Sage Journals: Discover world-class research

Abstract

Natural fibre-reinforced polymer composites have gained significant attention as sustainable alternatives to conventional synthetic materials. This study investigates the mechanical properties enhancement of polypropylene (PP) composites through hybrid reinforcement using oil palm empty fruit bunch (EFB) fibres and carbon nanospheres (CNS) derived from sago bark waste. The CNS was synthesized via pyrolysis at 500°C under nitrogen atmosphere and characterized using TEM, XRD and FTIR which reveals spherical nanoparticles with diameters ranging from 15 to 50 nm and partially crystalline structure. PP composites were fabricated with varying EFB fibre content (0–50 wt%) and CNS loading (0–1.5 wt%) using twin-screw extrusion followed by injection molding. FTIR analysis confirmed successful interfacial interactions between PP matrix, EFB fibres, and CNS through peak shifts and intensity modifications. Thermal analysis demonstrated improved thermal stability with 20°C higher degradation onset temperature for the nanocomposite. Mechanical testing revealed that the optimal composition of 20 wt% EFB fibre with 1 wt% CNS achieved remarkable property enhancements compared to neat PP/EFB composite in which the tensile strength increased by 17%, flexural strength improved by 20% and impact strength enhanced by 32%. An artificial neural network (ANN) model was developed to predict mechanical properties and achieve excellent correlation (R = 0.985) between predicted and experimental values, enabling efficient optimization of composite formulations. This research demonstrates a viable approach for converting agricultural waste into high-performance composites with enhanced mechanical properties suitable for automotive components and consumer product applications.

Keywords

Empty fruit bunch fibre composite carbon nanosphere mechanical properties

Introduction

The increasing global demand for environmentally sustainable materials has driven significant research interest toward polymer composites reinforced with natural fibres. This shift from conventional synthetic reinforcements to bio-based alternatives is motivated by growing environmental concerns, stricter regulations on non-biodegradable waste and the need to reduce carbon footprints across manufacturing sectors.¹ Polypropylene (PP) has emerged as a preferred thermoplastic matrix due to its excellent balance of properties including good chemical resistance, low density, high thermal stability, and cost-effectiveness. PP-based composites find extensive applications in automotive components, consumer products, packaging and construction materials that making improvements in their mechanical properties particularly valuable for industrial applications.²

In the pursuit of sustainable composite materials, natural fibres derived from agricultural waste streams have become increasingly attractive for reinforcing polymeric matrices like PP. These natural reinforcements offer distinctive advantages including abundance, renewability, biodegradability and low cost compared to traditional synthetic fibres.³ Oil palm empty fruit bunch (EFB) fibres, a major by-product of palm oil processing has garnered particular attention as reinforcing agents in this context. With global palm oil production exceeding 73 million metric tons annually and Malaysia and Indonesia accounting for approximately 85% of this production, enormous quantities of EFB waste (approximately 20–22% of fresh fruit bunch weight) are generated and creating significant disposal challenges.⁴ These lignocellulosic fibres possess favorable characteristics including low density, reasonable tensile properties, high specific strength and natural abundance of hydroxyl groups that facilitate surface modifications. However, despite these advantages, the hydrophilic nature of EFB fibres often results in poor compatibility with hydrophobic polymer matrices like PP which leading to inadequate interfacial adhesion and subsequently compromised mechanical properties in the resulting composites.⁵

The challenge of interfacial incompatibility between natural fibres and polymer matrices has prompted researchers to investigate various enhancement strategies for these composites. To address this fundamental limitation, approaches including chemical treatments, coupling agents and more recently the incorporation of nanofillers have been extensively explored.^6,7 Among emerging nanomaterials, carbon nanospheres (CNS) have shown promising potential as interfacial modifiers and secondary reinforcements in polymer composites. These nanomaterials characterized by their spherical morphology with diameters typically ranging from 10 to 500 nm, possess exceptional properties including high surface area-to-volume ratio, excellent mechanical strength, good thermal stability and electrical conductivity.^8,9 Furthermore, CNS can be synthesized from renewable biomass sources through pyrolysis processes, aligning with sustainability objectives while potentially enhancing the mechanical, thermal and functional properties of the resulting composites. The unique spherical geometry of CNS presents a significant advantage over other carbon nanomaterials like nanotubes or nanoplatelets as it allows for reduced agglomeration tendency and enables better dispersion within polymer matrices even at low loading levels.¹⁰

The strategic incorporation of CNS into natural fibre-reinforced polymer composites represents a synergistic approach to advanced material design that addresses multiple challenges simultaneously. These nanospheres can potentially serve as bridging elements at the interface between hydrophilic natural fibres and hydrophobic polymer matrices, thereby enhancing stress transfer mechanisms and interfacial adhesion crucial for superior mechanical performance. Additionally, when properly dispersed, CNS may effectively fill micro-voids within the composite structure, reducing stress concentration points and providing crack-bridging effects that improve mechanical properties and fracture resistance. This multifunctional reinforcement mechanism is particularly valuable in natural fibre composites where interfacial weaknesses often limit performance capabilities.¹¹ Previous studies have demonstrated that carbon-based nanofillers can significantly enhance the mechanical properties of polymer composites at relatively low loading levels with improvements attributed to their ability to restrict polymer chain mobility and act as nucleating agents for crystallization as well as create three-dimensional reinforcing networks within the matrix.¹² Furthermore, the functionalization of carbon nanomaterials through various treatments can introduce reactive groups that further promote chemical bonding with both natural fibres and polymer matrices and resulting in more robust interfacial interactions and superior load transfer efficiency.

Despite the growing interest in natural fibre-polymer composites, comprehensive studies investigating the synergistic effects of oil palm EFB fibres and biomass-derived carbon nanospheres in PP matrix systems remain limited. Most existing research has focused on individual reinforcement strategies without exploring the potential of hybrid reinforcement systems that combine natural fibres with carbon nanofillers. This study addresses this research gap by systematically investigating the influence of varying EFB fibre content (0–50 wt%) and CNS loading (0–1.5 wt%) on the mechanical and thermal properties of PP composites. The CNS were synthesized from sago bark waste through controlled pyrolysis that provides an additional pathway for agricultural waste valorization while creating value-added nanomaterials. Additionally, an artificial neural network (ANN) model was developed and validated to predict the mechanical properties of the composites which facilitates optimization of formulation parameters without extensive experimental work. The research aims to establish optimal composition ranges for maximum mechanical performance, elucidate structure-properties relationships in these hybrid nanocomposites and demonstrate the potential for developing sustainable, high-performance materials suitable for automotive and consumer product applications.

Methodology

Material

The study utilized Oil Palm Empty Fruit Bunch (EFB) fibres with measuring of 1–3 mm in length and with a density of 1.07 g/cm³ sourced from Kilang Sawit Panching in Kuantan, Malaysia. Additionally, chemicals such as sodium hydroxide and acetic acid were procured from Ize Solution Sdn. Bhd.

Preparation and characterization of carbon nanosphere (CNS)

The sago bark used in this study was sourced from a sago palm plantation in Malaysia which is known for its large-scale production of sago-based products and significant generation of sago bark waste. The fibrous residues were isolated and oven-dried at 110°C for 48 hours to ensure complete moisture removal. The dried bark was then crushed and pulverized at 12,000 rpm using a grinder (NL Scientific 1009 X, Malaysia). Subsequently, the grounded material was sieved to achieve a uniform particle size of approximately 60–70 μm. The processed sago bark underwent pyrolysis in a tube furnace (Nabertherm R 120/500/13, Germany) under a nitrogen atmosphere at flow rate of 150 cm³/min. The temperature was raised to 500°C at a heating rate of 5°C/min for 2 hours and then gradually cooled to room temperature under N² to yield the pyrolyzed material. To purify the product, the pyrolyzed carbon was washed with 1 M hydrochloric acid (HCl) followed by deionized water which resulting in the final carbon nanospheres (CNS). The surface morphology of CNS was investigated using Transmission Electron Microscopy (TEM). The analysis was performed on a Phillips Tecnai G220 microscope. Crystallinity analysis was carried out via X-ray Diffraction (XRD) on a PANalytical X’pert3 Powder diffractometer operating at 40 kV and 30 mA with scans recorded from 10° to 80° two-theta at a rate of 8.2551° per minute. Chemical bonding characterization was conducted through Fourier Transform Infrared Spectroscopy (FTIR) using a Thermo Scientific Nicolet 6700 instrument (Germany) with the standard KBr pellet method scanning from 4000 to 700 cm⁻¹. Comparative FTIR analysis between CNS, Raw EFB fibres, and PP/EFB/CNS composites was performed to investigate potential interfacial interactions.

Composite fabrication and characterization

The oil palm empty fruit bunch (EFB) fibres were initially washed with deionized (DI) water to remove impurities, followed by alkali treatment using 1 wt% sodium hydroxide (NaOH) at a fibre-to-water ratio of 1:20. The treatment was conducted under ultrasonication for 120 minutes at room temperature to enhance fibre-matrix adhesion. After treatment, the fibres were thoroughly rinsed with DI water until a neutral pH (∼7) was achieved followed by neutralization with diluted acetic acid to remove residual alkali to produce treated EFB fibres (TEFB). For composite preparation, PP and TEFB were blended in 500 g batches at varying fibre loading (0-50 wt%) and compounded using a co-rotating twin-screw extruder (TSE-25). The extrusion was performed at a screw speed of 80 rpm, with the barrel temperature maintained at 180°C throughout the process. In composites incorporating CNS, 1 wt% CNS was introduced during extrusion to ensure homogeneous dispersion. The extruded strands were pelletized and subsequently dried at 80°C for 24 hours to eliminate moisture. The dried pellets were stored in sealed plastic bags to prevent moisture absorption prior to further processing. The composite pellets were then injection-molded into dumbbell-shaped (ASTM D638) and rectangular bar specimens (ASTM D790 and EN ISO 179) using an injection molding machine. The processing parameters using a melt temperature of 180°C, an injection pressure of 100–200 bar and a mold temperature of 40°C. Mechanical characterization involved tensile testing according to ASTM 638-08. A Shimadzu Universal testing machine was used at a cross-head speed of 10 mm/min to determine tensile strength. Flexural testing was conducted following ASTM D790, using the same Shimadzu machine. The specimens were tested at a cross-head speed of 2 mm/min with a span-to-depth ratio of 16:1 to determine flexural strength. Impact testing followed EN ISO 179, using a Ray-Ran Pendulum Charpy Impact System. Each type of sample was tested five times. One-way analysis of variance (ANOVA) was applied to mechanical properties to determine the significance and relative importance of the main factors as per conducted by Ahmad, et al.¹³ The significance of each mean property value was determined at p < .05 confidence level.

Composite performance prediction using artificial neural network (ANN)

The artificial neural network (ANN) model was developed using MATLAB software to predict the mechanical properties of PP/TEFB/CNS composites. The neural network toolbox (nntool) in MATLAB was employed to construct and train the predictive model. The ANN architecture consisted of three layers including input layer, hidden layer, and output layer. The input layer contained three neurons corresponding to the three input parameters of PP content, EFB fibre ratio, and CNS concentration. The output layer comprised five neurons representing the target mechanical properties: tensile strength, flexural strength, impact strength, tensile modulus, and strain at break. The number of neurons in the hidden layer was determined using the 2n + 1 formula, where n represents the number of input parameters. The network architecture employed a feed-forward backpropagation (FFP) network type with Levenberg-Marquardt training algorithm (TRAINLM). The adaptation learning function was set to LEARNGDM which helps accelerate convergence and improve training stability. The performance function used was mean squared error (MSE) to evaluate the network prediction accuracy during training. The transfer function for all layers was set to hyperbolic tangent sigmoid (TANSIG). The data division strategy employed random division (dividerand) to ensure unbiased distribution of training, validation and testing datasets.

Results and discussion

Morphological characterization of carbon nanosphere (CNS)

Figure 1 shows the TEM image of CNS derived from sago bark by revealing spherical particles with diameters ranging from approximately 15-50 nm. The nanospheres exhibit varying degrees of dispersion across the field, with larger spheres (40–50 nm) maintaining reasonable separation while smaller particles (15–25 nm) show some clustering. The observed morphology displays well-defined boundaries and relatively smooth surfaces which is the characteristics typical of carbonaceous materials derived from biomass precursors. The heterogeneous size distribution observed may be attributed to the specific pyrolysis conditions employed during synthesis while still falling within the range of carbon nanospheres reported in similar studies.^14,15

Figure 1.

TEM image of CNS.

XRD analysis

Figure 2 displays the XRD pattern of carbon nanospheres (CNS) derived from sago bark which featuring two distinctive diffraction peaks at 17.81° and 22.26°. The peak at 17.81° corresponds to the (110) crystallographic plane while the more prominent peak at 22.26° is attributed to the (120) plane indicative of the microcrystalline cellulosic structure in the carbonaceous material. The observed broad diffraction peaks with gradually decreasing intensity beyond 30° suggest the presence of amorphous carbon structure in the synthesized nanospheres. This pattern aligns with typical XRD profiles of carbon materials derived from biomass precursors through pyrolysis methods.^16,17 The relative intensities of the peaks, with the 22.26° peak showing higher intensity compared to the 17.81° peak indicate the predominance of the (120) planes in the crystalline domains of the CNS. These structural characteristics confirm the successful transformation of the cellulosic components of sago bark into carbon nanospheres during the pyrolysis process and consistent with previous reports on biomass-derived carbon nanomaterials.¹⁸

Figure 2.

XRD pattern of sago bark derived CNS.

TGA analysis

Figure 3 shows the thermal degradation behavior of PP, TEFB, PP/TEFB, and PP/TEFB/CNS. As can be seen, PP reveals an onset at 350°C and complete decomposition by 450°C and leaving 3% residue which in line with characteristic of clean-burning thermoplastic polymers.^19,20 TEFB demonstrated single-stage degradation beginning at 250°C with 23% char residue at 500°C, showing one continuous weight loss curve despite its lignocellulosic nature containing multiple components. TEFB demonstrated single-stage degradation beginning at 250°C with 23% char residue at 500°C which showing one continuous weight loss curve characteristic of lignocellulosic materials where the degradation of multiple components such as moisture, hemicellulose, cellulose and lignin) occur.^21,22 The PP/TEFB composite exhibited multi-stage degradation with onset at 300°C and 7% residue as well as two distinct degradation steps. The first degradation step occur between 300 and 380°C corresponds to the thermal decomposition of the TEFB fibre component, while the second degradation step at 380 °C–450 °C represents the degradation of the PP matrix that demonstrate each component maintains its characteristic thermal decomposition behavior within the composite.²³ In case of PP/TEFB/CNS, the sample demonstrated enhanced thermal stability with degradation onset at 320°C which represents a 20°C improvement compared to PP/TEFB composite. In case of PP/TEFB/CNS, the sample exhibited thermal degradation onset at 320°C and complete decomposition by 480°C with thermal degradation occurring through the sequential decomposition of fibre components followed by polymer matrix degradation. The nanocomposite retained 5% ash content at 500°C which is lower than PP/TEFB composite that indicates modified degradation behavior due to CNS incorporation. The addition of CNS improved the thermal stability by 20°C compared to PP/TEFB composite probably due to multiple stabilization mechanisms including physical barrier effect that restricts volatile diffusion during degradation. It also enhanced thermal conductivity and provides uniform heat dissipation throughout the matrix and promotes protective char formation creating insulating layers as well as interfacial interactions between CNS particles and both polymer matrix and fibre components. Similar findings have shown that nanofillers helps in improving thermal stability of polymer composites through these synergistic mechanisms with carbon nanofillers acting as thermal stabilizers that delay degradation onset and modify decomposition pathways.^24–26

Figure 3.

TGA curves of various sample.

FTIR analysis

The FTIR spectra of TEFB, PP, CNS, and PP/TEFB/CNS are presented in Figure 4 which showing characteristic absorption peaks corresponding to various functional groups within the wavenumber range of 4000-400 cm⁻¹. The broad absorption peak observed around 3400-3200 cm⁻¹ in TEFB spectrum corresponds to O-H stretching vibrations from hydroxyl groups present primarily in cellulose and remaining hemicellulose components as the NaOH treatment removes most lignin content from the lignocellulosic fibre.^27–29 The peak at 2920 cm⁻¹ represents C-H stretching vibrations from aliphatic methyl and methylene groups which is prominent in both TEFB spectra.^30,31 The reduced intensity or absence of absorption bands at 1740 cm⁻¹ and 1600–1500 cm⁻¹ region in TEFB confirms the effective removal of lignin and acetyl groups through alkali treatment as these peaks typically correspond to C = O stretching from acetyl groups in hemicellulose and aromatic C = C stretching vibrations characteristic of lignin structure.^32,33 The absorption bands at 1370 cm⁻¹ represent C-H bending vibrations, while the peaks at 1240-1030 cm⁻¹ region correspond to C-O stretching vibrations from cellulose and remaining hemicellulose components.^27,34,35 PP spectrum shows characteristic peaks at 2850 cm⁻¹ for C-H stretching vibrations while peaks at 1460 and 1221 cm⁻¹ for C-H bending vibrations and absorption bands below 1000 cm⁻¹ corresponding to C-C stretching and skeletal vibrations of the polymer backbone.^30,36–38 CNS spectrum exhibits relatively fewer and weaker absorption peaks due to its carbon-based structure with some peaks in the fingerprint region below 1600 cm⁻¹ representing C-C and C = C vibrations.^39,40 PP/TEFB/CNS spectrum displays a combination of characteristic absorption peaks from all three components showing the broad O-H stretching at 3400 cm⁻¹ from TEFB, C-H stretching and bending vibrations from both PP and TEFB in the 2920-1380 cm⁻¹ region, C-O stretching peaks from TEFB in the 1240–1030 cm⁻¹ region and carbon-related vibrations from CNS in the fingerprint region below 1000 cm⁻¹.

Figure 4.

FTIR spectra of TEFB, PP, CNS and PP/TEFB/CNS.

The intensity analysis reveals that TEFB exhibits the most intense absorption peaks especially in the broad O-H stretching peak at 3400 cm⁻¹ and the C-O stretching peaks in the 1240–1030 cm⁻¹ region, reflecting the high concentration of hydroxyl and ether functional groups in the alkali-treated cellulosic material. PP shows moderate intensity peaks primarily in the aliphatic C-H region (2920–2850 cm⁻¹) and bending vibrations (1460–1380 cm⁻¹), consistent with its saturated hydrocarbon structure.⁴¹ CNS demonstrates the lowest overall peak intensity due to its predominantly carbon-based composition with fewer polar functional groups which resulting in weaker infrared absorption. The PP/TEFB/CNS spectrum exhibits a combination of characteristic peaks from all components with the O-H stretching peak from TEFB being significantly reduced in intensity compared to pure TEFB which suggest possible hydrogen bonding interactions between the fibre and matrix components. The aliphatic C-H peaks remain prominent indicating the presence of both PP and TEFB components. Notably, the PP/TEFB/CNS spectrum shows some peak shifts and broadening compared to individual components particularly in the C-O stretching region which may indicate interfacial interactions between TEFB, PP, and CNS. The fingerprint region below 1500 cm⁻¹ in the PP/TEFB/CNS shows overlapping peaks from all components with the overall intensity being intermediate between the pure components and confirming successful incorporation of all materials in the composite.

Mechanical properties of composite

Figure 5(a) shows the tensile strength of PP/TEFB/CNS composites with different EFB fibre content (0–50 wt%) and CNS concentration (0, 0.5, 1.0, and 1.5 wt%). The results show that both fibre loading and CNS content significantly affect the mechanical properties. Neat PP composites without CNS (0% CNS) showed tensile strength of 32 MPa at 0% fibre content. The strength increased to 35 MPa at 20% fibre loading. Beyond 20% fibre content the strength decreased continuously to 28 MPa at 50% fibre loading. This pattern indicates that low fibre concentrations provide reinforcement benefits. Higher fibre loadings cause properties deterioration due to fibre agglomeration and poor adhesion.⁴² The addition of 0.5% CNS improved tensile strength across all fibre concentrations. The strength ranged from 33 MPa at 0% fibre to 37.0 MPa at 20% fibre content. The strength then decreased to 30.0 MPa at 50% fibre loading. This improvement demonstrates the positive effect of CNS addition. The 1% CNS loading achieved the best tensile strength enhancement with a maximum value of 41 MPa at 20% fibre content. This represents a 17% improvement compared to the neat PP/TEFB composite at the same fibre loading. The superior performance occurs due to optimal nanofiller dispersion. CNS enhances interfacial bonding between PP matrix and TEFB fibres.⁴³ The nanoparticles also improve stress transfer efficiency and provide crack resistance.⁴⁴ However, 1.5% CNS loading showed reduced performance compared to 1% CNS. The maximum strength reached only 39 MPa at 20% fibre content. This reduction indicates nanofiller agglomeration at higher CNS concentrations. Too much CNS reduces reinforcement efficiency. The optimal combination is 20% TEFB fibre with 1% CNS due to synergistic effects. The natural fibre provides structural reinforcement while CNS enhances compatibility and mechanical properties. Beyond 20% fibre content all composites showed declining strength. This suggests that CNS benefits cannot overcome excessive fibre loading problems. These problems include poor fibre dispersion and increased void content. The results demonstrate that both fibre and nanofiller concentrations must be carefully optimized. Proper balance is crucial for achieving maximum mechanical performance in PP-based natural fibre nanocomposites.

Figure 5.

Mechanical testing of PP/TEFB/CNS composite for (a) tensile strength; (b) flexural strength; (c) impact strength; (d) representative tensile stress-strain curves for PP/TEFB/CNS composites at 20 wt% EFB fibre content with varying CNS loading.

Figure 5(b) shows the flexural strength of PP/TEFB/CNS composites with different EFB fibre content (0–50 wt%) and CNS concentration (0, 0.5, 1.0, and 1.5 wt%). PP composites without CNS showed flexural strength of 20 MPa at 0% fibre content. The strength increased to 25 MPa at 20% fibre loading. Beyond 20% fibre content, the strength decreased dramatically to 13 MPa at 50% fibre loading. This trend indicates that fibre reinforcement is effective only at moderate loadings. Excessive fibre content causes severe properties deterioration. The addition of 0.5% CNS improved flexural strength across all fibre concentrations. The strength ranged from 21 MPa at 0% fibre to 27.0 MPa at 20% fibre content. The strength then decreased to 18 MPa at 50% fibre ratio. This improvement shows that CNS addition provides better fibre-matrix interfacial bonding in terms of assisting to enhance the flexural strength of the composite.⁴⁵ The 1% CNS loading achieved the highest flexural strength enhancement with a maximum value of 30 MPa at 20% fibre content. This represents a 20% improvement compared to the neat PP/TEFB composite at the same fibre loading. The superior flexural performance occurs due to enhanced matrix-fibre adhesion and improved stress distribution. CNS particles act as coupling agents between the hydrophilic EFB fibres and hydrophobic PP matrix.⁵ The nanoparticles also provide additional resistance to crack initiation and propagation under bending loads.

The impact strength of PP/TEFB/CNS with various EFB fibre and CNS content is shown in Figure 5(c). Neat PP composites without CNS showed impact strength of 24 J/m at 0% fibre content. The strength increased to 31 J/m at 20% fibre loading. Beyond 20% fibre content the strength decreased continuously to 23 J/m at 50% fibre loading. This pattern indicates that moderate fibre loading improves impact resistance through energy absorption mechanisms. Higher fibre loadings reduce toughness due to stress concentration and poor interfacial bonding.⁴⁶ The addition of 0.5% CNS improved impact strength across all fibre concentrations. The strength ranged from 25 J/m at 0% fibre to 34 J/m at 20% fibre content. The strength then decreased to 25 J/m at 50% fibre loading. This improvement demonstrates that CNS enhances the energy absorption capacity of the composite. The 1% CNS loading achieved the highest impact strength enhancement with a maximum value of 41 J/m at 20% fibre content. This represents a 32% improvement compared to the neat PP/TEFB composite at the same fibre loading. The superior impact performance occurs due to enhanced crack deflection and energy dissipation mechanisms. CNS particles create multiple crack propagation paths and provide bridging effects during impact loading.⁴⁷ The nanoparticles also improve fibre pullout resistance and matrix deformation capacity.⁴⁸ The 1.5% CNS loading showed good performance with a maximum strength of 37 J/m at 20% fibre content. However, the overall performance was lower than 1% CNS across most fibre concentrations. This indicates that optimal CNS loading is around 1% for maximum impact resistance. Higher concentrations may cause nanofiller agglomeration that reduces toughening effectiveness. The optimal combination is 20% TEFB fibre with 1% CNS due to synergistic toughening effects. The natural fibres provide crack deflection and fibre pullout mechanisms while CNS enhances matrix toughness and interfacial strength. Beyond 20% fibre content, all composites showed declining impact strength. The decline is most severe for composites without CNS reinforcement. This confirms that CNS addition helps maintain impact properties even at higher fibre loadings. The impact strength results show the most significant improvements with CNS addition compared to tensile and flexural properties. This behavior is characteristic of nanoparticle reinforcement where small additions can dramatically enhance toughness through multiple energy absorption mechanisms. The results confirm that 1% CNS with 20% TEFB fibre provides optimal impact performance in PP-based natural fibre nanocomposites.

Figure 5(d) shows the representative tensile stress-strain curves for PP/TEFB/CNS composites at optimal 20 wt% EFB fibre content with different CNS loadings (0%, 0.5%, 1%, and 1.5% CNS). The curves exhibit nonlinear behavior characteristic of semi-crystalline thermoplastic composites with natural fibre reinforcement. All compositions display an initial linear elastic region followed by yielding and plastic deformation before final fracture. The neat PP/TEFB composite (0% CNS) shows a maximum stress of 35 MPa with strain at break of 6.4%. With addition of 0.5% CNS, the tensile strength slightly increases to 37 MPa while maintaining similar ductility. The optimal composition with 1% CNS demonstrates the highest tensile strength of 41 MPa, representing a 17% improvement compared to the neat PP/TEFB composite. The 1.5% CNS loading shows slightly reduced performance with maximum stress of 39 MPa. The stress-strain behavior reveals that CNS incorporation affects both the elastic modulus (slope of initial linear region) and the post-yield behavior. The curves demonstrate that optimal CNS loading (1%) enhances load-bearing capacity while maintaining reasonable elongation, indicating improved stress transfer efficiency between the fibre and matrix through enhanced interfacial adhesion. Furthermore, the stress-strain curves exhibit nonlinear behavior characteristic of semi-crystalline thermoplastic composites with natural fibre reinforcement, displaying continuous deformation from the elastic region through to failure.^49,50 The composites demonstrate a gradual yielding behavior rather than a sharp distinct yield point, which is typical for ductile thermoplastic matrices. Analysis of the stress-strain curves reveals that the onset of yielding occurs at 28 MPa for the neat PP/TEFB composite (0% CNS), progressively increasing to 30 MPa with 0.5% CNS, 32 MPa with 1% CNS, and 31 MPa with 1.5% CNS at the optimal 20% fibre content. This enhancement in yield stress with optimal CNS incorporation (1%) demonstrates improved load-bearing capacity during the elastic-to-plastic transition which attributed to enhanced interfacial bonding and more efficient stress transfer between the fibre and matrix. The gradual transition from elastic to plastic deformation is typical for PP-based composites reinforced with natural fibres, where the yielding process is influenced by fibre content, orientation and interfacial bonding.

The tensile modulus and strain at break were found to be primarily governed by fibre content rather than CNS loading. As shown in Table 1, modulus increases from 1200 MPa to 1950 MPa with increasing EFB content (0–50 wt%), while strain at break decreases from 8% to 4%. CNS loading (0–1.5 wt%) showed negligible effect on these properties, indicating that the nanofiller primarily enhances interfacial bonding and strength rather than bulk stiffness.

Table 1.

Tensile modulus and strain at break as a function of EFB fibre content.

EFB fibre content	Modulus (MPa)	Strain at break (%)
0	1200	8.0
10	1350	7.2
20	1500	6.4
30	1650	5.6
40	1800	4.8
50	1950	4.0

One-way ANOVA was applied to mechanical properties to determine the significance between different compositions as presented in Table 2. The statistical analysis confirmed significant differences between composite formulations, with the addition of CNS showing statistically significant improvements in mechanical properties compared to neat PP/TEFB composites. The one-way ANOVA results demonstrate that there were statistically significant differences between the mean values of mechanical properties for different composite formulations when the P-value was less than 0.05. This indicates that the observed variations in tensile strength, flexural strength, and impact strength with different EFB fibre contents and CNS loadings are statistically meaningful.

Table 2.

Summary of one-way ANOVA for mechanical properties.

Variables	df	Tensile strength	Flexural strength	Impact strength
Mixture	3	0.02*	0.01*	0.00*

*Note: Significantly different at p ≤ .05.

Table 3 presents a comprehensive comparison of the mechanical properties of PP/TEFB/CNS composites with recent natural fibre-reinforced polymer composite studies from related literature. The optimal formulation (20 wt% EFB +1 wt% CNS) achieved tensile strength of 41 MPa, demonstrating competitive performance across different enhancement strategies. Compared to PP/maple wood composites using conventional MAPP coupling agents,⁵¹ the current study achieved 51% higher tensile strength (41 vs 27.13 MPa), demonstrating CNS effectiveness in enhancing interfacial adhesion at significantly lower loading (1% vs 3%). Sullins, et al.⁵² investigated PP/hemp fibre composites with various treatments and reported tensile strength of 33-38 MPa at 40% fibre content with 5% MAPP, indicating that the current PP/EFB/CNS composites achieves comparable or superior tensile performance at half the fibre loading and one-fifth the additive content. Nejat et al.⁵³ achieved flexural strength of 30.21 MPa using recycled PP/bagasse composites with 50% fibre and 4% nanoclay combined with alkali treatment, remarkably similar to the current study’s 30 MPa flexural strength achieved at only 20% fibre content. This demonstrates that biomass-derived CNS provides comparable reinforcement efficiency to conventional nanoclay while using substantially lower fibre loadings. However, Basiji et al.⁵¹ achieved higher flexural strength (48.74 MPa) with 15% maple wood fibre, reflecting the inherently superior mechanical properties of wood fibres compared to agricultural residue fibres. Rhodes et al.⁵⁴ reported the highest specific properties through basalt/hemp hybridization, though this incorporated inorganic fibres rather than the fully bio-based composite presented here. The tensile modulus of 1500 MPa demonstrates effective stiffness enhancement, positioning between Rhodes et al.⁵⁴ 1000 MPa (15% wood) and Sullins et al.⁵² 3000–3500 MPa range (40% hemp), as expected for the intermediate fibre loading employed. Nejat et al.⁵³ achieved substantially higher modulus (5660 MPa) using 50% bagasse with rigid nanoclay platelets, confirming the stiffness-fibre content relationship. The impact strength of 41 J/m (0.041 kJ/m²) represents 32% enhancement compared to neat PP/TEFB, positioning competitively against Sun et al.⁵⁵ at 0.022 kJ/m² for enzyme-treated coir/PLA composites. These results position the nanocomposites as viable candidates for automotive interior components and consumer products where moderate mechanical properties, sustainability and cost-effectiveness are prioritized.

Table 3.

Comparison of mechanical properties with PP/natural fibre composites and related literature.

Study	Matrix	Fibre/filler type	Fibre content (wt%)	Additive	Tensile strength (MPa)	Flexural strength (MPa)	Impact strength (kJ/m²)	Modulus (MPa)
Current study	PP	EFB	20	1% CNS	41	30	0.041	1500
Basiji et al.⁵¹	PP	Maple wood	15	3% MAPP	27.13	48.74	-	1000
Nejat et al.⁵³	Recycled PP	Bagasse	50	4% nanoclay + alkali	-	30.21	-	5660
Rhodes et al.⁵⁴	PP	Basalt/Hemp	30	5% MAPP	54.3	86.6	-	4920
Sullins et al.⁵²	PP	Hemp	40	5% MAPP	33–38	-	-	3000–3500
Meekum⁵⁶	rPET/PE blend	Rice husk	20	Peroxide/silane	35–45	-	-	-
Sun et al.⁵⁵	PLA	Coir	4-6	Amylase enzyme	22.57	∼40	0.022	-

Performance prediction using artificial neural network (ANN)

The performance of the developed ANN model for predicting mechanical properties of PP/EFB/CNS nanocomposites was evaluated using correlation coefficient (R) values for training, validation, testing and overall datasets. Figure 6 shows the regression plots comparing predicted outputs against target values for all datasets. The ANN model demonstrated excellent predictive capability across all evaluation phases. The training dataset achieved a correlation coefficient of R = 0.99875. This indicates strong learning performance during the network training process. The validation dataset showed R = 0.99972. This confirms that the model maintained high accuracy on unseen data during training and avoided overfitting. The testing dataset yielded R = 0.9986. This represents the highest correlation among all phases and demonstrates superior generalization capability of the trained network. The overall performance across all datasets achieved R = 0.99886. This indicates excellent agreement between predicted and experimental values with 98.5% correlation. This high correlation coefficient confirms that the ANN model successfully captured the complex nonlinear relationships between input parameters and output mechanical properties. The regression plots show data points closely aligned with the ideal Y = T line across all datasets. The training and validation plots display consistent scatter patterns around the regression line. This indicates stable learning without significant bias. The testing plot shows particularly tight clustering around the ideal line. This confirms excellent predictive accuracy on completely independent data. The comprehensive comparison between experimental and predicted values for all five mechanical properties across 24 compositions is presented in Supplemental Table S1, demonstrating the model’s excellent predictive capability for tensile strength, flexural strength, impact strength, modulus and strain at break.

Figure 6.

ANN regression plots showing correlation between predicted and experimental values for all five mechanical properties (tensile strength, flexural strength, impact strength, modulus and strain at break).

Conclusion

This study successfully demonstrated the synergistic effects of EFB fibres and CNS on enhancing the mechanical properties of PP composites. The CNS synthesized from sago bark waste through pyrolysis at 500°C exhibited spherical morphology with diameters ranging from 15 to 50 nm and partially crystalline structure. This confirms successful biomass conversion into valuable nanomaterials. FTIR analysis confirmed successful interfacial interactions between all components through peak shifts and intensity modifications. This was particularly evident in the C-O stretching region. Thermal analysis demonstrated improved thermal stability with the PP/TEFB/CNS nanocomposite showing 20°C higher degradation onset temperature at 320°C compared to PP/TEFB composite at 300°C. This indicates enhanced thermal performance through multiple stabilization mechanisms including physical barrier effects and protective char formation. The mechanical properties evaluation revealed that the optimal composite formulation of 20 wt% EFB fibre with 1 wt% CNS achieved remarkable enhancements compared to neat PP/EFB composite. Tensile strength increased by 17% from 35 to 41 MPa. Flexural strength improved by 20% from 25 to 30 MPa. Impact strength enhanced by 32% from 31 to 41 J/m. These improvements were attributed to enhanced interfacial adhesion between hydrophilic EFB fibres and hydrophobic PP matrix facilitated by CNS acting as coupling agents. The improvements also resulted from improved stress transfer mechanisms and crack bridging effects. The extended artificial neural network model successfully predicted all mechanical properties including tensile strength, flexural strength, impact strength, tensile modulus and strain at break with excellent overall correlation of R = 0.99886, demonstrating its effectiveness as a comprehensive optimization tool for composite formulation design without extensive experimental work. The model successfully identified optimal compositions and can serve as a valuable tool for future composite development. However, fibre loadings exceeding 20 wt% and CNS concentrations above 1 wt% resulted in decreased mechanical properties. This occurred due to fibre agglomeration and nanofiller clustering and poor interfacial adhesion. These results emphasize the importance of optimal formulation design. This research contributes to sustainable material development by demonstrating effective valorization of two agricultural waste streams. Oil palm EFB and sago bark were successfully converted into high-performance composites. The developed materials show significant potential for applications in automotive components and consumer products where enhanced mechanical properties and sustainability are crucial. Future work should focus on long-term durability studies and environmental impact assessment and scale-up production feasibility to facilitate commercial implementation of these bio-based nanocomposites.

Supplemental Material

Supplemental Material - Effect of biomass-derived carbon nanosphere on mechanical properties of polypropylene/empty fruit bunch fibre composites

Supplemental Material for Effect of biomass-derived carbon nanosphere on mechanical properties of polypropylene/empty fruit bunch fibre composites by Huei Ruey Ong, Yifei Shi, Chi Shein Hong, Md Maksudur Rahman Khan, Wan Mohd Eqhwan Iskandar in Journal of Thermoplastic Composite Materials.

Footnotes

Acknowledgements

The authors acknowledge the support of Geely University of China and DRB-HICOM University of Automotive Malaysia for their valuable contributions to this project. Their support and resources have been instrumental in the successful completion of this research.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Huei Ruey Ong

Wan Mohd Eqhwan Iskandar

Supplemental Material

Supplemental material for this article is available online.

References

Morais

Drew

, et al. Handbook of Climate Change Mitigation and Adaptation. Heidelberg: Springer, 2022, pp. 1987–2029.

Hossain

Shahid

Mahmud

, et al. Research and application of polypropylene: a review. Discover Nano 2024; 19(1): 2.

Maiti

Islam

Uddin

, et al. Sustainable fiber‐reinforced composites: a review. Advanced Sustainable Systems 2022; 6(11): 2200258.

Dolah

Karnik

Hamdan

. A comprehensive review on biofuels from oil palm empty bunch (EFB): current status, potential, barriers and way forward. Sustainability 2021; 13(18): 10210.

Ermeydan

. Interfacial Bonding Characteristics in Natural Fiber Reinforced Polymer Composites: Fiber-matrix Interface In Biocomposites. Heidelberg: Springer, 2024, pp. 23–45.

Jagadeesh

Puttegowda

Mavinkere Rangappa

, et al. Influence of nanofillers on biodegradable composites: a comprehensive review. Polym Compos 2021; 42(11): 5691–5711.

Devnani

Sinha

. Effect of nanofillers on the properties of natural fiber reinforced polymer composites. Mater Today Proc 2019; 18: 647–654.

Zhang

Qiao

Z-A

Dai

. Recent advances in carbon nanospheres: synthetic routes and applications. Chem Commun 2015; 51(45): 9246–9256.

Yao

Huang

, et al. Synthesis and applications of carbon nanospheres: a review. Particuology 2024; 87: 325–338.

10.

Ali

Rahimian Koloor

Alshehri

, et al. Carbon nanotube characteristics and enhancement effects on the mechanical features of polymer-based materials and structures–A review. J Mater Res Technol 2023; 24: 6495–6521.

11.

Han

Wang

Jin

, et al. Multiscale carbon nanosphere–carbon fiber reinforcement for cement-based composites with enhanced high-temperature resistance. J Mater Sci 2015; 50: 2038–2048.

12.

Díez-Pascual

AM.

Carbon-based polymer nanocomposites for high-performance applications. Polymers. 2020; 12(4): 872.

13.

Ahmad

Ishak

Mohammad Taha

, et al. Mechanical, thermal and physical characteristics of oil palm (Elaeis Guineensis) fiber reinforced thermoplastic composites for FDM–type 3D printer. Polym Test 2023; 120: 107972.

14.

Kang

Saito

. Synthesis of structure-controlled carbon nano spheres by solution plasma process. Carbon 2013; 60: 292–298.

15.

Guan

Yang

Zhu

, et al. The controllable porous structure and s-doping of hollow carbon sphere synergistically act on the microwave attenuation. Carbon 2022; 188: 1–11.

16.

Karna

. Synthesis and characterization of carbon nanospheres. Open Access Library Journal 2017; 4(05): 1–7.

17.

Zhang

Zong

Zhang

, et al. Preparation of hollow mesoporous carbon spheres by pyrolysis-deposition using surfactant as carbon precursor. J Power Sources 2021; 484: 229274.

18.

Weng

Guan

Wang

, et al. Hollow carbon nanospheres derived from biomass by-product okara for imaging-guided photothermal therapy of cancers. J Mater Chem B 2019; 7(11): 1920–1925.

19.

Majewsky

Bitter

Eiche

, et al. Determination of microplastic polyethylene (PE) and polypropylene (PP) in environmental samples using thermal analysis (TGA-DSC). Sci Total Environ 2016; 568: 507–511.

20.

Abbas-Abadi

Van Geem

Fathi

, et al. The pyrolysis of oak with polyethylene, polypropylene and polystyrene using fixed bed and stirred reactors and TGA instrument. Energy 2021; 232: 121085.

21.

Ruksathamcharoen

Chuenyam

Ajiwibowo

, et al. Thermogravimetric analysis of combustion characteristics and kinetics of hydrothermally treated and washed empty fruit bunch. Biofuels. 2021; 12(8): 913–925.

22.

Nyakuma

Wong

Oladokun

. Non-oxidative thermal decomposition of oil palm empty fruit bunch pellets: fuel characterisation, thermogravimetric, kinetic, and thermodynamic analyses. Biomass Convers Biorefin 2021; 11(4): 1273–1292.

23.

Radzi

NAM

Sofian

Jamari

. Formulation of recycle polypropylene reinforced with oil palm empty fruit bunch and fly ash composite for high thermal properties through D-optimal mixture design. Melville, NY: AIP Publishing, 2024, p. 3128.

24.

Batakliev

Georgiev

Kalupgian

, et al. Physico-chemical characterization of PLA-based composites holding carbon nanofillers. Appl Compos Mater 2021; 28(4): 1175–1192.

25.

Heidarian

Hassan

. Characterization and comparison of fluoroelastomer unfilled, filled with carbon nanotube (unmodified, acid or base surface modified) and carbon black using TGA-GCMS. J Elastomers Plast 2021; 53(7): 861–885.

26.

Chandel

Kapurderiya

Kanse

, et al. Synergistic effect of single wall carbon nanotubes/graphene oxide hybrid nano‐filler on thermal stabilization and tensile strength of polyacrylonitrile copolymer. Polym Compos 2023; 44(12): 8589–8600.

27.

Iskandar

WME

Ong

Khan

MMR

, et al. Effect of ultrasonication on alkaline treatment of empty fruit bunch fibre: Fourier transform infrared spectroscopy (FTIR) and morphology study. Mater Today Proc 2022; 66: 2840–2843.

28.

Sayakulu

Soloi

. The effect of sodium hydroxide (NaOH) concentration on oil palm empty fruit bunch (OPEFB) cellulose yield. J Phy Conf Series 2022; 2314: 012017, IOP Publishing.

29.

Ong

Khan

MMR

Prasad

DMR

, et al. Palm kernel meal as a melamine urea formaldehyde adhesive filler for plywood applications. Int J Adhesion Adhes 2018; 85: 8–14.

30.

Ong

Prasad

Rahman Khan

, et al. Effect of palm kernel meal as melamine urea formaldehyde adhesive extender for plywood application: using a Fourier transform infrared spectroscopy (FTIR) study. Trans Tech Publ 2012; 121-126: 493–498.

31.

Caban

. FTIR-ATR spectroscopic, thermal and microstructural studies on polypropylene-glass fiber composites. J Mol Struct 2022; 1264: 133181.

32.

Chin

Tee

Tong

, et al. Thermal and mechanical properties of bamboo fiber reinforced composites. Mater Today Commun 2020; 23: 100876.

33.

Ong

Khan

Yousuf

, et al. Effect of waste rubber powder as filler for plywood application. Pol J Chem Technol 2015; 17(1): 41–47.

34.

Ong

Rahman Khan

Ramli

, et al. Effect of CuO nanoparticle on mechanical and thermal properties of palm oil based alkyd/epoxy resin blend. Procedia Chem 2015; 16: 623–631.

35.

Ong

Khan

MMR

Ramli

, et al. Tailoring base catalyzed synthesis of palm oil based alkyd resin through CuO nanoparticles. RSC Adv 2015; 5(116): 95894–95902.

36.

Tong

Chin

Mustafa

, et al. Influence of alkali treatment on physico-chemical properties of Malaysian bamboo fiber: a preliminary study. Malays J Anal Sci 2018; 22(1): 143–150.

37.

Ramli

Khan

MMR

Yunus

, et al. In-situ impregnation of copper nanoparticles on palm empty fruit bunch powder. Adv Nanoparticles 2014; 03: 65.

38.

Fan

Huang

Y-Z

Lin

J-N

, et al. Microplastic constituent identification from admixtures by Fourier-transform infrared (FTIR) spectroscopy: the use of polyethylene terephthalate (PET), polyethylene (PE), polypropylene (PP), polyvinyl chloride (PVC) and nylon (NY) as the model constituents. Environmental Technology & Innovation 2021; 23: 101798.

39.

Bohus

Lukács

, et al. Comparative study of carbon nanosphere and carbon nanopowder on viscosity and thermal conductivity of nanofluids. Nanomaterials 2021; 11(3): 608.

40.

Sivamaran

Balasubramanian

Gopalakrishnan

, et al. Combined synthesis of carbon nanospheres and carbon nanotubes using thermal chemical vapor deposition process. Chem Phys Impact 2022; 4: 100072.

41.

Morgado

Gomes

Bettencourt da Silva

RJN

, et al. Validated spreadsheet for the identification of PE, PET, PP and PS microplastics by micro-ATR-FTIR spectra with known uncertainty. Talanta 2021; 234: 122624.

42.

Bartos

Kócs

Anggono

, et al. Effect of fiber attrition, particle characteristics and interfacial adhesion on the properties of PP/sugarcane bagasse fiber composites. Polym Test 2021; 98: 107189.

43.

Olonisakin

Fan

Xin-Xiang

, et al. Key improvements in interfacial adhesion and dispersion of fibers/fillers in polymer matrix composites; focus on pla matrix composites. Compos Interfaces 2022; 29(10): 1071–1120.

44.

Liu

, et al. Metal–organic framework-derived hollow carbon nanosphere@ Ni/reduced graphene oxide composites for supercapacitor electrodes with enhanced performance. ACS Appl Nano Mater 2023; 6(3): 1582–1591.

45.

Shahapurkar

Venkatesh

Tirth

, et al. Influence of graphene nano fillers and carbon nano tubes on the mechanical and thermal properties of hollow glass microsphere epoxy composites. Processes 2021; 10(1): 40.

46.

Mustapha

SNH

Norizan

CWNFCW

Roslan

, et al. Effect of Kenaf/empty fruit bunch (EFB) hybridization and weight fractions in palm oil blend polyester composite. J Nat Fibers 2022; 19(5): 1885–1898.

47.

Zhao

Wang

Bie

, et al. The effect of impact load on the atomistic scale fracture behavior of nanocrystalline bcc iron. Nanomaterials 2024; 14(4): 370.

48.

Intarabut

Sukontasukkul

Phoo-Ngernkham

, et al. Influence of graphene oxide nanoparticles on bond-slip reponses between fiber and geopolymer mortar. Nanomaterials 2022; 12(6): 943.

49.

Gehring

Bouchart

Dinzart

, et al. Microstructure, mechanical behaviour, damage mechanisms of polypropylene/short hemp fibre composites: experimental investigations. J Reinforc Plast Compos 2012; 31(22): 1576–1585.

50.

Dewidar

Bakrey

Hashim

, et al. Mechanical properties of polypropylene reinforced hemp fiber composite. Mater Phys Mech 2012; 15(2): 119–125.

51.

Basiji

Erchiqui

Koubaa

, et al. Influence of fiber concentration and length on the dielectric, mechanical, and thermal properties of maple wood fiber-reinforced polypropylene. J Thermoplast Compos Mater. 2025; 38(9): 4386–4409.

52.

Sullins

Pillay

Komus

, et al. Hemp fiber reinforced polypropylene composites: the effects of material treatments. Compos B Eng 2017; 114: 15–22.

53.

Nejat

Naghdi

Nadali

, et al. Effects of nanoclay cloisite 20A and alkali treatments on structure-property relationships of bagasse/recycled polypropylene nanocomposites. J Thermoplast Compos Mater 2024; 37(1): 310–335.

54.

Rhodes

Pelaez-Samaniego

Kiziltas

, et al. Mechanical and viscoelastic properties of basalt-hemp hybrid reinforced polypropylene. J Thermoplast Compos Mater 2024; 37(1): 336–362.

55.

Sun

Yang

Wang

, et al. Properties of coir fibre/polylactic acid composites based on amylase processing. J Thermoplast Compos Mater 2024; 37(1): 426–439.

56.

Meekum

. Manufacturing of wood plastic composite from polyethylene/poly (ethylene terephthalate) waste/ultra high molecular weight polyethylene blends and rice husk fiber: effect of blends ratio and peroxide/silane crosslink system. J Thermoplast Compos Mater 2024; 37(1): 28–65.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.16 MB