Super Toughened Poly(lactic acid)-Based Ternary Blends via Enhancing Interfacial Compatibility.

Novel super toughened bioplastics are developed through controlled reactive extrusion processing, using a very low content of modifier, truly a new discovery in the biodegradable plastics area. The super toughened polylactide (PLA) blend showing a notched impact strength of ∼1000 J/m with hinge break behavior is achieved at a designed blending ratio of PLA, poly(butylene succinate) (PBS), and poly(butylene adipate-co-terephthalate) (PBAT), using less than 0.5 phr peroxide modifier. The impact strength of the resulting blend is approximately 10 times that of the blend with the same composition without a modifier and ∼3000% more than that of pure PLA. Interfacial compatibilization among the three biodegradable plastics took place during the melt extrusion process in the presence of a controlled amount of initiator, which is confirmed by scanning electron microscopy and rheology analysis. The synergistic effect of strong interfacial adhesion among the three blending components, the decreased particle size of the most toughened component, PBAT, to ∼200 nm, and its uniform distribution in the blend morphology result in the super tough biobased material. One of the key fundamental findings through the in situ rheology study depicts that the radical reaction initiated by peroxide occurs mainly between PBS and PBAT and not with PLA. Thus, the cross-linking degree can be controlled by adjusting renewable sourced PLA contents in the ternary blend during reactive extrusion processing. The newly engineered super toughened PLA with high stiffness and high melt elasticity modulus could reasonably serve as a promising alternative to traditional petroleum plastics, where high biobased content and biodegradability are required in diverse sustainable packaging uses.

Introduction

Research on toughness of polymer-based materials has been attracting attention from scientists and industry for decades because it is one of the important properties in the real application of these materials.[1] Impact resistance can be used to measure the ability of a material to withstand the stress of a sudden load, which is probably the most important mechanical property because it is closely related to the safety, liability, and service life of the plastic product.[2] Therefore, toughening modification of traditional brittle plastics such as polystyrene (PS)[3] and polypropylene (PP)[4] has been extensively researched and reported. However, with the increasing attention on global environmental issues and shortages of our finite petroleum resources, sustainable biodegradable polymers with superior properties have drawn much interest for their advantage of replacing the century’s traditional nonbiodegradable polymers in various applications.[5] Many novel biobased and biodegradable polymers with different properties, such as polylactide (PLA), poly(butylene succinate) (PBS), polycaprolactone (PCL), poly(butylene adipate-co-terephthalate) (PBAT), and polyhydroxyalkanoate (PHA), have been developed and modified extensively in recent years.[6] Nevertheless, most of the inherent inferior properties of these polymers have limited their applications for many structural materials in the market when used alone.

For example, PLA, an important 100% biobased and biodegradable polyester on a commercial scale, possesses high modulus and strength comparable to commercial polystyrene (PS). However, its brittleness and low impact resistance greatly limit its applications. Therefore, different kinds of modifications on these biopolymers, including incorporating thermal, mechanical, processability, and gas-barrier properties, have been investigated in the past years.[7] As mentioned in the beginning, toughening modification with acceptable stiffness and strength is so important in the application, which holds an absolutely large part of the research, especially on brittle polymers such as PLA.[8] Toughening the brittle polymers, incorporating a rubber phase or melt blending with other tough elastomers,[9] optimizing morphologies of the blends such as core–shell structures,[10] and adding a third polymer[11] or filler (microfibers or nanofillers) to the polymer matrix[6] to increase the blend compatibility have been used as main approaches.

As important elastomers, ethylene-based rubbers, such as poly(ethylene-glycidyl methacrylate) (EGMA),[12] ethylene methyl acrylate–glycidyl methacrylate (EMA–GMA), and poly(ether-b-amide) copolymer (PEBA), have been widely used in improving the toughness of biopolymers and given positive results. The toughening mechanisms of PLA by EMA–GMA and EMAA-Zn have already been studied systematically.[13] The studies found that the type of compatibilizer, blending temperature, composition ratio, phase morphology, and suitable interfacial adhesion influence the final impact strength of the blends. Yuryev et al.[14] reported the toughening of PLA by EMA–GMA with Joncryl ADR-4368C as the compatibilizer and found that the processing temperature had an influence on the final impact strength of the PLA blends. Zhang et al.[15] prepared a toughened PLA/EMA–GMA/PEBA ternary blend by controlling the morphology. Other studies on natural rubbers[16] and epoxidized rubbers[17] have also been investigated. Recently, fully biobased thermoplastic vulcanizates consisting of PLA and biobased elastomers such as Poly (lactate/butanediol/sebacate/itaconate) (PLBSI)[18] and vulcanized unsaturated aliphatic polyester elastomers (UPEs)[19] by dynamic vulcanization with high toughness have also been prepared and researched by others. Two issues are worth noticing from the above research: (i) the toughening efficiently worked only when interfacial compatibilization between PLA and the other polymers is improved on blending with rubbery polymers (elastomer), and (ii) the incorporation of the toughened elastomer or a third polymer such as poly(methyl methacrylate) (PMMA)[11b] sacrificed the biodegradability of PLA adversely.

Biodegradability should be taken into consideration when using modified biobased polymers, especially for sustainable packaging uses. Melt blending with fully biodegradable polymers to fabricate toughened green blends seems to be more important and interesting in research and industry. Therefore, various kinds of biodegradable polymers have been researched. Many studies on improving the physical properties of neat polymers by binary blending or ternary blending methods have been done previously.[20] In the melt blending of different polymers, the compatibility between the phases is a vital factor that influences the performance of the end products.[21] Recently, Guo et al[11b] reported an increase in the compatibility of 70PLA/30PBAT by structuring nanoscale interfacial PMMA, where the impact strength can be improved to 188 J/m. The resulting impact is not significant to substitute high-impact plastics such as acrylonitrile–butadiene–styrene (ABS). This is because the interface adhesion for high-impact strength should be much stronger compared to that required in improving tensile strain, which is ascribed to the debonding with suitable interface interactions.[13d]

On the basis of the above discussion, fully biodegradable polymers are selected to fabricate super tough green blends with increased interfacial compatibility between the different polymers. Ravati and Favis have shown that a complete wetting tricontinuous morphology formed in the 33PBS/33PLA/33PBAT ternary blends can be achieved.[22] They found that balanced toughness and strength are obtained in the blends with three highly interacting phases where all components are fully continuous.[20a] By adjusting the composition ratios of the three biodegradable polymers in our studies, we aimed to increase the biocontents.

The novelty of our studies includes the following: (i) the fully biodegradable matrix design with improved impact strength to realize super toughness (notched impact strength >530 J/m) of the final blends. (ii) The mechanical properties of the PLA/PBS/PBAT blends were governed by a tailored morphology and controlled viscosity of the toughened PBAT component during the reactive extrusion processing. (iii) Also, by using PBS in the ternary blend, which possesses high compatibility with the toughening phase, for example, PBAT, to decrease the amounts of the toughening phase used, we could balance the stiffness–toughness in the final products. With significantly improved super toughness and melt viscosity, the fully biodegradable blends can find extensive application to replace some conventional petroleum-based plastics, creating a sustainable future.

Results and Discussion

Property Performance Evaluation

The toughness performance of the blends with and without peroxide is evaluated in our work by notched impact strength firstly and demonstrated in Figure . It can be found that the notched Izod impact strength of the blends is increased significantly because of the compatibility effect brought about by the in situ reaction. The notched impact strength of the compositions is shown in Figure a. Although PBAT shows high toughness with nonbreak impact performance, the impact strength of the ternary blends without peroxide is still as low as 100 J/m, much lower than commercial ABS with 450J/m.[24] Interestingly, a sharp brittle–ductile transition occurred when peroxide was added into the ternary blends with increasing PBS and PBAT amounts to 40 wt %. The impact strength of the blends with peroxide is almost 10 times that without peroxide, from break to hinge as shown in Figure . Compared to that of pure PLA, the level of impact strength of the blends is almost 30 times and even higher than that of commercial ABS, showing super toughness. The impact strength of the samples showed little change with increasing peroxide concentrations, remained at high levels (>700 J/m), and reached 1000 J/m, which is rarely reported previously. As we know, the impact strength is important for the application because it can be used to evaluate the ability of a material to withstand sudden fracture or the amount of energy required to propagate a crack.[25] It is believed that the super tough biodegradable ternary blends can find an application in fields requiring high toughness.

Figure 1

Impact performance of the ternary blends (a) reported in this work compared to (b) other works: (a) notched Izod impact strength of the PLA ternary blends and (b) comparison of stiffness–toughness properties between this work and other reported super tough PLA blends, PLA/UPE/DCP,[19] as well as a petro-based polymer, ABS (Magnum 1150 EM).[23]

Normally, when increasing the toughness of a material by using high contents of rubber, the stiffness of the material decreases dramatically. Maintaining the high stiffness while improving the toughness is of a significant value in polymer research. The data in Figure b address the advantage of our reported blends compared to the traditional petrol-based plastics, such as ABS, and most of the earlier reported super tough PLA blends. Compared to PBAT, a more toughened bioplastic but with a low modulus value of ∼70 MPa, the PBS component has a higher modulus value (∼700 MPa) that helps in improving the stiffness of the PLA-based ternary blend during reactive blending.

The typical tension stress–strain curves of the ternary blends are shown in Figure , and the relative mechanical properties are obtained and depicted in Figure . The stiffness (tensile modulus in Figure a) is mainly decided by the PBS and PBAT composition ratios. Without peroxide, the modulus decreased drastically with increasing PBS and PBAT amounts, and this occurrence is fairly common because the additional PBS and PBAT are both rather tough materials with low moduli, especially PBAT with a modulus about 70 MPa. However, the modulus increased at high peroxide contents, especially for P-S10T10-I0.75 and P-S20T20-I0.75. It is suggested that the increased modulus probably results from the improved crystallinity of PLA as the branching chains can act as a nucleation agent of PLA, which can be explained by the thermal properties as shown in Figure . The tensile strength (Figure b) lies in the ranges of 33 to 38 MPa, within 10–20 wt % PBS and reaches the highest at 40 wt % PBS content, 0.75 phr peroxide. This is ascribed to the strain hardening in tensile stretching, because of the high contents of PBS and PBAT, as shown in Figure b.

Figure 2

Typical tension stress–strain curves of neat PLA and its compounding blends with different compositions: (a) the stress–strain curves of the PLA and its blends with and without peroxide and (b) the stress–strain curves of the PLA ternary blends without peroxide.

Figure 3

The effect of the peroxide amounts on the tensile mechanical properties of the PLA/PBS/PBAT blends: (a) tensile modulus (automatic Young’s), (b) tensile strength at break, (c) elongation at yield and break (%), and (d) flexure modulus and strength.

Figure 4

Effect of the peroxide contents on the thermal properties of the blends as calculated from the second heating DSC curves (heating at 10 °C/min): (a) the dependence of crystallinity of PBS on peroxide contents, (b) the dependence of crystallinity of PLA on peroxide contents, and (c) the dependence of melting point of the PBS and PLA phases on peroxide contents.

The incorporation of the tough PBS and PBAT increased the elongation at break of the blends significantly, as shown in Figure c. It is around 167% for the P-S40T20 samples, much higher than that of the pure PLA. It should also be noted that the increase in peroxide concentrations tended to result in a decrease in the tensile strain, totally different from the impact strength change. This difference could be due to the differences in the crack-initiation mode and strain rate between these two tests. With the introduction of peroxide into the ternary blends, the molecular chains are highly branched or cross-linked. First, the existing high-entanglement molecular chains greatly limited the chain movement during the tensile stretch of the samples, leading to the decreased elongation at break, especially for the P-S40T20-I samples with high gel contents. Second, to achieve good impact strength, a high level of interfacial adhesion is required.[26] In this sense, the impact strength is more sensitive to the interfacial compatibility. However, the very strong interfacial adhesion is unfavorable for the debonding of the phase, which can lead to the microvoiding during the tensile stretch[13d,27] as the debonding of the toughening phase from the matrix requires very strong external force, such as high-speed impact. As a slow deformation process, the force in the tensile testing is not strong enough to pull out the toughening phase when the adhesion between PBAT and the matrix is very strong. As a result, the absence of the microvoiding will delay the occurrence of matrix yielding, resulting in the inferior tensile elongation at break. Meanwhile, the very strong interfacial adhesion means that the polymer chains are interacting with each other with high entanglement. The relaxation and chain alignment in the stretching direction during the tensile testing will be suppressed once the chains are highly entangled, which leads to the decreased elongation at break of the plastics. However, for samples with zero or low gel contents, P-S10T10-I0.3 and P-S20T20-I0.3, the elongation at break increased because of the enhanced compatibility between the ternary components as the branching or cross-linked copolymers form during the extrusion reaction. Especially for the P-S10T10-I0.3, the elongation of the samples with 0.3 phr peroxide is double that of the P-S10T10 without peroxide, revealing the enhanced compatibility effects caused by the peroxide. Meanwhile, the flexural modulus and strength shown in Figure d are mainly influenced by the PBS and PBAT amounts in the blends, no matter how much peroxide is used. The flexural strength and modulus both decreased with increasing PBS and PBAT contents.

The related thermal parameters calculated from the differential scanning calorimetry (DSC) cooling and following heating curves are summarized in Table , and the crystallinity and melting point of the blends are shown in Figure , to investigate the effect of peroxide concentrations on the thermal properties. From the listed glass-transition temperature (Tg) of the blends, it can be found that the compatibility between PLA and PBS/PBAT has been improved with this method as the Tgvalues are close to each other after reactive extrusion. The dependence of the crystallinity (Xc) on peroxide contents is completely opposite for the PBS and PLA. With the introduction of 0.3 phr peroxide, the Xcof PBS decreases while the Xcof PLA increases. The Xc of PLA has been increased from about 2.5% (amorphous) to 15% for the P-S40T20-Iseries. Two reasons account for this opposition. First, PBS possesses better reactivity with peroxide compared to PLA and peroxide. During the process, the PBS chains are highly branched or even cross-linked, and the cross-linking would greatly limit the chain movement, which suppressed the crystal growth of PBS, whereas the PLA chains are only slightly branched because of the low reactivity between the PLA and peroxide. Second, the crystallization of PLA is mainly controlled by the nucleation efficiency so that different materials, such as high-melting-point PLA,[28] nucleation agent TMC-328,[29] and stereocomplex PLLA/PDLA,[30] can be used as nucleation agents of PLA to improve its crystallization rate and final crystallinity. In our ternary blends, PBS melts itself,[31] and slightly branched PLA chains[32] act as nucleation sites to improve the crystallization of PLA. The melting point of PBS and PLA both decreased with increasing peroxide concentrations, as shown in Figure c. The reduction of the melting points of the two phases indicated that the crystals formed in PBS and PLA are imperfect compared to those of the pure polymers. This is because the molecular segmental mobility of the polymer is greatly restricted by the enhanced branching/cross-linking reaction. Thus, the crystal growth and the lamella thickness increase would be suppressed by the branched/cross-linked chains, resulting in the imperfect crystals in different phases. Finally, the imperfect and thinner crystals result in the decrease of the melting temperature (Tm) of PBS and PLA. On the other hand, the suppressed crystal growth because of the enhanced entanglement structures also indicated that the molecular chains in the different phases intercalated, meaning that the good compatibility resulted from the peroxide interactions.[33]

Table 1

Related Thermal Parametersa

sample	Tg (°C)	Tc (°C)	Tcc (°C)	Xc (%)	Tm (°C)
neat PLA	60.5		114.3	0.9	148.0
neat PBS	–32.0	89.2		33.8	114.1

Glass transition temperature (Tg), thermal Crystallization temperature (Tc) of PBS and PLA, cold crystallization temperature (Tcc) of PLA, crystallinity (Xc), and melting point (Tm) of PLA and PBS.

Tg of PBS cannot be detected by DSC for the samples with low percentage of PBS.

Cold crystallization of PLA overlapped with the melting of PBS with increasing PBS contents.

Besides the super toughness of the biodegradable blends, the materials also possess high melt elasticity because of the in situ reaction between the components and peroxide. The storage and loss moduli of the blends with 0.3 phr and without the peroxide initiator are shown in Figure a,b, respectively. For the neat PLA, PBS, and PBAT, a typical Newtonian liquid behavior (G′ ∼ ω2, G″ ∼ω) could be observed at low frequencies. For the ternary blends, the terminal relaxation will deviate from this Newtonian liquid behavior because of the multirelaxation behavior in the ternary blends. This deviation becomes more obvious with increasing PBS and PBAT contents. A tanδ peak can be found in the P-S20T20 and P-S40T20 blends from the loss tangent (tanδ = G″/G′) shown in Figure c, meaning that a network structure appears with the introduction of higher PBS and PBAT amounts. This may be ascribed to the structure evolution of the blends from the sea-island to co-continuous morphology at higher PBS and PBAT amounts. Undoubtedly, the storage and loss moduli increased with increasing PBS content in the blends, even without peroxide, which is ascribed to the enhanced effect of PBS, which possesses higher melt moduli compared to the other two components as shown in Figure a,b.

Figure 5

Frequency dependence of the rheology properties for the compounding samples with different blends: (a) storage modulus (G′), (b) loss modulus (G″), (c) loss tangent (tan δ), and (d) complex viscosity (η).

The elastic effects of the ternary blends became more significant with the introduction of peroxide, reflected by the increased storage moduli (Figure a) and decreased tanδ (lower than 1) as shown in Figure c. With the introduction of peroxide, the non-Newtonian low-frequency plateau of G′, which reflects some difficult relaxation internal structures within a long-time scale, is observed, revealing that some molecular networks formed in our melt blends, which is caused by the branching/cross-linking reactions that happened in the system. The cross-linked or branching networks in the ternary blends caused the improvement in the viscosity of the blends, as shown in Figure d, especially in the low frequencies. Meanwhile, the same viscosity of the samples with and without peroxide at high frequencies ensures the processability of this material in high-shear processing such as extrusion. The blends with peroxide exhibited shear-shinning behavior in all testing frequencies, meaning the relaxation or disentanglement of the molecular network in the whole time scale, from short- to long-time relaxation. With the introduction of peroxide, the blends experience a transition in behavior from liquid-like to gel-like, as revealed by the tan δ in Figure c, indicating that the melt elasticity of the blends has been greatly improved because of the branching or cross-linking of the samples. More importantly, the enhanced melt modulus and viscosity effect improved with increasing PBS/PBAT amounts in the blends, meaning that the branching/cross-linking reaction mainly happened between PBS/PBAT and peroxide. This is also verified by the gel contents as shown in Figure S1; the gel contents increased with PBS/PBAT amounts in the blends at the same peroxide concentration. The gel content is as low as 0.9 wt % for the P-S40T20 blends with 0.3 phr peroxide, ensuring the processability of the designed materials. As the gel content reaches 17.5 wt % with 0.75 phr peroxide, low dosage of peroxide should be used in the designed blends to avoid the over-cross-linking. Considering the important role of chain entanglement in the stretching shape process and crack growth resistance, the enhanced elastic effect and melt viscosity properties of the developed toughened PLA materials show promising applications in film forming and thermoforming.

Toughening Mechanism

To visualize the effects of peroxide contents on the energy dissipation modes involved in the toughened PLA materials, the impact fracture surfaces beyond the notch are examined by SEM, as shown in Figure . As the blends with and without peroxide show huge differences in impact strength, the SEM images of PLA/PBS/PBAT and PLA/PBS/PBAT/peroxide 0.3phr are shown here to clarify the reason for this. Compared to the smooth fracture surface of the samples without peroxide, which show lower impact strength, the blends with 0.3 phr peroxide exhibit a rough fracture surface with improved signs of plastic deformation (as shown in Figure ). The observations of the impact surface are well consistent with the corresponding impact strengths. It is suggested that the plastic deformation in the matrix is mainly caused by the tough PBAT to absorb the impact energy. This is the main reason why super tough blends can be obtained in our study, the same as the observation in the previous research of Ma et al.[34] From the impact fracture surface, no obvious PBAT debonding can be found in the tough samples; this is because of the strong interfacial force between PBAT and the matrix from the branching/cross-linking reaction. Previously, it has been reported that the optimal particle size to induce internal rubber cavitation is larger than 200 nm,[35] and because of the higher intrinsic brittleness of PLA, it is reported that the optimal particle size to toughen PLA is 0.5–0.9 μm.[13d] However, from the etching SEM photos shown in Figure a, we calculate that the mean diameter of the PBAT phase in the toughened blends is around 210 nm (count from Image-Pro Plus software), which is close to the critical particle size for internal rubber cavitation but far lower than the optimal particle size for toughening PLA. However, from the impact fracture surface shown in Figure , some tiny voids can be found in the toughened samples without any sign of PBAT debonding, especially for P-S20T20-I0.3. It is suggested that the optimal particle size for internal cavitation may decrease when we have some PBS to adjust the intrinsic brittleness of the matrix. On the other hand, it is worth noting that in the processing–structure–property relationship of polymer blends, the interplay of factors (e.g., phase interface adhesion, dispersed phase size, and composite ratios) could result in an enhancement or deterioration of the mechanical properties.[36] To clarify the effect of composite ratios and peroxide on the phase interface and structure changes of the ternary blends, we purposely fracture the samples in liquid nitrogen to avoid any interference from the plastic deformation and take an SEM observation.

Figure 6

SEM images of an impact fracture surface of the ternary blend with various weight compositions.

Figure 7

Morphology characterization by SEM and rheology testing: (a) SEM images of the ternary blends with and without peroxide, (b) Cole–Cole plots for the neat polymer and their ternary blends with 0.3 phr and without peroxide, and (c) the effect of the peroxide contents on the Cole–Cole plots of the blends.

At 0.3 phr peroxide, the impact strength jumps from 37 J/m for P-S10T10-I0.3 to 728 J/m for P-S20T20-I0.3. Our further studies show that the impact strength of PLA70/PBS20/PBAT10/peroxide 0.3 phr and PLA70/PBS10/PBAT20/peroxide 0.3 phr blends is 44 and 60 J/m (the data are shown in Figure S4), respectively, exhibiting low toughness. It means that the toughness of the blends is also influenced by the composition ratios in addition to the improved interfacial compatibility. To clarify this, we etch the PBAT in tetrahydrofuran (THF) and then observe the morphologies of the blends by SEM as shown in Figure a. It is found that the dispersed PBAT changes from droplet in P-S10T10 to co-continuous in P-S20T20; the rationality on the dispersion of 20 wt% PBAT as a co-continuous structure in the ternary blend is based on the partial compatibility of PBS and PBAT because of their chemical structures.[22] However, the SEM observation shows the co-continuous blend morphology in the absence of peroxide versus droplet morphology in the presence of peroxide, resulting from the increased viscosity of the dispersed phase. The size of the PBAT droplet decreased dramatically to around 200 nm for P-S20T20-I0.3 and P-S40T20-I0.3. As the main toughening agent, the PBAT phase with decreased dispersion size is believed to be able to maximize the toughness enhancement role of PBAT and lead to the sudden increase of the toughness in P-S20T20-I0.3. However, for P-S10T10-I0.3, the particle size of PBAT shows no sign of decreasing and remains to be around 1–2 μm. That is why the toughness of the P-S10T10 cannot be improved even after peroxide curing. Through decreasing the size of the toughening droplet and taking advantage of the partial compatibility between the toughening phases, for example, PBAT and PBS, high toughness materials can be obtained with stiffness–toughness balance.

The morphology structure evolution can also be verified from the Cole–Cole plots from the rheology testing, which has been proven to be sensitive to the phase structures in polymer blends.[37] By depicting η″ versus η′, the Cole–Cole plots of the different blends are shown in Figure b,c. As shown in Figure b, the only circular arc in the curve of the pure polymers corresponds to the homogeneous structure relaxation of PLA, PBS, and PBAT.[37a] After blending the three polymers together, a second circular arc or a tail appears on the right-hand side of the curves based on different composition ratios, corresponding to the different relaxation mechanisms. The two peaks in the Cole–Cole plots indicated that the mixture has undergone phase separation.[38] At low frequencies, the second circular arc of P-S10T10 is the sign of the relaxation stemming from the deformation of the suspended droplets,[39] meaning that a drop–matrix phase morphology formed in the blend, whereas the tails in the high frequency of P-S20T20 and P-S40T20 correspond to a co-continuous phase relaxation, meaning that a co-continuous phase structure has been formed in these blends.[37a] These rheological studies are well consistent with the morphology observation by SEM. Interestingly; the Cole–Cole plots did not show any relaxation arc or tail after peroxide is incorporated, as shown in Figure c. They all show good linearity and are independent of the peroxide concentrations. This is because the relaxation of the whole molecular chain, whatever the composition or structure, has been greatly limited after highly branching or cross-linking the polymer chains in the presence of peroxide. The relaxation of the different phases cannot be detected in our testing frequency ranges.

In summary, the most probable toughening mechanism for PLA/PBS/PBAT/peroxide with high impact strength is singled out to be tough PBAT-induced plastic deformation in which the interface adhesions are significantly improved in the presence of peroxide. With the increasing viscosity of the toughening phase, which resulted from the reactive extrusion, the PBAT phase transformed from co-continuous structures into “droplet” structures with a decreased diameter. The decreased size of the toughening phase is believed to improve the toughness of the blends dramatically. According to Jalali’s research, a co-continuous structure formed in the PLA/PBAT binary blends with 40/60, 50/50, and 60/40 composite ratios.[37c] By structuring the PLA ternary blends with PBS and PBAT, the co-continuous structure is found at 20 wt % PBAT via taking advantage of the compatibility between PBS and PBAT. The decreased amount of toughening agent allows us to design and fabricate materials with balanced stiffness–toughness properties. Benefitting from the structure tailoring via reactive extrusion, our research shows that a small amount of peroxide (0.3 phr) could improve the toughness of the ternary blends with co-continuous structures by improving the interface interactions and transforming the co-continuous structure into a small-size droplet with strong adhesion to the matrix.

Radical Reaction Mechanism

The content of PLA and PBS/PBAT in the gel (or blends) can be easily distinguished via thermogravimetric analysis (TGA) because of the large difference between the decomposition temperatures of PLA (ca. 370 °C) and PBS/PBAT (ca. 400 °C) and the decomposition temperature sensitivity of the PLA and the other two polymers, as shown in Figure d. The content of PBS/PBAT in the initial blends of P-S10T10 (no peroxide) is only 20 wt % (Figure a); however, it is much higher (ca. 55 wt %) in the extracted gel of the P-S10T10-I0.5 blend (Figure a), indicating that the main reaction happened between peroxide and PBS/PBAT. The TGA results of the extracted gels indicated that only 45% PLA exists in the gels, which decreased a lot compared to the initial blending ratio of PLA (80 wt %). The same phenomenon can be observed in the P-S20T20-I0.5 and P-S40T20-I0.5 blends, as shown in Figure b,c. The PLA contents in the extracted gels are 50 wt % for P-S20T20-I0.5 and 32 wt % for P-S40T20-I0.5. Unfortunately, the contents of the PBS and PBAT in the gels cannot be measured because of the close decomposition temperature of these two materials. It can be found that only one DTG peak is observed for PBS and PBAT in all blends. The undissolved gel part is believed to be the highly entangled cross-linking polymer chains. This means that PLA, PBS, and PBAT chain segments are highly entangled together interacting as copolymers. The Fourier transform infrared (FTIR) spectrum of the extracted gel of P-S40T20-I0.75 is given in Figure S5. It can be found that the extracted gel shows characteristic peaks of PLA (−C=O at 1715 cm–1), PBS (−C–O at 806 cm–1), and PBAT (out-of-plane bending of =C–H in the benzene ring at 731 cm–1),[40] verifying the existence of the copolymers. Except for the gel content calculation, the DTG curves also provide us thermal stability information of the blends. It can be found that the thermal degradation temperatures of the blends, including the initial and maximum degradation temperatures, are almost unchanged with or without peroxide, independent of peroxide concentrations. This is because even though the polymer chains are branched/cross-linked with the introduction of peroxide, the bond energy of −C–C– or −C–O– is not influenced. Meanwhile, the thermal degradation temperatures of the polymers in TGA testing are mainly decided by their bond energy.

Figure 8

TGA curves of the compounding samples with different composition ratios and pure PLA, PBS, and PBAT: (a) P-S10T10-I, (b) P-S20T20-I, (c) P-S40T20-I, and (d) neat PLA, PBS, and PBAT.

The above phenomena of the dependence of gel contents on composition ratios by gel calculation as shown in Figure S1 and TGA testing indicated that the reactions initiated by peroxide during the extrusion mainly happened between the PBS and PBAT components. To demonstrate the reaction priorities and possible reaction mechanism, we conducted a qualitative testing on the gel formation between these three polymers and peroxide by rheology testing. An in situ reaction between the polymer and the peroxide is conducted on a rheometer at 175 °C under small dynamic frequencies, and the dependence of the moduli on reaction time is depicted in Figure . The gel point here is qualitatively defined as the instant time at which the complex moduli, G′ and G″, cross each other.[41] It can be found that the gel reaction happened quickly in PBS/peroxide (∼88 s) and PBAT/peroxide (∼156 s), whereas no gel formation occurred in the PLA/peroxide sample in a small-amplitude oscillatory shear, meaning the higher capacity of PBS and PBAT for developing radical reactions with peroxide than that of PLA. This is ascribed to the higher content of mobile H atoms on the CH2 groups of PBS and PBAT, whereas PLA has only a methyl, which is not active, or methyne hydrogen, which is sterically hindered and gives a rather stable free radical. However, the PLA gels and copolymer of PLA–PBS–PBAT still exist in the final reaction products according to the TGA analysis and compatibility by SEM. This is supposed to the chain scission of PLA under high shear processing involving oxygen. The formation of the PLA gels can occur through the reaction of macroradicals, such as PLA–O–O·, and that of the copolymers can occur through the reaction of macroradicals PBS·, PBAT· or PBS–O–O·, and PBAT–O–O· with PLA–O–O. This is a qualitative discussion because the gel point can only be detected by this method only when the relaxation exponent of the gel is equal to .[41] However, from the qualitative comparison, the possible reaction mechanism in this ternary blends is given by us as follows:

Figure 9

Modulus vs reaction time of the PLA, PBS, and PBAT with 0.3 phr peroxide at 175 °C under time sweeping by a rheometer: sweeping at 1% strain and 1 Hz frequency.

Peroxide preferentially reacts with PBS and PBAT chains, extracting the secondary hydrogen from the CH2 group of PBS/PBAT to yield free radicals (PBS· and PBAT·) on the main chain of PBS or PBAT.

The reaction between PLA and peroxide mainly happens between the fragmentation chains of PLA because of the high shear and thermal decomposition in the presence of oxygen and free radicals of peroxide to form PLA–O–O·.

Finally, the cross-linked/branched structures of PLA, PBS, or PBAT via homogeneous radical coupling and copolymers via radical coupling reactions between the macroradicals PBS·, PBAT· or PBS–O–O·, and PBAT–O–O with PLA–O–O are formed.

Conclusions

Super tough biodegradable ternary blends from PLA, PBS, and PBAT were successfully prepared by a simple melt reactive blending method. The super tough PLA/PBS/PBAT ternary blend exhibiting a high impact strength of ∼900 J/m with hinge break impact behavior was obtained by optimizing the blending ratio of PBS and PBAT with a small amount of peroxide. The small amount of peroxide can effectively improve the impact strength of the blends with 40% PBS and 20% PBAT to ∼1000 J/m, 10 times that of the blend with the same composition without peroxide and 30 times that of pure PLA. Meanwhile, the existence of PLA ensures the high bio-contents of the blend with modulus above 1 GPa and strength above 30 MPa. The interfacial interaction and compatibility of the different phases have been greatly improved because of the robust reactivity between the biodegradable polymer and peroxide through the SEM and rheology analysis. The gap between the different phases disappeared with the incorporation of the optimized content of peroxide. More importantly, by structuring such ternary blends, we can obtain co-continuous structures at low PBAT contents (20 wt %) via taking advantage of the partial compatibility between PBS and PBAT. The reduced amount of toughening agent allows us to design and fabricate materials with balanced stiffness–toughness properties. The influence of peroxide on the thermal properties of the blends was studied by DSC. The slight decline of Tg of the PLA and increase of the Tg of PBS/PBAT in the presence of peroxide indicated the improved compatibility by the reactive extrusion. On the other hand, the melting point of PLA and PBS both decreased because of the imperfect crystals formed in the blends, which resulted from the limited chain movement by branching/cross-linking. On the basis of the gel content calculations, a qualitative study on the gel formation of the prepared bioblends by rheology testing revealed the possible reaction mechanism between these three biopolymers with peroxide. It is supposed that peroxide preferentially reacted with the CH2 group of PBS/PBAT to yield free radicals, whereas the reaction between the PLA and peroxide is very difficult and mainly happens between the decomposition chains of PLA, oxygen, and free radicals of peroxide during the process. Discussion on the reaction mechanism will allow us to further optimize the in situ extrusion reaction, which is widely used in industrial applications. This sustainable super toughened PLA bioblends with significantly high impact strength holds a promising application to substitute petroleum-based plastics.

Materials and Methods

Materials

In this work, we use NatureWorks PLA with the brand name Ingeo Biopolymer 4043D. The PBS (Tunhe TH803s) and PBAT (Tunhe TH801t) used in this study are obtained from Xinjiang Blueridge Tunhe Chemical Industry Co., Ltd. The peroxide initiator, that is, 2,5-bis(tert-butyl peroxy)-2,5-dimethylhexane (Luperox 101) with a half-life of 1 min at 177 °C is purchased from Sigma-Aldrich and used here. PLA, PBS, and PBAT pellets are dried in an oven at 80 °C for 24 h before melt blending. Chloroform (CHCl3), tetrahydrofuran (THF), and acetone are purchased from Sigma-Aldrich and used as received.

Bioblend Fabrication

Melt blending is performed using a co-rotating twin screw extruder (Leistritz Micro-27, Germany) with a screw diameter of 27 mm and an L/D ratio of 48. The screw speed is set at 100 rpm, and the melt-zone temperatures are set to 175 °C. For all ternary blends, the feeding rate is fixed at 5 kg/h to make sure that the residence time of the materials in the extruder is about 60 s. For homogeneous mixing and effective use of the small amount of peroxide, we used a solution of peroxide in acetone, added it to the dried polymer matrix, and then mechanically mixed them, followed by feeding to the extruder for melt processing. The blend formulations and codes are shown in Table . Here, we use “P” to represent PLA, “S” for PBS, “T” for PBAT, and “I” for the peroxide initiator, which are followed by numbers indicating the contents of PBS, PBAT, and peroxide. For example, P-S10T10-I indicates a formulation of 80 wt % PLA, 10 wt % PBS, and 10 wt % PBAT with x (x = 0.3, 0.5, and 0.75) phr peroxide initiator. After being oven-dried, the extruded pellets are injection-molded using a micro 15 cm3 co-rotating twin screw compounder and a micro 12 cm3 injection molding machine (manufactured by DSM Research, Netherlands) into ASTM-standard specimens at a melting temperature of 190 °C and a molding temperature of 30 °C. The processing is conducted at a filling pressure of 8 bar and a packing pressure of 8 bar with a screw speed of 100 rpm for 2.5 min (including feeding time). All mechanical test specimens are conditioned for 2 days at 23 °C and 50% RH before mechanical testing and characterization.

Table 2

Compositions of the Reactive Blends for All Specimens

Characterization Methods

Rheology

The rheology behaviors of the compounded materials are studied using a rheometer (Anton-Paar MCR-302 Instrument). The frequency-dependent rheology behavior of all of the samples is monitored at 175 °C and 1% strain under N2 protection. The samples for rheological studies are injected into a disk with a diameter of 25 mm and a thickness of 1 mm at 180 °C using a micro 15 cm3 DSM injection molding machine (DSM Research, Netherlands).

The in situ reaction among the biodegradable polymers (PLA, PBS, and PBAT) and peroxide initiator in the rheometer is conducted by time sweeping at 175 °C, 1% strain, and 1 Hz under N2 protection. The samples are prepared by dispersing the diluted peroxide/acetone solutions (0.2 g/30 mL) on the injected plastic disks, followed by the evaporation of acetone at 25 °C in a fume hood.

DSC

The thermal properties of the blends are determined on a DSC instrument (TA Q200, USA) under a N2 atmosphere. The samples are heated to 180 °C at 10 °C/min and maintained for 3 min before cooling to −70 °C at 10 °C/min. The second heating scans are monitored between −70 to 180 °C at 10 °C/min for determining the glass-transition temperature (Tg), cold crystallization temperature (Tcc), melting temperature (Tm), and crystallinity (Xc). Such studies are targeted to investigating the influence of peroxide on the crystallization behavior of the blends.

The crystallinity of PBS (XPBS) is calculated based on the melting enthalpy of PBS (ΔHPBS) occurring at ∼112 °C and the weight ratio of PBS (WPBS), as illustrated in eq .where ΔHm is the melting enthalpy of PBS, ΔHm° is the melting enthalpy assuming 100% crystalline PBS (220 J/g here),[42] and wf is the weight fraction of PBS in the blend.

The crystallinity of PLA (XPLA) is calculated based on the melting enthalpy (ΔHPLA), the cold crystallization enthalpy of PLA (ΔHcPLA), and the weight ratio of PLA (WPLA), as illustrated in eq .where the ΔHm′ and ΔHc are the melting enthalpies and cold crystallization enthalpies of PLA, ΔHm° is the melting enthalpy assuming 100% crystalline PLA (93.7 J/g here),[43] and wf′ is the weight fraction of PLA in the blend.

Thermal Stability Testing

The thermal stability of the blends and extracted gels are tested by TGA (TA Q500) from 30 to 500 °C at 10 °C /min under N2. The first derivative of the TGA curves (DTG curves) is obtained from the TA2000 software.

SEM

The morphology of the blends is observed by SEM with an accelerating voltage of 10 kV (Hitachi S-570). The samples are fractured in liquid nitrogen and obtained from the notched Izod impact test, which are sputtered with gold first and then observed by SEM.

Gel Content Calculations

The gel contents of the blends are calculated by a dissolution–extraction method according to ASTM D2765. The polymer is sealed in stainless steel wire mesh and dissolved in chloroform for 24 h. After the extraction, the undissolved parts are left in the fume hood for 3 h and then transferred into a vacuum oven at 40 °C to remove the chloroform. The residue of the insoluble polymer is weighed and reported as wt % gel content.

Mechanical Property Tests

The tensile and flexure properties of the samples are tested using an Instron 3382 according to the ASTM standards D638 and D790, respectively. The Izod impact strength of notched samples is tested using a Testing Machine Inc. (TMI) instrument according to ASTM D256. In all mechanical tests, five samples were tested, and the mean and standard deviation results are reported.

FTIR Spectroscopy

The gel extracted from the dissolution testing is characterized using a Fourier transform infrared (FTIR) spectrometer equipped with a smart orbit ATR (Nicolet 6700, Thermo Scientific).