Bioinspired Tannic Acid-Modified Coffee Grounds as Sustainable Fillers: Effect on the Properties of Polybutylene Adipate Terephthalate Composites

Preparing composites from gricultural waste with biodegradable polymers is one of the strategies used to ensure the long-term sustainability of such materials. However, due to the differences in their chemical properties, biomass fillers often exhibit poor interfacial adhesion with polymer matrices. Inspired by mussel foot silk, this work focused on the surface modification of coffee grounds (CGs) using a combination of tannic acid (TA) and alkali treatment. CGs were used as a biomass filler to prepare polybutylene adipate terephthalate (PBAT)/CG composites. The modification of CGs was demonstrated by Fourier transform infrared spectroscopy (FTIR), the water contact angle, and scanning electron microscopy (SEM). The effect of CGs on the rheological, tensile, and thermal properties of the PBAT/CG composites was investigated. The results showed that the addition of CGs increased the complex viscosity, and the surface modification enhanced the matrix–filler adhesion. Compared with unmodified CG composites, the tensile strength and the elongation at break of the composite with TA-modified alkali-treated CGs increased by 47.0% and 53.6%, respectively. Although the addition of CGs slightly decreased the thermal stability of PBAT composites, this did not affect the melting processing of PBAT, which often occurs under 200 °C. This approach could provide a novel method for effectively using biomass waste, such as coffee grounds, as fillers for the preparation of polymer composites.


Introduction
As a popular drink, coffee has gradually become an indispensable part of people's lives. Based on the latest statistics of the International Coffee Organization (ICO), the global consumption of coffee is approximately 10 million tons [1]. Thus, coffee grounds (CGs) are generated in huge quantities after coffee processing, accounting for more than 50% of coffee beans [2,3]. Currently, they are mostly disposed of by being stored and incinerated; this produces methane and carbon dioxide, which are very harmful to the environment [4,5]. Therefore, there is an urgent need to develop new strategies to address the environmental pollution and disposal difficulties associated with CGs. In this context, many researchers have focused on valorizing CGs to obtain high-value products, such as biofuels [6][7][8], adsorbents [9,10], antioxidants [11][12][13], etc. Furthermore, as a biomass filler, they not only reduce the cost of producing biodegradable polyesters, but also preserve the degradable properties of the composite [14,15]. Research on eco-friendly biomass-based composites has recently become a hot topic.
However, their different chemical properties constitute a real barrier to the incorporation of CGs in polymer matrices, thereby leading to the deterioration of the composites' (Shanghai, China). Sodium chloride (NaCl) and sodium hydroxide (NaOH) were o from Shanghai Macklin Biochemical Co., Ltd. (Shanghai, China).

Preparation of CG
The CG was dried in oven at 80 °C and then pulverized into powder by mec crusher (800Y, Yongkang Boou Hardware Products Co., Ltd., Yongkang, China). powders were sieved with 100 meshes and collected.

Surface Modification of CG
The 4.0 g TA and 30.0 g NaCl were dissolved in 2 L of distilled water to achie TA solution, and the pH was adjusted to 8.5 through Tris. After that, 100.0 g CG wa to TA solution by stirring at room temperature for 24 h. The modified CG was a by extraction, washing, drying, and recorded as CG-TA. The mechanism diagram surface-modified CG is shown in Figure 1. TA forms tris-complex with metal ions 7. In this paper, the TA-Na + complexes were deposited on the surface of CG, intro highly reactive polyphenol hydroxyl groups through hydrogen bonding [28,32]. hesion component of mussel foot silk protein was simulated on the CG surface, wh expected to enhance the interfacial interaction with the polymer matrix. CG-OH tained by impregnating dried CG in 2 wt% NaOH solution for 24 h, and CG-OHmodified by TA on the basis of CG-OH.

Preparation of PBAT/CG Composites
CG and PBAT were fully dried, weighed by 30:70 (w/w), and manually mixe The mixture was melt blended through internal mixer at temperatures of 140 °C fo and the speed of the screws was 60 rpm. The PBAT/CG blends were compression at 140 °C with a holding pressure of 10 MPa for 5 min. All composites were fast between two platens.

Characterization
The absorption peaks of CG-related functional groups were recorded by Transform Infrared Spectroscopy (FTIR) (Nicolet 6700, American Thermo Compa der transmission mode using the potassium bromide compression method. The

Preparation of PBAT/CG Composites
CG and PBAT were fully dried, weighed by 30:70 (w/w), and manually mixed well. The mixture was melt blended through internal mixer at temperatures of 140 • C for 7 min and the speed of the screws was 60 rpm. The PBAT/CG blends were compression molded at 140 • C with a holding pressure of 10 MPa for 5 min. All composites were fast cooled between two platens.

Characterization
The absorption peaks of CG-related functional groups were recorded by Fourier Transform Infrared Spectroscopy (FTIR) (Nicolet 6700, American Thermo Company) under transmission mode using the potassium bromide compression method. The spectra were gained by 32 scans with resolution of 4 cm −1 in the wavenumber range of 4000 to 500 cm −1 .
The water contact angle was tested at room temperature with contact angle tester (DSA100, Germany). The CG powders were pressed into sheets by infrared sheet press with water as the liquid phase. The scanning electron microscopy (SEM) (JSM-6510, Japan) was used to observe the surface morphology of CG powder and the frozen fracture morphology of PBAT/CG composites at 10 kV accelerating voltage. Prior to observing the phase morphology, the samples were gold sprayed to enhance conductivity. The elemental content of CG surface could be obtained using Energy-dispersive X-ray spectroscopy (EDS) (OXFORD INCA250, Oxford, UK).
The rheological properties of the neat PBAT and its composites were investigated through rotary rheometer (MCR302, Graz, Austria). Fixed strain values of 1% were used to verify the linear viscoelastic zone. The dynamic frequency scanning was performed at 160 • C and 0.1-100 rad/s shear frequency with fixed strain amplitude. All samples were circular plates that were 25 mm in diameter and 1.0 mm in thickness.
The thermal properties of PBAT and composites were determined using differential scanning calorimetry (DSC) (DSC 204fl Phoenix). All samples (5-8 mg) were heated from 25 • C to 200 • C at a rate of 10 • C/min, held constant for 2 min, and cooled to 25 • C at the same rate. The whole process was under N 2 atmosphere. The crystallinity (X c ) of PBAT was obtained according to Equation (1).
where ∆H m 0 and ∆H m were the melt enthalpy of 100% crystalline PBAT (114 J·g −1 ) and the melt enthalpy of composites, respectively. ω PBAT was the mass fraction of PBAT in the composites.
According to ASTM D882-2018, the tensile properties of composite sheets were tested with universal testing machine (CMT5254, Shenzhen Sans Measurement Technology Co., Ltd., Shenzhen, China). For each formulation, the five dumbbell-type specimens were tested with stretching rate of 50 mm/min. Thermogravimetric analysis (TGA) (Q500, TA Instruments, New Castle, DE, USA) was carried out to determine the thermal stability of PBAT and composites. The experiments were performed in the temperature range of 25-600 • C under N 2 atmosphere at heating rate of 10 • C/min.

Characterization of CG
In order to demonstrate that TA effectively modified CGs and interacted with the CG surface, FTIR was performed to gain insight into the functional groups of TA, pristine CG, and modified CG (Figure 2a). The results show that TA has strong absorption peaks at 3390 cm −1 and 1720 cm −1 , which are attributed to the stretching vibrations of hydroxyl (-OH) and carbonyl groups (C=O), respectively [35]. As for CGs, the appearance of a broad peak in the wavelength range of 3000 to 3600 cm −1 corresponds to the stretching vibrations of the O-H and N-H groups present in the lignocellulosic components and proteins. The two sharp peaks near 2850 cm −1 and 2930 cm −1 correspond to the symmetric and asymmetric stretching of C-H bonds, respectively. Combined with the carbonyl peak at 1740 cm −1 , these two peaks are associated with ester groups in lipids [36,37]. Since the characteristic peak of TA overlapped with CGs, it was necessary to perform an alkali pretreatment of the CGs. The results showed that the carbonyl group peak of CG-OH disappeared, while the carbonyl group peak of CG-OH-TA reappeared at 1720 cm −1 and was consistent with the position of the TA carbonyl peak.
According to Moraczewski [38], the shifting or broadening of hydroxyl groups' peak position was usually a sign of hydrogen bond formation. Moreover, Guan et al. [39] believed that the stretching vibrations of the hydroxyl group appearing at a lower wavenumber reflected the presence of hydrogen bonding interactions, which was observed in TA-modified CGs. Indeed, the peak position shifted from 3450 cm −1 for CG to 3380 cm −1 for CG-TA and 3370 cm −1 for CG-OH-TA, revealing the formation of new hydrogen bonds between CG and TA. position was usually a sign of hydrogen bond formation. Moreover, Guan et al. [39] believed that the stretching vibrations of the hydroxyl group appearing at a lower wavenumber reflected the presence of hydrogen bonding interactions, which was observed in TA-modified CGs. Indeed, the peak position shifted from 3450 cm −1 for CG to 3380 cm −1 for CG-TA and 3370 cm −1 for CG-OH-TA, revealing the formation of new hydrogen bonds between CG and TA.  Table 1. Due to the high lipid content in CGs and interparticle interactions [36], it resulted in CG particle agglomeration and a water contact angle value of 112.2°. The alkali treatment mainly removed lipids and disrupted the adhesion between CG particles; therefore, the structure of CG-OH became loose, and the water contact angle value decreased to 100.6°. The TA modification of CGs did not only improve the compact particle structure, but it also significantly reduced the contact angle value. This phenomenon is due to the modification process and the molecular structure of TA. Indeed, stirring facilitated the dispersion of CG particles, and TA introduced sufficient polar groups on the CG surface to give activity as well as to form TA-Na + complexes, which increased the roughness of the CG surface [34], thus improving the interfacial adhesion. Furthermore, the hydrophilic-modified CGs could improve the interfacial adhesion with the hydrophobic PBAT matrix. In addition, a higher O/C ratio implies a higher TA content on the CG surface [40,41]. The O/C ratio of pristine CGs was 42.2% and increased to 47.5% for CG-TA and 53.9% for CG-OH-TA. The increase in Na, Cl content, and O/C ratio proves the successful modification of CGs by TA.  Table 1. Due to the high lipid content in CGs and interparticle interactions [36], it resulted in CG particle agglomeration and a water contact angle value of 112.2 • . The alkali treatment mainly removed lipids and disrupted the adhesion between CG particles; therefore, the structure of CG-OH became loose, and the water contact angle value decreased to 100.6 • . The TA modification of CGs did not only improve the compact particle structure, but it also significantly reduced the contact angle value. This phenomenon is due to the modification process and the molecular structure of TA. Indeed, stirring facilitated the dispersion of CG particles, and TA introduced sufficient polar groups on the CG surface to give activity as well as to form TA-Na + complexes, which increased the roughness of the CG surface [34], thus improving the interfacial adhesion. Furthermore, the hydrophilic-modified CGs could improve the interfacial adhesion with the hydrophobic PBAT matrix. In addition, a higher O/C ratio implies a higher TA content on the CG surface [40,41]. The O/C ratio of pristine CGs was 42.2% and increased to 47.5% for CG-TA and 53.9% for CG-OH-TA. The increase in Na, Cl content, and O/C ratio proves the successful modification of CGs by TA.

Morphology
By observing the fracture morphologies of PBAT/CG composites acquired by quenched liquid nitrogen, SEM revealed the dispersion and interfacial adhesion of CG in the PBAT matrix ( Figure 3). As seen in the fracture morphology of PBAT/CG composites, the addition of CG increased the heterogeneity of PBAT blends owing to the different sizes of CG particles and uneven dispersion [42]. The formation of larger CG particles was attributed to CG self-agglomeration.

Morphology
By observing the fracture morphologies of PBAT/CG composites acquired by quenched liquid nitrogen, SEM revealed the dispersion and interfacial adhesion of CG in the PBAT matrix ( Figure 3). As seen in the fracture morphology of PBAT/CG composites, the addition of CG increased the heterogeneity of PBAT blends owing to the different sizes of CG particles and uneven dispersion [42]. The formation of larger CG particles was attributed to CG self-agglomeration. In general, the existence of CG decreased the tensile properties of composites in various ways. There were several explanations for this phenomenon. Firstly, the incorporation of CGs interrupted the continuity of PBAT chains and decreased the amount of the matrix, which obstructs the stress transfer and decreases the stress support [43]. Secondly, the CG agglomeration causes stress concentrations, creating weak points in the composites [42]. Moreover, Obasi [44] suggested that the decrease in mechanical properties of composites was also related to the polarity difference between the filler and polymer matrix and their poor interfacial adhesion.
SEM results show that the fracture morphologies of PBAT/CG-TA and PBAT/CG-OH composites exhibited smaller-sized CG particles than PBAT/CG composites; however, exposed CG particles were still observed. Furthermore, CG-OH-TA significantly enhanced the dispersion and mechanical adhesion of CG in the PBAT matrix. Additionally, no evidence of pull-out or separation of CG particles was observed in fracture morphology of the PBAT/CG-OH-TA composite. In general, the existence of CG decreased the tensile properties of composites in various ways. There were several explanations for this phenomenon. Firstly, the incorporation of CGs interrupted the continuity of PBAT chains and decreased the amount of the matrix, which obstructs the stress transfer and decreases the stress support [43]. Secondly, the CG agglomeration causes stress concentrations, creating weak points in the composites [42]. Moreover, Obasi [44] suggested that the decrease in mechanical properties of composites was also related to the polarity difference between the filler and polymer matrix and their poor interfacial adhesion.
SEM results show that the fracture morphologies of PBAT/CG-TA and PBAT/CG-OH composites exhibited smaller-sized CG particles than PBAT/CG composites; however, exposed CG particles were still observed. Furthermore, CG-OH-TA significantly enhanced the dispersion and mechanical adhesion of CG in the PBAT matrix. Additionally, no evidence of pull-out or separation of CG particles was observed in fracture morphology of the PBAT/CG-OH-TA composite.

Rheological Properties
The rheological properties of PBAT and its composites were analyzed to confirm the adhesion of CGs in the PBAT matrix. All samples were in a completely molten state at 160 • C. The complex viscosity (η*) of PBAT appeared to plateau at low and medium frequencies, while it exhibited shear thinning with increasing frequency. This phenomenon was more significant for the composite samples. Furthermore, the incorporation of CGs would hinder the movement of PBAT molecular chains [45]. Compared to the PBAT/CG composite, the η* of modified-CG composites slightly increased, which shows that the modification enhanced the adhesion of CGs and PBAT. As shown in Figure 4, CGs have a reinforcing effect in the PBAT matrix, as the η*, storage modulus (G ), and loss modulus (G ) of PBAT/CG composites were higher than those of neat PBAT [46]. Under the action of TA, the interfacial adhesion between CGs and PBAT is due to the hydrogen bonds and the interfacial adhesion with weak interaction forces. Therefore, the η*, G , and G of modified composites did not significantly increase in rheological curves.
°C. The complex viscosity (η*) of PBAT appeared to plateau at low and medium frequencies, while it exhibited shear thinning with increasing frequency. This phenomenon was more significant for the composite samples. Furthermore, the incorporation of CGs would hinder the movement of PBAT molecular chains [45]. Compared to the PBAT/CG composite, the η* of modified-CG composites slightly increased, which shows that the modification enhanced the adhesion of CGs and PBAT. As shown in Figure 4, CGs have a reinforcing effect in the PBAT matrix, as the η*, storage modulus (G′), and loss modulus (G″) of PBAT/CG composites were higher than those of neat PBAT [46]. Under the action of TA, the interfacial adhesion between CGs and PBAT is due to the hydrogen bonds and the interfacial adhesion with weak interaction forces. Therefore, the η*, G′, and G″ of modified composites did not significantly increase in rheological curves.  Figure 5 shows the DSC results of PBAT and PBAT/CG composites, and Table 2 summarizes the specific thermal property data. The heating curve of neat PBAT displayed an onset melting temperature of 97.4 °C and a low crystallinity of 8.3%. The incorporation of CGs did not induce any obvious effect on the melting temperature (Tm) of PBAT, while the crystallization temperature (Tc) shifted from 55.6 °C for neat PBAT to higher temperatures (74-76 °C). This shift suggests that CGs may be acting as nucleating agents [47]. The crystallinity of PBAT/CG composites was slightly lower compared to neat PBAT, which may be explained by the enhanced adhesion between modified CGs and PBAT, which hinders the mobility of PBAT chains.   Figure 5 shows the DSC results of PBAT and PBAT/CG composites, and Table 2 summarizes the specific thermal property data. The heating curve of neat PBAT displayed an onset melting temperature of 97.4 • C and a low crystallinity of 8.3%. The incorporation of CGs did not induce any obvious effect on the melting temperature (T m ) of PBAT, while the crystallization temperature (T c ) shifted from 55.6 • C for neat PBAT to higher temperatures (74-76 • C). This shift suggests that CGs may be acting as nucleating agents [47]. The crystallinity of PBAT/CG composites was slightly lower compared to neat PBAT, which may be explained by the enhanced adhesion between modified CGs and PBAT, which hinders the mobility of PBAT chains. more significant for the composite samples. Furthermore, the incorporation of CGs would hinder the movement of PBAT molecular chains [45]. Compared to the PBAT/CG composite, the η* of modified-CG composites slightly increased, which shows that the modification enhanced the adhesion of CGs and PBAT. As shown in Figure 4, CGs have a reinforcing effect in the PBAT matrix, as the η*, storage modulus (G′), and loss modulus (G″) of PBAT/CG composites were higher than those of neat PBAT [46]. Under the action of TA, the interfacial adhesion between CGs and PBAT is due to the hydrogen bonds and the interfacial adhesion with weak interaction forces. Therefore, the η*, G′, and G″ of modified composites did not significantly increase in rheological curves.  Figure 5 shows the DSC results of PBAT and PBAT/CG composites, and Table 2 summarizes the specific thermal property data. The heating curve of neat PBAT displayed an onset melting temperature of 97.4 °C and a low crystallinity of 8.3%. The incorporation of CGs did not induce any obvious effect on the melting temperature (Tm) of PBAT, while the crystallization temperature (Tc) shifted from 55.6 °C for neat PBAT to higher temperatures (74-76 °C). This shift suggests that CGs may be acting as nucleating agents [47]. The crystallinity of PBAT/CG composites was slightly lower compared to neat PBAT, which may be explained by the enhanced adhesion between modified CGs and PBAT, which hinders the mobility of PBAT chains.

Tensile Properties
The tensile properties of neat PBAT and PBAT composites are shown in Figure 6. PBAT has excellent tensile properties with good tensile strength (~25.9 MPa) and an excellent elongation at break (~867.1%). In contrast, the direct incorporation of 30 wt% CGs into PBAT would reduce the overall tensile properties of the prepared composites. Indeed, the tensile strength and elongation at the break of the PBAT/CG composite decreased to 7.1 MPa and 331.2%, respectively. This may be explained by the agglomeration of CGs, the weak interfacial adhesion of CG in PBAT matrix, and reductions in the continuous region of PBAT, as shown in SEM images. The tensile properties of neat PBAT and PBAT composites are shown in F PBAT has excellent tensile properties with good tensile strength (~25.9 MPa) and a lent elongation at break (~867.1%). In contrast, the direct incorporation of 30 wt% C PBAT would reduce the overall tensile properties of the prepared composites. Ind tensile strength and elongation at the break of the PBAT/CG composite decrease MPa and 331.2%, respectively. This may be explained by the agglomeration of C weak interfacial adhesion of CG in PBAT matrix, and reductions in the continuou of PBAT, as shown in SEM images. Although the PBAT/modified CG composite exhibited worse tensile properti pared to neat PBAT, these properties were still considerably improved comp PBAT/CG composite. Firstly, the surface treatment weakened the intermolecular tion of CGs. Secondly, alkali treatment disrupted the interparticle adhesion of C promoted their migration [48]. In comparison with the PBAT/CG composite, the strength and elongation at break of PBAT/CG-OH-TA composite increased by 47. 53.6%, respectively. The alkali-treated CG surface favors the TA-Na + complexes dep and increases the CGs' surface roughness. Moreover, TA-Na + complexes can ac interface, interlocking pins to generate higher friction with CG and PBAT, which a tributed to the enhancement of interfacial adhesion [34]. Improvements in the tensile properties could be attributed to the improved sion of CGs and the enhanced physical adhesion between CG and PBAT because t tallinity of these samples was similar [46]. The improved interfacial adhesion pro better stress transfer from PBAT to CG, which results in a more compact fracture a ter tensile properties. The mechanisms of alkali treatment, TA modification, and al synergistic modification of CG are shown in Figure 7. Although the PBAT/modified CG composite exhibited worse tensile properties compared to neat PBAT, these properties were still considerably improved compared to PBAT/CG composite. Firstly, the surface treatment weakened the intermolecular interaction of CGs. Secondly, alkali treatment disrupted the interparticle adhesion of CGs and promoted their migration [48]. In comparison with the PBAT/CG composite, the tensile strength and elongation at break of PBAT/CG-OH-TA composite increased by 47.0% and 53.6%, respectively. The alkali-treated CG surface favors the TA-Na + complexes deposition and increases the CGs' surface roughness. Moreover, TA-Na + complexes can act as the interface, interlocking pins to generate higher friction with CG and PBAT, which also contributed to the enhancement of interfacial adhesion [34].
Improvements in the tensile properties could be attributed to the improved dispersion of CGs and the enhanced physical adhesion between CG and PBAT because the crystallinity of these samples was similar [46]. The improved interfacial adhesion provides a better stress transfer from PBAT to CG, which results in a more compact fracture and better tensile properties. The mechanisms of alkali treatment, TA modification, and alkali/TA synergistic modification of CG are shown in Figure 7.
Compared to neat PBAT, the PBAT/CG composites showed decreased tensile properties even with modified CG, which was maybe closely related to micro-sized CGs. Therefore, the effect of the CG particle size on the tensile properties of the PBAT/CG composites was also investigated. As shown in Figure 8, the CG particle size was reduced by a mechanical crusher, sieved, and finally, it acquired three types of CGs with average particle sizes of 48.23, 28.19, and 21.25 µm under different mesh sieves of 60, 100, and 200 mesh. Compared to neat PBAT, the PBAT/CG composites showed decreased ten erties even with modified CG, which was maybe closely related to micro-s Therefore, the effect of the CG particle size on the tensile properties of the PBA posites was also investigated. As shown in Figure 8, the CG particle size was r a mechanical crusher, sieved, and finally, it acquired three types of CGs with av ticle sizes of 48.23, 28.19, and 21.25 µm under different mesh sieves of 60, 10 mesh. The effects of particle size on the tensile properties of PBAT composites ar Figure 9. Fixing the CG content at 30 wt%, the tensile strength and elongation a of the composites with CGs with an average size of 48.23 µm (using a 60-mesh s only 4.5 MPa and 94.3%. By decreasing the particle size of CGs, the tensile st composites with CGs with an average size of 28.19 µm and with CGs with an av of 21.25 µm were 7.1 MPa and 10.3 MPa, respectively, and their elongation at b 331.2% and 499.8%, respectively. The results confirmed that smaller CG particle tribute to the tensile properties of PBAT composites. The smaller the particle s  Compared to neat PBAT, the PBAT/CG composites showed decreased tensile properties even with modified CG, which was maybe closely related to micro-sized CGs. Therefore, the effect of the CG particle size on the tensile properties of the PBAT/CG composites was also investigated. As shown in Figure 8, the CG particle size was reduced by a mechanical crusher, sieved, and finally, it acquired three types of CGs with average particle sizes of 48.23, 28.19, and 21.25 µm under different mesh sieves of 60, 100, and 200 mesh. The effects of particle size on the tensile properties of PBAT composites are shown in Figure 9. Fixing the CG content at 30 wt%, the tensile strength and elongation at the break of the composites with CGs with an average size of 48.23 µm (using a 60-mesh sieve) were only 4.5 MPa and 94.3%. By decreasing the particle size of CGs, the tensile strengths of composites with CGs with an average size of 28.19 µm and with CGs with an average size of 21.25 µm were 7.1 MPa and 10.3 MPa, respectively, and their elongation at break were 331.2% and 499.8%, respectively. The results confirmed that smaller CG particle sizes contribute to the tensile properties of PBAT composites. The smaller the particle size of CGs, the larger the specific surface area that facilitates effective stress transfer. In this study, CG particles processed by a 100-mesh sieve were used, considering their easier processability and yield. To further reduce the CG particle size, it would be difficult to achieve using current mechanical crusher. According to the literature, ball milling [49] and steam blasting [50] could more effectively reduce the filler size down to nano size, which may be beneficial for improving the mechanical properties of biomass composites in the future. The effects of particle size on the tensile properties of PBAT composites are shown in Figure 9. Fixing the CG content at 30 wt%, the tensile strength and elongation at the break of the composites with CGs with an average size of 48.23 µm (using a 60-mesh sieve) were only 4.5 MPa and 94.3%. By decreasing the particle size of CGs, the tensile strengths of composites with CGs with an average size of 28.19 µm and with CGs with an average size of 21.25 µm were 7.1 MPa and 10.3 MPa, respectively, and their elongation at break were 331.2% and 499.8%, respectively. The results confirmed that smaller CG particle sizes contribute to the tensile properties of PBAT composites. The smaller the particle size of CGs, the larger the specific surface area that facilitates effective stress transfer. In this study, CG particles processed by a 100-mesh sieve were used, considering their easier processability and yield. To further reduce the CG particle size, it would be difficult to achieve using current mechanical crusher. According to the literature, ball milling [49] and steam blasting [50] could more effectively reduce the filler size down to nano size, which may be beneficial for improving the mechanical properties of biomass composites in the future. Polymers 2023, 15, x FOR PEER REVIEW 10 of 14 Figure 9. Tensile properties of PBAT/CG composites with different CG particle sizes. Figure 10 compares the filler content and elongation at the break of the prepared PBAT/CG-OH-TA composites in this study with other PBAT-based composites [2,26,46,[51][52][53][54][55][56]. In most of those studies, the loading of the incorporated biomass affected the mechanical properties, especially at high filler contents. The tensile properties of the composites could be improved by filler modification, but the used method often affects the elongation at break. However, in our study, compared to PBAT/CG composites, the obtained PBAT/CG-OH-TA composites here improved the tensile strength as well as the elongation at break. This was probably due to the increase in surface roughness of CG modified by TA, which facilitated CG distribution and the mechanical interlocking between CG and PBAT. In conclusion, the modification method used in this study preserved the excellent toughness of PBAT and provided a valuable reference for biomass-based fillers for the preparation of composites.  Figure 11 shows the thermal stability results of PBAT and corresponding PBAT/CG composites using TGA. Thermal parameters such as T5%, Td-max, and residual mass at 600 °C are concluded in Table 3. The presence of terephthalic moieties of PBAT molecular chains enables it to have a better thermal stability [2] and to decompose in a narrow temperature range. Indeed, the PBAT decomposition started at about 330 °C and was almost   [2,26,46,[51][52][53][54][55][56].

Thermal Stability
In most of those studies, the loading of the incorporated biomass affected the mechanical properties, especially at high filler contents. The tensile properties of the composites could be improved by filler modification, but the used method often affects the elongation at break. However, in our study, compared to PBAT/CG composites, the obtained PBAT/CG-OH-TA composites here improved the tensile strength as well as the elongation at break. This was probably due to the increase in surface roughness of CG modified by TA, which facilitated CG distribution and the mechanical interlocking between CG and PBAT. In conclusion, the modification method used in this study preserved the excellent toughness of PBAT and provided a valuable reference for biomass-based fillers for the preparation of composites.
Polymers 2023, 15, x FOR PEER REVIEW 10 of 14 Figure 9. Tensile properties of PBAT/CG composites with different CG particle sizes. Figure 10 compares the filler content and elongation at the break of the prepared PBAT/CG-OH-TA composites in this study with other PBAT-based composites [2,26,46,[51][52][53][54][55][56]. In most of those studies, the loading of the incorporated biomass affected the mechanical properties, especially at high filler contents. The tensile properties of the composites could be improved by filler modification, but the used method often affects the elongation at break. However, in our study, compared to PBAT/CG composites, the obtained PBAT/CG-OH-TA composites here improved the tensile strength as well as the elongation at break. This was probably due to the increase in surface roughness of CG modified by TA, which facilitated CG distribution and the mechanical interlocking between CG and PBAT. In conclusion, the modification method used in this study preserved the excellent toughness of PBAT and provided a valuable reference for biomass-based fillers for the preparation of composites.  Figure 11 shows the thermal stability results of PBAT and corresponding PBAT/CG composites using TGA. Thermal parameters such as T5%, Td-max, and residual mass at 600 °C are concluded in Table 3. The presence of terephthalic moieties of PBAT molecular chains enables it to have a better thermal stability [2] and to decompose in a narrow temperature range. Indeed, the PBAT decomposition started at about 330 °C and was almost  Figure 11 shows the thermal stability results of PBAT and corresponding PBAT/CG composites using TGA. Thermal parameters such as T 5% , T d-max , and residual mass at 600 • C are concluded in Table 3. The presence of terephthalic moieties of PBAT molecular chains enables it to have a better thermal stability [2] and to decompose in a narrow temperature range. Indeed, the PBAT decomposition started at about 330 • C and was almost completed at about 490 • C. Nearly 90% of the mass loss occurred between 350 • C and 430 • C. Moreover, the maximum decomposition rate occurred at 402 • C, corresponding to a mass loss rate of 21.3%·min −1 . All PBAT/CG composites had similar pyrolysis curves, and their thermal degradation could be divided into two steps. The first step was related to the degradation of hemicellulose (~200 • C) within CGs [22], and the second step was the same as the thermal degradation of PBAT. Therefore, the incorporation of CGs decreased the T 5% and T d-max ; however, this would not affect the melt processing, since the processing temperature of PBAT is often under 200 • C. completed at about 490 °C. Nearly 90% of the mass loss occurred between 350 °C and 430 °C. Moreover, the maximum decomposition rate occurred at 402 °C, corresponding to a mass loss rate of 21.3%·min −1 . All PBAT/CG composites had similar pyrolysis curves, and their thermal degradation could be divided into two steps. The first step was related to the degradation of hemicellulose (~200 °C) within CGs [22], and the second step was the same as the thermal degradation of PBAT. Therefore, the incorporation of CGs decreased the T5% and Td-max; however, this would not affect the melt processing, since the processing temperature of PBAT is often under 200 °C.

Conclusions
In conclusion, this study introduced a novel and green method to improve the interface adhesion between CG and PBAT. Under mild conditions, TA deposited on the CG surface by forming complexes with the metal ion (Na + ), which increased the surface's wettability and roughness. This was confirmed using multiple techniques such as FTIR, water contact angle, SEM, and EDS. Fracture morphology and rheological property results of the prepared composites indicated that the adhesion of modified CG to PBAT was enhanced. Moreover, compared to the PBAT/CG composite, the tensile strength and elongation at the break of the PBAT/CG-OH-TA composite were enhanced by 47.0% and 53.6%, respectively. Furthermore, the addition of CGs slightly decreased the thermal stability of PBAT composites; however, this did not affect the melt processing of PBAT, which often occurred under 200 °C. This approach could provide a new method for the effective use of biomass waste as fillers, which could reduce the cost of polymer-based products by adding a large amount of biomass waste, particularly into relatively expensive biodegradable polymers.

Conclusions
In conclusion, this study introduced a novel and green method to improve the interface adhesion between CG and PBAT. Under mild conditions, TA deposited on the CG surface by forming complexes with the metal ion (Na + ), which increased the surface's wettability and roughness. This was confirmed using multiple techniques such as FTIR, water contact angle, SEM, and EDS. Fracture morphology and rheological property results of the prepared composites indicated that the adhesion of modified CG to PBAT was enhanced. Moreover, compared to the PBAT/CG composite, the tensile strength and elongation at the break of the PBAT/CG-OH-TA composite were enhanced by 47.0% and 53.6%, respectively. Furthermore, the addition of CGs slightly decreased the thermal stability of PBAT composites; however, this did not affect the melt processing of PBAT, which often occurred under 200 • C. This approach could provide a new method for the effective use of biomass waste as fillers, which could reduce the cost of polymer-based products by adding a large amount of biomass waste, particularly into relatively expensive biodegradable polymers.