Arrowroot starch-based films incorporated with a carnauba wax nanoemulsion, cellulose nanocrystals, and essential oils: a new functional material for food packaging applications

Arrowroot starch (AA)-based films incorporated with a carnauba wax nanoemulsion (CWN), cellulose nanocrystals (CNCs), and essential oils (EOs) from Mentha spicata (MEO) and Cymbopogon martinii (CEO) were produced using the casting technique and then characterized in terms of their water barrier, tensile, thermal, optical, and microstructural properties and in vitro antifungal activity against Rhizopus stolonifer and Botrytis cinerea. Whereas the incorporation of CNCs decreased the moisture content and water vapor permeability of the AA/CWN/CNC film, the additional incorporation of either EO decreased the transparency and affected the microstructure of the AA/CWN/CNC/EO nanocomposites. MEO and CEO incorporation improved the thermal stability of the films and provided excellent protection against fruit-spoiling fungi. Because of their excellent barrier properties against fungal growth, water vapor permeability, and ultraviolet and visible light, these AA/CWN/CNC/EO films have promising potential for application as active food packaging or coating materials.


Introduction
Food packaging is traditionally based on polymers of petrochemical origin, such as polypropylene, polyethylene, and polystyrene, owing to their low costs and well-established production. However, the nonbiodegradability of these materials has caused serious environmental problems. For this reason, immense efforts are being made in the research and development of biodegradable films for food packaging applications, with one new approach being the use of biopolymers extracted from food sources (Mahcene et al. 2020).
Among the various biopolymers, starch is a good raw material for the production of edible and biodegradable films because of its transparency, good gas barrier property, high availability, and low production cost (Thakur et al. 2019). Arrowroot (Maranta arundinacea), an unconventional food species native to Latin America, has a high amylosebased starch content in its rhizome. Amylose-rich starches are an interesting resource for the production of biodegradable films with good mechanical properties (Valadares et al. 2020). Although the hydrophilic nature of starch-based films makes them poor water barriers (Thakur et al. 2019), this disadvantage can be minimized through the incorporation of hydrophobic compounds, such as oils, fats, fatty acids, and waxes (e.g., carnauba wax) (Rodrigues et al. 2014;Syahida et al. 2020).
However, the addition of these compounds may create discontinuities in the film structure, negatively affecting its optical, mechanical, and barrier properties (Syahida et al. 2020;Zhang et al. 2018). In recent years, nanotechnology has been used as a new strategy for improving the properties of films (Espitia et al. 2019). For example, the incorporation of a nanoemulsion of carnauba wax into arrowroot starch (AA)based films lowered the water vapor permeability, improved the transparency and tensile strength, and smoothened the microstructure of the nanocomposites in comparison with those same features in films made with a microemulsion of the wax (Oliveira Filho et al. 2020a). Thus, the formulation of starch-based films incorporated with lipid nanoemulsions is a promising approach for the development of new film materials with improved properties.
Cellulose nanocrystals (CNCs), which are needlelike nanomaterials measuring 4-25 nm in diameter and 100-1000 nm in length (Jonoobi et al. 2015), have been proposed as a new strategy for improving the mechanical and water barrier properties of new coating or film materials, reinforcing the biopolymer matrix, and allowing the development of nanocomposite materials (Azeredo et al. 2012;Dai et al. 2020;Pereda et al. 2014). Additionally, CNCs have low or no cytotoxic effects in humans (Dong et al. 2012) and can improve the stability of lipid compounds in emulsion-based systems via noncovalent physical adsorption (Hubbe et al. 2017).
Moreover, other ingredients with biological properties can be added to nanocomposite films. Indeed, nanoemulsion-based films and coatings have shown promise for transporting natural bioactive compounds, such as essential oils (EOs) (Aswathanarayan and Vittal 2019), which are natural antimicrobials and antioxidants valued for their safe, biodegradable, and non-toxic properties (Abdollahi et al. 2013;Atarés and Chiralt 2016;Sánchez-González et al. 2011b).
The objective of this study was to develop a novel functional material for food packaging applications. To this end, we fabricated films based on an AA matrix, into which a carnauba wax nanoemulsion (CWN) (to improve the water vapor barrier property), CNCs (to enhance the tensile properties), and EOs from a green mint plant and palmarosa grass (to provide antifungal properties) were incorporated. Characterization of the AA/CWN/CNC/EO nanocomposite properties revealed the films to be promising materials for the primary packaging or coating of a variety of perishable food products, such as fresh fruits, vegetables, breads, and cheeses.

Carnauba wax nanoemulsion
The CWN (droplet size: 39.3 ± 0.7 nm; zeta potential: -40.32 ± 1.0 mV) was prepared according to the method described by Hagenmaier and Baker (1994) and adapted by Campos et al. (2019). Characterization of the nanoemulsion was carried out as previously described in another study by our group (Oliveira Filho et al. 2020a).

Film preparation
To prepare the films, CNCs and AA (5:95, w/w) were first dissolved in water, under agitation with a magnetic stirrer (150 rpm) in a thermostatic bath (TE-2005, TECNAL, Piracicaba, Brazil) at 85 ± 2°C for 5 min, to produce a 2% (w/w) aqueous solution. Then, the CWN (15% on a dry starch basis) was added to the aqueous mixture (Oliveira Filho et al. 2020a) and the suspension was homogenized. Thereafter, glycerol was added as the plasticizer to a level of 0.17 mL/g AA.
After the dispersions had been cooled to 40°C, MEO or CEO was added at the concentrations of 0.1%, 0.2%, and 0.3% (v/v) and the mixtures were stirred on a high-speed mixer (UltraTurrax T25, IKA Werke GmbH & Co, Staufen, Germany) for 5 min. The film-forming dispersions (25 mL) were cast on Petri dishes (Ø12 cm) and dried at 35°C for 24 h, following which the resultant films were detached from the plates and conditioned at ambient temperature (relative humidity (RH): 50%) for 48 h before analysis.

Film characterization
Thickness, density, moisture content, and water solubility The film thickness was determined with a portable digital micrometer (Mitutoyo Co., Kawasaki-Shi, Japan) to the nearest 0.001 mm. The measurements were carried out at a minimum of five points of each film (Oliveira Filho et al. 2020a). The film density was determined directly from the film weight and dimensions (volume) and the values considered were the average of ten determinations (Yadav and Chiu 2019).
The moisture content was measured according to the method described by Oliveira Filho et al. (2020a). Square (2 cm 2 ) film samples were weighed before and after drying at 105°C for 24 h, and the moisture content was calculated using Eq. (1).
The water solubility was measured as described by Kavoosi et al. (2014) with modifications. The initial dry weight of 2 cm 2 square film samples was determined by weighing them before and after drying at 100 ± 5°C for 24 h. The samples were immersed in 50 mL distilled water for 24 h at 23 ± 2°C and dried at 100 ± 5°C for 24 h, and the final dry weight was measured. The water-soluble content was then calculated as the percentage weight that remained after water immersion. The water solubility of the film was calculated using Eq. (2).
Water vapor permeability The water vapor permeability (WVP) of the films was determined using the gravimetric method (E96/E96M-16, 2016). In brief, the films were sealed in permeation cups (diameter: 35 mm) containing 6 mL of distilled water and placed in an air circulation oven (Solab SL-102, Piracicaba, SP, Brazil) at 40°C with activated silica gel (RH: 0%). The cups were then weighed at least 10 times over the next 34 h. The WVP (g mm h -1 cm -2 Pa -1 ) was calculated according to Eq. (3): where m is the water loss from the permeation cup, A is the film sample area, t is the time of analysis, and Dp is the difference in water vapor pressure between the inside and outside of the cup.

Tensile properties
The tensile properties of 50 mm 9 10 mm film specimens (n = 10) was determined using the D882-12 method (ASTM, 2012). This analysis was performed using the TA.XTplus texture analyzer (Stable Micro Systems, Surrey, UK) equipped with A/TG Tensile Grips and a 50 N load cell, with an initial grip separation of 20 mm and a crosshead speed of 80 mm/ min.

Thermogravimetric analysis
Thermal degradation profiles were obtained by thermogravimetric analysis (TGA), using a TGA Q500 analyzer (TA Instruments, New Castle, DE, USA) with heating from 10 to 600°C at a rate of 10°C/min and a nitrogen flow rate of 40 mL/min. The percentage weight loss (%) and the first derivative of the TGA curve (%/°C) were determined as a function of temperature.
The film opacity, which was based on the fractional transmittance at 600 nm (T 600 ) and film thickness (x, mm), was calculated using Eq. (6) (Hamdi et al. 2019).
An ultraviolet-visible (UV-Vis) spectrophotometer (Shimadzu 1600, Portland, OR, USA) was used to measure the optical barrier properties of the film against UV and visible light, with scanning carried out between 250 and 800 nm.

Scanning electron microscopy
A scanning electron microscope (JEOL-JSM 6510 model, Jeol, Tokyo, Japan), set at an acceleration voltage of 5 kV, was used to evaluate the microstructure of the films as described previously (Oliveira Filho et al. 2020a).

In vitro antifungal activity
The antifungal activity of the films (diameter: 10 mm) against R. stolonifer and B. cinerea was evaluated in vitro using the method described by Oliveira Filho et al. (2019). In brief, 100 lL of fungal spore suspension (adjusted to 10 5 spores/mL) was cultured on potato dextrose agar plates. The sample films were then placed over the fungal mat on the agar surface and the plates were incubated at 25°C for 72 h. The result was expressed as the diameter of the zone of inhibition measured with a caliper.

Statistical analysis
The results are expressed as the average of three replicates (with triplicate analyses for each repetition) ± standard deviation. The data were analyzed using one-way analysis of variance followed by the Tukey post hoc test, with statistical significance of differences set at p B 0.05.

Results and discussion
Physical properties of the films The density of the films (Table 1) varied from 0.80 to 0.95 g/cm 3 and decreased significantly with the addition of EOs, indicating that the films became less dense. Similar behavior was reported by Sánchez-González et al. (2011a), who reported a decrease in the density of films based on hydroxypropylmethylcellulose and chitosan when the EOs of bergamot, lemon, and tea tree were added. Similar behavior was reported by Yadav and Chiu (2019) for j-carrageenan films reinforced with cellulose nanocrystals. Table 1 shows the water-related properties of the films. The moisture contents varied from 7.03 to 7.89%, with no significant difference arising from the incorporation of the CNCs, MEO, and CEO. As EOs are hydrophobic in nature, it was expected that the moisture content of the films would be reduced. However, this was not observed in the present work, probably as a result of the low concentrations of EOs tested.
The solubility of the films in water varied from 12.1 to 26.4%, decreasing significantly with the addition of CNCs and increasing concentrations of either of the EOs, corroborating the density of films (Table 1). The reduction in water solubility of the AA-based films following the incorporation of CNCs was a result of the formation of a three-dimensional (3D) cellulose network through hydrogen bonding between the starch and CNC molecules. Three-dimensional networks reduce the solubility of biopolymers, reinforcing the structure and restricting the interactions between the polymer and water molecules (Noshirvani et al. 2018).
The reduction in water solubility of the films by the addition of the EOs was probably due to the hydrophobic nature of these molecules and their low affinity to water molecules (Ma and Wang 2016). The same observation has been reported for starch films incorporated with Syzygium aromaticum EO (Sousa et al. 2019) and chitosan films incorporated with Citrus limonia EO (Oliveira Filho et al. 2020b).
The addition of CNCs to the AA/CWN film reduced its WVP (from 3.98 to 3.21 10 -7 g H 2 O m -1 h -1 -Pa -1 ), similar to results reported in the literature (Abdollahi et al. 2013;Pereda et al. 2014;Sogut 2020). According to El Miri et al. (2015), CNCs limit the mobility of water molecules through the film matrix, resulting in a reduction in the WVP of the nanocomposite film.
The incorporation of the various concentrations of CEO or MEO did not alter the WVP of the films. This was similar to the results reported for whey protein films incorporated with oregano EO (Zinoviadou et al. 2009) and AA-based films incorporated with Piper aduncum EO (Valadares et al. 2020).
The addition of EOs was expected to reduce the WVP of the films owing to the hydrophobic nature of the oils, as previously observed for the water solubility property (Table 1). However, because WVP is a function of both solubility and diffusivity (Santos et al. 2014), the lack of a significant variation in the WVP may be due to a concomitant increase in water molecule diffusivity resulting from discontinuities in the matrix caused by the incorporated EO molecules (as shown by the scanning electron micrographs discussed below).

Tensile properties of the films
The film thickness increased significantly with the addition of CNCs, CEO, and MEO, similar to the results reported for films composed of whey protein isolate, CNCs, and bergamot EO (Sogut 2020) and those prepared with chitosan, CNCs, and palm oil (Pereda et al. 2014). The increase in thickness of the films may be related to the increase in the amount of solids present in the nanocomposites (de Souza Coelho et al. 2020), with differences in homogeneity within the biopolymer matrices, and could also be due to interactions between the components used in the formulation of the nanocomposites (Sogut 2020). Table 2 shows the stress properties of the films. The incorporation of CNCs increased the tensile strength of the AA/CWN film from 3.0 to 5.3 MPa. This increase can be due to interactions between the CNCs and starch molecules and the reinforcement effect from voltage transference at the CNC-starch interface (de Mesquita et al. 2010;Khan et al. 2012). The interaction between -OH groups and hydrogen bonds generated between the CNCs and the starch can favor a good interface between the matrix and the filling, which in turn result in high TS values for nanocomposite films (de Mesquita et al. 2010). To explain the effect of CNC reinforcement in the present study, the medium-field mechanical model can be adopted. The medium field model is based on the concept that nanocrystals are dispersed homogeneously in the polymeric matrix, but there is no interaction between nanocrystals (Favier et al. 1995). The stress reinforcing effect can result from efficient load transfer to a nanocrystal network, leading to a more uniform stress distribution and minimization of the stress concentration area (Kanagaraj et al. 2007;Š turcová et al. 2005).These interactions strengthen the 3D network of the nanocomposite film by creating nanofillers, which improve the mechanical properties of the film and limit the movement of the biopolymer chains (Jouyandeh et al. 2019).
The tensile strength of the films with EOs incorporated was lower than that of the AA/CWN/CNC film. However, all films were superior in tensile strength to the control film (AA/CWN without CNCs and EOs). The changes in the mechanical properties were likely due to the presence of discontinuities in the polymer matrix caused by the EO molecules (Atarés and Chiralt 2016), corroborating the film structures seen in the scanning electron micrographs and the theory of an increase in diffusivity as being responsible for the nonreduction in WVP (Table 1). Thus, EOs increase the extensibility, flexibility, and mobility of films and decrease their cohesive strength (Mahcene et al. 2020). Whereas the elongation at break was significantly reduced from 247 to 125.7% with the addition of CNCs, it was not affected by the addition of EOs. The CNC-mediated increase in tensile strength and reduction in elongation at break of the films have also been reported by other authors (Dai et al. 2020;de Souza Coelho et al. 2020;Yadav et al. 2016).
Thermogravimetric analysis Figure 1 shows the TGA curves and their first derivatives. The starting and maximum decomposition temperatures (T onset and T max ) are shown in Table 2. Weight loss of the samples occurred in three major stages. The first stage occurred during the temperature range of 25-250°C and was related to the evaporation of water molecules (52-92°C), glycerol (200-250°C), and other volatile low-molecularweight components. The second stage occurred between 270 and 350°C and was related to the thermal degradation of starch and CNCs, which occur at similar temperatures (Rico et al. 2016). The last stage occurred from 350 to 490°C (Freitas et al. 2016;Milanovic et al. 2010) and was caused by the degradation of CWN and thermally stable compounds Values in the same column followed by at least one common letter (or not followed by any letters) are not significantly different according to the Tukey test (p \ 0.05) AA arrowroot starch, CWN carnauba wax nanoemulsion, CNC cellulose nanocrystals, MEO Mentha spica essential oil, CEO Cymbopogon martini essential oil, T onset starting decomposition temperature, T max maximum decomposition temperature *Essential oils were added at the concentrations of 0.1% (MEO1 or CEO1), 0.2% (MEO2 or CEO2), or 0.3% (MEO3 or CEO3) Values of thickness, tensile strength, and elongation at break in the same column followed by at least one common letter are not significantly different according to the Tukey test (p \ 0.05) present in the EOs (Alizadeh et al. 2017;Sousa et al. 2019). The films with MEO and CEO incorporated showed higher T max values than the other films, indicating that the addition of these oils had improved the thermal stability of the nanocomposites (Table 2). These results were corroborated by other studies that showed that the improvement in thermal properties led to the higher homogeneity observed in the biopolymer matrix (Noshirvani et al. 2017;Sousa et al. 2019).

Optical properties
The optical properties of the films are listed in Table 3. The incorporation of EOs and CNCs did not change the L* parameter (luminosity) of the films. Moreover, the hue values ranged from 89.34°to 90.30°(between red and yellow), indicating that the films were yellow in color (Table 3).
The C* value decreased with the addition of CNCs, indicating that the nanocrystals lowered the color intensity of the films. By contrast, the value increased with the addition of 0.2% and 0.3% EOs, indicating that the oils made the film coloring more intense. Similar phenomena have been reported for agar films (Shankar et al. 2015) and starch films (de Souza Coelho et al. 2020). In another study, the yellowish coloration of corn and wheat starch films was attributed to the lemon EO added (Song et al. 2018).
The addition of CWN, CNCs, and EOs decreased the transparency of the AA-based films. The opacity values increased from the AA/CWN film to the AA/ CWN/CNC/MEO3 and AA/CWN/CNC/CEO3 films (1.22 to 2.70 and 2.99, respectively). This increase may be due to the strong interaction between the CNCs and the starch matrix as well as light scattering by the nanocrystals (Li et al. 2018).
The increased opacity could also be due to the hinderance of light passage through the film as a result of CNC accumulation within the matrix (Abdollahi et al. 2013), as evidenced by the CNC aggregates observed in the scanning electron micrographs (Fig. 2). Similar results have been reported by other authors (de Souza Coelho et al. 2020;Thomas et al. 2020). Meanwhile, the decrease in film transparency caused by the addition of EOs was probably due to the dispersion of light by the oil droplets in the film matrix, as previously described for other films Sousa et al. 2019). Figure 3 shows the light transmission rate of the films. All of the fabricated films were found to be strong barriers against UV light (200-350 nm) not exceeding 0.1%; that is, they provided a 100% barrier to UV light. In the visible light region (380-780 nm), the light transmission rate of the control film (AA/ CWN) was 24.0-65.4%, but this decreased to a range of 26.0-55.9% in the AA/CWN/CNC film. The addition of EOs also caused a slight reduction in light transmittance rates compared with that of the AA/ CWN/CNC film (Fig. 3). The best light barrier performance was observed for films incorporated with CNCs and EOs at the highest concentration (0.3%), which was due to the increased opacity (reported in  Table 3). Reductions in the UV-Vis light transmission rate have also been observed for chitosan films incorporated with Citrus limonia EO (Oliveira Filho et al. 2020b) and potato starch films incorporated with CNCs (Oliveira et al. 2017). Therefore, AA/CWN/ CNC films with MEO or CEO can be used as food packaging materials, as they have excellent light barrier function. Figure 2 shows the surface and cross-sectional microstructures of the AA-based films containing CWN, reinforced with CNCs, and supplemented with MEO or CEO. The AA/CWN film had a dense and regular surface with some lipid clusters present. In the film containing CNCs, the surface was rougher and more opaque, which was attributed to the nanocrystal aggregates, as observed in other studies of starch films containing CNCs (Johar and Ahmad 2012;Silva et al. 2019). None of the films had obvious cracks or discontinuities in their microstructures, and the addition of CNCs positively impacted the traction and barrier properties of the films (Tables 1, 2, and 3).

Characterization of the film microstructures
As shown in Fig. 2, compared with the AA/CWN/ CNC film, the films with EOs had lower amounts of CNC aggregates and regular and compact structures, probably as a result of the uniform distribution of the droplets within the emulsion and good compatibility between the matrices. These characteristics indicated that the emulsion was stable and there was no phase separation or droplet aggregation during the preparation and drying of the films. This may have been due to interactions between the CNCs and EOs that electrostatically stabilized the oil droplets, giving rise to Pickering emulsions (Zhang et al. 2017;Zhou et al. 2018).
With higher concentrations of EOs in the films, the microstructure of the cross-sections was slightly Fig. 2 The light transmission rate of films. The formulation codes refer to the main components, namely: AA, arrowroot starch; CWN: carnauba wax nanoemulsion; CNC: cellulose nanocrystals; MEO1: Mentha spicata essential oil at 0.1%; MEO2: Mentha spicata essential oil at 0.2%; MEO3: Mentha spicata essential oil at 0.3%; CEO1: Cymbopogon martini essential oil at 0.1%; CEO2: Cymbopogon martini essential oil at 0.2%; CEO3: Cymbopogon martini essential oil at 0.3%  (Pastor et al. 2013;Zhou et al. 2018)]. The characteristics of oil droplets in a Pickering emulsion can affect the immobilization of the emulsion (Ribeiro- Santos et al. 2017). Overall, the findings of the film microstructure corroborated the results previously described for the properties that were improved with the addition of CNCs and EOs.
Antifungal activity Table 4 shows the antifungal activities of the films against R. stolonifer and B. cinerea. As expected, the AA/CWN and AA/CWN/CNC films did not show antifungal activity against the two fungi studied, corroborating previous results obtained with films based on starch, waxes, and nanocellulose (Ochoa et al. 2017;Raigond et al. 2019;Salmieri et al. 2014). By contast, the films with EOs incorporated showed obvious antifungal activity that was directly proportional to the EO concentration used. The diameters of the inhibition zones against R. stolonifer increased from 16.0 to 25.2 mm for films with MEO and from 18.7 to 29.8 mm for films with CEO. Those against B. cinerea increased from 19.0 to 33.7 mm for films with MEO and from 24.3 to 29.8 mm for films with CEO. Thus, B. cinerea was more sensitive than R. stolonifer to the EOs studied. The antifungal effect of MEO is related to its chemical composition, mainly of carvone, which has high antimicrobial activity (Soković et al. 2009). By contrast, the antifungal activity of CEO is related to the synergistic effects of its major compounds: geraniol, linalool, neral, and mirceno (da Rocha Neto et al. 2019). Taken together, these results confirmed that MEO and CEO act as antifungal agents. Therefore, these EOs in combination with a film composed of AA, CWN, and CNCs provide a functional film material.

Conclusions
The incorporation of CNCs into an AA-based matrix increased the thickness and tensile strength and decreased the WVP of the AA/CWN/CNC film. Furthermore, the addition of CNCs with or without EOs decreased the transparency of the films and improved their visible light barrier property. Additionally, the incorporation of EOs improved the thermal stability of the films. The microstructure of the films was affected by the CNCs, becoming rougher and more opaque. The addition of EOs provided the films with excellent antifungal activity against postharvest fruit-spoiling fungi. Overall, films composed of AA, CWN, CNCs, and MEO or CEO were excellent barriers against WVP, UV light, and fungal growth. Thus, the AA/CWN/CNC/MEO and AA/CWN/CNC/ CEO nanocomposite films represent novel materials with potential application as active packaging materials or coatings for fresh fruits and vegetables as well as for other food products that are easily spoiled by surface fungal growth, such as breads and cheeses.