Reactive Compatibilization of Poly(Butylene Succinate)/Soy Protein Isolate Bio-Composites By Peroxide And Diacrylate Via Melt Blending

Poly(butylene succinate) (PBS)/soy protein isolate (SPI) bio-composites were reactive compatibilized by adding dibenzoyl peroxide (BPO) and hexanediol diacrylate (HDDA) via melt blending in an internal mixer. The structure and properties of composites were studied by scanning electron microscopy (SEM), Fourier transform infrared spectroscopy (FTIR), Soxhlet extraction experiments, dynamic mechanical analyses (DMA), rheological tests, contact angle measurements, thermogravimetric analyses (TGA), differential calorimetry (DSC), wide-angle X-ray diffraction (WAXD), mechanical tests and water absorption tests. The results show that a branched graft copolymer (SPI-g-HDDA-g-PBS) was produced with the aid of BPO and HDDA via melt blending, which served as the compatibilizer and improved the compatibility and enhanced the adhesion between PBS matrix and SPI phase by forming a network structure in the composites. The crystal form of PBS was not changed in the composites. The melt viscocity and elasticity, hydrophobicity, thermal stability, crystallinity, tensile strength and water resistance of PBS/SPI-HDDA/BPO composites were improved compared with PBS/SPI composites.


Introduction
Petroleum-based plastics have been extensively applied in various fields due to their excellent properties. With depleting petroleum resources and rising global concern for climate change, the development of biobased and biodegradable polymers has attracted much attention. [1,2] It is necessary to replace traditional petroleum-based commodity polymers such as polyolefins and poly(ethylene terephthalate) (PET) by biodegradable and biobased plastics, e.g., poly(lactic acid) (PLA), poly(butylene succinate) (PBS) and poly(hydroxy butyrate) (PHB). [3,4] Among biodegradable polymers, PBS is one of the most attractive polyester owing to outstanding mechanical properties, preferable heat and chemical resistance, melt-processability and biodegradability. [5] However, PBS shows the disadvantages of high cost and slow degradation rate, which limit its wide application. [6,7] Blending PBS with low-cost natural biodegradable polymers, e.g., starch [8] and protein [9], is a valid solution to this problem.
As a byproduct of soybean oil-extraction, soy protein (SP) has attracted considerable attention because of its renewability, biodegradability, abundance and low cost. [10] Soy flour (SF), soy protein concentrate (SPC) and soy protein isolate (SPI) are three grades of SP, containing ca. 50%, 70% and 90% protein, respectively.
[11] SP contains various amino acids, with glycinin (11S, ca. 52%) and β-conglycinin (7S, ca. 35%) being the predominant ones. Both have quaternary structures with disulfide bonds binding polypeptide subunits together [12,13]. SP is characterized by numerous reactive groups, such as -COOH, -OH and -NH2, which could be used as sites for chemical modifications and polymerizations. [14,15] Indeed, cross-linking of SP could results in an increased molecular weight, and reduced elasticity and solubility. [16] Wet or dry (melt) processing are two methods to produce soy-based plastics, and the dry processing is preferable due to industrial friendly and easily adaptable. Plasticization and destructurization [17], graft copolymerization [18], oxidation [19], blending and compositing [20] are often used to modify the surface hydrophobicity, solubility, water-holding capacity, gelation and mechanical properties of SP-based plastics.
Bonham et al. [23] compared the mixing methods of co-rotation (CR) and counter-rotation (CTR) twin-screw extrusions for a blend of plasticized SPI (PSPI) and PBS in a 30: 70 wt.-% ratio. CTR extrusion provided enhanced interfacial adhesion, increased tensile elongation and prolonged thermal degradation temperatures because CTR allowed for better destruction and mixing of PSPI in the PBS matrix. Reddy et al. [24] studied the synergistic effect of simultaneous plasticization and denaturation of the reactive blending of thermoplastic SM (TSM) with polycaprolactone (PCL) and PBS, which had yielded biodegradable blends with improved tensile properties and elongation. Renoux et al. [25] successfully prepared plasticized isolated soy protein (PISP)/poly [(butylene succinate)-co-adipate] (PBSA) blends via single-step and two-step melt compounding methods, respectively.
Compared with the two-step method, one-step method was an effective and timesaving method to enable thorough mixing and thus better dispersion of PISP into the PBSA. Chen and Zhang [26] processed SPC as a plastic by adjusting water content and blended it with poly(butylene adipate-co-terephthalate) (PBAT).
Percolated thread structures were formed in SPC prior to compounding. Tensile properties were greatly improved in the blends with percolated SPC thread structures.
Anstey et al. [27] investigated the processability and biodegradability of PBS/bio-based fillers (SM, canola meal (CM) and corn gluten meal (CGM)) composites using twin-screw compounding. The results showed that the biodegradation rate of PBS was increased due to the existence of meal-based fillers in the bio-composites, which enhanced the hydrolytic biodegradation of the material and facilitated micro-organism growth. Coltelli et al. [28] found that soy lecithin and the Schotten-Baumann modified whey were effective in compatibilizing the plasticized whey protein (PWP)/PBS blends. Significant increases in elastic modulus, tensile strength, elongation at break and crystallinity were observed with respect to the not compatibilized blend.
Soy protein shows strong hydrophilicity due to the polar groups, while PBS presents hydrophobicity due to the repeated aliphatic units. Generally, the poor interfacial adhesion between two incompatible polymers having different polarity will bring about inferior physical and mechanical properties. It is necessary to improve the compatibility of PBS and SPI by modification and compatibilization. [29,30] The reactive blending is a simple, effective and low-cost method to modulate interfacial interactions by in-situ chemical reaction, such as crosslinking and grafting, and thus enhances the properties of immiscible blends. [31] The reactive processing methods had been adopted in many studies, such as PBS/waxy starch [32], PBAT/thermoplastic starch [33], PWP/PBS [28], NTP/PBS [22], TSM/PCL/PBS [24] and SPC/Eastar Bio Copolyester (EPE) [34] composites.
Herein, the reactive blending method is adopted to compatibilize PBS/SPI composites by using dibenzoyl peroxide (BPO) as the initiator, hexanediol diacrylate (HDDA) as the grafting monomer. The compatibility, microstructures and properties of the composites were systematically investigated.

Experimental Materials
Bio-PBS TM (FZ91PB, produced from polymerization of bio-based succinic acid and Next, the sheets were cooled at room temperature and 10 MPa for 10 min in another press. Finally, they were taken out from the mold and stored in a desiccator for further measurements.

Scanning electron microscopy (SEM) observations
An SEM (JSM-6510, Tokyo, Japan) was used to observe the microstructures of PBS composites. The samples were frozen and fractured in liquid nitrogen. The newly fractured surfaces were coated with thin gold layers before SEM observations.

Fourier transform infrared spectroscopy (FTIR) characterizations
The functional groups in SPI, PBS and PBS composites before and after extractions were identified by an FTIR instrument (Vector 22, Bruker, Billerica, MA, USA). At room temperature, the spectra were scanned over the wavenumber range of 400 to 4000 cm -1 , with 64 scans at a resolution of 4 cm -1 . The ATR accessory was adopted on the surfaces of the samples.

Crosslinked grafts contents determination
In order to determine the content of crosslinked grafts in the composite, solvent extraction was used. To dissolve the PBS component, the sample (about 0.2 g, cut into small pieces) was Soxhlet extracted with chloroform (150 ml) at 70°C for 24 hours.
The residue was considered as SPI-g-HDDA-g-PBS and SPI (SPI is insoluble in CHCl3), which was vacuum-dried to constant weight. The contents of crosslinked SPI-g-HDDA-g-PBS were calculated by Equation (1): where WR is the residue weight, WSPI is the SPI weight in the raw sample and W0 is the raw sample weight, respectively. Three samples were used for each recipe and the average value was calculated.

Dynamic mechanical analyses (DMA)
A dynamic mechanical analyzer (Q800, TA Instruments, USA) was used to test PBS and PBS composites. The specimens (25×10×1 mm 3 ) were tested using a tensile mode at a 1 Hz vibration frequency. The temperature range was -85 to 100 °C and the heating rate was 3 °C/min.

Rheological tests
The rheological tests were carried out in a rotational rheometer (MCR302,

Contact angle measurements
The contact angles of water, diiodomethane on the surfaces of PBS and PBS composites were measured by a contact angle analyzer (YIKE-360B, Chengde, China). Before testing, absolute ethyl alcohol was used to clean the surfaces of the samples. The liquid (4 μl) was dropped on the surfaces of the specimens with a micro syringe. The dynamic water contact angles (θwater) were recorded every 20 s for 120 s.
Five replicates were tested for each sample to obtain the average value and the variance. The Owens and Wendt theory [30] was used to calculate surface free energies.

Thermogravimetric analyses (TGA)
The thermal stability of PBS, SPI and PBS composites were investigated by a TGA instrument (Netzsch STA 449C, Selb, Germany). The samples (10~15 mg) were heated from 40 to 800 °C with the heating rate of 10 °C/min under the nitrogen flow.
In addition, derivative thermogravimetry (DTG) curves were acquired to examine the weight loss rate as a function of temperature.

Differential scanning calorimeter (DSC) measurements
A DSC apparatus (ZF-DSC-D2, Shanghai Zufa Industry Co., Ltd., Shanghai, China) was used to evaluate the melting and crystallization behaviors of the specimens under nitrogen protection. Firstly, each specimen (~5 mg) was heated from room temperature to 150 °C at 10°C/min. It was maintained at 150 °C for 5 minutes to eliminate the thermal history. Subsequently, the samples were cooled to 30°C at 10 °C/min, and kept at 30 °C for 5 minutes. Finally, the samples were reheated to 150 °C at 10 °C/min.
Data was obtained from the first cooling and the second heating curves. The cooling curve provided the peak (T c p ) and onset (T c on ) crystallization temperatures. The melting temperature ( T m ) was determined by the second melting curve. The crystallinity (X c ) of the specimens was calculated by the following Formula (2): where ω PBS is the PBS wt% in the composite, ∆H m 0 represents the enthalpy of 100% crystalline PBS and ∆H m 0 =200 J/g. [35] Wide angle X-ray diffraction (WAXD) analyses An X-ray diffractometer (SmartLab, Rigaku, Tokyo, Japan) was used to characterize the crystal structures of SPI, PBS and PBS composites at room temperature, operating with Cu Kα (1.5418 Å), from 5° to 55° at a scan speed of 10 °/min.
The Scherrer's equation (3) was used to calculate the grain size of PBS in the composites.
where Dhkl (nm) represents the crystallite size (hkl) perpendicular to the crystal plane; k symbols the crystal shape factor and k=0.89; λ is the X-ray wavelength (nm); θ is the diffraction angle; β represents the peak width at half-height.

Mechanical tests
Tensile tests were carried out by using a universal testing machine (MZ-4000D1, Jiangsu Mingzhu Test Machine Co., Ltd, Jiangdu, China) according to ISO 527-2012.
The test sheets were cut into the specimens (ISO 527-1BA) and stored in a desiccator at 23 °C for 24 hours. The crosshead speed was set at 10 mm/min. Five replicates were tested for each recipe to obtain the average value and the variance.

Water absorption (WA) tests
The samples were dried in a vacuum oven at 50°C until constant weights (m1) were obtained. Subsequently, the samples were immersed in the distilled water at room temperature, and they were taken out at specific time intervals. Tissue papers were used to gently absorb the excess water on the surfaces of the samples and then they were weighed as m2. WA values were calculated by Equation (4): It is well known that the physical, chemical, mechanical, rheological and processing properties of polymer blends are dependent on their phase morphologies, especially for mechanical properties. [36] Figure 1 presents the phase morphologies of SPI and the cryogenically fractured surfaces of the composites. As shown in Fig. 1(a),

SEM observations
the appearance of most SPI particles is elliptical or spherical, and their sizes are between 10 and 100 µm, except for a few irregular particles are beyond 100 µm.  [11] reported that the compatibility between SPC and PLA was improved by addition of a compatibilizer (PLA-g-maleic anhydride (MA)). Therefore, the adhesion between SPI and PBS is greatly enhanced by adding BPO and HDDA, attributing to the generation of the crosslinked grafts between two phases, serving as the compatibilizer.  The FTIR is an effective tool to investigate the reactions and interactions between functional groups of the composites. Figure 2 shows FTIR spectra of PBS, SPI and PBS composites before and after extraction. On Curve (1)  The proposed reactive formation mechanism of SPI-g-HDDA-g-PBS is shown in  Fig. 3 is a simple example, and the actual situation is more complicated, for example, it is possible to produce a branched entangled network connected with multiple PBS and SPI macromolecules.  The effect of BPO and HDDA on the mechanical and compatibility of PBS/SPI composites is evaluated by the DMA. The storage modulus (E') and loss factor (tan δ) of the composites as a function of temperature are shown in Fig. 4, and Table 2 lists the DMA parameters. Figure 4(    However, it can be seen that the changes of G' is more obvious than G'' for each specimen, indicating that the composite structure is more responsive to G' than G''. [32] Contact angle characterizations Generally, the water contact angle (θwater) characterizes the hydrophilic/hydrophobic properties of the material's surface. Figure 6 shows the θwater curves of PBS and PBS composites. The θwater of the pure PBS is higher than those of PBS composites, indicating the hydrophobicity of PBS. In addition, the presence of SPI in the blends results in a decrease of the θwater with the time, especially for the composites with higher SPI contents, because the hydrophilic SPI is rich in polar groups (e.g., -OH, -COOH and -NH2). [47] Moreover, the θwater values of PBS/SPI-HDDA/BPO composites are higher than those of PBS/SPI composites.
Because the existence of SPI-g-HDDA-g-PBS enhances the interaction between PBS and SPI and facilitates to encapsulate SPI particles by PBS molecules, resulting in PBS-rich surfaces, as shown in Fig. 2 (Curve (5)). Therefore, the hydrophobicity of PBS/SPI-HDDA/BPO composites is increased.
In order to further study the polarity and interactions within the composites, the Owens and Wendt's method is selected, and the dispersive (γ s d ) and polar (γ s p ) components of the surface energy (γ s ) can be calculated according to Equations (5-7).
where γ s is the surface free energy (N/m), θ 1 and θ 2 represent the initial θ values of probe liquids (°), γ L1 d , γ L1 p , γ L2 d and γ L2 p are the dispersion, and polarity components of the probe liquids, respectively.  Table 3 presents the components of surface energies of the detect liquids. Water (H2O) and diiodomethane (CH2I2) are selected as experimental liquids due to the different polarities of them. Table 4   The TGA is very useful for quantitatively determining the degradation behavior and composition of composites, and it also exhibits the interaction between the components. Figure 7 furnishes TG curves and derivative thermogravimetric (DTG) curves of SPI, PBS and PBS composites. Table 5 provides the initial degradation temperature (T5%, the temperature at the 5% weight loss), the temperature at maximum degradation rate (Td,max) and the activation energy for thermal decomposition (Et) of SPI, PBS and PBS composites.
The TGA data are used to calculate the Et of the composites with Equation (8).
where θ is (T-Tmax), T is the temperature, α is the decomposed fraction, R is the gas constant. Et can be calculated from the slope of ln[ln(1-α) -1 ] vs θ. [48] As shown in Fig. 7, the TG curve of the pure PBS has one step, corresponding to the single peak on DTG curve. In addition, the thermal degradation of SPI can be divided into four stages. The first transition is the evaporation of the residual moisture in SPI before 120 °C. The second step from 180 to 330 °C due to the degradation of small molecules and the breakdown of some unstable chemical bonds in SPI. The third stage is the degradation of the backbone peptides between 330 and 480 °C. The last stage above 520 °C represents the carbonization and charring of SPI phase. Two obvious peaks located at 282.55 and 570.71 °C can be seen on the DTG curve of SPI. [49] The TG curves of the composites exist between those of SPI and PBS. With the increase of SPI content, they approach to SPI curve. However, there is single obvious degradation peak on the DTG curves of PBS composites, corresponding to the PBS phase in the blends. Two DTG peaks of SPI phase cannot be observed for the composites. Only when SPI content is 40%, a small peak of the loss of moisture appears on the DTG curves of the blends.
As shown in Table 5, the T5% of SPI (73.10 °C) is lower than that of the pure PBS (342.62 °C). The T5% of PBS composites is lower than that of the pure PBS, and it decreases with SPI content, indicating that SPI decreases the initial thermal stability of the blends. Moreover, with the addition of SPI, the Tmax, PBS values of the composites are lower than that of the pure PBS, due to poor thermal stability of SPI.
Furthermore, the Tmax, PBS values of PBS/SPI-HDDA/BPO composites are larger than those of PBS/SPI composites, indicating that the presence of SPI-g-HDDA-g-PBS enhances the adhesion between PBS and SPI, leading to improved thermal stability.
It is meaningful to study the thermal degradation kinetics and degradation mechanisms of materials by calculating the Et. [50] The Et of the pure PBS is the largest, meaning that the pure PBS is thermally stable than others. The addition of SPI decreases the Et of the composites, due to its poor thermal stability. In addition, the presence of SPI-g-HDDA-g-PBS in PBS/SPI-HDDA/BPO composites may improve the thermal stability, contributing to higher Et than that of PBS/SPI composites. These results are consistent with Tmax, PBS values.  The DSC curves of PBS and PBS composites are shown in Fig. 8 and DSC data are listed in Table 6. The second heating curves of the specimens are presented in Fig. 8(b). Double melting peaks exist on curves of the pure PBS and PBS/SPI composites, attributing to the melting-recrystallization mechanism. During the cooling process, two stacks of lamellae with different sizes are formed, namely metastable and perfect lamellae.

DSC measurements
During the heating program, the imperfect lamellae melt firstly, and then, they rearrange into perfect lamellae. Therefore, the low and high temperature melting peaks are attributed to the melting of metastable and perfect lamellae, respectively.    The crystal structures of SPI, PBS and PBS composites are studied using WAXD analyses (Fig. 9). For the pure PBS, there are two strong diffraction peaks locate at 19.6° and 22.66°, and an inconspicuous peak locates at 21.8°, corresponding to (020),  Table 7 shows the grain sizes of PBS and PBS composites. The grain sizes of (020) and (110)   between two phases. The SEM micrographs (Fig. 1) show that the fractured surfaces are coarse, accompanying with large holes and interstices for PBS/SPI composites.
Zhou et al. [55] also found the reduction in mechanical properties of PBAT/SM blends due to the poor interfacial adhesion.
Moreover, the tensile strength curve of PBS/SPI-HDDA/BPO composites is higher than that of PBS/SPI composites, indicating that the crosslinked graft copolymer enhances the adhesion between PBS and SPI. As shown in Fig. 1  that the crosslinked graft copolymer enhances the adhesion between PBS and SPI, and thus improves the water resistance of the composites. Xu et al. [58] found that the water resistance of SPI films was increased due to the formation of crosslinked network structure between 1,2,3-propanetriol diglycidyl ether (PTGE) and SPI. Das et al. [16] also found that the water absorption of crosslinked SPC decreased due to the increase of the percentage of formaldehyde or furfural.
For PBS/SPI (60/40) composite, the water absorption decreases significantly with time, showing that SPI dissolves in water. As shown in SEM micrographs (Fig. 1), many holes and cracks are observed on the fractured surface of PBS/SPI composites, and thus SPI particles may detach easily from the matrix at 40% content. The crosslinked graft copolymer forms a network structure to enhance the adhesion between PBS and SPI phases, showing that the water absorption of PBS/SPI (60/40) composite is lower than that of PBS/SPI-HDDA/BPO (60/40/0.1) composite.

Conclusions
The reactive blending method is used to produce PBS/SPI bio-composites due to its low-cost and facile production. SEM, FTIR, DMA and Soxhlet extraction experiments indicate the existence of the branched graft copolymer (SPI-g-HDDA-g-PBS) in the PBS/SPI-HDDA/BPO composites, which improves the compatibility and adhesion between the PBS matrix and SPI phase. The rheological analyses indicate that SPI-g-HDDA-g-PBS enhances the entanglement between two phases and increases the melt viscosity and elasiticity. The contact angle results show that the hydrophobicity of PBS/SPI-HDDA/BPO composites is higher than that of PBS/SPI composites because the existence of SPI-g-HDDA-g-PBS facilitates to form PBS-rich surfaces. It is found that the thermal stability of the composites is lower than the pure PBS due to the poor thermal stability of SPI. In addition, the presence of SPI-g-HDDA-g-PBS results in improved thermal stability. The DSC results show that the crosslinked graft copolymer serves as the nucleating agent to promote the crystallization of PBS in the composites. The WAXD results show that the addition of SPI, BPO and HDDA does not modify the crystal form of PBS. The tensile strength of PBS/SPI-HDDA/BPO composites is higher than that of PBS/SPI composites whereas the elongation at break of the former is lower than the latter due to the crosslinked graft copolymer enhances the adhesion between PBS and SPI. In addition, the enhanced adhesion between PBS and SPI by the grafts also improves the water resistance of the composites.