Optimization of Maleinized Linseed Oil Loading as a Biobased Compatibilizer in Poly(Butylene Succinate) Composites with Almond Shell Flour

Green composites of poly(butylene succinate) (PBS) were manufactured with almond shell flour (ASF) by reactive compatibilization with maleinized linseed oil *MLO) by extrusion and subsequent injection molding. ASF was kept constant at 30 wt %, while the effect of different MLO loading on mechanical, thermal, thermomechanical, and morphology properties was studied. Uncompatibilized PBS/ASF composites show a remarkable decrease in mechanical properties due to the nonexistent polymer‒filler interaction, as evidenced by field emission scanning electron microscopy (FESEM). MLO provides a plasticization effect on PBS/ASF composites but, in addition, acts as a compatibilizer agent since the maleic anhydride groups contained in MLO are likely to react with hydroxyl groups in both PBS end chains and ASF particles. This compatibilizing effect is observed by FESEM with a reduction of the gap between the filler particles and the surrounding PBS matrix. In addition, the Tg of PBS increases from −28 °C to −12 °C with an MLO content of 10 wt %, thus indicating compatibilization. MLO has been validated as an environmentally friendly additive to PBS/ASF composites to give materials with high environmental efficiency.


Introduction
Nowadays, there is growing interest in the development of biopolymers that could substitute for, or at least compete with, conventional petroleum-based polymers. This need is much more pronounced in the packaging industry due to the high volume of waste it generates. Among different alternatives such as proteins, polysaccharides, bacterial polymers, and so on, biopolyesters (either from petroleum origin or bio-derived) are gaining relevance as they are biodegradable (disintegrable in controlled compost soil conditions). One of these polyesters is poly(butylene succinate) (PBS), which presents interesting possibilities for manufacturing wood plastic composites (WPCs) or natural fiber-reinforced plastics (NFRPs) [1]. In general, the biodegradable synthetic polymers are mainly aliphatic polyesters produced by microbiological and chemical synthesis, natural polymer-based products, and their blends such as poly(lactic acid) (PLA), poly(ε-caprolactone) (PCL), poly(glycolic acid) (PGA), poly(butylene succinate) (PBS), poly(butylene succinate-co-adipate), among others. PBS offers interesting possibilities in the packaging industry due to an excellent combination of flexibility and biodegradability (more correctly, disintegration in controlled compost soil). PBS is produced through the condensation reaction of glycols such as 1,4-butanediol and succinic acid. Currently, PBS is obtained from the end chains and starch provides enhanced mechanical properties and thermal stability. It has also been proven to have its in poly(3-hydroxybutyrate) (PHB) formulations [36].
In a previous work with PBS composites and almond shell flour (ASF), maleinized linseed oil was revealed as the best compatibilizer compared to other families such as acrylates, anhydrides, and epoxidized oils [1]. This work is focused on the study of the effect of MLO loading on the final properties of PBS-ASF composites with the aim of improving interface phenomena between the PBS matrix and the embedded ASF particles.

Materials
A commercial grade of poly(butylene succinate) (PBS, Bionolle 1020MD, supplied by Showa Denko, Tokyo, Japan) was used as the base material for composites. This commercial grade possesses a melt flow index (MFI) comprised between 20 and 34 g/(10 min) and a density of 1.26 g cm −3 .
Almond shell powder/flour (ASF) was purchased from Jesol Materias Primas (Valencia, Spain). The supplied powder was sieved in a vibrational sieve RP09 CISA®(Barcelona, Spain) to obtain a homogeneous particle size of 150 µm.
Maleinized linseed oil (MLO) was supplied by Vandeputte (Mouscron, Belgium) under the trade name of VEOMER LIN. It possesses a viscosity of 10 dPa, measured at 20 • C, and an acid value of 105-130 mg KOH g −1 .

Manufacturing of PBS/ASF/MLO Composites
Initially, PBS and ASF were dried at 50 • C for 24 h to remove moisture since polyesters are highly sensitive to hydrolysis at high temperatures. MLO was heated to 40 • C to reduce its viscosity and enable good pre-mixing with PBS and ASF. Pre-mixing was carried out using ziploc bags with the appropriate compositions of each component (see Table 1). After this pre-mixing stage, the mixtures were compounded in a twin-screw co-rotating extruder from Dupra S.L. (Alicante, Spain) with a screw diameter of 25 mm and a length (L) to diameter (D) ratio (L/D) of 24. The screw speed was set to 40 rpm and the temperature profile was as follows: 120 • C, 125 • C, 130 • C, and 130 • C from the hopper to the die. After extrusion, the compounded materials were pelletized and dried at 50 • C for 24 h before further processing by injection molding in a Meteor 270/75 injection machine from Mateu & Solé (Barcelona, Spain). The temperature profile was set to 110 • C (hopper), 115 • C, 120 • C, and 125 • C (injection nozzle). with an energy of 1 J was used to obtain the impact energy on standard notched samples ("V" type notch with a radius of 0.25 mm), according to ISO 179-1:2010. At least five different samples were used for each test to ensure reproducible results and average values were calculated for each property.

Morphological Characterization
Fractured samples from impact tests were observed by field emission scanning electron microscopy (FESEM) in a ZEISS ULTRA 55 microscope from Oxford Instruments (Abingdon, UK). All samples were previously subjected to a sputtering process with an Au-Pd alloy to enhance electrical conductivity. The sputtering was carried out in a EMITECH SC7620 sputter coater from Quorum Technologies (Lewes, UK).

Thermal Characterization
The main thermal transitions were obtained by differential scanning calorimetry (DSC) using a Mettler-Toledo 821 calorimeter (Schwerzenbach, Switzerland). Samples with an average size of 5-7 mg were subjected to a dynamic thermal program consisting on three stages, a first heating from 25 • C to 200 • C, then a controlled cooling from 200 • C to −50 • C and, finally, a second heating process from −50 • C to 300 • C. The scanning rate was 10 • C min −1 for all three stages. All DSC runs were conducted in triplicate under a nitrogen atmosphere with a flow rate of 66 mL min −1 . Standard sealed aluminum crucibles with a total volume of 40 µL were used. The degree of crystallinity was calculated by the following equation: where ∆H m stands for the measured melt enthalpy. ∆H 0 m (J g −1 ) represents the melt enthalpy of a theoretically fully crystalline PBS polymer, with a value of 110.3 J g −1 for PBS [49]. With regard to w, it represents the weight fraction of all added components (except for PBS), including both ASF and MLO. The thermal stability at elevated temperatures was followed by thermogravimetry (TGA) in a TGA/SDTA 851 from Mettler-Toledo. Standard alumina crucibles with a total volume of 70 mL were charged with approx. 6 mg and subsequently subjected to a dynamic heating program from 30 • C to 700 • C at a constant heating rate of 20 • C min −1 in air atmosphere.

Thermomechanical Characterization
Dynamic mechanical thermal analysis (DMTA) was conducted on a DMA1 analyzer from Mettler-Toledo. Single cantilever flexural tests were carried out at a frequency of 1 Hz at a heating rate of 2 • C min −1 and a maximum deflection of 10 µm. Rectangular samples with dimensions of 10 × 7 × 1 mm 3 were subjected to a dynamic heating program from −50 • C to 80 • C.
Thermal stability of the developed composite materials was studied by thermomechanical analysis (TMA) in a Q400 TMA analyzer from TA Instruments (New Castle, DE, USA). A constant force of 0.02 N was applied to squared samples of 10 × 10 × 4 mm 3 and subjected to a temperature sweep from −50 • C up to 80 • C at a constant heating rate of 2 • C min −1 . The coefficient of linear thermal expansion (CLTE) was calculated both below and above the glass transition temperature (T g ). All thermal runs were done in triplicate. Table 2 summarizes the main parameters obtained from mechanical characterization as a function of the MLO content. As expected, the addition of ASF into the PBS matrix leads to an increase in stiffness together with a clear decrease in ductility. PBS possesses a σ t of about 31.5 MPa; this is dramatically reduced by almost half (14.8 MPa) by filling the PBS matrix with 30 wt % ASF. As described previously, the highly hydrophilic nature of ASF does not allow interaction with the highly hydrophobic PBS matrix and this is responsible for poor load transfer from the filler to the matrix. In fact, the filler offers the typical stress concentration phenomenon due to a lack of interaction with the polymeric matrix. This lack of interaction contributes to poor material cohesion (small gaps between the particle filler and the surrounding matrix) and, therefore, the mechanical properties are worse than neat PBS. This stress concentration phenomenon is also evident when observing the extremely high decrease in ε b , which drops from 215.6% to 6.3% (a decrease of about 97%). In contrast, the material becomes stiffer, as shown by the E t values. Neat PBS is a flexible polymer with a relatively low modulus of 417.4 MPa; this is almost doubled by the addition of 30 wt % ASF (787.9 MPa). This increase is evident as the modulus represents the ratio of the applied stress (σ) and the obtained elongation (ε) in the linear region. As has been shown in Table 2, the tensile strength decreases (by 50%) but the elongation at break decreases by 97%. Therefore, the σ/ε ratio gives higher values as the decrease in ε is much pronounced than that of σ.

Effect of MLO Loading on Mechanical Properties of PBS/ASF Composites
MLO has a dual effect on PBS/ASF composites. As the MLO load increases, both the tensile modulus and the tensile strength decrease due to a plasticizing effect, as reported by Ferri et al. [50] for PLA formulations containing 10-13 wt % MLO. Tensile strength and tensile modulus decrease in all formulations with increasing MLO content. In contrast, ductility is remarkably improved, as can be seen from the elongation at break values, which change from 6.3% (uncompatibilized PBS/ASF composite) to almost 26% for the PBS/ASF composite compatibilized with 4.5 wt % MLO. The elongation at break is not remarkably improved with higher MLO content. Quiles-Carrillo et al. [23] reported similar findings in PLA composites with ASF compatibilized with MLO. As an efficient plasticizer, MLO remarkably improved the ductility of the PLA/ASF composites. Elongation at break increased by 292% and 84% in relation to the uncompatibilized PLA/ASF and to the neat PLA, respectively. Ferri et al. [50] also reported a remarkable increase in elongation at break in PLA formulations with 10-13 wt % MLO. This increase in ductility is related to the lubricant effect of MLO, which increases chain mobility. The plasticizer increases the free volume and, subsequently, chain interactions decrease, leading to improved intermolecular mobility. Chieng et al. [35] also reported similar findings in PLA biocomposites plasticized with epoxidized vegetable oils (EVOs).
Regarding Shore D hardness, its evolution is identical to the tensile modulus. Neat PBS shows a Shore D value of 60.1. The addition of 30 wt % ASF leads to stiffer material with a Shore D value of 71.2. Then, as the MLO content increases, Shore D progressively decreased to values of 62.0 for the PBS/ASF composite with 10 wt % MLO.
The second effect of MLO can be inferred by observing the impact strength values in Table 2. Neat PBS is characterized by good energy absorption, which gives an impact strength of 16.5 kJ m −2 . The addition of 30 wt % ASF causes a dramatic decrease in impact strength down to 1.8 kJ m −2 . The impact strength is directly related to both ductility and resistance. As previously indicated, both the tensile strength and the elongation at break are remarkably reduced due to poor material cohesion, which does not allow stress transfer and, hence, there is low deformation with low applied stresses. Obviously, the ability to absorb energy is directly linked to the deformation capability, together with the need for high stress to promote deformation. Therefore, ASF addition leads to extremely low impact energy values. The addition of MLO provides a plasticizing effect, as indicated previously; in addition, reactions between MLO and both PBS and ASF could be expected as the absorbed energy increases to almost 4 kJ m −2 with 4.5 wt % MLO (which represents a percentage increase of more than 120% with respect to the uncompatibilized PBS/ASF composite). In fact, the maleic anhydride pendant group can react with hydroxyl groups in both PBS (end chains) and ASF particles (cellulose and hemicellulose). This effect has been previously reported by Ernzen et al. with polyamides and maleinized soybean oil [51], and Quiles-Carrillo et al. with polyesters [23,47] with maleinized linseed oil. Figure 1 gathers FESEM images corresponding to PBS/ASF composites (uncompatibilized and MLO-compatibilized composites). As can be observed in Figure 1a, polymer-particle adhesion is very poor. It is possible to find a small gap (1-3 µm) between the ASF particle and the surrounding PBS matrix. In addition, this lack of interaction is evident upon a detailed observation of the PBS fracture surface. PBS shows a porous pattern (1 µm) that is the reverse of that of ASF, which is identical to the typical surface of an almond shell at the macroscale [52]. The addition of 2.5 wt % MLO (Figure 1b) provides some interaction as the gap has been reduced. Nevertheless, some ASF particles have been pulled out without any interaction. Once again, it is possible to observe a negative copy of the almond shell microstructure on the PBS matrix in some areas. Compatibilization is more evident with higher MLO loads, as can be seen in Figure 1c, 1d, and 1e with 4.5, 7.5, and 10 wt % MLO, respectively. The typical PBS surface (negative copy of the ASF microstructure) after particle removal is not detected. The gap is almost inexistent and no signs of particle removal can be observed. All these issues indicate the good compatibilizing effect of MLO as it can react with both PBS and ASF. Table 3 and Figure 2 show a summary of the main thermal parameters obtained by DSC characterization of PBS/ASF composites. The melt peak temperature remains almost invariable after the addition of ASF and MLO and changes in a very narrow range from 113 • C to 116 • C [53]. A slight decrease in the melt peak temperature can be observed. This could be related to the nucleant effect a lignocellulosic filler exerts on semicrystalline polymers. Accordingly, the overall degree of crystallinity increases after the addition of ASF due to an interfacial phenomenon as the cellulose crystals act as nucleant points for PBS [54,55]. On the other hand, the addition of MLO provides a decrease in crystallinity down to values close to those of neat PBS. This effect is particularly noticeable for high MLO contents in the 7.5-10.0 wt % range. The plasticizing effect of MLO at higher concentrations provides increased chain mobility, which contributes to a slight decrease in the melt peak temperature down to 113 • C. It is worth noting the role of the melt temperature and crystallinity in the overall processing of polymers and their composites [56].   Table 3 and Figure 2 show a summary of the main thermal parameters obtained by DSC characterization of PBS/ASF composites. The melt peak temperature remains almost invariable after the addition of ASF and MLO and changes in a very narrow range from 113 °C to 116 °C [53]. A slight decrease in the melt peak temperature can be observed. This could be related to the nucleant effect a lignocellulosic filler exerts on semicrystalline polymers. Accordingly, the overall degree of crystallinity increases after the addition of ASF due to an interfacial phenomenon as the cellulose crystals act as nucleant points for PBS [54,55]. On the other hand, the addition of MLO provides a decrease in crystallinity down to values close to those of neat PBS. This effect is particularly noticeable for high MLO contents in the 7.5-10.0 wt % range. The plasticizing effect of MLO at higher concentrations provides increased chain mobility, which contributes to a slight decrease in the melt With regard to thermal degradation at elevated temperatures, Figure 3a shows the TGA thermogram with the weight loss in function of increasing temperature. It has been reported that PBS decomposes by the following mechanisms, which cover typical random chain scission of aliphatic polyesters such as PLA [57], and specific chain scission. Shih et al. reported that the diffusion effect becomes important at elevated temperatures during degradation of PBS [58].

Effect of MLO Loading on Thermal Properties of PBS/ASF Composites
With regard to thermal degradation at elevated temperatures, Figure 3a shows the TGA thermogram with the weight loss in function of increasing temperature. It has been reported that PBS decomposes by the following mechanisms, which cover typical random chain scission of aliphatic polyesters such as PLA [57], and specific chain scission. Shih et al. reported that the diffusion effect becomes important at elevated temperatures during degradation of PBS [58].  Addition of 30 wt % ASF into the PBS matrix provides a slight change in the TGA curves as the ASF degradation occurs in different stages. ASF shows the typical degradation profile of a lignocellulosic filler. At about 90-100 °C it is possible to observe a first weight loss that corresponds to the residual moisture contained in the ASF filler [59]. Over this temperature, ASF shows thermal stability up to 210-220 °C. Degradation of cellulose and hemicelluloses starts at about 220 °C and proceeds until 330-340 °C. Decomposition of cellulose and hemicellulose involves complex reactions (dehydration, decarboxylation, etc.) and the breakage of C-H, C-O, and C-C bonds. Lignin is more thermally stable and its degradation occurs in a broad temperature range starting at about 250 °C and ending at 450 °C in a progressive weight loss process. This higher thermal stability of lignin versus cellulose and hemicellulose is due to the presence of aromatic rings. Quiles-Carrillo et al. [60] showed similar degradation behavior in PLA/ASF composites. As has been reported by many authors, the degradation of lignocellulosic materials is a complex issue that mainly involves the degradation of hemicelluloses, cellulose, and lignin [61]. The components that most readily degrade due to a temperature rise are hemicelluloses, which decompose (depolymerize) in the temperature range 180-350 °C [62]. Crystalline areas of cellulose are more thermally stable and start to degrade at 275 °C; the decomposition takes place up to temperatures of 350-370 °C [63]. With regard to lignin, it shows a broader degradation range that starts at about 200 °C and lasts up to 700 °C, showing a typical shoulder at elevated temperatures in TGA thermograms [64]. The TGA curves corresponding to PBS/ASF composites show a slight decrease at the onset degradation temperature since the ASF filler Addition of 30 wt % ASF into the PBS matrix provides a slight change in the TGA curves as the ASF degradation occurs in different stages. ASF shows the typical degradation profile of a lignocellulosic filler. At about 90-100 • C it is possible to observe a first weight loss that corresponds to the residual moisture contained in the ASF filler [59]. Over this temperature, ASF shows thermal stability up to 210-220 • C. Degradation of cellulose and hemicelluloses starts at about 220 • C and proceeds until 330-340 • C. Decomposition of cellulose and hemicellulose involves complex reactions (dehydration, decarboxylation, etc.) and the breakage of C-H, C-O, and C-C bonds. Lignin is more thermally stable and its degradation occurs in a broad temperature range starting at about 250 • C and ending at 450 • C in a progressive weight loss process. This higher thermal stability of lignin versus cellulose and hemicellulose is due to the presence of aromatic rings. Quiles-Carrillo et al. [60] showed similar degradation behavior in PLA/ASF composites. As has been reported by many authors, the degradation of lignocellulosic materials is a complex issue that mainly involves the degradation of hemicelluloses, cellulose, and lignin [61]. The components that most readily degrade due to a temperature rise are hemicelluloses, which decompose (depolymerize) in the temperature range 180-350 • C [62]. Crystalline areas of cellulose are more thermally stable and start to degrade at 275 • C; the decomposition takes place up to temperatures of 350-370 • C [63]. With regard to lignin, it shows a broader degradation range that starts at about 200 • C and lasts up to 700 • C, showing a typical shoulder at elevated temperatures in TGA thermograms [64]. The TGA curves corresponding to PBS/ASF composites show a slight decrease at the onset degradation temperature since the ASF filler is less thermally stable than PBS [11]. Sanchez-Jimenez at al. studied the thermal decomposition of cellulose and the typical degradation temperature range they reported overlaps with that of PBS [65]. It has been reported that the reactive extrusion of aliphatic polyesters such as PLA with styrene-epoxy acrylic oligomers can provide increased thermal stability due to the branching effect reactive extrusion provides [66]. As has been suggested in this study, MLO provides two overlapping processes: on one hand, a plasticizing effect and, on the other hand, a compatibilization effect due to the reaction of maleic anhydride with hydroxyl groups in PBS and ASF filler. Despite this, the thermal stability at moderate temperatures is not improved, as can be seen in Table 4, with a decreasing values of T 5% with increasing MLO. The same tendency can be observed for the maximum degradation rate temperature (T max ) from 415.64. Despite this, it seems that the typical degradation peak of lignin at elevated temperatures is moved to higher temperatures with MLO addition, as can be observed in Figure 3b, which gathers the first derivative TGA curves for all PBS/ASF composites as well as raw ASF and neat PBS.  Figure 4 shows the evolution of the loss modulus (G") and the damping factor (tan ∂). The glass transition temperature (T g ) can be estimated (among other methods) by the peak temperature corresponding to the loss modulus (G"). It is important to remark that the T g value could not be observed by DSC due to the low signal it gives. Neat PBS shows a T g value of about −29 • C, which is in accordance with other T g values reported in the literature [67]. The addition of 30 wt % ASF does not promote any change in the T g value, thus indicating no interaction. In contrast to the expected behavior suggested by mechanical characterization, as the MLO increases the T g also increases, in a slight way, but detectable by DMTA characterization. The compatibilizing effect of MLO by reaction with both PBS and ASF filler is evidenced by an increase in the T g value since the reaction of MLO with both polymer matrix and particle filler restricts chain mobility and, hence, the T g is increased. As reported by Quiles-Carrillo et al. [23] in a previous work, the effect of MLO is evident when comparing the FTIR spectra of uncompatibilized and compatibilized PLA/ASF composites. Comparison of these FTIR spectra indicate that new esters and carboxylic acids are obtained after reactive extrusion with PLA and ASF. They propose the formation of a cellulose-g-PLA by the reaction of the multiple anhydride groups in MLO with the hydroxyl groups in both PLA end chains and the lignocellulosic filler. Similar behavior could be expected for this system with a more flexible polymer (PBS) and the same lignocellulosic filler without previous ultraviolet (UV) surface treatment. PBS/ASF composites with 2.5 wt % MLO possess a T g of −27 • C, which means a slight increase of 2 • C with respect to the PBS (or uncompatibilized PBS/ASF composite). As has been reported, MLO could potentially provide several overlapped effects such as plasticization, chain extension, branching, and even crosslinking. All these phenomena could overlap and, depending on the polymer, one mechanism could stand out over the others [50].

Effect of MLO Loading on Thermomechanical Properties of PBS/ASF Composites
In this system, some plasticization is evident, as suggested by mechanical characterization, but, together with this phenomenon, compatibilization could occur, due to the reaction of maleic anhydride with PBS end chains and the lignocellulosic filler that restricts the Brownian motion of long-chain molecules [68]. On the other hand, it has also been reported that MLO could lead to some polymer crosslinking that could contribute to a slight increase in the T g . This increase in T g is clear evidence of the compatibilization effect [69]. As the MLO content increases, a slight increase in T g is observed as well. Therefore, for PBS/ASF composites compatibilized with 10 wt % MLO, the T g reaches values of about −23.5 • C. To obtain a complete characterization of the thermomechanical properties, the change in dimensions as a function of temperature was obtained by thermomechanical analysis (TMA). In particular, the coefficient of linear thermal expansion (CLTE) was calculated both below and above the glass transition temperature, as shown in Table 5. As PBS is extremely flexible, the comparison material has been the uncompatibilized PBS/ASF composite, which shows a CLTE of 59.3 µm m −1 °C −1 . As previously indicated, the addition of MLO provides an increase in elongation due to a combined effect of plasticization and compatibilization through reaction. This increase in ductility is directly related to an increase in the CLTE. In fact, the CLTE shows a clear increasing trend with the MLO content, up to values of 84.8 µm m −1 °C −1 below the Tg. This same tendency can be observed for the CLTE measured above the Tg, as seen in Table 5.

Conclusions
Maleinized linseed oil (MLO) has been validated as a good compatibilizer in green composites of PBS and ASF, with the additional feature of being environmentally friendly. Uncompatibilized PBS/ASF composites show a dramatic decrease in both tensile strength and elongation at break due to the lack of polymer-filler interaction. Accordingly, the impact strength of the uncompatibilized PBS/ASF composite is also reduced. MLO provides a combination of plasticization and compatibilization through the reaction of maleic anhydride pendant groups in the triglyceride structure with the hydroxyl groups contained in both PBS (end chains) and ASF (mainly cellulose and hemicelluloses). This leads to a noticeable increase in elongation at break as well as in the impact strength. MLO reaction with both PBS and ASF is also assessed by an increase in the Tg from −28 °C To obtain a complete characterization of the thermomechanical properties, the change in dimensions as a function of temperature was obtained by thermomechanical analysis (TMA). In particular, the coefficient of linear thermal expansion (CLTE) was calculated both below and above the glass transition temperature, as shown in Table 5. As PBS is extremely flexible, the comparison material has been the uncompatibilized PBS/ASF composite, which shows a CLTE of 59.3 µm m −1 • C −1 . As previously indicated, the addition of MLO provides an increase in elongation due to a combined effect of plasticization and compatibilization through reaction. This increase in ductility is directly related to an increase in the CLTE. In fact, the CLTE shows a clear increasing trend with the MLO content, up to values of 84.8 µm m −1 • C −1 below the T g . This same tendency can be observed for the CLTE measured above the T g , as seen in Table 5.

Conclusions
Maleinized linseed oil (MLO) has been validated as a good compatibilizer in green composites of PBS and ASF, with the additional feature of being environmentally friendly. Uncompatibilized PBS/ASF composites show a dramatic decrease in both tensile strength and elongation at break due to the lack of polymer-filler interaction. Accordingly, the impact strength of the uncompatibilized PBS/ASF composite is also reduced. MLO provides a combination of plasticization and compatibilization through the reaction of maleic anhydride pendant groups in the triglyceride structure with the hydroxyl groups contained in both PBS (end chains) and ASF (mainly cellulose and hemicelluloses). This leads to a noticeable increase in elongation at break as well as in the impact strength. MLO reaction with both PBS and ASF is also assessed by an increase in the T g from −28 • C (neat PBS) up to −12 • C (PBS/ASF composite compatibilized with 10 wt % MLO). The compatibilization effect is also observed by scanning electron microscopy as the gap between the filler and the surrounding matrix almost disappears. Therefore, the herein-developed composites represent an interesting approach to reduce the overall cost of using PBS by using industrial waste coming from the almond industry, thus leading to highly environmentally friendly materials. These PBS/ASF composites can be successfully compatibilized with a vegetable-oil-derived additive, which contributes to the high environmental efficiency of these materials.