Tuning APTES silanization window in PLA/cellulose nanofiber biocomposites: Optimizing interfacial adhesion, mechanical strength, and thermal stability

Abstract

Polylactic acid (PLA) is limited by inherent brittleness and poor thermal stability, hindering its engineering applications. This study systematically investigated cellulose nanofibers (CNFs) silanized with 3- aminopropyl-triethoxysilane (APTES) at mass ratios of 2:1 (R5) and 4:1 (R7), incorporated into PLA at 0.5 wt% to 1.5 wt% levels. Unmodified 1.0 wt% CNFs (R2) enhanced impact strength (10.15 ± 0.50 kJ·m⁻²) via pull-out toughening. Moderate silanization (R5) improved interfacial adhesion, achieving balanced tensile strength (50.2 ± 1.4 MPa), modulus (321 ± 16 MPa), and elongation (5.0 ± 0.2%). Stronger silanization (R7) increased modulus (545 ± 26 MPa) but reduced elongation (4.04 ± 0.18%), inducing brittleness. Fourier transform infrared analysis confirmed reduced hydroxyl groups and Si–O–Si/Si–O–C bond formations. Uniform interphases were observed in R5, while R7 exhibited voids and heterogeneity in scanning electron microscopy. Thermogravimetric analysis revealed higher onset (T_onset 352 ± 2 °C) and maximum decomposition (T_max 360 ± 2 °C) temperatures for R5 compared to R2 (335 ± 2 °C, 342 ± 2 °C). This study validates an “optimized silanization window” (2:1 ratio), enabling simultaneous enhancements in stiffness, toughness, and thermal stability for sustainable PLA/ CNFs biocomposites, suitable for industrially compostable packaging and biomedical applications.

Full Article

Tuning APTES Silanization Window in PLA/Cellulose Nanofiber Biocomposites: Optimizing Interfacial Adhesion, Mechanical Strength, and Thermal Stability

Ramtin Ghorbanpour  , Seyed Eshagh Ebadi  ,* Valiullah Moosavi,

Mojtaba Soltani  , and Mohsen Saffari

Polylactic acid (PLA) is limited by inherent brittleness and poor thermal stability, hindering its engineering applications. This study systematically investigated cellulose nanofibers (CNFs) silanized with 3- aminopropyl-triethoxysilane (APTES) at mass ratios of 2:1 (R5) and 4:1 (R7), incorporated into PLA at 0.5 wt% to 1.5 wt% levels. Unmodified 1.0 wt% CNFs (R2) enhanced impact strength (10.15 ± 0.50 kJ·m⁻²) via pull-out toughening. Moderate silanization (R5) improved interfacial adhesion, achieving balanced tensile strength (50.2 ± 1.4 MPa), modulus (321 ± 16 MPa), and elongation (5.0 ± 0.2%). Stronger silanization (R7) increased modulus (545 ± 26 MPa) but reduced elongation (4.04 ± 0.18%), inducing brittleness. Fourier transform infrared analysis confirmed reduced hydroxyl groups and Si–O–Si/Si–O–C bond formations. Uniform interphases were observed in R5, while R7 exhibited voids and heterogeneity in scanning electron microscopy. Thermogravimetric analysis revealed higher onset (T_onset 352 ± 2 °C) and maximum decomposition (T_max 360 ± 2 °C) temperatures for R5 compared to R2 (335 ± 2 °C, 342 ± 2 °C). This study validates an “optimized silanization window” (2:1 ratio), enabling simultaneous enhancements in stiffness, toughness, and thermal stability for sustainable PLA/ CNFs biocomposites, suitable for industrially compostable packaging and biomedical applications.

DOI: 10.15376/biores.21.1.1172-1191

Keywords: PLA biocomposites; Cellulose nanofibers (CNFs); APTES silanization; Interphase engineering; Mechanical properties; Thermal stability; Fracture micromechanics; Optimized silanization window

Contact information: Department of Wood and Paper Science and Technology, Cha.C., Islamic Azad University, Chalus, Iran;

* Corresponding author: eshaghebadi@iau.ac.ir

INTRODUCTION

The growing global demand for sustainable and eco-friendly materials has drawn considerable attention to bio-based and biodegradable polymers. Among them, poly(lactic acid) (PLA) is recognized as a key candidate for applications such as industrially compostable packaging, medical devices, and lightweight structures, owing to its renewable origin, degradability under industrial composting conditions, and favorable mechanical strength (Yang et al. 2020). These attributes align well with the United Nations’ Sustainable Development Goal 12 on responsible consumption and production. However, the inherent brittleness, low elongation at break, and limited thermal stability of PLA restrict its engineering applications (Mandal et al. 2021). Therefore, the simultaneous improvement of strength, toughness, and thermal stability of PLA—without compromising its intrinsic properties—remains both a scientific and industrial necessity.

To enhance PLA performance, several strategies have been developed, including blending with flexible polymers (Wang et al. 2022), incorporation of bio-based plasticizers (Srisawat et al. 2022), and the addition of inorganic nanoparticles such as SiO₂ or ZnO (Kaseem 2021). The authors’ previous study also demonstrated that graphene nanosheets could significantly improve the elastic modulus and tensile strength of PLA/beech-wood flour composites (Ghorbanpour et al. 2022). Despite these advances, challenges such as modulus reduction or loss of biodegradability remain. Consequently, the use of natural nanofillers such as CNFs has emerged as an effective approach due to their biocompatibility and ability to enhance both mechanical and thermal performance.

CNFs are considered ideal reinforcements for PLA because of their high specific modulus, low density, and efficient derivation from naturally abundant cellulose (Shi et al. 2025; Abdul Khalil et al. 2020). Uniform dispersion of CNFs within the PLA matrix facilitates effective stress transfer across the interphase and activates fracture mechanisms such as fiber pull-out and crack deflection (Nazrin et al. 2021). Nevertheless, the intrinsic incompatibility between hydrophilic CNFs and hydrophobic PLA often leads to fiber agglomeration and weak interfacial adhesion (Mandal et al. 2021). Surface modification of CNFs with the silane coupling agent 3-aminopropyl-triethoxysilane (APTES) can overcome this limitation. The coupling agent APTES reacts with hydroxyl groups on CNFs and ester groups in PLA, forming stable Si–O–Si and Si–O–C linkages, while its amino groups promote stronger interfacial compatibility (Magee et al. 2023; Zhang et al. 2023).

The amino groups of APTES play a specific role in polar PLA matrices by forming hydrogen bonds with PLA carbonyl groups (NH₂ … O=C) and potentially enabling nucleophilic interactions during melt processing (Li et al. 2020; Panaitescu et al. 2022). While non-amino silanes (e.g., vinyl or alkyl) are effective in non-polar polymers (PP, PE) via hydrophobic matching, they exhibit weaker adhesion to polar PLA (Zhang et al. 2023). Thus, APTES was selected for its enhanced interfacial compatibility with PLA’s chemistry.

Recent studies have confirmed that silanization improves the mechanical and thermal stability of PLA/CNFs composites (Liu et al. 2022; Yang et al. 2020). However, the effect of silanization intensity has received limited attention (Srisawat et al. 2022; Zhang et al. 2025). Insufficient silanization leaves unreacted hydroxyl groups, restricting adhesion, whereas excessive silanization produces heterogeneous layers that cause stress concentration and brittle fracture (Varghese et al. 2023). This contrast highlights the existence of an “optimized silanization window,” which has seldom been systematically examined.

This study systematically defines and investigates the concept of an optimized silanization window in PLA/CNFs nanocomposites. Unlike previous works that primarily compared silanized and unsilanized systems (Qian et al. 2018; Liu et al. 2022; Lozano Fernandez and Miskolczi 2022), the present work analyzes the intensity of surface modification as a critical variable. The findings demonstrate that balanced interfacial adhesion, mechanical strength, and thermal stability are achieved only within a moderate silanization range-identified herein as the optimized silanization window.

The selection of APTES is further justified by its established biocompatibility and suitability for bio-based systems. The aminopropyl functionality is non-toxic at low concentrations and has been widely employed in biomedical coatings, drug delivery systems, and tissue-engineering scaffolds owing to its ability to promote cell adhesion and protein immobilization without inducing adverse immune responses (Panaitescu et al. 2022). In contrast to halogenated or mercapto-silanes, APTES does not generate cytotoxic byproducts, making it a preferred coupling agent for sustainable and health-related applications of PLA/CNF composites (Goddard and Hotchkiss 2007)

MATERIALS AND METHODS

Materials

The main materials used in this study included PLA as the polymer matrix, CNFs as the reinforcing agent (either unmodified or surface-modified), and APTES as the surface coupling agent. Auxiliary solvents included ethanol and acetic acid (Table 1). APTES was selected due to its relative biocompatibility for bio-based applications (Li et al. 2020; Panaitescu et al. 2022).

Table 1. Materials Used in this Study

Surface Modification of Cellulose Nanofibers (CNFs)

Surface modification of CNFs with APTES was performed following the procedure reported by Shi et al. (2022), with two different silanization intensities. In the mild treatment (M1-CNF), a silane-to-cellulose mass ratio of 2:1 was used in an ethanol/water medium (pH 4.0, adjusted with 1% acetic acid) at 25 °C under magnetic stirring (300 rpm) for 24 h. In the intense treatment (M2-CNF), the same conditions were applied using a silane-to-cellulose ratio of 4:1. The modified products were centrifuged (8000 rpm, 5 min), washed three times with deionized water, and freeze-dried at −50 °C for 48 h to prevent agglomeration (Dong et al. 2014; Meng et al. 2022). To examine the effects of silanization intensity and CNFs loading, ten formulations (R0 to R9) were designed (Table 2).

Table 2. Compositions of PLA/CNFs Nanocomposite Samples (R0 to R9)

Preparation of Nanocomposites

Unmodified and APTES-modified CNFs (2:1 for M1 and 4:1 for M2) were incorporated into PLA at loading levels of 0.5 wt% to 1.5 wt%. Silanization was conducted in an ethanol/water medium under acidic conditions with magnetic stirring (Piorkowska et al. 2022). The PLA/CNFs nanocomposites were processed using a laboratory twin-screw extruder at 190 °C, 150 rpm, with a residence time of 8 min. The extruded strands were subsequently hot-pressed at 180 °C and 5 MPa to form composite sheets (Zhang et al. 2021; Shi et al. 2022).

Mechanical Testing

Tensile tests were conducted according to ASTM D638 on dumbbell-shaped specimens (gauge length 50 mm, crosshead speed 5 mm/min) with four replicates (n = 4) (Zhang et al. 2023). Izod impact tests were carried out on notched specimens following ASTM D256 (2014b). All data were reported as mean ± standard deviation and statistically analyzed using one-way ANOVA followed by Duncan’s multiple range test (P < 0.05) (Srisawat et al. 2022).

Chemical and Thermal Analysis

Samples R2, R5, and R7 were selected for Fourier transform infra-red (FTIR), thermogravimetric analysis (TGA), and scanning electron microscopy with energy dispersive X-ray spectrometric (SEM/EDX) analyses, since the properties of neat PLA have been previously reported (Yang et al. 2020). FTIR spectra were obtained using a Bruker Tensor spectrometer (Germany) in the range of 400 to 4000 cm⁻¹ with a resolution of 4 cm⁻¹ (Gitari et al. 2019). TGA was carried out on a Mettler Toledo instrument (Switzerland) under nitrogen atmosphere at a heating rate of 10 °C/min from 25 to 600 °C, using approximately 5 mg of each sample (Yun et al. 2023). The SEM/EDX analyses were performed using a SNE-4500M microscope (Korea) on gold-coated specimens to examine fracture morphology and confirm the presence of Si and N elements. A summary of instruments used for processing and characterization is provided in Table 3.

Table 3. Instruments Employed for Processing and Characterization

RESULTS AND DISCUSSION

Tensile and Impact Tests

Tensile and impact tests were conducted to evaluate the effects of CNFs and surface modification with APTES on the performance of PLA-based biocomposites. The data were expressed as mean ± standard deviation (four replicates) and statistically analyzed using one-way ANOVA followed by Duncan’s multiple range test (α = 0.05). The results are presented in Figs. 1–4.

Tensile Strength

Neat PLA (R0) exhibited a tensile strength of 46.6 ± 2.0 MPa, showing a typical brittle behavior. The incorporation of unmodified CNFs (R1–R3) enhanced the tensile strength; the R2 sample (1 wt% CNFs) achieved the highest improvement (50.2 ± 1.4 MPa), which was statistically different from neat PLA. At a higher CNFs loading (R3), the strength further increased to 55.5 ± 2.6 MPa, though variations were observed in other mechanical properties. Mild surface modification (M1-CNF, APTES-to-cellulose ratio 2:1) in samples R4–R6 improved fiber–matrix compatibility. The sample R5 (1 wt% M1-CNF) exhibited a tensile strength between 49 and 52 MPa, which was statistically distinct from that of unmodified CNFs composites. In contrast, severe modification (M2-CNF, ratio 4:1) in R7 to R9 led to a substantial decrease in tensile strength (R7: 24.3 ± 0.5 MPa). This reduction is attributed to the formation of thick and heterogeneous silane layers, which caused stress concentration and brittle fracture (Feng et al. 2020) (Fig. 1).

Fig. 1. Tensile strength (mean ± SD, n = 4) of PLA/CNF composites (R0 to R9) with Duncan’s grouping (letters above bars, α = 0.05). Specimen codes: R0 = neat PLA; R1 to R3 = unmodified CNF (0.5 to 1.5 wt%); R4 to R6 = M1-CNF (APTES:cellulose = 2:1, 0.5 to 1.5 wt%); R7 to R9 = M2-CNF (APTES:cellulose = 4:1, 0.5 to 1.5 wt%).

Elastic Modulus

The tensile modulus results are shown in Fig. 2. Neat PLA exhibited a modulus of 293 ± 13 MPa. The addition of unmodified CNFs (R1–R3) had a limited effect, with the R2 sample showing a modulus of 261 ± 17 MPa, indicating unstable reinforcement due to the absence of strong chemical bonding (Lu et al. 2022). In contrast, mild surface modification (M1-CNF, 2:1 APTES-to-cellulose ratio) in R5 increased the modulus to 321 ± 16 MPa, suggesting effective stress transfer at the interfacial region owing to the formation of silanol linkages and Si–O–Si networks (Yang et al. 2020). Severe modification (M2-CNF, 4:1 ratio) in R7 resulted in a modulus of 545 ± 26 MPa, nearly twice that of neat PLA. However, this increase in stiffness was accompanied by a significant loss of toughness, indicating excessive interphase hardening.

Fig. 2. Elastic modulus (mean ± SD, n = 4) of PLA/CNFs composites (R0–R9) with Duncan’s grouping (letters above bars, α = 0.05)

Elongation at Break

Elongation at break, as an indicator of flexibility, is presented in Fig. 3. Neat PLA exhibited an elongation at break of 7.25% ± 0.93%. The addition of unmodified CNFs slightly increased the elongation, reaching 9.18% ± 0.77% for R3. This improvement is attributed to fiber pull-out and microcrack deflection mechanisms (Khan et al. 2021). In the M1-CNF group (2:1 APTES-to-cellulose ratio), the elongation decreased to 5.25% ± 0.62% in R5, likely due to restricted polymer chain mobility caused by stronger interfacial bonding. In the M2-CNF group (4:1 ratio), this reduction was more pronounced, with R7 showing only 4.04% ± 0.18%, indicating a loss of flexibility resulting from excessive interphase stiffness.

Fig. 3. Elongation at break (mean ± SD, n = 4) of PLA/CNFs composites (R0 to R9) with Duncan’s grouping (letters above bars, α = 0.05)

Impact Strength

Impact strength, an indicator of toughness, is shown in Fig. 4. Neat PLA exhibited a low impact resistance of 3.59 ± 0.23 kJ/m². In the unmodified CNFs group, sample R2 achieved the highest improvement (10.15 ± 0.50 kJ/m²), which was nearly three times that of neat PLA.

Fig. 4. Impact resistance (mean ± SD, n = 4) of PLA/CNFs composites (R0 to R9) with Duncan’s grouping (letters above bars, α = 0.05)

This enhancement is attributed to fiber pull-out and crack deflection mechanisms. In the M1-CNF group (2:1 APTES-to-cellulose ratio), sample R5 recorded an impact strength of 7.72 ± 0.89 kJ/m², which, although lower than R2, provided a balanced combination of strength and toughness along with increased stiffness. In the M2-CNF group (4:1 ratio), impact resistance decreased; particularly, R9 showed 1.14 ± 0.05 kJ/m² and R7 7.08 ± 0.40 kJ/m², indicating that the system’s toughness was highly sensitive to the intensity of surface modification.

Mechanical tests revealed that the quality of the interfacial region was the key factor governing the performance of PLA/CNFs composites. In sample R2 (PLA + 1 wt% unmodified CNFs), weak interfacial adhesion promoted reinforcement through a fiber pull-out mechanism, leading to an increase in impact strength to 10.15 ± 0.50 kJ/m². In R5 (PLA + 1 wt% M1-CNF), improved interfacial bonding enhanced both tensile strength and modulus to 50.2 ± 1.4 MPa and 321 ± 16 MPa, respectively, achieving a desirable balance between stiffness and toughness. Conversely, R7 (PLA + 0.5 wt% M2-CNF) exhibited a higher modulus (545 ± 26 MPa) but lower elongation at break (4.04 ± 0.18%) and impact strength (7.08 ± 0.40 kJ/m²), indicating a more brittle behavior caused by excessive surface modification.

Based on Figs. 1 to 4 and Table 3, R2, R5, and R7 were selected as representative samples for unmodified, mildly modified, and heavily modified CNFs composites, respectively. These samples were subjected to FTIR, TGA, and SEM/EDX analyses to further elucidate the relationship between interfacial structure and mechanical performance.

Pearson correlation analysis was performed on the mean mechanical properties (Table 4). A strong negative correlation was observed between the elastic modulus and elongation at break (r = –0.872, p < 0.001), which was attributed to over-silanization effects in samples R7 to R9. When these formulations were excluded, strong positive correlations emerged among tensile strength, elongation, and impact strength (r = 0.76–0.92), confirming the synergistic reinforcement and optimized interfacial bonding achieved under moderate silanization conditions (e.g., R5). This statistical analysis supports the structure–property relationships discussed earlier.

Table 4. Pearson Correlation Coefficients (r) between Mean Mechanical Properties of PLA/CNF Composites (n = 10 formulations)

FTIR Analysis

To examine interfacial chemical changes and confirm the surface modification of CNFs, FTIR analysis was performed on samples R2 (PLA + 1 wt% unmodified CNFs), R5 (PLA + 1 wt% M1-CNF, APTES ratio 2:1), and R7 (PLA + 0.5 wt% M2-CNF, APTES ratio 4:1). The normalized FTIR spectra are presented in Fig. 5.

All FTIR spectra were normalized to the PLA carbonyl (C=O) stretching peak at ~1750 cm⁻¹ to account for minor differences in CNF loading and ensure that band intensity variations reflect silanization degree, not filler content (Fortunati et al. 2012; Shi et al. 2022).

Fig. 5. Normalized FTIR spectra of PLA/CNFs composites: R2 (PLA + unmodified CNFs), R5 (PLA + M1-CNF, silanized with APTES at a 2:1 ratio), and R7 (PLA + M2-CNF, silanized with APTES at a 4:1 ratio). Key absorption bands are annotated.

Normalized FTIR spectra of R2, R5, and R7 in the range of 400 to 4000 cm⁻¹ were analyzed to assess structural variations induced by CNFs surface modification. A broad band in the 3200 to 3500 cm⁻¹ region, corresponding to O–H stretching vibrations, was observed in all samples. The higher intensity of this band in R7 indicates an increase in polar groups or altered hydrogen bonding interactions as a result of APTES treatment (Hongyu et al. 2024). The peaks near 2900 cm⁻¹ (–CH₂ and –CH₃ stretchings) and 1735 to 1750 cm⁻¹ (C=O stretching) confirm the stability of the PLA backbone across all samples (Fortunati et al. 2012; Moscoso et al. 2019). In R5 and particularly in R7, new bands appeared in the 1500 to 1600 cm⁻¹ region, corresponding to N–H bending vibrations, confirming the covalent bonding of APTES amino groups to CNFs—absent in R2 (Tee et al. 2013). The bands at 1000 to 1100 cm⁻¹, attributed to Si–O–Si and Si–O–C stretching, emerged in the modified samples with greater intensity in R7, suggesting higher silanization efficiency at the 4:1 ratio (Li et al. 2014; Yang et al. 2017). The 1080 to 1180 cm⁻¹ region (C–O–C stretching of PLA) overlapped with the silane-related bands, requiring complementary analysis for precise interpretation (Plackett et al. 2010). R7 (0.5 wt% M2-CNF) was selected, as it exhibited the highest stiffness and lowest toughness among the intense silanization group (4:1 APTES ratio), representing the critical over-silanization condition. Normalized spectra and comparison with R8 (1.0 wt%) confirmed that Si–O–Si intensity increased with APTES ratio, independent of CNF loading. In summary, the appearance of Si–O–Si, Si–O–C, and N–H bands in R5 and R7—particularly their stronger intensity in R7—confirmed the successful chemical modification of CNFs. These spectral changes facilitate improved interfacial adhesion and are expected to enhance the mechanical integrity and structural homogeneity of the composites, consistent with previous studies (Suryanegara et al. 2010; Mohammed et al. 2022).

SEM Analysis

To elucidate the interfacial micromechanisms underlying the mechanical behavior (strength, modulus, elongation, and impact) and the FTIR observations, the surface and fracture morphologies were examined using SEM, while elemental content was obtained from EDX analysis. The objective was to establish a direct correlation between the intensity of APTES silanization, the interphase architecture, and the fracture pattern, in order to explain the observed “stiffness enhancement versus toughness reduction” at different silane loadings. SEM observations focused on R5 (mild silanization, 2:1 ratio) and R7 (intense silanization, 4:1 ratio), since R5 exhibited a balanced interphase and R7 demonstrated a high modulus but increased brittleness. Sample R2 (unmodified CNFs) had been previously characterized, and the morphology of untreated CNFs within the PLA matrix has been well documented; thus, the emphasis here was placed on silanized samples to assess the effect of coating intensity (Kamran et al. 2022).

Surface morphology

At low magnifications (100× and 300×), R5 displayed a smoother surface with a more homogeneous dispersion of the reinforcing phase (Figs. 6a, 6b).

Fig. 6. SEM micrographs of sample R5 (PLA/M1-CNF, silane-to-cellulose ratio 2:1): (a) surface at 100×, (b) surface at 300×, (c) fracture surface at 1000×, and (d) fracture surface at 5000×. These images reveal relatively uniform dispersion of CNFs

Traces of flow-induced alignment were visible, but voids surrounding the fibrils were sparse and small—consistent with improved wettability and reduced surface energy in the presence of a thin silane layer. These observations correspond well with FTIR results (reduced O–H band and appearance of Si–O–Si/Si–O–C vibrations), confirming the formation of chemical bridges and a decrease in interfacial porosity (Panaitescu et al. 2021).

In R7 (Figs. 7a, 7b), a greater presence of surface particles, patchy regions, and irregular microvoids was observed, indicating a thicker and more heterogeneous silane coating. This morphological pattern correlates well with the reduced elongation at break and impact strength, suggesting that excessive silanization led to a brittle interphase structure. In summary, R5 exhibited a cleaner surface with fewer voids, corresponding to better wettability and interfacial compatibility, whereas R7 showed a more spotted and porous morphology, consistent with the formation of a thicker and non-uniform silane layer.