Structure–Property Relationships of Hydroxypropyl Methylcellulose Films Reinforced with Cellulose Nanocrystals Enriched in Cellulose I and Cellulose II from Different Agricultural Residues

Abstract

Converting agricultural waste into cellulose nanocrystals provides a sustainable route for producing biodegradable packaging materials. This study investigated how the crystalline polymorphism of CNCs from different agricultural residues influences the structure–property relationships of hydroxypropyl methylcellulose (HPMC) nanocomposite films. Cellulose I-rich CNCs from durian rind (CNC I) and cellulose II-rich CNCs from sugarcane bagasse (CNC II) were directly compared under identical HPMC film-forming conditions at a fixed CNC loading to clarify their effects on film structure and performance. CNC I and CNC II were characterized by extraction yields of 41.8% and 32.7%, crystallinity indices of 37.4% and 53.1%, and aspect ratios of 18.3 and 10.5, respectively. Dynamic light scattering revealed bimodal apparent hydrodynamic diameters for CNC I (30.8 and 139.8 nm), whereas CNC II showed a single apparent diameter of 102.4 nm. These differences in crystalline structure, morphology, size distribution, aspect ratio, crystallinity, and aggregation behavior contributed to distinct mechanical, barrier, and water-stability properties of the resulting HPMC–CNC films. Compared with neat HPMC, HPMC–CNC I exhibited improved mechanical properties, with tensile strength and Young’s modulus increasing from 34.6 to 53.2 MPa and from 1097.9 to 3302.5 MPa, respectively. In contrast, HPMC–CNC II showed lower oxygen transmission rates of 76.0 and 142.4 cm³·m^–2·day^–1·bar^–1 at 50% and 80% RH, respectively, together with reduced water solubility to 10.8–30.4%. These findings demonstrate that CNCs from agricultural residues can be strategically selected based on their source-dependent structural characteristics to tailor the mechanical, barrier, and water-stability properties of biodegradable HPMC films while supporting the value-added utilization of agricultural residues.

1. Introduction

Over the years, synthetic polymers have been widely used in the packaging industry because of their low cost, ease of processing, and good mechanical and barrier properties. However, most petroleum-based polymers are resistant to degradation and can persist in the environment after disposal, leading to plastic waste accumulation and serious environmental concerns. Consequently, the growing demand for sustainable and biodegradable packaging materials has promoted increasing interest in biopolymer films as alternatives to petroleum-based plastics. , Among biopolymer-based materials, cellulose and cellulose derivatives are particularly promising for sustainable packaging applications because they are derived from renewable resources, are biodegradable, and can be processed into films and coatings with tunable properties. In addition, cellulose derivatives have been reported to be nontoxic and biodegradable, with nonharmful degradation by-products, supporting their suitability for food-contact and pharmaceutical-related applications. In particular, cellulose ethers have been widely investigated due to their excellent film-forming ability and compatibility with various industrial applications. Hydroxypropyl methylcellulose (HPMC) is a prominent cellulose ether that is extensively used in pharmaceutical and biomedical applications because of its biocompatibility, transparency, and flexibility. However, previous comparative studies have shown that cellulose derivatives, including HPMC, exhibit relatively poor oxygen barrier performance compared with other polysaccharides such as chitosan, low-methoxyl pectin, carrageenan, pullulan, cationic starch, and sodium alginate, while their hydrophilic nature also limits the water vapor barrier performance. These limitations indicate that HPMC-based films require further modification to improve their mechanical strength and barrier properties for practical packaging applications. , To overcome these limitations, the incorporation of nanomaterials has proven to be an effective approach for enhancing the performance of biopolymer films. Recent reviews have shown that nanofillers such as nanocellulose, nanoclays, graphene oxide, and metal oxide nanoparticles can improve the mechanical and barrier properties of biodegradable polymer films. These enhancements are mainly associated with the high surface area, high aspect ratio, and possible interfacial interactions of the nanomaterials, which may improve load transfer behavior within the polymer matrix and create more tortuous diffusion pathways for permeating molecules.

Cellulose nanocrystals (CNCs) have attracted increasing attention as sustainable reinforcing agents. They are rod-like nanoparticles typically produced by acid hydrolysis of cellulose and exhibit attractive properties such as high stiffness, low density, high crystallinity, and strong hydrogen-bonding ability with polymer matrices. Accordingly, CNCs have been widely applied in packaging, biomedical, and advanced functional materials. − Previous reports demonstrated that waste-paper-derived CNCs successfully reinforced HPMC/cationic starch films, enhancing their mechanical performance by promoting hydrogen-bond formation between the nanocrystals and the polymer chains. Similarly, Bigi et al. demonstrated that orange peel-derived CNCs effectively reinforced CS/HPMC films by improving tensile strength, light-barrier performance, water vapor barrier properties, and water resistance. Consistently, CNC-reinforced chitosan/HPMC films showed higher tensile strength and elastic modulus, while water absorption decreased from 22% to 9–11%, confirming improved polymer-network cohesion after CNC incorporation. In addition, HPMC films reinforced with bacterial cellulose nanocrystals exhibited increased tensile strength and modulus. These studies indicate that CNC loading strongly affects the final properties of HPMC-based films. Moreover, Larsson et al. showed that nanocellulose type and morphology strongly influence swelling and mass transport in HPMC films by comparing nanofibrillated cellulose (NFC)- and CNC-reinforced systems. However, their study did not address the effects of different CNC polymorphs, particularly cellulose I and cellulose II, on HPMC film properties under the same formulation.

Cellulose used for CNC production exists in different crystalline structures depending on its source and processing conditions. , Native cellulose typically occurs as cellulose I, whereas cellulose II is formed through regeneration and exhibits greater thermodynamic stability. , These polymorphs differ in chain arrangement, hydrogen-bonding networks, and crystalline packing. Cellulose I has a parallel-chain structure that typically yields nanocrystals with high aspect ratios. , In contrast, cellulose II possesses an antiparallel-chain arrangement with denser crystalline packing and higher thermodynamic stability. Such structural differences are expected to influence the reinforcing efficiency of CNCs and their effects on moisture resistance and gas barrier properties. For example, studies on CNC polymorphs in poly­(lactic acid) composites have shown that the crystalline form of CNC significantly affects composite stiffness and barrier behavior, highlighting the importance of polymorph selection in nanocomposite design. These findings suggest that CNC crystalline types (cellulose I or II) can be used to tailor film properties for specific packaging applications. CNCs isolated from different nonwood agricultural residues may vary considerably in polymorphism, morphology, crystallinity, size distribution, aspect ratio, aggregation, and possible surface characteristics. − These variations reflect differences in the initial fiber architecture and cellular organization inherent to each biomass source, as well as differences in biomass chemical composition and the extent of nanofibrillation achieved during processing, which influence the dimensions, aggregation, and dispersion behavior of the resulting nanocellulose. Such source-dependent variation can, in turn, affect their reinforcing behavior within the polymer matrices. Therefore, durian rind and sugarcane bagasse are distinct biomass sources suitable for studying how structurally different CNCs affect HPMC films. This study investigated how cellulose I-rich and cellulose II-rich CNCs, with different source-dependent characteristics, influence the structure and performance of HPMC nanocomposite films. Cellulose I-rich and cellulose II-rich CNCs were extracted from durian rind and sugarcane bagasse fibers, respectively, and incorporated into HPMC films at an identical composition to establish a controlled comparative system. This ratio was selected to enable a direct comparison between the two CNC systems rather than determine the optimum nanofiller loading. The prepared films were comprehensively evaluated in terms of their microstructure, molecular interactions, mechanical behavior, oxygen and moisture barrier properties, sorption characteristics, and water stability. The results provide insight into how the CNC crystal structure and source-dependent characteristics regulate film functionality and establish a basis for optimizing biodegradable HPMC films for packaging applications.

2. Materials and Methods

2.1. Materials

Sodium hydroxide, sodium chlorite, acetic acid, and sulfuric acid, all of AR grade, were purchased from RCI Labscan Ltd. (Bangkok, Thailand). HPMC (Hypromellose 2910, USP/NF grade, viscosity of approximately 4,000 cP in a 2% aqueous solution at 20 °C; Product No. H3785, Sigma-Aldrich, St. Louis, MO, USA) was used as the film-forming polymer. Two types of fibers, from durian rind (Durio zibethinus Murray cv. Monthong) and sugarcane bagasse (Saccharum officinarum L.), were obtained from market waste in Chiang Mai, Thailand. The fibers were dried, ground, and milled into a powder with an approximate particle size of 450 μm.

2.2. Preparation of Celluloses and CNCs

The raw fibers were treated with sodium hydroxide and bleached with sodium chlorite following a method modified from our previous report, as shown in Figure . In brief, 100 g of each fiber was treated in an 18% w/v sodium hydroxide solution using a 20:1 solution-to-fiber ratio and heated at 85 °C for 5 h. The resulting samples were then washed with distilled water. Next, the samples were bleached twice by mixing sodium chlorite, sodium hydroxide, and acetic acid, heating the mixture at 85 °C for 3 h, and then washing with distilled water. The resulting cellulose was dried in a hot-air oven and then sieved.

1.

Schematic illustration of CNC I and CNC II extraction and the fabrication of HPMC-based nanocomposite films. The figure was newly designed and prepared by the authors.

Different types of CNCs were extracted from durian rind (CNC I) and sugarcane bagasse (CNC II) according to the method described by Thipchai et al., with slight modifications (Figure ). Cellulose samples (50 g) were hydrolyzed with 32% (v/v) sulfuric acid at 50 °C for 5 h. The reaction was quenched by adding 5000 mL of distilled water. The suspension was then centrifuged (Biobase, China) and repeatedly washed with distilled water until the pH reached approximately 5.0. Finally, the suspension pH was adjusted to 7.0 using 0.25 M sodium hydroxide solution. The suspension was blended for 30 min with an SBD12A blender (Anitech, Thailand) and subsequently ultrasonicated for 30 min using a BEM-900A ultrasonicator (Bueno-Biotech, China). Finally, the suspension was freeze-dried and stored in a desiccator. The CNC yield was calculated on a cellulose dry-weight basis as the ratio of the freeze-dried mass of purified CNCs to the initial dry mass of cellulose used for acid hydrolysis and was expressed as a percentage using the following eq :

$$\mathrm{CNC\; yield}(\%)=\frac{{\mathit{W}}_{\mathrm{CNC}}}{{\mathit{W}}_{\mathrm{cellulose}}}\times 100$$ 1

where W _CNC is the freeze-dried mass of purified CNCs and W _cellulose is the initial dry mass of cellulose used for acid hydrolysis.

2.3. Preparation of HPMC-Based Nanocomposite Films

HPMC films reinforced with different types of CNCs (HPMC–CNC) were prepared according to the method of Ureña et al. with minor modifications (Figure ). CNC I and CNC II were dispersed in distilled water (1% w/v, 300 mL) and sonicated using an ultrasonicator (Bueno-Biotech, China) in an ice bath for 30 min to obtain homogeneous CNC suspensions. HPMC (1% w/v) was dissolved in 300 mL of distilled water, followed by the addition of glycerol (0.9 g) and mixing for 5 min. The CNC suspension was then combined with the HPMC solution at a CNC:HPMC ratio of 1:1 (w/w) and stirred at 800 rpm and 25 °C for 2 h. The mixture was left overnight to remove entrapped air bubbles. Finally, the casting mixture was poured into square Petri dishes and kept at 25 °C for 24 h to evaporate water and form films with an average thickness of approximately 56 μm. Neat HPMC films without CNC were prepared using the same procedure as the control. After drying, the thickness of each film specimen was measured at five random positions using a digital micrometer gauge (resolution: 0.0001 mm; model 293-130-10, Mitutoyo Corporation, Kawasaki, Japan) before mechanical and barrier property testing. The average measured thickness was used for mechanical and barrier property calculations, and the thickness values were reported as mean ± standard deviation.

2.4. Crystal Structure Characterization

The crystalline characteristics of CNC I and CNC II were analyzed by X-ray diffraction (XRD) using a SmartLab diffractometer (Rigaku, Tokyo, Japan) to evaluate their polymorphic forms. Before measurement, freeze-dried CNC powders were placed on a background sample holder, ensuring a flat and compact surface for accurate diffraction analysis. The diffraction patterns were recorded over a 2θ range of 5°–50° at a scan rate of 2° min^–1. The crystallinity index (X_c ) was calculated following the method reported by Kargarzadeh et al., using eq :

$$X_c(\%)=\frac{A_c-A_a}{A_c}\times 100$$ 2

where A_c represents the total integrated area under the diffraction pattern, and A_a corresponds to the integrated area of the amorphous contribution.

The crystalline interplanar spacings (d-spacings) of the CNCs were calculated from the XRD peak positions using Bragg’s equation (eq ):

$$d_{hkl}=\frac{\lambda }{2\sin \theta }$$ 3

where λ is the wavelength of the incident X-rays (Cu Kα, λ = 1.5406 Å) and θ is half of the diffraction angle (2θ).

Crystallographic models of cellulose I and cellulose II were constructed using crystallographic information files (CIFs) derived from experimentally determined crystal structures reported by Nishiyama et al. and Langan et al. The CIFs were obtained from crystallographic databases or were reconstructed from published crystallographic data following the methodology described by French. Structural visualization and schematic illustration of the cellulose crystal structures were performed using the VESTA software.

2.5. Hydrodynamic Size Distribution and Zeta Potential Analysis of CNC Suspensions

The hydrodynamic size distribution and zeta potential of CNC suspensions were analyzed using dynamic light scattering (DLS) with a Zetasizer system (Malvern Panalytical, UK), which has a detectable size range of 0.6–6000 nm. Mean particle diameters were obtained using DTS software (version 5.00), following Grondahl et al. Zeta potential values were measured using the same instrument and reported as mean ± standard deviation. Each measurement was performed in triplicate for each sample.

2.6. Morphological Characterization of CNCs and HPMC-Based Nanocomposite Films

Transmission electron microscopy (TEM) observations of the CNCs were conducted using a JEM-2100Plus microscope (JEOL, Tokyo, Japan) at an accelerating voltage of 100 kV. CNC suspensions (0.01% w/v) were ultrasonicated for 20 min, dropped onto copper grids, and negatively stained with ammonium molybdate before observation. The surface features of neat HPMC and CNC-reinforced HPMC films were examined by field-emission scanning electron microscopy (FE-SEM; Hitachi, Japan). Film specimens (5 × 5 mm²) were mounted on aluminum stubs with double-sided carbon tape and sputter-coated with a thin gold layer (∼5 nm). SEM images were obtained at magnifications of 2000–15,000×.

2.7. Fourier Transform Infrared Spectroscopy (FTIR) Analysis

The functional groups present in CNCs, neat HPMC, and CNC-reinforced HPMC films were analyzed using FTIR spectroscopy (FT/IR-4700, JASCO, Pfungstadt, Germany). All samples were analyzed using the same FTIR spectrometer in ATR mode. The dried CNCs and film samples, cut into 10 mm × 10 mm pieces, were directly placed on the ATR crystal. FTIR spectra were recorded over the wavenumber range of 4000–400 cm^–1 using 64 scans.

2.8. Thermal Properties

Thermogravimetric analysis (TGA) was performed to evaluate the thermal stability of neat HPMC and CNC-reinforced HPMC films using a TGA/DSC thermal analyzer (Mettler Toledo, Schwerzenbach, Switzerland). Dried samples weighing approximately 5–10 mg were heated from 20 to 500 °C at a heating rate of 10 °C/min under a nitrogen atmosphere with a flow rate of 20 mL/min.

2.9. Mechanical Properties

Uniaxial tensile tests were performed to assess the mechanical performance of the films using a texture analyzer (TA-HD Plus, Stable Micro Systems, UK) under controlled conditions of 25 °C and 50% RH. Tests were performed with a 30-kg load cell at a crosshead speed of 50 mm min^–1. Young’s modulus, tensile strength, and elongation at break were calculated from the stress–strain curves. Film specimens (7 × 2.5 cm²) were conditioned at 25 °C and 50% RH for 48 h before testing, and six to ten replicates were analyzed.

2.10. Oxygen Barrier Properties

The oxygen barrier properties of HPMC-based nanocomposite films were measured using a Gas Transmission Tester (GTT) permeameter (Brugger Feinmechanik GmbH, Munich, Germany) at 25 °C. The instrument was evacuated under primary vacuum to eliminate residual air before testing. One side of the film was then exposed to oxygen gas (27 cm·min^–1), and the pressure increase on the opposite side was recorded. Measurements were carried out under two RH conditions (semi-dry at 50% RH and humid at 80% RH) using a partial pressure difference of 1000 hPa. Samples were conditioned for 48 h at the corresponding humidity levels before testing, and all measurements were performed in triplicate. The oxygen transmission rate (OTR) was calculated from the steady-state gas flow using eq :

$$OTR=\frac{1}{A·\mathrm{\Delta }p}\left(\frac{dV}{dt}\right)$$ 4

where dV/dtrepresents the steady-state volumetric gas flow rate (cm³·day^–1), Ais the effective test area of the film (m²), and Δpis the applied partial pressure difference across the film (bar).

2.11. Water Vapor Permeability (WVP)

WVP was determined by a gravimetric method at 25 °C under RH gradients of 0–50% and 50–100%. Silica gel and distilled water were used to maintain approximately 0% and 100% RH inside the permeability cells. The cells were sealed with silicone O-rings and placed in a climatic chamber at 25 °C and 50% RH. The mass change was monitored using an analytical balance (±0.1 mg), and WVP was calculated from the steady-state WVTR according to eq . Film samples were conditioned at 25 °C and 50% RH for 48 h before testing. All measurements were performed in quadruplicate.

$$WVP=\frac{WVTR\times L}{\mathrm{\Delta }P}$$ 5

Here, L is the film thickness (mm), WVTR is the water vapor transmission rate (g·m^–2·day^–1), and ΔP is the partial water vapor pressure difference across the film (kPa).

2.12. Moisture Sorption Isotherm

Moisture sorption isotherms of HPMC and CNC-reinforced HPMC films were determined using a gravimetric method. Film samples (30 × 30 mm) were dried to constant weight and equilibrated at 25 °C in desiccators containing saturated salt solutions (P₂O₅, LiCl, KC₂H₃O₂, MgCl₂, K₂CO₃, NaBr, SrCl₂, KCl, and KNO₃), providing RH levels from 0 to 93%. Samples were periodically weighed until equilibrium was reached (mass change <0.0005, dry basis). All measurements were conducted in triplicate. After equilibration, the samples were oven-dried at 105 °C for 24 h, and the equilibrium moisture content was then calculated according to eq :

$$X_{w,exp}=\frac{m_{eq}-m_{dry}}{m_{dry}}$$ 6

where X_w,exp is the experimental equilibrium moisture content (g water/g dry solid), m _eq is the equilibrium mass (g), and m _dry is the dry mass (g). For graphical presentation, equilibrium moisture content was expressed as EMC (% dry basis).

The predicted equilibrium moisture content (X_w , _pred ), g water/g dry solid), was calculated using the GAB model, as described by eq :

$$X_{w,pred}=\frac{M_{0}CK{a}_w}{(1-K{a}_w)[1-K{a}_w+CK{a}_w]}$$ 7

where M ₀ is the monolayer moisture content, C is the Guggenheim constant, K is the multilayer correction factor, and a_w is the water activity.

The GAB model parameters were evaluated using nonlinear regression, and the goodness-of-fit was assessed using the coefficient of determination (R ²). The model was fitted to the experimental sorption data over a water activity range of 0.11 ≤ a_w ≤ 0.85. Data points at a_w = 0and a_w = 0.93 were excluded from the fitting procedure, as they fall outside the assumptions of the GAB model. Specifically, the point at a _w= 0 corresponds to residual bound moisture, whereas the high-water-activity point (a _w= 0.93) is associated with polymer swelling and free water uptake, which cannot be adequately described by the GAB model.

2.13. Water Solubility

HPMC-based nanocomposite films (20 × 20 mm) were dried at 45 °C for 24 h and then conditioned at 25 °C and 0% RH for 48 h. Each film specimen (∼0.05 g) was placed in 100 mL of distilled water for 24 h under both agitated conditions (shaking at 500 rpm) and non-agitated conditions. After immersion, the insoluble residues were separated by filtration and subsequently dried at 105 °C for 24 h. Water solubility was calculated using eq . Three replicates of each film were tested,

$$\mathrm{Water\; solubility}(\%)=\frac{W_i-W_f}{W_i}\times 100$$ 8

where W_i and W_f represent the initial dry weight and the final dry weight of the insoluble fraction, respectively.

2.14. Statistical Analysis

Results are presented as mean values with their corresponding standard deviations. Differences among treatments were analyzed using one-way ANOVA, and multiple comparisons were conducted using Tukey’s test at a significance threshold of p < 0.05. Statistical processing was carried out using SPSS software (version 27; IBM Corp., USA).

3. Results and Discussion

3.1. Crystal Structure Characterization

The crystalline structures of CNC I and CNC II were examined using XRD, as shown in Figure a. The XRD patterns confirmed the polymorphic structures of the nanocellulose samples. CNC I revealed peaks at 2θ values of 14.7°, 16.4°, 22.5°, and 34.5°, corresponding to the (1–10), (110), (200), and (004) planes of cellulose I, respectively. In contrast, CNC II showed characteristic peaks at 12.2°, 20.1°, 22.1°, and 34.5°, which were assigned to the (1–10), (110), (200), and (004) crystalline planes of cellulose II, respectively. , Cellulose I is the predominant crystalline form in native cellulose, with glucan chains arranged in a parallel orientation, whereas cellulose II generally forms after regeneration or chemical conversion of cellulose I and is characterized by antiparallel chain packing. Weak reflections at 12.2° and 20.1° in CNC I also suggested the presence of a minor cellulose II phase within the dominant cellulose I structure.

2.

(a) XRD patterns of CNC I and CNC II showing characteristic diffraction peaks corresponding to different crystallographic planes; (b) crystal structure of cellulose I with parallel chain packing; and (c) crystal structure of cellulose II with antiparallel chain packing. The crystal structures were visualized using VESTA software.

The X_c values from XRD analysis indicated the relative proportions of crystalline and amorphous regions within the CNC samples. In the present work, CNC II showed a higher X_c value (53.1%) than CNC I (37.4%), indicating a higher degree of crystalline ordering in the cellulose II-rich CNC sample. The higher apparent crystallinity of CNC II may be associated with structural rearrangements during chemical regeneration, which promote antiparallel chain packing and a reorganized hydrogen-bonding network. In contrast, the lower X_c value of CNC I suggests that some amorphous regions remained after sulfuric acid hydrolysis. In addition, the higher apparent crystallinity of cellulose II compared with cellulose I is consistent with previous studies using the Segal method, which reported higher apparent crystallinity for cellulose II than for cellulose I.

The d-spacing values calculated from the experimental peak positions (as described in the XRD analysis section) supported the identification of polymorphs, as shown in Figure b,c. For CNC I, the d-spacing values for the (1–10), (110), and (200) planes were 0.602, 0.540, and 0.395 nm, respectively. These values reflect the compact packing associated with the parallel chain orientation of native cellulose I. Conversely, CNC II exhibited a larger interplanar spacing for the (1–10) plane (0.725 nm) than CNC I, which is characteristic of antiparallel chain packing and structural rearrangement during cellulose regeneration. However, the d-spacing of the (110) plane in CNC II (0.441 nm) was smaller than that of CNC I, indicating contraction along this crystallographic direction due to reorganization of the hydrogen-bonding network in cellulose II. The d-spacing values of the (200) plane in CNC I (0.395 nm) and CNC II (0.402 nm) were similar, indicating that this plane is relatively insensitive to polymorphic transformation. In summary, the expansion of the (1–10) plane together with the contraction of the (110) plane support the transition from the parallel chain packing of cellulose I to the antiparallel arrangement of cellulose II, consistent with molecular dynamics simulations of regenerated cellulose.

3.2. CNC Yield, Hydrodynamic Size Distribution, Zeta Potential, and Morphology of CNCs

The CNC yields of CNC I and CNC II were 41.8 ± 0.7% and 32.7 ± 0.6%, respectively (Table ), based on the initial dry mass of cellulose used for acid hydrolysis. Although CNC II exhibited a higher crystallinity index than CNC I (Figure ), its lower CNC yield indicates that CNC recovery was not governed solely by crystallinity. Instead, the observed yield difference likely arises from a combination of factors, including the biomass origin, cellulose polymorphic form, cellulose-chain accessibility, and the differential vulnerability of specific cellulose domains to acid attack. This interpretation is consistent with the general understanding that amorphous regions are preferentially hydrolyzed under acidic conditions, whereas crystalline domains remain largely intact. The yields reported here fall within the range documented for CNCs derived from other lignocellulosic feedstocks. For instance, Deepa et al. reported yields of 28.6% to 40.1% using biomass sources such as banana rachis, sisal, kapok, pineapple leaf, and coir. In addition to CNC yield, hydrodynamic size distribution and surface charge were analyzed to further understand the suspension behavior of CNC I and CNC II.

1. CNC Yield Calculated on a Dry-Weight Basis and Dimensions of CNCs Determined from TEM Images .

Type of CNCs	CNC yield (%)	Diameter (nm)	Length (nm)	Aspect ratio (Length/Diameter)
CNC I	41.8 ± 0.7	12.0 ± 4.2	219.6 ± 53.8	18.3
CNC II	32.7 ± 0.6	14.1 ± 4.2	148.2 ± 20.9	10.5

^aData are presented as mean ± standard deviation. Dimensions were measured from 100 nanocrystals using TEM images, including additional micrographs not shown in the manuscript. Length was measured only for fully visible, clearly distinguishable, and nonoverlapping nanocrystals.

The zeta potential values of CNC I and CNC II were −52.2 ± 3.9 and −61.3 ± 2.3 mV, respectively, indicating negatively charged CNC surfaces in aqueous suspension. These highly negative values are consistent with electrostatic repulsion among CNC particles and colloidal stability in suspension. Previous studies have also reported that higher absolute zeta potential values are associated with stronger interparticle repulsion and improved dispersion stability of nanocellulose suspensions. However, TEM (Figure b) observations showed some bundled or aggregated nanocrystals, which may be partly associated with drying during sample preparation. Such aggregation or bundling may occur during TEM sample drying, where water evaporation on the grid can promote capillary-force-induced assembly and CNC–CNC hydrogen bonding. ,

3.

(a) DLS hydrodynamic size distributions with photographs of CNC suspensions and (b) TEM images of CNC I and CNC II at ×20,000 and ×50,000 magnifications.

The hydrodynamic size distribution of CNC suspensions, measured by DLS, is shown in Figure a. CNC I exhibited a bimodal distribution with hydrodynamic diameters of 30.8 and 139.8 nm, whereas CNC II showed a single peak with a hydrodynamic diameter of 102.4 nm. The bimodal distribution of CNC I indicates a heterogeneous suspension containing both smaller nanocrystal populations and larger hydrodynamic domains, which may be associated with the partial aggregation of CNC structures. This variation in the particle size distribution may be linked to the initial fiber architecture, biomass source, and extent of fibrillation during sulfuric acid treatment. , However, DLS was used to determine the apparent hydrodynamic size distribution and possible aggregation behavior of CNC suspensions, whereas the actual CNC diameters, lengths, and aspect ratios were determined from TEM images.

The morphologies of CNCs from different sources were examined by TEM. Both samples exhibited the typical rod-like morphology of CNCs produced by acid hydrolysis. CNC I displayed long, needle-like nanocrystals that were densely packed and interconnected, forming bundles and network-like aggregates, which may be associated with the parallel chain packing of cellulose I and possible interparticle interactions between nanocrystals. This observation is consistent with the bimodal DLS data and suggests that CNC I exhibited a heterogeneous size distribution. In contrast, CNC II showed shorter nanocrystals with a more loosely entangled and dispersed morphology. This is consistent with the antiparallel chain arrangement of cellulose II, which modifies the hydrogen-bonding network and reduces interparticle attraction. However, differences in aggregation and dispersion were observed between the two CNC types. In particular, CNC I exhibited a distinctive morphology at various locations in the TEM images, consisting of both short and long nanofibers with small diameters. This structural heterogeneity may be attributed to the complex anatomical structure of the durian rind.

The dimensional characteristics of the CNCs determined from TEM images, including length and diameter, are summarized in Table . The average diameter of CNC I was 12.0 ± 4.2 nm, which was smaller than that of CNC II (14.1 ± 4.2 nm). In contrast, CNC I exhibited a greater average length (219.6 ± 53.8 nm) than CNC II (148.2 ± 20.9 nm). Consequently, CNC I showed a higher aspect ratio (18.3) than CNC II (10.5). In comparison, a previous study reported that onion-residue-derived CNCs had an average diameter of approximately 5.6 nm, an average length of 341 nm, and an aspect ratio of approximately 61. These differences may be attributed to variations in biomass source, chemical composition, extraction method, and hydrolysis conditions. Nevertheless, the higher aspect ratio of CNC I increased its effective surface contact with the HPMC matrix and provided more accessible hydroxyl groups for interaction. In contrast, the more organized crystalline domains of CNC II contributed to greater crystallinity and a reduced swelling tendency. These differences indicate that CNC hydration and rheological behavior are governed by biomass source, processing conditions, crystal structure, and surface chemistry. This interpretation is consistent with recent research on waste-derived cellulose and chitin nanofibers, which showed that feedstock origin and processing route strongly influenced nanofiller morphology, crystallinity, surface chemistry, and reinforcement performance. These morphological differences agree with the XRD results (Figure a), confirming the presence of cellulose I and cellulose II structures in CNC I and CNC II, respectively. Consequently, these structural variations are expected to influence the reinforcing efficiency and microstructural uniformity of CNC-reinforced HPMC films.

3.3. FTIR Analysis

The FTIR spectra of CNCs, HPMC, and HPMC-based composite films reinforced with CNC I and CNC II are shown in Figure , while the assignments of the characteristic bands are summarized in Table . The chemical purification process, including alkaline treatment and bleaching, removed non-cellulosic components such as hemicellulose and lignin before CNC formation through acid treatment. This treatment selectively degraded the amorphous cellulose regions because crystalline domains show greater resistance to acid attack. , However, FTIR analysis confirmed that the cellulose chemical structure was maintained after processing, as evidenced by the characteristic C–H stretching band at 2890 cm^–1. The bands at 3490 and 3440 cm^–1 were assigned to O–H stretching vibrations characteristic of cellulose I and cellulose II, respectively, reflecting differences in hydrogen bonding within the crystalline regions. These spectral features support the structural distinction between cellulose I and cellulose II, as illustrated in Figure . In cellulose I, the parallel chain arrangement is mainly stabilized by intramolecular hydrogen bonds, including O3H···O5 and O2H···O6, whereas cellulose II, which adopts an antiparallel chain arrangement, exhibits a different hydrogen-bonding network with dominant intermolecular O6H···O2 interactions. The band near 1635 cm^–1 is mainly attributed to the bending vibration of adsorbed water molecules (H–O–H) within the hydrophilic cellulose structure.

4.

FTIR spectra of CNCs, HPMC, and HPMC-based composite films reinforced with CNC I and CNC II (HPMC–CNC I and HPMC–CNC II).

2. FTIR Band Assignments of CNCs, HPMC, and HPMC-CNC Films.

Wavenumber (cm–1)	Chemical assignment	References
3490, 3440	Hydrogen-bonded O–H groups in cellulose I and cellulose II
2890	C–H stretching vibration of cellulose, indicating preservation of the cellulose backbone after processing	,
1635	H–O–H bending vibration of absorbed water associated with the hydrophilic cellulose structure
1375	C–H bending vibration of cellulose and HPMC
1160–1110	Asymmetric C–O–C stretching vibration of glycosidic linkages in cellulose	,
1056, 1030	C–O stretching vibration of polysaccharide chains, including HPMC and cellulose
950	Band associated with the HPMC backbone vibration
895	β-glycosidic linkage vibration of cellulose, indicating preservation of the cellulose backbone	,

5.

Schematic representation of the hydrogen-bonding arrangements and chain packing in cellulose I and cellulose II.

The persistence of this band across all samples reflects the hygroscopic nature of cellulose and HPMC. Characteristic cellulose bands appear at 1160–1110 cm^–1 (C–O–C stretching) and 1050–1030 cm^–1 (C–O stretching). The β-glycosidic band at 895–950 cm^–1 indicates the preservation of the main cellulose backbone after acid treatment. , The absorption band at 1375 cm^–1 corresponds to C–H bending vibrations of cellulose and its derivatives, including HPMC. The removal of hemicellulose and lignin leads to a dominant cellulose phase with two C–O stretching bands at 1056 and 1030 cm^–1, whereas HPMC shows only the 1056 cm^–1 band. The emergence of a band at 1030 cm^–1 after the incorporation of CNCs into HPMC corresponds to C–O stretching of cellulose, confirming nanocomposite formation. In addition, the shift of the O–H stretching band to lower wavenumbers may indicate possible intermolecular hydrogen-bonding interactions in the HPMC–CNC composite films.

3.4. Thermal Properties

The thermal degradation behavior of neat HPMC and CNC-reinforced HPMC films was evaluated using TGA/DTG analysis, as shown in Figure . All films exhibited minor weight loss at low temperatures (50–130 °C), mainly due to the evaporation of absorbed moisture and volatile components. A similar initial weight loss has been reported for starch-based composites reinforced with CNC, in which the first degradation stage was attributed to water evaporation and solvent removal from the polymer matrix. , In the present study, the major thermal degradation occurred in the range of approximately 325–383 °C, corresponding to the decomposition of the HPMC matrix and cellulose-based components. Neat HPMC exhibited a maximum degradation temperature (T _max) of 361.42 °C, while HPMC–CNC I and HPMC–CNC II showed comparable T _max values of 361.11 and 362.16 °C, respectively. Although HPMC–CNC II presented the highest T _max, the small differences among the films indicate that the main thermal degradation behavior was primarily governed by the HPMC matrix and was not markedly affected by CNC incorporation. This result is consistent with previous reports on HPMC/nanocellulose nanocomposite films, in which the incorporation of nanocellulose did not drastically alter the main thermal degradation profile of the HPMC-based matrix but influenced the degradation stages and residue formation. Similar behavior was also reported for chitosan–CNC composite films, in which cellulose-related degradation peaks appeared between 314 and 364 °C after CNC addition, while the overall degradation temperature of the composites remained nearly unchanged at approximately 279 °C. Although the onset degradation temperatures (T _onset) of the CNC-reinforced films were slightly lower than that of neat HPMC, the residual weight at 500 °C increased from 8.24% for neat HPMC to 11.05% and 12.06% for HPMC–CNC I and HPMC–CNC II, respectively. This increase indicates a higher residual mass at 500 °C following CNC incorporation. The higher residual weight of the CNC-reinforced films may be associated with the presence of crystalline cellulose domains and possible hydrogen-bonding interactions between HPMC chains and CNC surfaces, as supported by the FTIR results (Figure ). In addition, previous studies have reported that the thermal residue of nanocellulose-based materials can be influenced by cellulose crystallinity, surface characteristics, raw material origin, and matrix–filler interactions, which may affect the thermal degradation pathway and the amount of residual mass after thermal degradation. , Among the composite films, HPMC–CNC II showed the highest residual weight at 500 °C. This result may be associated with the higher crystallinity and distinct chain packing arrangement of cellulose II-rich CNCs, which could contribute to the higher residual mass within the HPMC matrix. Previous studies have reported that cellulose polymorphs and crystallinity can influence the thermal degradation behavior and final residue of CNCs. In addition, cellulose II is known to possess an antiparallel chain arrangement and a distinct hydrogen-bonding network compared with cellulose I. The TGA/DTG results suggest that CNC incorporation did not markedly alter the main thermal degradation behavior of HPMC films but increased the residue after degradation. These findings indicate that CNC I and CNC II can act as thermally compatible reinforcing nanofillers for HPMC-based films. ,

6.

Thermal characteristics of HPMC-based films: (a) TGA curves and (b) DTG curves of HPMC, HPMC–CNC I, and HPMC–CNC II films.

3.5. Mechanical Properties

Before evaluating the mechanical and barrier properties, the film thickness was measured. The average thicknesses of HPMC, HPMC–CNC I, and HPMC–CNC II films were 55.8 ± 2.8, 55.9 ± 1.8, and 55.9 ± 0.8 μm, respectively. The comparable thickness values indicate that the observed differences in mechanical and barrier properties were mainly associated with CNC incorporation rather than film thickness variation. The mechanical properties of HPMC-based composite films (HPMC–CNC I and HPMC–CNC II) are shown in Figure . The tensile strength, elongation at break, and Young’s modulus of the nanocomposite films depended on the CNC type. The addition of CNCs significantly improved the tensile strength and modulus of HPMC-based films (Figure a,b). Pure HPMC showed the lowest tensile strength (34.6 MPa) and Young’s modulus (1097.9 MPa), whereas HPMC–CNC I reached the highest values (53.2 and 3302.5 MPa). HPMC–CNC II also exhibited improvements (48.2 and 2763.3 MPa) but remained lower than CNC I. Conversely, elongation at break (Figure c) decreased markedly for both CNC-reinforced films (3.6–4.9%), indicating increased brittleness. These results demonstrate that CNC I is a more effective reinforcing agent within the HPMC matrix than CNC II. Although CNC II had higher crystallinity based on XRD analysis (Figure a), the improvement in the mechanical properties of HPMC was lower than that of CNC I. This behavior may be attributed to the lower aspect ratio of CNC II and reduced polymer-chain mobility, leading to increased brittleness and lower elongation at break. , This trend is consistent with previous findings for chitosan–CNC composite films, where CNC incorporation increased crystallinity and improved tensile strength and Young’s modulus, but reduced elongation at break. Similar behavior was reported for PVA-CNC films, where tensile strength increased from 55.96 to 63.75 MPa, accompanied by a slight decrease in elongation at break. In CNC-reinforced starch-based films, higher CNC contents also reduced elongation at break because of CNC aggregation and restricted polymer-chain mobility. Furthermore, Juncus effusus-derived CNCs with higher aspect ratio, crystallinity, and surface charge provided stronger reinforcement in κ-carrageenan films. At 8 wt % loading, the tensile strength increased from 23.43 to 39.89 MPa, corresponding to an improvement of approximately 70%.

7.

Mechanical response of the films: (a) tensile strength, (b) Young’s modulus, and (c) elongation at break of the nanocomposite films. Different lowercase letters above the bars in each panel (a–c) indicate significant differences among samples (p ≤ 0.05).

The difference in mechanical properties between HPMC–CNC I and HPMC–CNC II suggests that reinforcement efficiency depends not only on crystallinity but also on particle morphology and possible CNC–matrix interactions. TEM analysis showed that CNC I possessed a smaller diameter and a higher aspect ratio than CNC II (Figure and Table ), which may favor the formation of a reinforcing network within the HPMC matrix. This network may promote more effective load-transfer behavior and contribute to mechanical reinforcement. Moreover, the improved mechanical properties of HPMC–CNC I films may be associated with the combined effects of a cellulose I-rich crystalline structure, CNC morphology, high aspect ratio, and possible CNC–HPMC interactions. The cellulose I-rich structure of CNC I may support reinforcing-network formation and possible load-transfer behavior along the nanocrystal axis. , In addition, the cellulose II-rich structure may contribute to CNC–HPMC compatibility through local chain packing and possible hydrogen-bonding interactions (Figure b,c). , Overall, CNC incorporation improved the mechanical performance of HPMC films. HPMC–CNC I increased tensile strength by 53.8% and Young’s modulus by 201% compared with neat HPMC, consistent with previous studies reporting significant improvements in strength and modulus with CNC addition. Similarly, Bilbao-Sainz et al. observed an approximately 22% increase in tensile strength in HPMC films reinforced with cellulose whiskers, which the authors attributed to improved load transfer and possible interfacial hydrogen-bonding interactions between the HPMC matrix and the rigid CNCs. Collectively, the XRD, TEM/DLS, FTIR, SEM, and mechanical results reveal a structure–property relationship in the HPMC–CNC films. Differences in CNC crystallinity, morphology, aspect ratio, aggregation behavior, and possible CNC–HPMC interactions help explain the superior mechanical reinforcement observed in HPMC–CNC I films.

3.6. Morphology of HPMC-Based Nanocomposite Films

The surface and cross-sectional morphologies of HPMC-based nanocomposite films incorporating different types of CNCs were investigated by SEM, as presented in Figure . HPMC films (Figure a,b) exhibited smooth surfaces with dense cross-sectional structures, indicating homogeneous polymer packing. After CNC incorporation, the HPMC–CNC I film showed a more homogeneous morphology than the HPMC–CNC II film. CNC I was distributed more uniformly in the HPMC matrix without noticeable agglomeration on either the film surface or in the cross-section (Figure c,d). This uniform morphology is consistent with the higher tensile strength and Young’s modulus of the HPMC–CNC I film (Figure ), suggesting better CNC–HPMC compatibility. Conversely, HPMC–CNC II films showed rougher surfaces with noticeable pores and agglomeration (Figure e,f). The cross-sectional images revealed a less compact internal structure, suggesting lower compatibility between CNC II and the HPMC matrix. These SEM observations agree with the morphological trends reported by Bilbao-Sainz et al. for HPMC films reinforced with cellulose nanostructures, where effective reinforcement was associated with more uniform dispersion and improved compatibility with the HPMC matrix. In contrast, heterogeneous dispersion and aggregation resulted in rougher surfaces and reduced film uniformity. Similarly, high CNC concentrations can promote particle agglomeration due to strong CNC–CNC interactions, reducing flexibility and decreasing elongation (Figure c). Similar embrittlement behavior has also been reported in CNC-reinforced epoxy adhesives, where increased CNC loading resulted in reduced toughness and strength due to nanoparticle agglomeration and localized stress concentration.

8.

SEM images of HPMC-based films: (a, b) HPMC, (c, d) HPMC–CNC I, and (e, f) HPMC–CNC II, showing surface morphology (a, c, e) and cross-sectional morphology (b, d, f).

3.7. Oxygen Barrier Properties

The OTR of HPMC and CNC-reinforced HPMC films at 50% and 80% RH is shown in Figure a. Pure HPMC exhibited the highest OTR values of 284.7 and 416.3 cm³·m^–2·day^–1·bar^–1 at 50% and 80% RH, respectively, indicating weaker oxygen-barrier performance compared with the composite films. This behavior is consistent with previous reports showing that cellulose derivatives such as HPMC exhibit relatively high oxygen permeability due to bulky side groups that increase the fractional free volume of the polymer matrix and facilitate gas transmission. Furthermore, methyl groups (−CH₃) increase the polymer-oxygen affinity, promoting oxygen transport through the film. This hydrophilic nature of HPMC also contributes to microstructural loosening, especially under humid conditions, resulting in limited resistance to gas permeation. , However, CNC incorporation reduced oxygen permeability, resulting in much lower OTR values than pure HPMC films. Notably, HPMC–CNC II films showed numerically lower OTR values than HPMC–CNC I films, at 76.0 versus 85.0 cm³·m^–2·day^–1·bar^–1 at 50% RH, and 142.4 versus 176.0 cm³·m^–2·day^–1·bar^–1 at 80% RH. This oxygen-barrier tendency of HPMC–CNC II may be associated with its higher crystallinity (Figure a), which can increase transmission resistance by creating more tortuous pathways for oxygen transport. , In addition, the antiparallel chain packing and higher crystallinity of CNC II may reduce swelling and restrict polymer-chain mobility (Figure b,c). thereby increasing diffusion-path tortuosity. Therefore, HPMC–CNC II films showed lower OTR values despite the less homogeneous morphology observed by SEM (Figure ). ,, An increase in RH from 50% to 80% led to higher OTR values for all films due to the plasticizing effect of absorbed moisture, which increases polymer-chain mobility and free volume.

9.

Barrier properties of HPMC and CNC-reinforced HPMC films: (a) OTR measured at 50% and 80% RH and (b) WVP measured under 0–50% and 50–100% RH gradients. Results are expressed as mean ± SD; different letters denote significant differences at p < 0.05.

3.8. WVP

Figure b presents the WVP of HPMC, HPMC–CNC I, and HPMC–CNC II films measured under two RH ranges (0–50% RH and 50–100% RH). WVP increased markedly when the RH range increased from 0–50% to 50–100% for all films. Under low-humidity conditions (0–50% RH), pure HPMC exhibited the highest WVP (1.7 × 10^–6 g·mm·h^–1·m^–2·kPa^–1), while the incorporation of CNC I and CNC II reduced the values to 1.2 and 1.3 × 10^–6 g·mm·h^–1·m^–2·kPa^–1, respectively. This reduction indicates that CNCs may hinder water vapor permeation by increasing crystalline domains, restricting polymer-chain mobility, and creating more tortuous diffusion pathways for water molecules. At high humidity (50–100% RH), WVP increased for all samples due to moisture absorption by the HPMC matrix. Pure HPMC showed the highest WVP (7.7 × 10^–6 g·mm·h^–1·m^–2·kPa^–1), whereas HPMC–CNC I and HPMC–CNC II films exhibited WVP values of 6.4 and 6.5 × 10^–6 g·mm·h^–1·m^–2·kPa^–1, respectively. Comparing the two types, HPMC–CNC I and HPMC–CNC II showed comparable reductions in WVP, particularly under high-humidity conditions, and the difference between the two samples was not statistically significant. This behavior contrasts with the oxygen barrier results (Figure a), where CNC II showed numerically lower OTR values. This difference arises from the different transport mechanisms of water vapor and oxygen. While oxygen permeation is mainly governed by crystallinity and diffusion-path tortuosity, water vapor transport is additionally influenced by hydrophilicity and moisture-polymer interactions. In addition, the abundant hydroxyl groups on CNC surfaces interact with water molecules, weakening the barrier effect at high RH. Overall, the XRD (Figure ), SEM (Figure ), OTR, and WVP results indicate that barrier properties are controlled by multiple structural factors rather than by SEM morphology alone. CNC incorporation improved oxygen and water vapor barrier properties mainly by increasing crystalline domains, restricting polymer-chain mobility, and creating more tortuous diffusion pathways.

3.9. Moisture Sorption Isotherm

The moisture sorption isotherms of HPMC and CNC-reinforced HPMC films as a function of RH at 25 °C are shown in Figure a. All films exhibited a typical sigmoidal (Type II) sorption behavior, which is characteristic of hydrophilic polymeric materials such as HPMC. This behavior indicates initial monolayer adsorption at low water activity, followed by multilayer water uptake at higher humidity levels. Furthermore, the Type II moisture sorption behavior observed in CNC-reinforced films agrees well with previous reports on CNC. HPMC showed the highest equilibrium moisture content (EMC) over the entire RH range, particularly above 75%. The high EMC of HPMC was attributed to its intrinsic hydrophilicity, , consistent with the higher WVP observed at elevated RH (Figure b). The moisture sorption observed in this study is in agreement with previous reports on cellulose-derivative films. As reported by Choi et al. and Su et al., HPMC films exhibit relatively high sensitivity to humidity due to their hydrophilic nature and the presence of bulky substituent groups. In addition, glycerol plasticization can further loosen the microstructure, thereby enhancing moisture absorption and gas permeability, and increasing the affinity of HPMC films for water. The humidity-dependent response of polysaccharide–nanocellulose systems depends on polymer-network organization and the physicochemical characteristics of the constituent polymers, which influence water sorption and moisture-induced changes in material properties. Meanwhile, the incorporation of CNCs reduced the EMC of the films, indicating improved resistance to moisture absorption. HPMC–CNC II exhibited the lowest moisture content, followed by HPMC–CNC I. This result indicates that CNC II is more effective in restricting moisture absorption, which may be associated with its higher crystallinity and lower swelling tendency rather than with the less homogeneous morphology observed by SEM (Figure ). , The GAB model parameters are presented in Table . The HPMC film exhibited the highest monolayer moisture content (M ₀ = 0.2225 g water/g solid). In contrast, CNC incorporation significantly reduced the M ₀ values to 0.1110 and 0.0884 g water/g solid for HPMC–CNC I and HPMC–CNC II, respectively. The lower M ₀ values indicate that the film has fewer sites available for water adsorption. The addition of CNCs may reduce the number of accessible hydrophilic sites and restrict water access through increased crystalline domains, CNC–HPMC interactions, and diffusion-path tortuosity. According to the GAB model, moisture sorption involves monolayer adsorption at low RH. At higher RH, multilayer water uptake occurs, driven by strong water-cellulose interactions arising from hydroxyl-rich CNC surfaces. A similar decrease in moisture sorption has been reported for cellulose nanoparticle-reinforced films by Bilbao-Sainz et al., supporting the result that CNC incorporation reduces the number of accessible primary sorption sites. The GAB constant C is related to the strength of the interaction between water and polymer at monolayer sites. This parameter increased upon CNC addition (C = 468.94–751.20) compared with neat HPMC (C = 215.29). This increase suggests stronger binding of the first layer of water molecules, likely due to hydrogen bonding between water and well-organized CNC surfaces. Meanwhile, the K values (0.7406–0.7852) remained within the typical range reported for biopolymer films, indicating physically reasonable multilayer sorption behavior. The high coefficients of determination (R ₂> 0.96 for HPMC-based nanocomposite films) confirm that the GAB model adequately describes the experimental sorption data. ,

10.

Moisture-related properties of HPMC and CNC-reinforced HPMC films: (a) EMC as a function of RH at 25 °C and (b) water solubility under non-agitated and agitated conditions. Mean ± SD is reported; different letter labels in panel (b) represent significant differences among treatments (p< 0.05).

3. GAB Model Parameters for the Moisture Sorption Isotherms of HPMC Films with and without Cncs .

Film	M 0 (g water/g solid)	C	K	R 2
HPMC	0.2225	215.29	0.5987	0.8722
HPMC–CNC I	0.1110	751.20	0.7406	0.9692
HPMC–CNC II	0.0884	468.94	0.7852	0.9749

^aGAB fitting was conducted using data in the aw range of 0.11–0.85.

3.10. Water Solubility

The water solubility of HPMC, HPMC–CNC I, and HPMC–CNC II films under nonagitated and agitated conditions is shown in Figure b. HPMC films exhibited very high solubility, ranging from 92.5% (non-agitated) to 95.9% (agitated), reflecting the hydrophilic nature of HPMC and its strong affinity for water. In contrast, CNC-reinforced films showed markedly lower solubility under both conditions. HPMC–CNC I exhibited solubility values of 13.3% (non-agitated) and 37.2% (agitated), while HPMC–CNC II showed the lowest values of 10.8% and 30.4%, respectively. This reduction may be attributed to the combined effects of CNC incorporation, the lower water solubility of the CNC fraction, increased crystalline domains, and CNC–HPMC interactions, which together restrict water penetration and film disintegration. This result is consistent with the WVP and moisture sorption isotherm results in Figures b and a. Notably, HPMC–CNC II was less soluble than HPMC–CNC I, possibly owing to its more organized crystalline structure, as shown by the XRD results (Figure ), which may limit water penetration. In addition, possible hydrogen-bonding interactions between HPMC chains and CNC surfaces may contribute to reduced chain mobility and suppressed film disintegration in aqueous environments. Agitation increased the solubility of all samples because it enhanced water penetration and accelerated mass transfer by promoting mechanical disruption and water diffusion within the polymer matrix. However, CNC-reinforced films exhibited lower water solubility than pure HPMC, indicating improved structural stability even under dynamic conditions. These results suggest that CNC incorporation, particularly CNC II, improves the water resistance of HPMC-based films. However, this improvement should be interpreted as the combined effect of CNC crystallinity, CNC–HPMC interactions, and the lower solubility of the CNC fraction.

4. Conclusions

This study demonstrates that the structure and functional properties of HPMC-based nanocomposite films are governed not only by CNC polymorphism but also by biomass source, morphology, size distribution, aspect ratio, crystallinity, aggregation behavior, and possible CNC–matrix interactions. Cellulose I-rich CNCs from durian rind (CNC I) and cellulose II-rich CNCs from sugarcane bagasse (CNC II) showed extraction yields of 41.8% and 32.7%, crystallinity indices of 37.4% and 53.1%, and aspect ratios of 18.3 and 10.5, respectively. CNC I exhibited a higher aspect ratio and more homogeneous dispersion in the HPMC matrix, which contributed to improved mechanical reinforcement. Compared with neat HPMC, HPMC–CNC I increased tensile strength from 34.6 to 53.2 MPa and Young’s modulus from 1097.9 to 3302.5 MPa, indicating more effective reinforcement and possible load transfer within the HPMC matrix. In contrast, CNC II exhibited higher crystallinity and a cellulose II-rich structure, which may contribute to more compact chain packing and restricted molecular transport. As a result, HPMC–CNC II showed lower OTR values of 76.0 and 142.4 cm³·m^–2·day^–1·bar^–1 at 50% and 80% RH, respectively, reduced water solubility by 10.8–30.4%, and the lowest monolayer moisture content (M ₀ = 0.0884 g water/g solid). Both CNC I and CNC II reduced water vapor permeability. However, under high relative humidity, the barrier effect was partially reduced because moisture-induced plasticization of the HPMC matrix increased chain mobility and facilitated water vapor transport. Thermal analysis showed that CNC incorporation did not markedly change the main decomposition temperature, with comparable T _max values of 361.42, 361.11, and 362.16 °C for neat HPMC, HPMC–CNC I, and HPMC–CNC II, respectively. However, the residual weight at 500 °C increased from 8.24% for neat HPMC to 11.05% and 12.06% for HPMC–CNC I and HPMC–CNC II, respectively, indicating a higher residual mass following CNC incorporation. Overall, CNC I was more effective for mechanical strengthening, whereas CNC II was more favorable for oxygen barrier performance, moisture resistance, and water stability. These findings highlight the importance of selecting agricultural-residue-derived CNCs with suitable source-dependent and polymorphic characteristics to design biodegradable HPMC-based packaging films with balanced mechanical, barrier, and water-stability properties.

Conceptualization, Data curation, Formal analysis, P.T. and P.R.; Funding acquisition, Investigation, Resources, Project administration, Supervision, P.R.; Methodology, P.T., T.K., and P.R.; Software, P.T. and P.R.; Validation, W.P., T.K., K.J., J.S., P.T., and P.R.; Visualization, P.T. and P.R.; Writing – original draft, P.T. and P.R.; Writing – review and editing, W.P., T.K., K.J., J.S., P.R., and P.T.; All authors have read and agreed to the published version of the manuscript.

The authors declare no competing financial interest.