Banana peel nanocellulose and soy protein hydrolysate complexed colloidal nanoparticles synthesis using ultrasonic interventions: characterization and stable pickering emulsion formation

Abstract

Pickering based emulsion system are been gaining interest in active delivery of encapsulated molecules in food system. In the present study, cellulose nanoparticles (CNPs) were isolated from food waste (banana peel) using acid hydrolysis followed by high-intensity ultrasonication. The complex colloidal nanoparticles (CNPSPH) were fabricated using hydrogen bonding and electrostatic interactions between cellulose nanoparticles and soy protein hydrolysates. With 400 W power level for 30 min, size of 53.11 ± 1.45 nm with polydispersity index of 0.21 ± 0.21 and Zeta potential of − 34.33 ± 0.77 were noted for generated CNPs. The three-phase contact angle (o/w) of CNPSPH at a mass ratio of 1:1 CNPs to SPHs (CNPSPH 1:1) was approximately 89.07°, indicating as effective Pickering emulsifiers. Furthermore, the stability of the Pickering emulsion stabilised by CNPSPH complex was investigated under various pH and temperature conditions. The findings will provide solution in development of nanocellulose-soy protein complex particles for a stabilized Pickering emulsion formation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-023-01477-w.

Introduction

In recent years, rapid increase of agricultural wastes has motivated efforts toward maximizing the efficiency of the utilization of agricultural waste for the production of different value-added products at industrial level. Cellulose a natural fiber are considered as the most abundant hydrophilic polymer that composed of glucose-glucose linkages (β-1,4-linked-glucopyranose unit) (Kasiri and Fathi, 2018). Cellulose nanoparticles have become a source of interest as a novel material due to its amorphous and crystalline nature (Costa et al., 2015). These fibres used as a modifier to alter the rheological properties in foods, paints, cosmetics and pharmaceutical products and can be obtained from agricultural waste sources, such as wood, wheat, straw, cotton, maize and non-plant sources such as bacteria, etc. (Abe et al., 2007; Morais Teixeira et al., 2010).

Emulsions based delivery systems have wide applications in the encapsulation and delivery of bioactive compounds in food industry for protecting them against chemical degradation, retaining their functional properties, enhancing bioavailability and controlled release (McClements, 2012). Surface modification is required to achieve the desired particle stability, which can be modified using a variety of treatment methods. Ultrasonication-based intervension are dependable, simple to use, provide reasonably priced and extremely accurate measurements. Acoustic spectroscopic analysis is commonly used in the preparation and characterization of various nanoparticles (Bielas et al., 2019). Acoustic radiation force, which scales proportionally to the square of acoustic pressure, should provide nanoparticles with the momentum they need to move through the dispersion while also overcoming energy barriers between the nanoparticle and emulsion interface. Particles in nano-range could form as a result of regionally high pressures, stresses, and temperatures brought on by cavitation, which could create and then violently collapse vapour cavities within the liquid (Lee et al., 2019). Low power ultrasonication treatment are typically used as a quick and nondestructive method for measuring the properties of materials, including composition and structure (Jozefczak and Wlazło, 2015).

“Pickering emulsions”, a type of emulsions, stabilized by solid colloidal particles instead of surfactants, which accumulate at the interface between two immiscible liquids (oil and water phase). In contrast to conventional emulsions, Pickering emulsions stabilize droplets against droplet flocculation and coalescence via a steric mechanism (Horozov and Binks, 2006). Conventional emulsions are stabilized by small molecular weight surfactants and amphiphilic polymers, are thermodynamically unstable and tend to breakdown over time due to physicochemical mechanisms such as gravitational separation, flocculation, coalescence, Ostwald ripening, and phase inversion (McClements, 2012). Pickering emulsion have attracted extensive interests in recent years as new dosage forms. Since, the classical emulsifier that stabilizes the oil droplets is replaced by solid particles due to their advantages of “surfactant-free”, high stability against coalescence and Ostwald ripening (Xiao et al., 2016). Cellulose selection is suitable for stable pickering emulsion formation because of its amphiphicity and nanoscale dimension structure (Bai et al., 2021). Pickering emulsions stabilized with nanocellulose from bamboo shoots exhibit improved physicochemical and elasticity behavior (Lei et al., 2022).

The polysaccharide-protein particles as natural particles stabilizer has been used as for food emulsion (Wong et al., 2021). Consequently, in the current work we studied, the adsorption and arrangement at interface of polysaccharide-protein complex particles prepared from cellulose nanoparticles and soy protein for a stabilized Pickering emulsion formation. Firstly, we explored the effect of the ultrasonication process on the cellulose nanoparticles extracted from a common agricultural waste, banana peel, locally known as Bhimkol in Northeast India. Secondly, the effect of concentration of polysaccharide and protein on the hydrodynamic diameter, contact angle and thermal property of the complexes. And thirdly, the effect of oil fraction (φ; 0.2–0.7) on the interfacial properties and stability of the Pickering emulsion was determined by creaming stability, droplet size and rheological behaviour.

Materials and methods

Materials and reagents

Whole Banana (Musa balbisiana BBB genotype) was obtained from the local market of Tezpur, India. The peels were only utilized for the study. Soy protein isolate was purchased from MYPROTEIN™, Noida, India and Olive oil was obtained from Olive Tree Trading Pvt. Ltd., Pune, India. All the chemicals and reagents were analytical grade and obtained form Merck, India; Sigma Aldrich, India and Himedia, India respectively.

Purified cellulose

Banana peels were dried in a hot air oven to a moisture content of 5.0 ± 1.0%, dry basis in an hot air circulating oven (SELEC TC344, Jain Scientific Glass Works, Ambala, India). The compositional analysis of dried banana peel was analysied by AOAC approved methods. The dried peels were ground and treated with aqueous NaOH (2% w/w) and continuously stirred for 4 h at 100 ± 5 °C. The slurry was washed several times with distilled water to remove NaOH completely and again re-dried at 50 ± 5 °C for 12 h in the oven. The dried material was bleached (dried material to solution ratio is 1:10 w/v) for 4 h at 80 ± 5 °C by adding equal parts (v:v) of acetate buffer (0.2 M, pH 4.5) and aqueous sodium chlorite (1.7 wt% NaClO₂ in water). The bleached material was repeatedly washed with distilled water until the pH of the material was neutral and subsequently dried at 45 ± 5 °C for 12 h. The final dried material obtained was purified cellulose.

Preparation of cellulose nanoparticles (CNP)

CNP were prepared by acid hydrolysis followed by utrasonication. In brief, purified cellulose was subjected to acid hydrolysis by adding H₂SO₄ (64%, w/w) with a sample to solvent ratio 1:20 and vigorous stirring in a water bath (BW-20G, Jeio-Tech, Seoul, South Korea) at 45 ± 5 °C for 60 min (Santos et al., 2013).

The hydrolysis was later arrested by diluting the mixture by 10-fold with cold distilled water and centrifuged in a laboratory refrigerated centrifuge (5430R, Thermo Fischer Scientific, Hampton, USA) at 4000g for 10 min at 15 °C to remove excess acid.

The precipitate obtained was then dialyzed (membrane cut off-3.5 kDa) with distilled water to remove soluble sugars, salt and non-reactive sulphur groups. Dialysis was continued till the pH of the pellet/precipitate became neutral (∼4 days). Subsequently, the resulting suspension of dialysis process was ultrasonicated in a probe ultrasonic processor (13 mm probe diameter, U500, Takashi, Tokyo, Japan) at different power and time levels (300 W: 15 min; 300 W: 30 min; 400 W: 15 min and 400 W: 30 min) to obtain CNP suspensions and later stored at 4 °C. About 3–4 drops of chloroform were added in the suspension to avoid any bacterial growth during storage. Some parts of individual suspensions were lyophilized for characterization.

Preparation of soy protein hydrolysates (SPH)

SPH were prepared using enzymatic hydrolysis of soy protein isolates (SPI) with Papain (≥ 10 units/mg protein). Soy protein isolates was hydrated by suspending in phosphate buffer (10 mM, pH 7.0; Cyber scan, EUTECH Instruments, Tuas, Singapore) at a concentration of 20.0 g/kg of SPI, for 2 h at 30 ± 5 °C along with constant stirring. Papain was added at a concentration of 1.5 g/kg of SPI and the suspension was incubated in an incubator at 50 ± 5 °C for 30 min with constant stirring. The enzymatic hydrolysis was immediately stopped after incubation by heating the resulting suspension at 95 ± 5 °C for 10 min. The suspensions were cooled to room temperature and lyophilized for further analysis.

Fabrication of cellulose nanoparticles and soy protein hydrolysates complex colloidal nanoparticles (CNPSPH)

Soy protein hydrolysates and cellulose nanoparticles complex was prepared by anti-solvent precipitation method (Dai et al., 2017). Briefly, 4.0 g SPH was dissolved in 100 mL ethanol (70%, v/v) with magnetic stirring (Remi Bharat Scientific World, Mumbai, India) at low speed. The stock solution was added drop-wise into 300 mL deionized water with constant stirring to obtain SPH dispersion. After 30 min of constant stirring, the dispersions were concentrated by rotary vacuum evaporator (45 ± 2 °C; W1200, Eyela Co. Ltd., Sahngahai, China) until the final concentration of the dispersion reached 2% (w/v). The pH of SPH particle dispersions was adjusted to 4.0 using 0.1 N HCl or NaOH. The purified CNP (15.0 g/kg) were dispersed in phosphate buffer (10 mM, pH 7.0). After dissolution the pH of the solution was also adjusted to 4.0. Then CNPSPH dispersions were formed with CNP to SPH ratios of 4:1, 2:1, 1:1, 1:2, 1:4 (w/w) termed as CNPSPH_4:1, CNPSPH_2:1, CNPSPH_1:1, CNPSPH_1:2, CNPSPH_1:4 respectively. Then the CNPSPHs dispersions were stored at 4 °C for further analysis and some parts of samples were lyophilized.

Formation of CNPSPH stabilized pickering emulsion

The Pickering emulsions were prepared using the same CNPSPH (1:1) with deionized water (1%, w/v) and different oil fractions (φ = 0.1, 0.3, 0.5, 0.7). The total volume of emulsions was set to 20 mL. Briefly, the oil was slowly added to the particles suspension with a high-speed homogenizer (UltraTurex 25, IKA, California, USA) at 7500 g. After the oil completely added, the mixtures were further homogenized for another 3 min to acquire oil in water type Pickering emulsions (Dai et al., 2017).

Characterization of cellulose nanoparticles, cellulose nanoparticles and soy protein hydrolysates complex

Dynamic light scattering measurements

The size distribution of CNP and CNPSPH complexes and their charge in aqueous dispersion measurments were carried out on a Malvern Zeta sizer (Nano S90, Worcestershire, UK) for determining dynamic light Scattering, polydispersity index (PDI) and ζ-potential parametres. The Dynamic Light Scattering system operating at a wavelength of 633 nm with a scattering angle of 90°, under a constant temperature of 25 °C was used. The size distribution of the samples was obtained in the fully automated mode of Zetasizer Nano S90. Samples were prepared by diluting CNP and CNPSPH complexes with distilled water to a concentration of 1 mg/mL prior to instrumental analysis.

Differential scanning calorimetry

Thermal properties of CNP and CNPSPH complexes were investigated using a differential scanning calorimeter (DSC-60 Shimadzu, Tokyo, Japan). The measurements were carried out under a inert atmosphere using nitrogen, and data pertaining to changes in enthalpy (ΔH) and degradation temperatures were obtained. Briefly, 5 mg of powdered samples were hermetically sealed in aluminium pan and subjected to heating at the rate of 10 °C/min from 25 to 250 °C.

Wettability of CNPSPH complexes

The three-phase contact angle (ϴ_o/w) of the CNPSNP complexes was measured using OCA25 (Dataphysics Instruments, Filderstadt, Germany) using the sessile drop method. For the experiments, freeze dried CNPSPH powders were first compressed to pellets of 10 mm diameter and 2 mm thickness. The pellets were then placed into a large liquid sample pool containing olive oil. Next, a drop of MilliQ water (2 µL) was deposited on the surface of the pellets employing a high-precision injector. After equilibration, the drop image was photographed using a digital camera and the contact angle was calculated based on droplet curvature profile fitted to the Laplace-Young equation (Zou et al., 2015). Contact angles were measured for three pellets per sample, and measured in triplicates for each pellet.

Emulsifying activity index (EAI) and emulsion stability index (ESI)

The emulsifying properties of the samples including EAI and ESI were assessed according to the procedure established by Pearce and Kinsella (1978). Briefly, 100 µL of the prepared emulsions were diluted with 5.0 mL 0.1% sodium dodecyl sulphate solution. Next, the absorbance of the diluted emulsion was recorded at 500 nm wavelength with a spectrophotometer (Eppendorf BioSpectrometer, Hamburg, Germany). EAI was calculated using the absorbance measured at zero elapse of time as per equation (Eq. 1). Further, ESI was calculated and reported as the time (in min) needed to reach a reduction in 50% of initial absorbance value (Eq. 2).

$$\text{EAI}\left(m^2,/,g\right)=\frac{2\times 2.303\times A\times N}{C\times \mathrm{\varnothing }\times 10000}$$ 1

where, C is the concentration of CNPSPH complex (g/mL), N is dilution factor, ɸ is oil volume fraction of the emulsion and A is the absorbance at zero elapse of time.

$$\text{ESI}\left(\text{min}\right)=\frac{A_o}{\mathrm{\Delta }A}\times t$$ 2

where, A₀ is the absorbance of the diluted emulsion of the diluted emulsion immediately after the homogenization ∆A is the change in absorbance between 0 and 10 min (A₀–A₁₀), t is the time interval which is 10 min in this case.

Stability evaluation of pickering emulsion with respect to pH and heat

The effect of pH and heat on the stability of Pickering emulsion was evaluated according to the method described by (Dai et al., 2017). For checking stability towards pH, the freshly prepared CNPSPH_1:1 stabilized emulsion gels (10 mL) were adjusted to pH 3.0, 5.0, 7.0, and 9.0 using either 0.1 M NaOH or 0.1 M HCl at 25 °C. Again, for heat stability, the prepared CNPSPH_1:1 stabilized emulsion gels were incubated in a water bath at different temperature ranges (20, 40, 50, 60, 70, and 80 °C) for 30 min and later cooling to 25 °C. The stability of the emulsions was evaluated based on rheological studies and emulsion microstructure evaluation using phase contrast microscopy as described further ahead.

Creaming stability

The creaming stability of the emulsions was determined using creaming index (CI) after 30 days of storage. The prepared emulsion samples were poured into glass tubes and sealed to prevent moisture loss. The CI values were calculated according to equation Eq. 3.

$$\text{CI}\left(\%\right)=\frac{H_s}{H_e}\times 100$$ 3

where H_s is the height of serum layer height and H_e is the total emulsion height.

Microscopic evaluation

Phase contrast microscopy

Microstructures of CNPSPH_1:1 stabilized Pickering emulsions were observed using phase contrast microscope (OLYMPUS CX41, Tokyo, Japan) equipped a digital camera. Briefly, the samples were placed on a glass microscopic slide without a cover-slip and observed at a magnification of 100×, the images were acquired by the computer. All the observations were carried out at room temperature The droplet sizes of the emulsions were measured using an open source ImageJ program (version 1.52c). The average droplet size was reported as the mean droplet size of at least fifty droplets from the images of different points (at least 4) of the three replicates.

Fluorescent microscopy

A fluorescence microscope IX 83 (Olympus, Tokyo, Japan) equipped with 488 nm Ar and 633 nm HeNe laser, was employed to visualize the microstructure of the CNPSPH_1:1 Pickering emulsion. The Pickering emulsions specimens were stained by adding fluorescent dye solution consisting of Nile Blue and Nile Red. Nile Blue was used to visualize CNPSPH complex and Nile Red was used to visualize the oil droplets. Briefly, Nile blue staining solution was prepared by dissolving 1 mg Nile blue dye in 10 mL of MilliQ water, and Nile red staining solution was prepared by 1 mg of Nile Red in 10 mL 1,2-propanediol. All the solutions were kept in dark at room temperature unit use. Next, a 20 µL aliquot of the Nile Blue or Nile Red solutions were thoroughly mixed with 5 mL of the prepared emulsion and approximately 80 µL of the stained sample was immediately placed onto a microscopic slide. The images of the stained samples were obtained after examining under the fluorescence microscope.

Rheological behaviour of Pickering emulsion gels

The rheological properties of Pickering emulsion gels were determined at 25 °C using a rheometer with a Peltier temperature control system (MCR 72, Anton Paar, Graz, Austria) with a steel parallel plate (25 mm diameter, gap 0.100 mm) (Dai et al., 2017). For each measurement, the emulsions were deposited onto the plate and waited for 5 min to allow temperature equilibrium before measurements. For the steady-state flow measurements, the shear rate was from 0.1 to 100 s⁻¹ and the apparent viscosity (η) was obtained from data analysis software. All the dynamic tests were performed within the linear viscoelastic region. The frequency was oscillated from 0.1 to 100 rad/s and the strain was made at 1%. The elastic modulus (G′) and loss modulus (G″) were recorded in respect to frequency.

Statistical analysis

Experiments were repeated at least three times (if not mentioned before), and results were analyzed through a one-way analysis of variance (SPSS 18.0, SPSS Inc., Chicago, USA), and significant differences within treatments (p < 0.05) were evaluated Duncan Multiple Range Test (DMRT).

Results and discussion

The compositional analysis of dried banana peel showed presence of carbohydrate, crude fibre, fat, protein, and ash as 42.20 ± 1.60, 30.29 ± 1.20, 4.35 ± 1.77, and 6.90 ± 0.24% respectively on dry matter basis. The cellulose at 29.46 ± 2.86%, hemi cellulose at 15.43 ± 2.20% and lignin at 17.84 ± 1.20% were present in the banana peel flour.

Characterization of CNP and CNPSPH complex

For characterization of CNPs and CNPSPHs, particle size distribution, zeta potential and polydispersibility index were measured and the results were reported in Fig. 1S and Table 1S (Supplementary materials) respectively. These physical characterisitics prameters are highly important to consider for development of food-grade products. Firstly, the important point was to verify the generation of cellulose nano particles (CNP) after treatment. From Fig. 1SA and Table 1S, it was observed that the unsonicated hydrolyzed and dialized cellulose i.e. CNP_0 showed a very broad particle size distribution with an mean diameter of 1076 ± 2.32 nm and a PDI of 0.94. Further, as observed it is evident that sonication had a major effect on reducing particle size of CNP. On analysing the mean particle size with respect to sonication it was found that increase of sonication power and time resulted in significant decrease in particle size. However, only with sonication at 400 W resulted in reaching the size required to define a material to be of nano-scale. Further, the maximum size reduction we could achieve was 53.11 ± 1.45 nm at 30 min of sonication at 400 W power level (Table 1S). Further, the PDI data (Table 1S) of the sonicated samples ranged between 0.43 and 0.21, indicating that the samples were monodispersed in nature, having a good uniformity in particle size distribution which is also evident from narrow particle size distribution range (Fig. 1SA). Overall, these results suggested that the cellulose nanoparticle (CNP) were successfully synthesized upon sonication at 400 W for 30 min. The conjugate of whey protein isolate and gellan gum with sonication power inputs up to 400 W improves the degree of glycation, showing that the sonication power had a positive effect on the intramolecular interaction and mobility of the reaction constituents (Dev et al., 2021). According to the study, the thermal, mechanical, and chemical effects of ultrasonication help to catalyse the protein-polysaccharide interaction in the Maillard reaction. The intense pulse ultrasonication causes cavitational bubbles in the reaction system to rise and fall. It ultimately increases the shear stress and temperature by repeatedly producing positive pressure cycles and incredibly turbulent conditions (Dev et al., 2021).

The average mean particle diameter for the variously complexed CNPSPHs ranged between 159.56 ± 3.08 to 499.00 ± 2.59 nm (Table 1S, Fig. 1SB). A 1:1 mass ratio of CNPs and SPHs (i.e. CNPSPH_1:1) exhibited the smallest particle size. Data from Table 1S also revealed that a increase or decrease in the mass ratio of either CNPs or SPHs the mean diameter of the complexes increased significantly. Although in all complexes the size distribution were rather narrow, the narrowest size distribution was obtained at the mass ratio of CNPs and SPHs 1:1 (CNPSPH_1:1), suggesting good stability and uniformity of the complexes. Also, the PDI values (Table 1S) suggested that CNPSPH_1:1 had a quite uniform particle size distributions as unlike the other prepared complexes.

The performance of nanoparticles as emulsifying or stabilizing agents is associated with its surface characteristics. Zeta potential is an indicator of the electrical charge near the surface of the nanoparticles. From Table 1S, it is clear that the zeta potential values increased significantly with respect to increasing ultrasonic power and sonication time. The values of the hydrolyzed cellulose changed from − 10.64 ± 1.92 to − 34.33 ± 0.77 mV, indicating that the resulting particles became electrically stable as also suggested by Tibolla and co-workers (2014). Earlier, Sejersen et al. (2007) reported that, the zeta potential of the nanosized particles should be high, so that the colloidal suspension can resist aggregation. The lower PDI value of the hydrolysed cellulose particles suggesting that CNP samples were monodispersed without significant aggregation (Ge et al., 2017). Overall, cellulose nanoparticles (CNP) obtained after ultrasonication treatment at 400 W for 30 min were of the smallest particle size, and having highest zeta-potential indicating best stability. Thereby, CNP_400_30 were chosen to be the best candidate for preparing the complexes with soy-protein hydrolysates (SPH).

Again the zeta potential of the complexed samples were between − 13.87 ± 1.92 and − 24.48 ± 1.17 mV (Table 1S). However, these values were lower in comparison to the CNPs. The plausible reason behind might be due to the absorption of CNPs on the surface of the protein molecule and possibly forming covalent attachments to the ionisable amino groups causing a reduction in zeta potential (Boostani et al., 2017). The complex with equal amount of polysaccharide and protein (CNPSPH_1:1) exhibiting highest value of zeta potential confirmed the electrical stability. Wang (2013) confirmed that image charge repulsion of highly charged particles regardless of their sign of charge can hinder the adsorption of particles at the interface and prevent the formation of Pickering emulsions. The PDI of the CNPSPH_1:4 and CNPSPH_4:1 attained a value higher than 0.5 (Table 1S), indicating that those samples were polydispersed due to aggregation or agglomeration of the particles, while the other complexes with CNPs to SPHs mass ratio 2:1, 1:1 and 1:2 were monodisperesed.

Thermal properties of cellulose nanoparticles soy-protein hydrolysate complexes (CNPSPH)

The DSC thermograms of Cellulose nano particles (CNP), native soy protein hydrolysate, (SPH) and CNPSPH complexes are shown in Fig. 1SC. The thermographs of an endothermic peak appeared from 7 to 150 °C due to water evaporation for CNP, similar results have also been observed for nanocellulose isolated from Brewer’s spent grain (Matebie et al., 2021). SPH exhibited two prominent endothermic peaks, typically such an characteristic is observed for soy proteins due to the denaturation of β-conglycinin (7s fraction) maximal peak temeprature and glycinin (11s fraction) at maximal peak temperature at 107 and 173 °C respectively. However, both the peaks observed are at higher temepratures which might be attributed to presence of crystals of salt remaining post hydrolysis and lyophilization. Similar peak shifting has been observed by Tang et al. (2007) when compared with native soy protein isolates. Even for the CNPSPH complexs the only the shifting of peaks has been observed compared to SPH, which might have resulted due to the conjugation of the CNP with SPH. This also suggests that CNP did not modify the SPH conformation chemically rather modified the structure and stability of SPH by physical interactions such as hydrophobic, hydrogen bonding interactions and Van der Waals’ force or by cross-linking (Liu et al., 2021). As a result the thermal stability of CNPSPH increased. This is alo evidenced by the lowering of the ΔH value of β-conglycinin and glycinin after conjugation (Table 1) compared to the uncomplexed SPH. Among the complexed samples, CNPSPH_1:1 and CNP_2:1 had the lowest value of ΔH (Table 1). This might be because of higher number of CNP and SPH involved in the complexation process with each other.

Table 1.

Changes in enthalpy of CNP, SPH and CNPSPH complexes

Sample	$\Delta \mathit{H}$1(mJ/mg)	$\Delta \mathit{H}$2(mJ/mg)
CNP	− 30.86	–
SPH	− 59.06	− 20.53
CNPSPH_4:1	− 29.44	− 15.11
CNPSPH_2:1	− 22.22	− 5.86
CNPSPH_1:1	− 22.82	− 16.96
CNPSPH_1:2	− 37.03	− 11.50
CNPSPH_1:4	− 35.77	− 13.50

*$\Delta H$₁ and $\Delta H$₂ corresponds to change in enthalpy for first and second peak respectively

Wettability and contact angle of CNPSPH complex

The interfacial wettability of the particles as an emulsifier plays significant role in the formation of stable Pickering emulsion. The wetting properties would be affected by the hydrophobic variation of the particles. Tzoumaki et al. (2011) concluded that the suitable wettability could form a steric hindrance in preventing the droplet coalescence by promoting the particles to absorb on the oil-water interface. The wettability of CNPSPHs were measured through investigating the oil-in-water three-phase contact angles (θ_o/w). Figure 1 showed the three-phase contact angle (θ_o/w) of CNPSPH at different CNPs to SPHs mass ratios. Li et al. (2012) reported that emulsions showed the greatest stability against coalescence when the contact angle was about 90°. Interestingly, the θ_o/w of CNPSPHs at a CNPs to SPHs mass ratio of 1:1 was 89.07°, suggesting that CNPSPH_1:1 was potentially suitable for acquiring stable Pickering emulsions. The θ_o/w of CNPSPHs at mass ratio 4:1 was around 106.72°. This result might be attributed hydrophobic nature of the cellulose nanoparticles. In contrast, the θ_o/w values of CNPSPHs at CNP to SPH mass ratios of 1:2 and 1:4 were obviously lower than 90°, suggesting that excessive SPHs were not suitable to form a stable Pickering emulsion (Deng, 2021).

Fig. 1.

Contact angle of CNPSPH complex; a CNPSPH_4:1, b CNPSPH_2:1, c CNPSPH_1:1, d CNPSPH_1:2, e CNPSPH_1:4.

Emulsifying properties of cellulose nanoparticles - soy protein hydrolysates (CNPSPH) complexes

The emulsifying properties of complexes are provided in Table 2. Reiffers-Magnani et al. (2000) stated that the quantity of proteins could affect the amount of proteins distributing at oil–water interface and also influenced emulsion performance. The complexes with cellulose nanoparticles and soy protein mass ratio 1:1 (CNPSPH_1:1) showed significantly (p ≤ 0.05) highest emulsifying activity (232.24 ± 0.88 m²/g) followed by CNPSPH_2:1 (200.03 ± 1.20 m²/g). The enhancement of emulsifying properties of complexed samples was due to the fact that a combining effect of stabilizing impact of the cellulose nanoparticles and emulsifying activity of soy protein, which might reduce the coalescence. In CNPSPH_1:1 and CNPSPH_2:1 there was no change in the Emulsion stability index (ESI) value absorbance after 10 min (Table 2) which clarified that those two samples showed the highest stability against separation due to increase in viscosity and enhancement of gel-like network formed because of covalent binding between cellulose nanoparticles and soy protein.

Table 2.

Emulsifying properties of CNPSPH complexed samples

Sample	EAI (m2/g)#	ESI (min)
CNPSPH_4:1	145.93 ± 0.20a	9.42 ± 0.18c
CNPSPH_2:1	200.03 ± 1.20d	**
CNPSPH_1:1	232.24 ± 0.88e	**
CNPSPH_1:2	167.47 ± 0.01b	6.14 ± 0.16b
CNPSPH_1:4	195.38 ± 0.78c	5.76 ± 0.26a

*Each value represents the meanoftriplicates. Values are the means ± S.D and ^a−d means in the same column having different superscripts are significantly different at p ≤ 0.05 by Duncan Multiple Range Test (DMRT)

**The emulsion was not separated within 10 min

^#The oil value fraction (ɸ) was 0.5

Characterization of the pickering emulsion stabilized by CNPSPH complex

Droplet size distribution of the pickering emulsion stabilized by CNPSPH complex

The emulsification efficiency of formulated Pickering emulsions stabilized by CNPSPH_1:1 was assessed with different φ values of oil 0.1–0.7, using droplet size and microscopy of the emulsion. This result was similar with a previous study by Dai et al. (2017) explained that at fixed volume of emulsions an increase of oil phase volume decreased the amount of protein particles available for stabilizing the oil/water interface and causing the formation of large droplets. No separation of phase was observed at intermediate oil contents (φ = 0.50) because the droplets were so close packed together that they could not move upwards. At high oil contents (φ = 0.70), an oil layer was observed, which might due to that the interfacial layers were disrupted because of closely packed droplets (Dai et al., 2017). The observations of the emulsions using optical microscope Fig. 2A revealed that the drop characteristics changes with pH. The emulsions showed highly stable against droplet aggregation at pH 5.0 and 7.0 and this might be due to that CNPSPH complex exhibited a better stability at range. For a particular oil fraction the droplet sizes of the emulsions were lower at pH 5 and 7.0 than at pH 3.0 and 9.0. At reduced pH 3.0, the mean droplet size was increased; this might be due to the coalescence of the emulsion droplets because of partial coverage of the surface resulting from the hydrolysis of the protein molecules. Further increasing the pH to 9.0, there was a phase separation in the emulsion which could be due to complete breakdown of the protein molecule at high pH. Lu et al. (2017) reported that cellulose nanocrystals could enhance emulsions stability against coalescence at pH 5.0 and 7.0 by promoting formation of networking system, which acted as a mechanical barrier to coalescence.

Fig. 2.

Microscopic images of CNPSPH_1:1 stabilized Pickering emulsions containing 50% oil at (A) different temperature (20, 30, 40, 50, 60, 70 and 80^oC), (B) different pH (3.0, 5.0, 7.0, 9.0) and (C) Fluorescence micrograph of emulsions droplets stabilized by CNPSPH_1:1 complex (oil and protein was stained with Nile Red and Nile Blue, respectively)

Influence of temperature on the stability of Pickering emulsions was also investigated (Fig. 2B). After emulsification, Pickering emulsions by CNPSPH_1:1 complex was incubated at the temperature ranging from 20 to 80 °C for 30 min. The microscopic images showed that the droplet size of the emulsions were increased slightly while the temperature was increased from 20 to 60 °C, indicating Pickering emulsion gels were kept stable against the aggregation. However, the mean diameter of droplets increased when the treatment temperature was above 60 °C, might be due to swelling of the CNPSPH complex at high temperature (Ge et al., 2017). Further Mwangi and co-worker (2016) also reported that this phenomenon of increasing mean droplet size of Pickering emulsion at higher temperature was interpreted as the Brownian motion of particles increased with the rise of temperature, which might lead to redistribute particles on the oil/water interface and expose the surfaces of droplets increasing the tendency of droplet to coalesce.

The Fluorescence microscopy micrographs clearly demonstrated the adsorption of CNPSPH complex on the droplet surface (Fig. 2). The oil and CNPSPH was stained with red and blue fluorescence dye, respectively. Both complete and partial surface coverage by CNPSPH were observed. Non-adsorbed CNPSPH particle was also visible in the continuous phase.

Creaming stability of the CNPSPH stabilized Pickering emulsion at different pH and temperatures

Temperature plays an important role in stability of Pickering emulsions. The pH is also an important factor in regulating the electrostatic interactions between adjacent nanoparticles at the oil-water interface (Zoppe et al., 2012). Results (Table 3) showed that the creaming stability of emulsions starts decreasing from 60 °C. It could be explained by that at higher temperature the cellulose nanoparticles soy protein complex started swelling and increased the droplet size, resulted in increasing the creaming index as explained by Ge et al. (2017). The Pickering emulsions showed highly stable against droplet coalescence at pH 5.0 and 7.0. For the emulsions produced at pH 3.0 and 9.0, a slightly yellow oil layer was detected at the glass vials after 45 days of storages, which indicate that phase separation occurred due to coalescence of dispersed droplets. Moreover, the emulsions with 30 and 50% oil showed the highest stability against creaming.

Table 3.

Creaming stability of the CNPSPH_1:1 stabilized emulsions at different pH and temperatures

	pH*				Temperature (°C)**
	3.0	5.0	7.0	9.0	20	40	50	60	70	80
φ = 0.1	62.70 ± 0.61c,o	23.86 ± 0.54a,m	31.89 ± 0.75b,m	73.02 ± 0.69d,n	19.88 ± 0.86a,n	20.08 ± 0.67a,n	25.41 ± 1.03b,n	27.02 ± 0.37c,n	60.88 ± 0.50d,n	77.58 ± 1.35e,n
φ = 0.3	59.32 ± 1.15c,n	23.52 ± 0.91a,m	30.84 ± 0.61b,m	68.52 ± 0.95d,m	7.05 ± 0.84a,m	5.90 ± 0.67a,m	8.72 ± 0.28b,m	10.42 ± 0.50c,m	45.63 ± 1.15d,m	52.45 ± 1.16e,l
φ = 0.5	56.74 ± 1.06c,m	20.30 ± 0.82a,l	23.53 ± 0.87b,l	63.61 ± 1.34d.l	2.21 ± 0.28a,l	2.58 ± 0.41a,l	3.21 ± 0.21a,l	5.59 ± 0.80b,l	33.57 ± 0.87c,l	76.97 ± 0.63d,n
φ = 0.7	40.26 ± 0.67a,l	40.36 ± 1.35a,n	40.78 ± 0.91a,n	63.51 ± 1.41b,l	21.38 ± 0.37a,o	19.81 ± 1.16a,n	25.17 ± 0.27b,n	26.85 ± 0.88b,n	61.98 ± 1.36c,n	71.11 ± 1.40d,m

*Each value represents the mean ± S.D of triplicates. Values are the means ± S.D and ^a−d means in the same column having different superscripts are significantly different at p ≤ 0.05 by Duncan Multiple Range Test (DMRT). **Each value represents the mean ± S.D of triplicates. Values are the means ± S.D and ^a−e means in the same column having different superscripts are significantly different at p ≤ 0.05 by Duncan Multiple Range Test (DMRT)

Rheological property of the CNPSPH stabilized emulsions

The rheological properties of emulsions were investigated by dynamic oscillatory measurements since rheological properties of emulsions have a significant effect on processing and application of the emulsions. The measurements showed that with increasing shear rate from 0.1 to 100s⁻¹ the apparent viscosity (η) was gradually decreased, suggesting that CNPSPH emulsions exhibits the shear-thinning behaviour (Fig. 3a,b). Binks and Lumsdon (2001) suggested that deflocculation behaviour of oil droplets in emulsion gels was reflected by this phenomenon. The magnitude of apparent viscosity of the emulsion with oil fraction 0.5 was higher than the others, moreover the emulsion with oil fractions 0.1 and 0.7 showed shear thinning behaviour at lower shear rate and exhibits Newtonian characteristics at higher shear rate may be due to bulk coalescence in emulsion (Ge et al., 2017). In all the emulsions the storage modulus (G′) were higher than the corresponding loss modulus (G″), indicating that the elastic gel-like structure were formed. CNPSPH_1:1 stabilized emulsion with oil fraction varied from 0.1 to 0.7 showed in Fig. 3a,b, indicating that the emulsion with oil fraction 0.1, 0.3 and 0.7 both the moduli were lower than the emulsion with oil fraction 0.5. Song et al. (2015) reported that the suitable oil fraction for Pickering emulsion is 0.5, elasticity of the emulsion decreased when the oil fraction was more than 50%.

Fig. 3.

Rheological properties of Pickering emulsion gels prepared by a, b CNPSPH with various oil fractions (ɸ=0.1, 0.3, 0.5, 0.7); (η: apparent viscosity, Gʹ: storage modulus, Gʹʹ: loss modulus

In this work, we demonstrated the CNPSPH complexes could be used as effective food-grade stabilizers of Pickering emulsion. Banana peel can serve as potential precursor for their conversion into CNPs. Ultrasonication treatment considerably reduced the CNP particle size. The emulsions having equal mass ratio of CNPs and SPHs exhibited better emulsification efficiency and stronger inter-particle structure, and the higher stability against coalescence and coacervation. As expected, the CNPSPHs based Pickering emulsion gels were successfully fabricated at high oil volume fractions (φ = 0.5). Furthermore, the fluorescence images fully confirmed efficient CNPSPHs absorbed at the oil/water interface forming steric barriers. The result of dynamic rheology showed that G′ was higher than G″, revealing the formation of elastic gel networks in the systems. Nevertheless, the gel structure prompted a long-term storage stability (45 days) against coalescence and Ostwald ripening of Pickering emulsion gels. These findings would likely be a major significance for not only develop a CNPSPH stabilized food-grade Pickering emulsion gels, but also for developing a delivery system for bioactive further.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Conflict of interest

The authors have no competing interests to declare that are relevant to the content of this article.