening experiment as the time passed from the monomers reaching the reaction temperature setpoint (180 or 150 °C) and the material changing to a rubbery state. At this point, the material wrapped around the mechanical stirrer, making it impossible to continue the reaction in melt conditions.[15] In hyperbranched polymer synthesis by polycondensation, the molecular weight of the products is increased exponentially in the later stages of polymerization until the gelation is achieved.[18,19] With this in mind, the synthesis reaction was repeated and stopped 5 min before the previously recorded gel time by removing the vessel from the heating element and stopping the mechanical stirring to obtain non-cross-linked liquid PGS polymers of the highest possible molecular weight before gelation. Although the selection of quenching 5 min before gelation is arbitrary, this synthesis strategy has been commonly used as an end point for gel-forming hyperbranched polymers.[45] The polyesters synthesized using this approach were fully soluble in THF, indicating the absence of gel macromolecules.

Blending of PLA and PGS or PGSMA Materials

The blends of PLA and PGS or PGSMA materials were fabricated using extrusion. A twin screw extruder (Xplore, DSM, the Netherlands) was used with a processing temperature of 180 °C at a screw speed of 100 rpm. For the extrusion, 10 g of materials were weighed in a preselected ratio of the blend components and added to the extrusion chamber simultaneously. For gel PGS, the materials were cut to a desired weight using a knife, whereas for liquid PGS or PGSMA, the material was kept in a plastic syringe that was heated to 80 °C using a temperature-controlled syringe heater (New Era, NY, USA) to melt PGS for weighing the desired amount. In REXs, the free radical initiator was added simultaneously with PLA and PGS or PGSMA at a fixed concentration of 0.2 phr. This concentration of initiator was selected based on previous studies for PLA reactive blends showing an optimum performance at initiator concentrations in the range of 0.1–0.3 phr.[27,32] After 2 min of extrusion, the material was collected in an injection device preheated to 180 °C and injected into molds kept at 30 °C for obtaining specimens for mechanical testing.

Mechanical Testing

Tensile and flexural testing of polymer samples were performed on an Instron universal testing machine (Instron, Canada). Tensile testing was carried at a crosshead speed of 5 mm/min following the procedure described on the ASTM D638 standard. Flexural testing was performed at a crosshead speed of 1.4 mm/min following the procedure outlined on the ASTM D790 standard. For each test, five specimens were tested, and results reported are the average and standard deviation of the original data. Notched Izod impact testing was performed on an impact testing machine equipped with a hammer of 0.5–5 ft lbs following the procedure from the ASTM D256 standard. The impact samples were notched 48 h before conducting the testing. The impact results reported are the average of six specimens tested.

Fractionation of PLA Blends in Solvents

The blends of PLA and PGS or PGSMA were subjected to fractionation in organic solvents to determine the soluble and gel contents. To determine the gel content, a sample of approximately 200 mg (mi) was dissolved in 30 mL of THF at 40 °C for 2 h. THF was able to dissolve both PLA and the un-cross-linked fraction of PGS or PGSMA materials. The cross-linked insoluble fraction was recovered by centrifugation at 8000 rpm for 10 min and washed twice with 30 mL of THF. After drying for 5 h at 80 °C, the recovered fraction was weighed (mf) and the gel fraction was calculated as

To determine the PGS or PGSMA un-cross-linked fraction in the PLA/PGS or PLA/PGSMA blends, a sample of approximately 200 mg (wi) was immersed in 30 mL of acetone and kept under agitation for 24 h at room temperature. Acetone can dissolve PGS or PGSMA un-cross-linked materials but does not dissolve PLA or cross-linked PGS or PGSMA materials. In this way, the un-cross-linked PGS or PGSMA fraction was extracted from the blends. After filtration and washing twice with 30 mL of acetone, the samples subjected to extraction were dried at 80 °C for 5 h and weighed (wf). The extracted fraction was calculated as

Characterization

Molecular weight was determined using gel permeation chromatography in a Viscotek GPCmax system (Malvern Instruments, UK) equipped with a refractive index detector. Two Styragel columns (HR1 and HR2, Waters Corporation, USA) and one PLgel-Mixed E column (Agilent Technologies, USA) were connected in series for the injection of samples. THF was employed as a solvent at a flow rate of 0.3 mL/min and a temperature of 40 °C. Before injecting the PGS samples, a calibration curve was constructed using a series of eight polyethylene glycol standards with molecular weights (Mn) in the range of 106–7830 Da (Easy Vial, Agilent Technologies, USA). FTIR spectra were collected using a Thermo Scientific Nicolet 6700 FTIR spectrometer equipped with an attenuated total reflection (ATR) accessory. For each sample, 64 scans were collected with a resolution of 4 cm–1. Scanning electron images were obtained from fractured impact specimens using an Inspect S50 scanning electron microscope (FEI, USA). The samples were coated with a thin layer of gold before imaging to prevent deformation of the samples upon observation. DSC traces were obtained under a nitrogen flow of 50 mL/min (DSC Q200, TA instruments) using samples of 10–20 mg. For the neat PGS or PGSMA, the materials were equilibrated at 100 °C following a ramp of 10 °C/min to −80 °C and a final temperature ramp of 5 °C/min to 100 °C, where the glass transition temperature (Tg) was observed. For the PGS or PGSMA reactive and unreactive blends, the samples were subjected to an initial heating cycle at 10 °C/min from 40 to 180 °C followed by a cooling cycle at 5 °C/min to −60 °C and a final heating cycle at 10 °C/min to 180 °C. The crystallinity percentage of the samples was calculated using the data from the first heating cycle aswhere ΔHm and ΔHcc correspond to the enthalpies of melting and cold crystallization and ΔH100 corresponds to the theoretical melting enthalpy of a 100% crystalline PLA (93 J/g).[36]

NMR experiments were recorded in [D8] THF. The blend samples were dissolved in [D8] THF and filtered through a nylon syringe filter of 0.2 μm pore size before injection to remove the gel fraction from the samples. The residual solvent signal at δ = 1.72 and 67.2 ppm, for 1H and 13C nuclei, respectively, were used as a standard reference. Experiments were conducted using a Bruker AVANCE III spectrometer with a 1H operating frequency of 600.0 MHz and a 13C operating frequency of 150 MHz, using a 5 mm TCI cryoprobe. The sample temperature was regulated at 22 °C. HMBC spectra (8 Hz coupling) were collected in PLA/PGSMA blends extruded in the absence and presence of free radical initiators to investigate the formation of the PLA-g-PGSMA copolymer.