Characterization of poly(lactic acid) biocomposites filled with chestnut shell waste

The commercial injection molding grade poly(lactic acid) (PLA) Ingeo™ 3001D with a melt flow rate (MFR) of 22 g/10 min (210 °C, 2.16 kg) supplied by Nature Works (USA) was used in our experiments.

Preliminary preparation of the chestnut shell waste filler (CN, Aesculus hippocastanum L.) included rinsing in running distilled water and drying at 50 °C for 24 h. First disintegration was processed in a low-speed mill cutter Shini SC-1411 and then milled in high-speed mill Retsch GM 200 (n = 2000 rpm). Application of the two-step milling allows the preparation of non-degraded natural filler for PLA-based biocomposites. The obtained waste chestnut shell powder was sieved by vibratory sieve shaker ANALYSETTE 3 Pro equipped with 200-µm-mesh size sieve. Characterization of particle size distribution was evaluated using laser particle sizer Fritsch ANALYSETTE 22 apparatus operated in the range of 0.08–2000 µm.

PLA pellets and CN powder were premixed using a high-speed rotary mixer Retsch GM200 (t = 3 min, n = 1000 rpm) with different amounts of filler (2.5, 5, 10, 20, and 30 wt%). Next, all blends were mixed in a molten state using a ZAMAK 16/40 EDH twin screw co-rotating extruder that operated at 190 °C and 100 rpm, and pelletized after cooling in a water bath. The normalized specimens for tensile and impact strength test were prepared with a Engel HS 80/20 HLS injection molding machine operated at 190 °C. Injection molding process was realized with the following parameters: mold temperature T _mould = 25 °C, injection speed V = 70 mm/s, forming pressure P _f  = 5 MPa and cooling time t = 45 s. Materials were every time dried in a vacuum at 50 °C for 24 h before being mixed in molten state. The samples were described as PLA, 2.5CN, 5CN, 10CN, 20CN, and 30CN, adequately to incorporated amount of chestnut shell waste powder.

In Fig. 1, particle size distribution function (Q3(x)) and its derivative (dQ3(x)) as a function of particle size (x) were presented. The arithmetic mean size of hybrid filler fraction after milling was equal to 10.71 µm. Additionally, characterization of organic filler particle size and shape was performed by means of scanning electron microscopy (SEM) (Fig. 2).

In Figs. 3 and 4, plots of weight loss vs. temperature and their derivatives (DTG) are presented. Temperatures recorded at two characteristic mass loss values (5 and 50%) were assigned as T ₅ and T ₅₀. The values of T ₅ and T ₅₀ for chestnut filler determined in nitrogen atmosphere were 202 and 357 °C, and in air atmosphere they were 169 and 302 °C, respectively. Peak temperature of DTG curves and maximum rate of degradation filler were 307 °C and 5.1%/min in inert atmosphere and 283 °C and 5.7%/min in oxygen atmosphere. The ΔW% was defined as 36.7% in nitrogen and 5.8% in air atmosphere. The thermal stability of chestnut filler was sufficient to proceed with melt processing.

The FT-IR spectra of ground chestnut shell, PLA, and composite containing 30 wt% of CN are presented in Fig. 5. In CN spectra the broad absorption band at 3000–3500 cm⁻¹ indicates bonded hydroxyl (–OH) groups existing in the natural filler [29]. Peak observed at 2900 cm⁻¹ corresponds to C–H stretching band originating from aliphatic moieties in cellulose and hemicellulose [30]. The peak at 1720 cm⁻¹ in CN spectra is related to C=O stretching band from ester linkage of carboxylic group in the lignin ferulic and p-coumaric acids [31]. The observed peak at 1606 cm⁻¹ corresponds to C=C aromatic stretching band and confirms the presence of lignin in the natural filler. Double absorption peak at wavenumbers of 1093 and 1068 cm⁻¹ is assigned to C–O stretching band from ether and alcoholic group [31, 32]. In case of PLA, absorption peaks at wavenumbers: 3510, 2995, 2950, 1749, 1450, 1362, and 1053 cm⁻¹ are related to C–O–O–H stretching, C–H aliphatic stretching, C–H aliphatic stretching (doublet), C=O bending, –CH₂ bending, C–H bending, and C–O stretching, respectively [33, 34]. The absorption bands at 1207 and 920 cm⁻¹ are associated with allyl-ketone chain vibration and flexural C–H bond vibration and are representatives of the PLA crystalline structure [35, 36]. In contrast to pure PLA, the 30CN composite sample exhibits distinct peaks at mentioned wavelengths which correspond to crystalline structure. This is due to a nucleating ability of ground chestnut shells incorporated into PLA. These results are in a good agreement with crystallinity degree values evaluated within DSC experiments, which will be discussed later. The poly(lactic acid)-based composites containing various amount of ground chestnut waste were prepared without coupling agents and preliminary chemical treatment of the natural filler. Therefore, no chemical interactions and additional peaks at FT-IR spectra of composite containing 30 wt% of CN were observed. The 30CN FT-IR spectrum presents peaks originating from both PLA and chestnut natural filler. Lowered absorbance at the wavelengths corresponding to poly(lactic acid) in case of the 30CN sample is attributed to lower amount of polymer in composite. What is important, despite the presence of 30 wt% amount of the natural filler with high ability for moisture and water adsorption, no significant effects of hydrolytic degradation of PLA in the composite during processing was observed. There was no shift of the peak at 1081 cm⁻¹, which may suggest the presence of hydrolysis due to hydrogen bonding or nucleophilic attack by water molecules at –C–O– ester bonds. Shift of the 1749 cm⁻¹ peak, corresponding to carbonyl group C=O band ascribed to ester bonds, was also negligible [37]. Therefore, it can be stated that it is not necessary to apply special methods of material drying or additional preprocessing of the composite pellets before melt processing, in comparison to pure PLA.

DSC curves of PLA and PLA–CN composites obtained during second heating process are presented in Fig. 6. Three transformations comprised by DSC data were observed and analyzed: enthalpy relaxation corresponding to glass transition (about 58 °C), cold crystallization (near 106 °C), and melting of polymer matrix (168 °C). Additionally, DSC parameters, such as glass transition temperature (T _g1), cold crystallization temperature (T _c), melting temperature (T _m), melting enthalpy (ΔH _m), and crystallinity level (X _c), are listed in Table 1. The crystallinity degree of pure PLA and PLA–CN composites was calculated using the following equation:

where ΔH _o is the melting heat of entirely crystallized PLA and its value is equal to 106 J/g [38], and Φ is the amount of filler. On the bases of the data presented, it is clearly visible that the incorporation of chestnut shell waste powder does not significantly influence glass transition temperature determined by DSC. Only in case of composite samples containing the highest amount of the filler small decrease of T _gl value was observed. Pure poly(lactic acid) reveals low crystallization ability; therefore, macromolecular chains are unable to arrange completely during rapid cooling. In effect, during PLA heating process, quenched polymer chains rearrange forming regions with higher crystallinity. What is more, cold crystallization temperature peak increases in comparison to pure PLA, and achieves maximum value of 106.7 °C for 5 wt% of CN. At the same time recorded T _c values for the highest filler concentration, i.e., 30 wt%, lowered and revealed value 104.4 °C. The lowered T _c for 30CN sample may have resulted from melt crystallization enhancement of PLA. High amounts of organic fillers with developed surface act as nucleating agents which promote regular arrangement of polymer chains and increase crystal growth during PLA-based composite cooling [10]. The heat of fusion values obtained during melting of PLA and PLA–CN composites remained at similar level. On the other hand, crystallinity level of PLA–CN composites calculated in accordance with Eq. 1 was significantly increased in comparison to pure PLA. For pure PLA, crystallinity level of 41.6% was measured; hence, for the highest CN content (30 wt% CN) X _c was equal to 49.2%. This phenomenon confirms nucleation ability of the incorporated organic filler particle. This is in good agreement with the work of Ouchiar et al. who proved that the use of adequate fillers may be a route to tuning the nanostructure of a polylactide [39]. Moreover, it can be stated that DSC results confirm FT-IR analysis, which reveals the presence of higher amount of crystalline phase in composite containing 30 wt% of the filler in comparison to pure PLA.

The results of mechanical properties obtained in the course of static tensile test and impact strength measurements are presented in Fig. 7. It may be seen that the incorporation of organic waste filler highly influenced the overall mechanical properties of PLA-based composites. The increasing content of ground chestnut shell led to gradual increase of the stiffness with simultaneous reduction of the composites’ tensile strength. Pure PLA samples showed elasticity modulus value of approx. 3.1 GPa. For lower amounts of CN (up to 5 wt%), the samples’ elasticity modulus values were similar to those recorded for pure PLA, and demonstrated average values of 3.1 GPa and 3.2 GPa, corresponding to 2.5CN and 5CN samples. Composites with higher filler concentrations, i.e., 10 wt% and more, showed stronger tendency to stiffness improvement, achieving the maximum of 30 wt% CN concentration, with E value of 3.8 GPa. According to the literature exploring this field, the Young’s modulus may increase with an increasing level of crystallinity of the composite samples [40]. Opposite tendency may be observed when analyzing the influence of filler loading on the tensile strength of PLA composites. This phenomenon is characteristic of composite materials modified by particle-shaped fillers with size above 1 µm [41]. As this research proceeded, the tensile strength decreased from 69 MPa (pure PLA) to 40.8 MPa for composite filled with 30 wt% of CN. This effect may be attributed to weak bonding between the hydrophilic natural filler and hydrophobic polymer matrix. Moreover, the application of unmodified particle-shaped filler with broad particle size distribution may act as supplement crack indentation point [41]. The only exception was observed for composite samples containing 2.5 wt% of CN, where tensile strength was not influenced by the presence of the CN filler. Considering elongation values recorded at samples break, behavior which is typical for filled polymeric materials was noticed. Average elongation at break value determined for pure PLA, equal to 4.8%, decreased to 1.5% for highest CN content. What is more, increasing content of the CN filler caused reduction of the elongation at break inversely proportional to increased brittleness of the PLA–CN composite samples determined by Dynstat impact test [42]. More than 50% reduction of impact strength of PLA–CN composites was observed in comparison to pure PLA samples. To conclude, incorporation of natural powder filler with broad particle size distribution, without preliminary chemical treatment, causes strong increase of the composite material stiffness and brittleness.

The dynamic mechanical thermal analysis was used to determine the influence of the filler content on the viscoelastic behavior of the composites, described by dynamic and damping properties. The storage modulus (G′) represents the elastic portion of viscoelastic behavior connected with solid-state behavior of sample, whereas loss factor (tanδ) is a very useful parameter for detecting molecular transitions of the polymers and allows to relate DMTA results to interactions between components of polymer composites and blends [43, 44]. Plots of the storage modulus and loss factor versus temperature T for all samples are shown in Figs. 8 and 9. The values of storage modulus at room temperature as well as detailed information about glass transition temperature of the composites are given in Table 2. The highest value of G′ (2340 MPa) was reported for the sample with the highest amount of ground chestnut shell in comparison with the reference sample PLA (1430 MPa). This phenomenon is in agreement with observed significant increase of Young’s modulus determined by tensile test and suggests that particle-shaped filler have strong influence on the elastic properties of composite. Hence, the tanδ peak intensity for the composites decreases with an increase of CN content and this may confirm that the introduction of this filler hinders mobility of the PLA chains [40, 45]. Incorporation of particle-shaped filler into polymer matrix causes increase of the elasticity and decreases ability for mechanical energy dissipation; therefore, less energy will be used to overcome the frictional forces between molecular chains as to decrease mechanical loss [46]. This fact may be also attributed to increased crystallinity level of the composites in comparison to pure PLA. The addition of chestnut shell led to the slight lowering of the composite glass transition temperature (T _g), from 70 °C for pure PLA samples down to 68 °C for composites containing 30 wt% of ground chestnut shell. Decrease of the T _g may be attributed to insufficient interfacial adhesion between polymer and filler [47]. These results are also in good agreement with DSC investigations where similar tendency was observed. The glass transition temperature determines the range of polymer material usage. It is noteworthy that the fact that there are no significant changes in T _g values is favorable for application of composite product in standard room temperature usage.

The SEM micrographs of PLA and PLA–CN composites containing from 2.5 up to 30 wt% of the natural filler, made with magnification 1000×, are presented in Fig. 10. The observed dispersion of the organic filler in polymeric matrix was sufficient. Particles with various sizes were uniformly dispersed in polymeric matrix, which may be also confirmed by the low values of standard deviations of mechanical parameters determined by uniaxial tensile test. However, for 30CN composite, pullout cavities around bigger sized chestnut particles were observed. To determine the adhesion between polymer matrix and the filler, additional SEM microphotograph is presented in Fig. 11. It can be noticed that the saturation of the 20-µm chestnut particle by poly(lactic acid) was not significant. Most of the particles observed in Fig. 10 were not fully pulled out from the bulk but broken on the fracture surfaces; however, in Fig. 11, cavities between filler and polymeric matrix can be clearly seen. Despite both these phenomena occurring simultaneously, dominant effect on the mechanical properties of PLA-based composites may be caused by insufficient adhesion between polymer and filler. Analysis results of SEM images are in agreement with described significant increase of the brittleness of the composites (increased stiffness with general deterioration of remaining mechanical properties of the composites) in comparison to pure PLA, as well as lowered glass transition temperature determined by DSC and DMTA.