Valorization of Oil Palm Empty Fruit Bunch for Cellulose Fibers: A Reinforcement Material in Polyvinyl Alcohol Biocomposites for Its Application as Detergent Capsules

Abstract

Cellulose fibers isolated from oil palm empty fruit bunches (OPEFB) have been studied as a potential reinforcement for polyvinyl alcohol (PVA) biocomposite. Analysis of variance (ANOVA) showed that all three parameters—hydrolysis temperature, time and acid concentration, as well as their interactions—significantly affected the yield of cellulose. Moving Least Squares (MLS) and Multivariable Power Least Squares (MPLS) models demonstrated good fitness. The model also proved that acid concentration was the dominant parameter, supported by the Fourier transform infrared spectroscopy (FTIR) analysis. Hydrolysis using 54% acid at 35 °C and 15 min achieved the highest cellulose yield of 80.72%. Cellulose-reinforced PVA biocomposite films demonstrated better mechanical strength, elongation at break, moisture barrier properties, thermal stability and poorer light transmission rate compared to neat PVA due to the high aspect ratio, crystallinity and good compatibility of cellulose fibers. These findings suggested the potential of cellulose fibers-reinforced PVA biocomposite film as water-soluble detergent capsules.

1. Introduction

Over the years, the demand for packaging materials has been rising rapidly due to the growing population worldwide and the demand for more convenient products with extended shelf life. Plastic materials offer outstanding qualities such as lightweight, hygienic, high strength, versatile, durable and flexible, making them the first choice for most packaging applications. Nevertheless, plastics are generally non-biodegradable and usually take thousands of years to decay. The associated environmental issues such as plastic pollution [1], increased emissions of greenhouse gasses [2] and global warming [3] have encouraged the development of “greener” or environmental-friendly materials to replace conventional petroleum-based plastics. These materials are usually biodegradable. In addition, they are extracted from renewable biomass sources such as crops and lignocellulosic biomass, promoting sustainability.

Polyvinyl alcohol (PVA) is a colorless polymer that is water-soluble and highly biodegradable. It is widely used in industries with good commercial value, given several advantages, such as good film formation, strong conglutination, high thermal stability and flexibility [4]. In addition, PVA-based films or composites possess high moisture absorption properties due to hydrophilic hydroxyl groups. Another unique feature of PVA is its low water barrier properties. These properties favor its application in the laundry detergent industry. In recent years, detergent capsules have gained interest in the laundry and home care sector [5]. The modern innovation of detergent capsules concentrates and packs the active ingredients of liquid detergent in a single unit dose. The capsule will dissolve upon contact with water and its content will release. Despite the favorable characteristics mentioned earlier, PVA films experienced low elongation at break, poor decomposition temperature and low glass transition temperature [6]. These make their usage as eco-friendly packaging materials very challenging. Much research proposed incorporating other polymers or fillers into the PVA matrix to improve its mechanical properties [7,8,9].

Cellulose fiber-reinforced biocomposites have attracted significant attention due to their appealing properties such as low density, non-toxic, low cost, non-abrasive and low resistance to biodegradability [10]. There are many applications of cellulose fibers in this context. To name a few, cellulose fibers can be oxidized for the preparation of poly(methyl methacrylate) (PMMA) nanocomposites [11], poly(2-acrylamido-2-methylpropanesulfonic acid/poly(acrylic acid (PAMPS/PAA) hydrogels [12], polygluronic acid [13] and bioadsorbent for paraquat [14]. In addition, various works have proven that adding cellulose fibers into PVA could provide a significant reinforcing effect to tackle the poor mechanical properties of the PVA film [15,16,17,18]. Cellulose can be extracted from agricultural wastes such as rice husk, corn stover, palm oil bunches, etc. [19,20,21]. In Malaysia, about 3.31 million hectares of land are under oil palm cultivation [22], and the amount of biomass wastes generated is an average of 231.5 kg dry weight/year [23]. Oil palm biomass wastes consist of the oil palm empty fruit bunches (OPEFB), oil palm trunk (OPT) and oil palm fronds (OPF). These wastes are of low commercial value and are usually landfilled. Therefore, incorporating cellulose fibers from agricultural waste into the PVA matrix to form cellulose fibers-reinforced PVA biocomposite film as water-soluble detergent capsules could promote sustainability.

In the present work, cellulose was isolated from the OPEFB and incorporated with PVA to form cellulose-reinforced PVA biocomposite film. OPEFB was selected as the raw material because of its high cellulose content, i.e., as high as 40% to 65% reported [24,25,26]. In order to maximize the recovery of cellulose fibers from OPEFB, it is vital to investigate the effect of different process parameters on the efficiency of cellulose fiber isolation during the hydrolysis reaction. The Moving Least Squares (MLS) and Multivariable Power Least Squares (MPLS) methods were applied in combination as the numerical tools for regression modeling to study the process parameters correlations in cellulose isolation from the biomass. The MLS method, proposed by Lancaster and Salkauskas [27], can produce a continuous equation in a successive series of arbitrary functions. Meanwhile, the MPLS method [28] could correlate the multivariate data as a power function. The MPLS method applied in this work aims to investigate the significance level of manipulated variables and provide a regression model to represent the hydrolysis reaction. The MLS method has been widely applied in finite element analysis and image processing, and the method has potential to be used for parameter study. Meanwhile, the MPLS method was recently proposed in our previous work [28] as an alternative for design of an experiment tool. In this article, we combined both methods to form the novel MLS–MPLS method, which is new yet robust in parameter study. Last but not least, the effect of different cellulose loadings in the PVA matrix was evaluated by comparing the physical, mechanical, moisture barrier and thermal properties of the biocomposite films formed.

2. Materials and Methods

2.1. Materials

OPEFB was donated by Kwantas Oil Sdn Bhd, Malaysia. The biomass was washed thoroughly with distilled water and dried in an oven (Carbolite AX 120, Verder Scientific, Derbyshire, UK) at 60 °C for 24 h. The dried OPEFB underwent size reduction and was screened through a mesh size of 0.5 mm. The sample was stored in an air-tight container before being used. Sodium hydroxide (99% purity), acetic acid glacial (99.8% purity), sodium chlorite (80% purity), sulphuric acid (95–98% purity) and polyvinyl alcohol (>95% purity) were the commercial source, and they were used without further purification.

2.2. Alkaline and Bleaching Treatment of OPEFB

Alkaline pretreatment of OPEFB was conducted with 4% (w/v) sodium hydroxide solution at a solid-to-liquid ratio of 1:30. The reaction mixture was heated at 70 °C for 3 h under continuous stirring, as suggested by Zailuddin and Husseinsyah [29]. This pretreatment process was repeated three times to ensure high lignin degradation to facilitate the extraction of cellulosic fibers from the OPEFB. In each pretreatment, the pretreated fibers were filtered and washed with distilled water to achieve pH 7 to ensure complete removal of residual alkali, followed by drying in the oven at 90 °C overnight. The resulting dried fibers were subjected to bleaching treatment by mixing the fibers with a solution of equal parts of acetic acid buffer (2.7 g sodium hydroxide and 7.5 mL acetic acid in 100 mL of distilled water) and aqueous sodium chlorite (1.7%, w/v), at a solid-to-liquid ratio of 1:20. The bleaching treatment was performed at 80 °C for 4 h under continuous stirring. The bleached fibers were filtered, washed and dried in the oven at 90 °C overnight.

2.3. Isolation of Cellulose through Acid Hydrolysis

Acid hydrolysis was performed by mixing 1 g of bleached OPEFB fibers with 10 mL sulphuric acid. The reaction was performed at three different acid concentrations: 54%, 60% and 64%, and temperatures: 35 °C, 45 °C and 55 °C, for 60 min with sampling at 15 min intervals. Hydrolysis was performed under constant stirring at 85 rpm. Soon after the hydrolysis reaction, 20 mL of cold distilled water was added to stop the reaction. The cellulose suspension was immediately subjected to the washing process under centrifugation (Universal 320R, Hettich, UK) at 8000 rpm for 15 min. The washing process was repeated until the supernatant showed pH 7. Lastly, the drying process was carried out in the oven at 40 °C overnight. The weight of extracted cellulose was recorded, and the yield was computed using Equation (1):

Yield of cellulose (%) = w_c/w_treatedOPEFB × 100

(1)

where w_c and w_treatedOPEFB represent the weight of dried cellulose after acid hydrolysis (g) and the weight of dried, bleached OPEFB (g), respectively. All experiments were conducted in duplicate.

2.4. Preparation of Cellulose-Reinforced PVA Biocomposite Films

Biocomposite films containing PVA and cellulose fibers were formed through solution casting method. PVA pellets were soaked in distilled water (10%, w/v) for 1 h to enhance their solubility, followed by heating the mixture at 97 °C for 30 min under continuous stirring. Cellulose fibers of different loadings (1, 2.5 and 5%, w/w) were added to the PVA solution, and the mixtures were stirred for 1 h. The mixed polymers solution of 5 g was cast on a Petri dish and dried at 40 °C for 24 h. A thin film of biocomposite was formed once the polymer solution was completely dried. A control sample (known as neat PVA) was formed by repeating the experiment without adding cellulose fibers.

2.5. Moving Least Squares (MLS) Method

MLS method is a method to construct a continuous function from a set of scattered data, as shown in Equation (2):

$$y^h={\displaystyle \sum _{j=1}^m{P}_j\left(\mathit{x}\right){\mu }_j}={\mathbf{P}}^{\mathrm{T}}\mathsf{\mu }$$

(2)

where y^h is the predicted response, P is the user-defined polynomial interpolants, μ is their associating coefficients, x = [x₁ x₂ … x_m] is the manipulated factors, while m is the number of manipulated variables. The discrete norm can be determined using Equation (3):

$$J={\displaystyle \sum _{i=1}^n{W}_i{R}^2}={\displaystyle \sum _{i=1}^n{W}_i{\left[{\left({\displaystyle \sum _{i=1}^m\left(P_j\left(\mathit{x}\right){\mu }_j\right)}\right)}_i-y_i\right]}^2}$$

(3)

where n is the number of variables within the specific set of manipulated variables, while y represents the actual response. W is the weight function, which can be defined using a quartic spline function as shown in Equation (4):

$$W_i=1-6r_i{}^2+8r_i{}^3-3r_i{}^4$$

(4)

where

$$r=\frac{\left|x-\mathbf{x}\right|}{\Vert \mathit{x}\Vert }$$

(5)

In the MLS method, the coefficients μ can be determined by taking the minimization of the discrete norm such that:

$$\frac{\partial J}{\partial \alpha }=0\to \mathbf{A}\mathsf{\mu }=\mathbf{B}\to \mathsf{\mu }={\mathbf{A}}^{-1}\mathbf{B}$$

(6)

where

$$\mathbf{A}={\displaystyle \sum _{i=1}^n{W}_i{P}_i\left(x_i\right)}{P}_i{}^T\left(x_i\right)$$

(7)

$$\mathbf{B}=W_i{P}_i\left(x_i\right)$$

(8)

2.6. Multivariate Power Least Squares (MPLS) Method

MPLS method was developed through modification of MLS model and presented as [28]:

$$y^h=a{x}_1{}^{b_1}{x}_2{}^{b_2}\cdots x_m{}^{b_m}$$

(9)

where a is the MPLS coefficient while b = [b₁ b₂ … b_m] is the index for x. The discrete norm can be obtained as:

$$J={\displaystyle \sum _{i=1}^n{\left(\mathrm{In}\left(y_i\right)-\mathrm{In}\left(y_i{}^h\right)\right)}^2}$$

(10)

Upon minimizing Equation (10), the value of a and b can be determined by solving Equation (11) and the matrix in Equation (12), respectively.

$$a=\exp \left[\frac{1}{n}\left(\left({\displaystyle \sum _{i=1}^n\mathrm{In}\left(y_i\right)}\right)-{\displaystyle \sum _{j=1}^m\left(b_j{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\mathrm{In}\left(x_{j,i}\right)}}\right)}\right)\right]$$

(11)

$$\left(\begin{array}{cccc}{\gamma }_1& -{\lambda }_{12}& \cdots & -{\lambda }_{1m}\\ -{\lambda }_{21}& {\gamma }_2& \cdots & -{\lambda }_{2m}\\ ⋮& ⋮& \ddots & ⋮\\ -{\lambda }_{m1}& -{\lambda }_{m2}& \cdots & {\gamma }_m\end{array}\right)\left(\begin{array}{c}{b}_1\\ b_2\\ ⋮\\ b_m\end{array}\right)=\left(\begin{array}{c}{\xi }_1\\ {\xi }_2\\ ⋮\\ {\xi }_m\end{array}\right)$$

(12)

where

$${\gamma }_j=n\left\{{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m{\left[\mathrm{In}\left(x_{j,i}\right)\right]}^2}}\right\}-{\left\{{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\left[\mathrm{In}\left(x_{j,i}\right)\right]}}\right\}}^2$$

(13)

$${\xi }_j=n\left[{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\mathrm{In}\left(x_{j,i}\right)\mathrm{In}\left(y_j\right)}}\right]-\left[{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\mathrm{In}\left(x_{j,i}\right)}}\right]\left[{\displaystyle \sum _{i=1}^n\mathrm{In}\left(y_j\right)}\right]$$

(14)

$${\lambda }_{j,k}=\left[{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\mathrm{In}\left(x_{j,i}\right)}}\right]\left[{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\mathrm{In}\left(x_{k,i}\right)}}\right]-n\left[{\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m\mathrm{In}\left(x_{j,i}\right)\mathrm{In}\left(x_{k,i}\right)}}\right]$$

(15)

The subscript k represents the value of the second subscript of the term λ.

A normalized x, X as expressed in Equation (16) can be used to obtain the normalized MPLS equation:

$$\mathit{X}=C+\frac{\mathit{x}-{\mathit{x}}_{\min }}{{\mathit{x}}_{\max }-{\mathit{x}}_{\min }}$$

(16)

where C is a positive integer, and in this paper, C = 1 is applied. The analysis of the significance level can be conducted by comparing the magnitude of index b. The larger the magnitude of b, the more significant the manipulated variable is. However, the significance level analysis is only valid if the coefficient of determination R² for Equation (9) exceeds 0.5 [28]. The coefficient of determination, R² can be defined mathematically as in Equation (17):

$$R^2=\frac{{\displaystyle \sum _{i=1}^n{\left(y^h{}_i-\overline{y}\right)}^2}}{{\displaystyle \sum _{i=1}^n{\left(y_i-\overline{y}\right)}^2}}$$

(17)

where $\overline{y}$ is the average value for y.

2.7. Scanning Electron Microscopy (SEM) Analysis

Surface morphology of raw OPEFB, bleached-OPEFB and isolated cellulose fibers was observed using a scanning electron microscope (Hitachi VP-SEM SU1510, Hitachi, Japan), operating at an accelerating voltage of 15 kV. Before the examination, the specimens were sputter-coated with gold using Hitachi E1010 Ion Sputter (Japan) to prevent electrostatic charging and poor resolution. Few images were taken at different magnifications, and the sizes of fibers were also measured for comparison purposes.

2.8. Fourier Transform Infrared Spectroscopy (FTIR) Analysis

Fourier transform infrared spectroscopy (FTIR) analysis was conducted using FTIR Spectrometer (PerkinElmer Spectrum 400, PerkinElmer, USA) to identify the variation in the chemical compositions in the OPEFB fibres after each treatment. About 0.5 mg to 1.0 mg sample was mixed with potassium bromide (KBr), and the mixture was pressed to form a disk. The sample was scanned with the spectra transmittance region between a wavenumber 4000 cm⁻¹ to 500 cm⁻¹ at resolution of 4 cm⁻¹.

2.9. Transparency Analysis

Transparency of the biocomposite films was measured with the aid of Secomam UviLine spectrophotometer 9400 (France). The biocomposite film was cut into a rectangular shape of 5 cm × 2.5 cm (length × width) and placed at the cuvette position. The light absorbance of the film was measured for wavelength ranging from 190 nm to 400 nm. Transparency value of the film was calculated using Equation (18) [30]:

Transpancy value = A_400/x

(18)

where A₄₀₀ is the light absorbance value at 400 nm wavelength, and x is the thickness of the film in mm.

2.10. Mechanical Properties

Mechanical properties for the biocomposite films were carried out according to the ASTM D882 standard using a tensile machine (Zwicki Z5.0TH, Zwick Roell, Germany). Each sample was cut into a rectangular shape of 15 cm × 2.5 cm (length × width). A stress-strain curve was obtained to study the mechanical performance of biocomposite films in terms of tensile strength, percentage of elongation at break, and Young’s Modulus. The values of the tensile properties are the averages of five measurements.

2.11. Water Vapor Transmission Rate (WVTR) Test

Water vapor transmission rate (WVTR) test was performed based on the standard testing method specified under ASTM E96. Before the test, the biocomposite films were placed in a drying cabinet for 5 days to remove the film’s moisture. A plastic cup was used as an impermeable dish, some silica gels were placed inside the cup, and the cup’s mouth was firmly sealed with biocomposite film. The whole assembly was weighed before placing in a controlled environment (desiccator) along with a plastic cup of distilled water. The assembly was taken from the desiccator and weighted every 24 h. The experiment was conducted for 5 consecutive days. Subsequently, the WVTR was calculated using Equation (19):

WVTR ((g/day)/m²) = G/(t × S)

(19)

where G is the weight gain of the biocomposite film (g), t is the testing duration (day), and S is the exposed surface area of the film (m²).

2.12. Thermagravimetric Analysis (TGA)

Thermagravimetric analysis (TGA) was carried out using the thermal analysis instrument (Q50, TA Instrument, USA). It was carried out under nitrogen atmosphere with a heating rate of 20 °C/min. The weight of the film was maintained, and heating scans were conducted in a temperature range from 30 °C to 600 °C.

3. Results and Discussion

3.1. Cellulose Isolation via Acid Hydrolysis

The compact arrangement of cellulose, hemicellulose and lignin within the OPEFB makes the biomass recalcitrance and limits the accessibility to its cellulose component. To isolate the high quality of cellulose from the biomass, pretreatment of OPEFB is essential. Alkaline pretreatment was applied to break the linkage between the lignin-carbohydrate complex to increase the accessibility of carbohydrates such as cellulose. Further treatment of the OPEFB with sodium chlorite increased the cellulose percentage in the OPEFB by removing the lignin and hemicellulose components. Subsequent acid hydrolysis removed the amorphous regions of the cellulose and retained its crystalline regions [31], resulting in agglomerated cellulose in micrometer size. The dissolution of cellulose in acid hydrolysis was affected by its process parameters. The following sections discuss their individual and interaction effects and significance level in detail.

3.1.1. Effect of Hydrolysis Temperature and Time on the Yield of Cellulose

The effect of acid hydrolysis temperature and time on the yield of cellulose was presented in

Table 1. Average yield of cellulose at various hydrolysis temperatures and times.

Acid Concentration (%)	Temperature (°C)	Time (min)	Yield of Cellulose (%)
54	35	15	80.72 ± 0.31
		30	79.09 ± 0.56
		45	78.36 ± 0.02
		60	75.63 ± 0.18
	45	15	78.18 ± 0.59
		30	72.64 ± 0.20
		45	73.14 ± 0.24
		60	71.93 ± 1.11
	55	15	68.42 ± 1.58
		30	50.29 ± 1.91
		45	58.41 ± 0.55
		60	50.51 ± 0.92

. At the temperature of 35 °C, the highest cellulose yield of 80.72% was obtained at the hydrolysis condition of 35 °C and 15 min. With an increment of 10 °C in hydrolysis temperature, the cellulose yield decreased to 78.18% and subsequently to 68.42% at 55 °C, at the same hydrolysis time and acid concentration. This observation implies that temperature favored the hydrolysis reaction, and as a result, cellulose isolation suffered due to extensive depolymerization. Similarly, the hydrolysis trend was reported by Zhang and his co-authors [32]. This phenomenon is because a higher hydrolysis temperature encourages the breakage of glycosidic bonds by protons (H⁺) within the cellulose structure. On top of that, high temperature not just hydrolyzed the disordered amorphous domains of cellulose but also has the potential to hydrolyze the highly ordered crystalline domains of cellulose [33].

For all the investigated hydrolysis temperatures, isolation of cellulose increased gradually with the reduction of hydrolysis time. The crystalline regions of cellulose were converted to amorphous cellulose during the acid hydrolysis. Therefore, the amount of cellulose isolated reduced with hydrolysis time. An analysis of variance (ANOVA) study was conducted to examine the significance level of each parameter and its interaction on the yield of cellulose isolated. ANOVA for the yield of cellulose based on hydrolysis temperature and time is presented in Supplementary Data (Table S1). F₀ for hydrolysis temperature, time and temperature–time interaction was compared to F_critical using α value of 0.05 as the significance level. All F₀ values were greater than their corresponding F_critical values. Therefore, it can be concluded that the main effect of hydrolysis temperature and time were significant, as well as the interaction between the hydrolysis temperature and time.

3.1.2. Effect of Acid Concentration and Time on the Yield of Cellulose

Table 2. Average yield of cellulose at various acid concentrations and hydrolysis times.

Temperature (°C)	Acid Concentration (%)	Time (min)	Yield of Cellulose (%)
35	54	15	80.72 ± 0.31
		30	79.09 ± 0.56
		45	78.36 ± 0.02
		60	75.63 ± 0.18
	60	15	69.37 ± 0.28
		30	68.30 ± 0.51
		45	63.37 ± 1.94
		60	60.38 ± 2.28
	64	15	60.36 ± 0.69
		30	34.19 ± 1.12
		45	16.10 ± 1.57
		60	13.10 ± 0.39

shows the effect of acid concentration and hydrolysis time on the yield of cellulose. Among the acid concentrations investigated, the lowest acid concentration of 54% yielded the highest amount of cellulose. Relatively lower cellulose yields were obtained at higher acid concentrations at all reaction times. As a result, the lowest cellulose yield of 13.10% was obtained for acid hydrolysis with 64% acid concentration for 60 min. The low yield of cellulose was suspected due to the extreme dissolution of the cellulose caused by high acid concentration. Concentrated acid has the potential to cleave the hydrogen bonds holding cellulose in the crystalline state [34], thus converting it to its monomer (sugars) and resulting in lower cellulose isolation. A similar report was found in Hamelinck et al. [35], which claimed that complete cellulose hydrolysis to sugar monomers at low temperatures was possible when a long reaction time and a highly concentrated acid were applied. The amorphous cellulose was degraded to undesired products of sugar monomers at a high acid concentration [36].

From Table 2, a similar trend of cellulose yield can be observed as shown in the effect of temperature discussed earlier, where the amount of cellulose isolated progressively decreased with hydrolysis time regardless of the acid concentration used. The effect of hydrolysis time was more pronounced at a higher acid concentration of 64%. The cellulose yield was drastically reduced from 60.36% to 13.10% when acid hydrolysis was performed at 15 min and 60 min, respectively. The longer the hydrolysis time, the more cellulose was converted into simple glucose [37], resulting in less cellulose isolated from the OPEFB. ANOVA on the yield of cellulose isolated based on acid concentration and hydrolysis time is available in Supplementary Data (Table S2). Similarly, all F₀ values were greater than their corresponding F_critical values, suggesting that the acid concentration, time and their interaction significantly affected cellulose yield.

3.2. Process Modeling Using MLS and MPLS Methods

Mathematical modeling was constructed via least squares methods, particularly the MLS and MPLS methods. The models obtained for these two regression methods are presented in Equations (20) and (21), respectively.

$$\begin{array}{l}{y}^h=-1650.7524+68.1638x_1-15.7560x_2+4.8021x_3-0.7042x_1{}^2\\ -0.0632x_2{}^2+0.0075x_3{}^2+0.3879x_1{x}_2-0.0118x_2{x}_3-0.0914x_1{x}_3\end{array}$$

(20)

$$y^h=1.1345x_1{}^{-1.4582}{x}_2{}^{-0.6287}{x}_3{}^{-0.6153}$$

(21)

where y^h, x₁, x₂, and x₃ denote yield of cellulose (%), acid concentration (%), hydrolysis temperature (°C) and time (min), respectively. The coefficient of determination R² for the MLS and MPLS models were 0.9725 and 0.7693, respectively. Since the R² obtained were larger than 0.5 indicated the goodness of the models in representing the reaction. The fitness of the models can also be visualized graphically using predicted value versus actual value plot, as shown in Figure 1. The closeness of the data to the diagonal lines signified the excellent representation of the regression models to the actual situation.

The significance of process parameters can be evaluated through the indices of the factors in the MPLS model. The higher the index of the factor, the greater the significance level of the parameter towards the response. Therefore, the most dominant parameter affecting the yield of cellulose was acid concentration (x₁), followed by hydrolysis temperature (x₂) and time (x₃). In acid hydrolysis, H⁺ ions tend to attack oxygen atoms in the β-1,4-glycosidic bond within cellulose polymers, resulting in a shorter chain of sugar molecules [38]. By using concentrated acid, the decomposition rate of the cellulose was accelerated as the concentration of H⁺ ions increased. Therefore, acid concentration was the most significant parameter affecting the cellulose isolation efficiency from the OPEFB. The MPLS model could also provide information on the interactions between parameters in affecting the cellulose yield. The positive coefficients are expected to cause a positive impact on the yield, while negative coefficients indicate that the parameters would negatively affect the cellulose yield. Based on Equation (21), all the three parameters are negative coefficients, which agreed with the discussion made earlier, whereby the cellulose yield was favored at low acid concentration, low hydrolysis temperature and short hydrolysis time.

3.3. Morphology of Cellulose Fibers

The surface morphology of cellulose isolated from the OPEFB through acid hydrolysis was compared with raw and bleached OPEFB. As shown in Figure 2, significant physical differences were observed from the SEM images after each chemical treatment. Initially, the raw OPEFB was coated with irregular deposition of residual waxes, lignin or other inorganic substances (Figure 2a), which was similar to other findings reported earlier [39,40]. After pretreated with an alkaline solution and bleached with acidified sodium chlorite, the surface of the fibers became cleaner and rougher, where more fibers were exposed, as shown in Figure 2b. The clean fibers’ surface was due to the removal of unwanted substances such as wax and cuticle through the interaction with sodium ions during the alkaline pretreatment [41]. The elimination of lignin increased the contact area, where the fibers became more exposed during the bleaching treatment.

Further treatment with sulphuric acid caused the formation of individual long cellulose fibrils, which were segregated from the thick fiber bundles (Figure 2c,e). The SEM images have revealed that highly purified cellulose was successfully isolated from the OPEFB with the removal of lignin and hemicellulose. A similar observation was reported by Chimentão and his co-authors [42]. It can be seen that the bundle sizes of fibers were gradually reduced after every treatment. Eventually, micro-size cellulose fibers were obtained. As depicted in Figure 2d, the fiber diameter of 739 nm was observed for cellulose isolated under the condition of 35 °C, 15 min and 54% acid concentration. At the higher acid concentration of 60% under the same hydrolysis temperature and time, the fibers obtained were of smaller diameter of approximately 317 nm, as shown in Figure 2f. This observation further confirmed that a stronger acid was more effective in cellulose hydrolysis and produced cellulose fibers of smaller size, but at the same time suffered from a lower cellulose yield due to extensive depolymerization and product degradation.

3.4. Structural Analysis of Cellulose Fibers

FTIR spectroscopy is a useful tool to monitor the change of functional groups in samples. FTIR spectra of raw OPEFB fibers, bleached OPEFB fibers and cellulose fibers are illustrated in Figure 3a. From the figure, few peaks (1735 cm⁻¹ and 1228 cm⁻¹) in the raw OPEFB fibers sample were not found in the bleached OPEFB fibers and cellulose fibers spectra. Peak at 1735 cm⁻¹ is associated with the presence of the acetyl and uronic ester groups of hemicelluloses or the ester carbonyl groups of lignin [43,44], while peak at 1228 cm⁻¹ is attributed to the syringyl ring unit and C–O stretching in lignin and xylan [45]. The disappearance of these peaks in the spectra of bleached OPEFB fiber and cellulose fiber verified the removal of hemicelluloses and lignin during the bleaching and hydrolysis of OPEFB fibers.

FTIR spectra of the treated and untreated fibers showed a broad absorption band at 3340 and 2900 cm⁻¹. These peaks correspond to the OH stretching vibration in cellulose and C–H stretching vibration of CH₂ and CH₃, respectively [44,46]. Peak at 3340 cm⁻¹ can also be attributed to the hydroxyl groups in cellulose [45]. Peak at 1035 cm⁻¹ appeared in all samples attributed to the C–O–C pyranose ring (anti-symmetric in phase ring) stretching vibration [47]. From the figure, sharper peaks were observed after each chemical treatment. This was due to the removal of non-cellulosic components in the OPEFB fiber during the treatment processes.

Figure 3b–d shows the FTIR spectra of cellulose fibers with variations in hydrolysis temperature, time and acid concentration, respectively. All samples displayed similar spectra, suggesting no changes in their chemical composition or new bonds formation during the acid hydrolysis. Same vibration bands at 3340, 2900, 1635, 1427, 1373, 1035, and 894 cm⁻¹ were found in all cellulose fibers. Peak at 1635 cm⁻¹ originates from the bending vibrations of the OH groups of cellulose [48,49]. Peaks at 1427 cm⁻¹ and 1373 cm⁻¹ reflect the symmetric bending of CH₂ and the bending vibrations of the C–H and C–O groups of the aromatic rings in cellulose, respectively [50,51]. Peak at 894 cm⁻¹ relates to the characteristic of cellulose with β-glycoside bonds of glucose ring [52]. By comparing the FTIR spectra of cellulose fibers at all investigated parameters, the difference in peaks intensity was more pronounced with variation of acid concentration. This finding was consistent with the earlier discussion, whereby acid concentration was the most dominant parameter for isolating cellulose fibers through acid hydrolysis.

3.5. Light Transparency of Cellulose-Reinforced PVA Biocomposite Films

Different formulations of cellulose-reinforced PVA biocomposite films were investigated to examine the suitability of the biocomposite films as packaging material for detergent capsules. Figure 4a shows the physical appearance of all formulations of PVA films. The blue background can be seen through for all the films, showing that these films were transparent. For better evaluation, the films’ transparency values were calculated, and the results were tabulated in

Table 3. Transparency value of neat PVA film and biocomposite films at different cellulose loadings.

Sample	Transparency Value
Neat PVA	19.4
PVA/1% cellulose	16.1
PVA/2.5% cellulose	14.8
PVA/5% cellulose	14.7

. Neat PVA film exhibited excellent transparency compared to cellulose-reinforced biocomposite films, whereby the transparency value of the biocomposite films gradually decreased with the increase in cellulose loading. This is due to the presence of opaque particles of cellulose fibers that blocked the light passage through the films.

Film with good UV-barrier properties reported longer shelf life of packed detergent products due to the delay of active ingredients degradation, which will diminish the quality of the product. For example, propane-1,2-diol, which is a common liquid laundry detergent ingredient, is known to discolor upon exposure to sunlight [53] and to form a colored complex with iron (III) [54]. UV-vis spectrophotometer was used to evaluate the light transmittance of the fabricated films. Spectrum study was conducted by measuring the light transmittance throughout the wavelength of 190 to 400 nm. The chosen wavelength range covers three UV light radiations, including UV-A (320–400 nm), UV-B (280–320 nm) and UV-C (190–280 nm).

Figure 4b illustrates the percentage of light transmittance for all the biocomposite films. Neat PVA was included as the control sample. Overall, light transmittances of biocomposite films were close to light transmittance of neat PVA film. A decrease in light transmittance can be observed with the increase in cellulose fibers loading. At the wavelength of 400 nm, neat PVA film was transparent with light transmittance of 84.9%. The biocomposite film with 1% cellulose fibers showed a slight decrease in light transmittance, resulting in 84.5% of light transmittance. The light transmittance continued leveling off to 83.0% and 81.3% for the biocomposite films containing 2.5% and 5% cellulose fibers, respectively. This result indicated that the presence of cellulose fibers in the biocomposite films had blocked the passage of light, which eventually reduced the percentage of light transmittance of the films. Another reason was light reflection caused by the non-uniformity dispersion of the cellulose fibers within the biocomposite films, which reduced the light transmittance through the cellulose-reinforced biocomposite films.

Naduparambath and his co-authors [15] reported similar observations, in which a lower light transmittance was observed at higher microcrystalline cellulose (MCC) loading in PVA/MCC biocomposite films. Likewise, Ching and his co-authors [55] observed that the percentage of light transmittance reduced for nanocellulose-reinforced PVA films when the amount of nanocellulose increased. It was due to the presence of agglomerated fibers in the biocomposites. In short, the cellulose-reinforced PVA biocomposite film is preferred as it is more resistant to light transmission, which could provide a better protective layer to the detergent capsule from UV light emitted from sunlight.

3.6. Mechanical Properties of Cellulose-Reinforced PVA Biocomposite Films

The packaging material should exhibit adequate mechanical strength to prevent and reduce the chances of detergent leaking and product contamination from broken detergent capsules within the same storage container. Inadequate mechanical strength may also lead to premature bursting of detergent capsules during packing, transportation and storage. Therefore, it is essential to evaluate the mechanical properties of cellulose-reinforced PVA biocomposite films by analyzing their tensile strength, elongation at break and Young’s modulus.

Table 4. Tensile strength, elongation at break and Young’s modulus of neat PVA film and biocomposite films at different cellulose loadings.

Sample	Tensile Strength (MPa)	Elongation at Break (%)	Young’s Modulus (MPa)
Neat PVA	28.65	97.18	3.05
PVA/1% cellulose	34.60	125.55	2.77
PVA/2.5% cellulose	36.09	119.96	2.96
PVA/5% cellulose	31.63	92.19	3.63

tabulates the tensile properties of different formulations of cellulose-reinforced PVA biocomposite films. The results showed a significant improvement in tensile strength when cellulose fibers were reinforced in the PVA matrix. The highest tensile strength of 36.09 MPa was obtained for the film that consisted of 2.5% cellulose fibers. The increase in tensile performance was due to the formation of hydrogen bonds between the cellulose fibers and the PVA, which contributed to the dispersion of fibers in the polymer network [56]. However, biocomposite film’s tensile strength was slightly reduced when its cellulose loading increased to 5%. The same finding was obtained by Li and his co-authors, where the 5% and 10% addition of cellulose nanofibers from pea hull waste into carboxymethyl cellulose film had a tensile strength of 32.95 MPa and 25.02 MPa, respectively. The decreased tensile strength may be due to the imperfection in fibers dispersion at high cellulose loading, which is also supported by the works reported by Asser et al. [57] and Oyeoka et al. [58]. The study has a higher tensile strength compared to the literature. For example, biocomposite produced from Areca nut husk cellulose fibers, PVA and chitosan film had a tensile strength of 9.46 MPa [59], and water hyacinth cellulose fibers, PVA-gelatin films had tensile strength ranges from 7.91 MPa to 13.83 MPa [58].

Similarly, elongation at break had greatly enhanced with the increment in fibers loading up to 2.5%. Among all, PVA/1% cellulose possessed the highest elongation at break at 125.55%. This observation might be due to the optimized compatibility between the cellulose and the PVA, which made the biocomposite film a highly toughened and flexible material. With the increase in cellulose loading to 5%, elongation at break was reduced due to particle saturation, which resulted in phase segregation. Heterogeneity of the sample affected its ability to be elongated. Similarly, this work has a better elongation at break than the literature. Biocomposite films produced from Areca nut husk cellulose fibers-PVA-chitosan film and water hyacinth cellulose fibers-PVA-gelatin films had an elongation at break of 19.35% [59], and 45.8% to 81.2% [58].

On the other hand, an increase in Young’s modulus with cellulose loading indicated that stress had been successfully transferred from the PVA matrix to the cellulose fibers due to effective dispersion and adhesion of cellulose fibers in the biocomposite film. This observation is due to the formation of a large interfacial area when many cellulose fibers of relatively smaller size were used as fillers [60]. The stress-strain behavior can also be illustrated using a stress-strain plot, as shown in Figure 5. It shows that PVA/5% cellulose samples had the highest stress-strain properties compared to neat PVA and biocomposite films containing 1% and 2.5% cellulose fibers.

In this study, the size of cellulose fibers used to reinforce the PVA was not consistent. Different fiber lengths have different effects on mechanical properties. For example, longer fibers could contribute to better tensile strength than shorter fibers [61]. This may be one reason for having results that were not aligned for all the three mechanical properties evaluated. Nevertheless, the best composition was PVA reinforced with 2.5% cellulose fibers because it had the best mechanical properties, except for a slight reduction in film stiffness. On top of that, tensile strength is the most crucial property among all the mechanical properties evaluated as it is used to measure the premature bursting of detergent capsules.

3.7. Moisture Barrier Properties of Cellulose-Reinforced PVA Biocomposite Films

The moisture barrier of biocomposite film is another critical property to be considered in packaging materials. This is because water vapor is one of the main permeates present in the environment, which could quickly diffuse through the packaging film—as a result, affecting the product quality and shelf life of content. The film should exhibit good moisture barrier properties as a detergent packaging material. By reducing the chances of water vapor being captured by the biocomposite film, dilution of detergent concentration and film solubility alteration could be minimized. WVTR is a standard indicator to measure the rate of water vapor permeates through the composite film at a specific thickness, temperature and relative humidity. Figure 6 shows the WVTR of all the samples over 4 days. It was observed that all the biocomposite films fabricated possessed better moisture barrier properties than neat PVA. Hydrophilic groups in the PVA increased their interactions with water molecules, leading to higher WVTR. On the other hand, the cellulose fibers in the biocomposite films served as impermeable physical barriers in the PVA matrix. These fibers prevented the water molecules from traveling and passing through the film by forming tortuous pathways, as shown in Figure 7. As a result, improvement in moisture barrier performance was observed in all the cellulose-reinforced PVA biocomposite films.

Among all the biocomposite films fabricated, the film with 1% cellulose fibers loading exhibited the best moisture barrier properties. The observation was due to fibers agglomeration at higher cellulose loading, which was ineffective in creating tortuous pathways for water vapor to pass through the film. As a result, the WVTR was significantly reduced with increased cellulose fibers loading in the biocomposite film. Similar results were reported by Shankar et al. [62] and Yang et al. [63]. Ali and his co-authors [6] also reported an increasing trend of WVTR with MCC loading in the PVA matrix. In another work by Patil et al. [64], a negative relationship was also obtained between the WVTR and cellulose fibers loading. The observation was due to the hydrogen bond interactions between the MCC and the PVA, which increased the hydrophilic groups in the biocomposite and, subsequently, the WVTR. These findings signified the importance of incorporating cellulose fibers into PVA in a suitable ratio so that favorable characteristics of biocomposite film could be formed to suit its application.

3.8. Thermal Stability of Cellulose-Reinforced PVA Biocomposite Films

TGA was used to evaluate the thermal performance of neat PVA and different formulations of cellulose-reinforced PVA biocomposite films. TGA thermogram in Figure 8a shows the percentage of weight loss with respect to the temperature for a heating profile of 30 °C to 600 °C. The figure shows that all the films exhibited three main weight loss regions. The first region appeared in a temperature range of 80 °C to 140 °C with weight loss of approximately 7%. It corresponded to the evaporation of the adsorped moisture in the films [58,65]. The significant weight loss of approximately 60% happened in the second region of 230 °C to 370 °C. The weight loss was due to the structural depolymerization, dehydration and decomposition of the biocomposite films, as confirmed by Salh and Raswl [66] and Oyeoka et al. [58]. The third region happened at temperatures above 370 °C. The weight lost observed could be due to the cleavage backbone of PVA composite films or the combustion of carbon materials [67,68]. Constant weight of 4% to 6% was observed at temperatures of 500 °C onwards.

As seen in Figure 8a, the incorporation of cellulose fibers in the PVA film had a negligible effect on the thermal profile of the PVA film. Cellulose-reinforced PVA film with 1% cellulose loadings exhibited the highest thermal stability. The observation will be more significant by presenting the thermogram in differential thermagravimetry (DTG), which plots the rate of material weight changes upon heating against temperature, as shown in Figure 8b. It can be seen that the curve for 1% cellulose-reinforced biocomposite film was moved to a higher temperature compared to the neat PVA. The increase in thermal stability in cellulose-reinforced PVA film could probably be due to the strong hydrogen bonding between the hydroxyl groups of cellulose and the PVA matrix, as reported by Lee and his co-authors [69]. Another plausible reason could be the insulation effect and mass transport barrier exhibited by the cellulose fibers, which inhibited the generation of volatile products during decomposition [70].

4. Conclusions

Cellulose fibers were successfully isolated from the OPEFB through sulphuric acid hydrolysis. The reaction was greatly affected by acid concentration, hydrolysis temperature and time in descending order. An apparent interaction effect was also observed between hydrolysis temperature and time as well as acid concentration and hydrolysis time. The cellulose yield was favored at low acid concentration, low hydrolysis temperature and short hydrolysis time, supported by both experiments and mathematical modeling. A maximum cellulose yield of 80.72% was attained at hydrolysis using 54% acid concentration at 35 °C for 15 min. The well-dispersed cellulose fibers showed good reinforcement effects in the PVA matrix. Tensile strength and elongation at break were greatly enhanced. In addition, the biocomposite films exhibited reduced light transmission rate, good absorption of UV rays and thermal stability, as well as reduced water vapor transmission rate. These properties are essential, especially in the detergent packing industry, to prevent the breakage of packaging material and product deterioration. Among all the biocomposite formulations investigated, PVA/2.5% cellulose was the most promising biocomposite film as it exhibited good mechanical and moisture barrier properties and blocked UV rays. The results indicated that cellulose-reinforced PVA biocomposite film is a potential material for biodegradable detergent capsules packaging.