Mango kernel starch films as affected by starch nanocrystals and cellulose nanocrystals

d to obtain components for nanocomposite films, namely, starch and starch nanocrystals (SNC) from seed kernels, and cellulose nanocrystals (CNC) from seed shells. Lignin was also recovered from shells. Starch-based films were prepared with different contents and combinations of SNC and CNC.

Carbohydrate Polymers 211 (2019) 209–216 Contents lists available at ScienceDirect Carbohydrate Polymers journal homepage: www.elsevier.com/locate/carbpol Mango kernel starch films as affected by starch nanocrystals and cellulose nanocrystals T Ana Priscila M Silvaa, Ana Vitória Oliveiraa, Sheyliane M.A Pontesb, André L.S Pereirab, ⁎ Men de sá M Souza Filhoc, Morsyleide F Rosac, Henriette M.C Azeredoc,d, a Federal University of Ceara, Department of Chemical Engineering, Campus Pici, Bl 709, 60455-760, Fortaleza, CE, Brazil Federal University of Ceara, Department of Organic and Inorganic Chemistry, Campus Pici, Bl 940, 60451-970, Fortaleza, CE, Brazil c Embrapa Agroindústria Tropical, R Dra Sara Mesquita, 2270, Fortaleza, CE, 60511-110, Brazil d Embrapa Instrumentaỗóo, R 15 de Novembro, 1452, Caixa Postal 741, São Carlos, SP, CEP 13560-970, Brazil b A R T I C LE I N FO A B S T R A C T Keywords: Biopolymers Polysaccharides Food packaging Nanoparticles Reinforcement Mango seeds have been used to obtain components for nanocomposite films, namely, starch and starch nanocrystals (SNC) from seed kernels, and cellulose nanocrystals (CNC) from seed shells Lignin was also recovered from shells Starch-based films were prepared with different contents and combinations of SNC and CNC SNC exhibited round-like rather than platelet-like morphology, and their effect on water vapor barrier was not as high as that of the needle-like CNC Also, CNC were more effective than SNC to increase elastic modulus On the other hand, CNC impaired more the elongation The optimized conditions (1.5 wt% CNC and 8.5 wt% SNC on a starch basis) resulted in a film with enhanced strength, modulus, and barrier to water vapor when compared to the unfilled film, although the elongation has been impaired Introduction In the last few decades, there has been an increasing focus on process and product development as part of a biorefinery approach, as well as a transition from petrochemistry to bioeconomy Those are aspects of global trends aiming to reduce wastes (particularly non-biodegradable wastes, but also food wastes) and to improve the economics of food processing Food processing byproducts can be converted into a variety of chemicals, including biopolymers that represent alternatives for fossil based polymers for some applications, including food packaging Mango is an important tropical fruit, very appreciated for its peculiar flavor The world mango production in 2016 has been about 46.5 million metric tons (Statista, 2018),‘Tommy Atkins’ being the most produced mango variety (UNCTAD, 2016) Mangoes have been recently suggested as potential feedstocks for integrated biorefineries (Arora, Banerjee, Vijayaraghavan, MacFarlane, & Patti, 2018; Banerjee et al., 2018; Matharu, Houghton, Lucas-Torres, & Moreno, 2016) The mango seed (stone) is a major by-product of mango processing, representing about 22% of the weight of Tommy Atkins mangoes (Plant-O-Gram, 2015); it contains a kernel containing more than 50% starch by weight (Kaur, Singh, Sandhu, & Guraya, 2004), and a shell whose major component (more than 50%) is cellulose (Henrique, Silvério, Flauzino Neto, & Pasquini, 2013) ⁎ Starch has a variety of food and non-food applications Conventional edible starch sources such as corn, potato, and cassava are usually of choice for food applications, whereas non-edible sources such as food byproducts have been frequently preferred for non-food applications, in order to avoid issues around food versus non-food competition, as well as to allow sustainable use of resources (Persin et al., 2011) Starch has been frequently used for production of biodegradable films for food packaging purposes However, since the tensile and barrier properties of starch films are usually poor, many studies have focused on the use of nano-reinforcements to improve their properties Cellulose nanocrystals (CNC) have been widely studied and proven to be effective to enhance the performance of starch films (Chen, Liu, Chang, Cao, & Anderson, 2009; Slavutsky & Bertuzzi, 2014; TerrazasHernandez et al., 2015) The effects of starch nanocrystals (SNC) on starch films have also been studied (Angellier, Molina-Boisseau, Dole, & Dufresne, 2006; García, Ribba, Dufresne, Aranguren, & Goyanes, 2011; Piyada, Waranyou, & Thawien, 2013), which are favored by the high affinity provided by the identical nature of matrix and filler A previous study has aimed to compare SNC and CNC effects on thermoplastic starch (González, Retegi, González, Eceiza, & Gabilondo, 2015), but only at the wt% reinforcement level Some recent studies have proposed mango kernel starch as a matrix for films or coatings, usually without nanollers (Nawab, Alam, Haq, & Corresponding author at: Embrapa Instrumentaỗóo, R 15 de Novembro, 1452, Caixa Postal 741, São Carlos, SP, CEP 13560-970, Brazil E-mail address: henriette.azeredo@embrapa.br (H.M.C Azeredo) https://doi.org/10.1016/j.carbpol.2019.02.013 Received November 2018; Received in revised form February 2019; Accepted February 2019 Available online 05 February 2019 0144-8617/ © 2019 Elsevier Ltd All rights reserved Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al 2.3 Isolation of starch nanocrystals (SNC) Hasnain, 2016, 2018) Our group have studied the properties of mango kernel starch films as affected by SNC from mango kernel starch (Oliveira et al., 2018), demonstrating that those nanocrystals have improved tensile and barrier properties of the films In this study, SNC (from mango kernels) and CNC (from mango kernel shells) have been combined as an attempt to further improve the physical properties of mango kernel starch films, as well as to compare the effects of SNC and CNC (both from mango seeds) on those films The production of starch nanocrystals (SNC) was based on an acid hydrolysis procedure adapted from the method described by Angellier, Choisnard, Molina-Boisseau, Ozil, and Dufresne (2004), as described by Oliveira et al (2018) 73.5 g of mango kernel starch were added to 500 mL of a mixture (volume ratio, 1:1) of HCl 3.16 M and H3PO4 3.16 M at 40 °C in a silicone bath under stirring (150 rpm) for days The suspensions were then submitted to six centrifugation-sonication cycles consisting of centrifugation (25,384 g, 20 min, °C), homogenization of the precipitate in a Walita Billy mixer (Walita, São Paulo, Brazil) for min, sonication on a Unique Cell Disruptor (Unique, Campinas, SP, Brazil) for 20 at 450 W The suspension was then dialysed until pH 6.5, then homogenized with an Ultra-Turrax T50 (5000 rpm, min), added with 0.01 vol% chloroform, and stored at °C The SNC yield was taken as the average yield from the three batches, based on the initial dry weight of starch The SNC suspension was the mixture of the suspensions from the three batches, and its solid content was determined by gravimetry after drying at 105 °C until constant weight Materials and methods 2.1 Materials and reagents Mangoes (Mangifera indica L cv ‘Tommy Atkins’) at ripening stage (GTZ, 1992) were purchased at a local market (CEASA, Maracanaú, CE, Brazil) Sodium hydroxide (NaOH), hydrochloric acid (HCl), phosphoric acid (H3PO4), sulfuric acid (H2SO4), acetic acid (C2H4O2), and hydrogen peroxide solution (H2O2 30 vol%) were purchased from Sigma-Aldrich Ltd (St Louis, USA), and absolute ethanol (C2H5O, > 99.9%), from Merck (Germany) 2.4 Processing of mango seed shells and isolation of cellulose nanocrystals (CNC) 2.2 Isolation and characterization of mango kernel starch (MKS) The processing of mango seed shells was carried out in three batches, and the yield of CNC and lignin were calculated as averages of the values from those batches Mangoes were manually depulped The starch isolation was adapted from the method described in the previous study by our group (Oliveira et al., 2018) The mango stones were decorticated, and the kernels were immersed in distilled water After all stones were decorticated, the kernels were removed from water, cut into small pieces (about cm3), and immersed into sodium metabisulfite 0.2 wt% (kernel: solution ratio, 1:2, wt/vol) for 48 h at °C, again to avoid browning The kernels were then divided into three batches for further processing The kernels were drained and ground in distilled water (kernel:water weight ratio, 1:1) with a cutter (R 502 VV., Roubot-Coupe, Vincennes, France) for at 1750 rpm, the resulting suspension was filtered in a 60 mesh sieve stirred (200 rpm) for 10 The retentate was again ground in water in the cutter and filtered, and the procedure was repeated until there was no retentate The starch slurry was allowed to settle for 40 min, separated, added with a 0.05 M NaOH solution (at a 1:2 slurry/solution weight ratio) in order to remove soluble fiber, and stirred at 200 rpm for h at 25 °C The suspension was then centrifuged (20,000 g, °C, 20 min) in a High-Speed Refrigerated Centrifuge CR22GIII (Hitachi, Tokyo, Japan) The starch precipitate was collected, repeatedly washed in distilled water (at a 1:2 precipitate/water weight ratio), stirred (200 rpm, min), and centrifuged, until the starch suspension reached pH Absolute ethanol was added to the suspension (2:1 ethanol/suspension volume ratio), which was stirred (200 rpm, h), then allowed to settle for h at 25 °C The decanted starch was washed with distilled water, vacuum filtered through 28 μm filter paper, allowed to dry in an oven (40 °C, 24 h), ground with an analytical mill (Ika A11, Staufen, Germany), and stored in sealed glass flasks The starch yield was expressed as extracted starch to initial raw kernels (wt%, on a dry basis), as an average value of the three batches The starch was analyzed for ashes (ASTM E1755-01, 2015ASTM E1755-01, 2015) and proteins (Bradford, 1976) in triplicate The amylose content was measured according to the standard method ISO (6647)-1 (2015), also in triplicate A scanning electron micrograph (SEM) was obtained from a starch powder placed on a stub and coated with a 20 nm layer of gold by using a Quorum QT150ES metallizer (Quorum, Laughton, UK) The micrograph was taken with a Quanta 450 FEG-FEI (Quanta, Hillsboro, USA), with an acceleration voltage of 15 kV, viewed at 2000× magnification The starch granule sizes were taken as averages of 50 measurements taken by using the software ImageJ 2.4.1 Isolation of CNC The CNCs were obtained from previously dried mango seed shells (in an oven at 50 °C until constant weight) The dried shells were ground in a Willye knife mill (STAR FT-80, Fortinox, Piracicaba, Brazil) with a 10 mesh (2.0 mm) sieve, forming the mango shell powder (MSP) The MSP was then submitted to acetosolv pulping The acetosolv solution (93 wt% acetic acid and 0.3 wt% hydrochloric acid in distilled water) was added to MSP at a 10:1 wt-to-volume ratio, and the mixture was cooked under reflux at 120 °C with stirring (150 rpm) for 90 in a flat-bottom flask in a mineral oil bath The mixture was filtered (in a 28 μm mesh) to separate the black liquor from the mango shell acetosolv pulp (MSAP) The MSAP was washed with acetic acid (99.7%) at 80 °C until the washing liquid was colorless and dried (60 °C, 24 h) Bleaching was then carried out For each 10 g of MSAP, 200 mL NaOH 4% (w/v) were heated to 40 °C, when the AP was added under stirring When the temperature of the mixture reached 65 °C, 60 mL H2O2 30 vol% were added After 60 at this same temperature, more 60 mL H2O2 30 vol% were added, and the reaction was kept for more 90 The material was then vacuum-filtered, washed with distilled water until its pH was equal to that of the washing water, and ovendried at 50 °C until constant weight, producing the bleached acetosolv pulp (MSBAP) The acid hydrolysis method used to obtain the CNC was based on a mixture of sulfuric acid 11.3 M with hydrochloric acid M at a volume ratio of 3:1, which was defined from a series of preliminary tests The acid solution (200 mL) was stirred in a round-bottom flask immersed in a 45 °C mineral oil bath When the solution temperature reached 45 °C, 10 g of MSBAP were added, and the hydrolysis was carried out for 60 under stirring The reaction was quenched by adding cold (4 °C) deionized water at a 1:3 volume ratio (suspension/water) The suspension was then centrifuged (26,400 g, 15 min, 20 °C), and the supernatant was discarded Deionized water (500 mL) was added again, and two more centrifugation cycles were conducted The final precipitate was added again with deionized water, and the suspension was sonicated with a Unique Cell Disruptor (60 W, in three cycles interspersed with intervals) The suspension was then dialysed against deionized water until neutrality, and the CNC suspension was stored at °C The CNC yield was based on the initial dry weight of 210 Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al MSBAP The CNC suspension was the mixture of the suspensions from the three batches, and its solid content was determined by gravimetry after drying at 105 °C until constant weight samples and conditioned for 48 h under controlled temperature and relative humidity (23 ± °C, 50 ± 3% RH) before analyses 2.4.2 Lignin recovery The lignin-rich black liquor (obtained as a byproduct of the CNC, as described in 2.4.1) was concentrated in a rotary evaporator (R-210/ 215, Buchi, Flawil, Switzerland) at 65 °C for lignin recovery The concentrated black liquor was diluted 10 times in hot (∼80 °C) distilled water, and kept for 24 h The mixture was then vacuum filtered (with μm filter paper), washed with distilled water until neutrality, and dried at 50 °C in air-circulating oven until constant weight The resulting material was then ground to a dark powder (lignin) with an analytical mill (Ika A11, Staufen, Germany) The lignin content (purity) of the powder was determined according to TAPPI T222 om-22 (2000), and the lignin yield was calculated on an MSP basis 2.7 Film characterization 2.7.1 Moisture content The moisture contents of the different films were measured by using a Marte ID50 infrared scale (Marte, São Paulo, SP, Brazil) 2.7.2 Tensile tests Tensile properties of 125 mm × 12.5 mm film strips (with at least five replicates) were measured according to D882-12 (ASTM, 2012), using a Universal Testing Machine (Emic DL-3000, São José dos Pinhais, Brazil) with a load cell of 100 N, initial grip separation of 100 mm, and crosshead speed of 12.5 mm/min Before the test, the sample thicknesses were measured with an Akrom KR1250 coating thickness tester (Akrom, São Leopoldo, RS, Brazil) to the nearest μm 2.5 Characterization of SNC and CNC SEM micrographs were obtained from starch granules and wt% SNC and CNC suspensions previously sonicated (Unique Cell Disruptor, 50 W, min), placed onto Formvar coated grids, and stained with 1% phosphotungstic acid The nanostructures were examined using a Tescan Vega microscope (Tescan, Brno, Czech Republic) equipped with a STEM detector at an acceleration voltage of 30 kV The SNC and CNC average dimensions were taken as an average of 100 measurements taken by using the software Gimp 2.8 The zeta potentials of SNC and CNC were measured in triplicate using a Zetasizer Nano ZS (Malvern Instruments Ltd., Worcestershire, UK), for suspensions diluted to 0.05 wt% with deionized water X-ray diffraction (XRD) analyses were carried out on MKS, SNC, and CNC by using a DMAXB diffractometer (Rigaku, Japan), using CuKα radiation, operated at 40 kV and 40 mA, at the 2ϴ range of 3–40°, with a step size of 0.05°, and scanning speed of 1° min−1 The diffractograms were deconvoluted by using Origin 9.0 software, and the crystallinity indexes (CrI) were calculated as the ratio between the sum of areas under the curves corresponding to the crystalline peaks and the sum of the areas of all deconvoluted peaks Fourier-transform infrared spectroscopy (FTIR) spectra of MKS, SNC, and CNC were recorded in a Schimatzu Prestige-21 spectrophotometer, using KBr pellets, in the wavenumber range from 4000 to 400 cm−1, with a resolution of cm−1 over 64 scans 2.7.3 Water vapor permeability (WVP) The water vapor permeability (WVP) determination, with at least five replicates, was based on the method E96/E96M-16 (ASTM, 2016) at 24 °C, using mL distilled water inside permeation cells (inner diameter, 24 mm; height, 10 mm) and silica gel as the desiccant material, in an Arsec DCV040 desiccator with air circulation (Arsec, Vargem Grande Paulista, SP, Brazil) Eight measurements were made within 24 h Before the test, the sample thicknesses were measured with an Akrom KR1250 coating thickness tester (Akrom, São Leopoldo, RS, Brazil) to the nearest μm 2.7.4 Fourier transform infrared (FTIR) Fourier-transform infrared spectroscopy (FTIR) spectra were recorded in a Schimatzu Prestige-21 spectrophotometer, using KBr pellets, in the wavenumber range from 4000 to 400 cm−1, with a resolution of cm−1 over 64 scans 2.7.5 Scanning electron microscopy (SEM) The scanning electronic microscopy (SEM) micrographs of the films were taken using a QUANTA FEG-FEI 450 microscope The samples were mounted on the aluminum stub using carbon-coated double sided adhesive tape, coated with a 20 nm-thick gold layer by using a Quorum QT150ES metallizer, and examined using an accelerating voltage of kV and a magnification of 2000 times 2.6 Formation of bionanocomposite films Twelve films were produced, according to a central composite design (CCD), with two variables: total nanofiller content (SNC + CNC, from to wt% on the starch matrix) and degree of replacement of SNC with CNC (from to 100 wt%), as presented in Table with the corresponding responses Each film was made from 12.5 g mango kernel starch previously gelatinized in 125 mL distilled water (95 °C, for 30 min, under stirring) Glycerol (3.13 g) was then added to the gelatinized starch, which was kept at 60 °C for 15 under stirring, then stabilized to 40 °C (also under stirring) The SNC and/or CNC suspensions (in volumes according to the set concentrations for each treatment), previously mixed and sonicated with a Unique USC-1400 bath sonicator (40 kHz, 135 W) in five 1-min sonication cycles interspersed with 30 s intervals, were then added to the gelatinized starch dispersion, and the mixture was stirred for more 30 at 40 °C The final film forming dispersion was then homogenized with an Ultra Turrax T25 (Ika, Staufen, Germany) at 10,000 rpm for 30 Air bubbles were removed under vacuum (using an MA 057/1 vacuum pump, Marconi, Piracicaba, SP, Brazil), and the films were cast on Mylar® substrates on leveled glass plates for a thickness of 1.5 mm The films were allowed to dry at room conditions (23 °C, 54% RH, 20 h) After being detached from the Mylar surfaces, films were cut into test 2.7.6 Statistical analyses Results of tensile tests and WVP were analyzed using the software Minitab® (Minitab Inc., State College, PA, USA) Full quadratic models were fit to the experimental responses by using the DOE (Design of Experiments)/Response Surface analysis, and the regressions were evaluated in terms of their determination coefficients (R2 values) and the significance of their F values The optimized conditions were defined by using the Response Optimizer function of the Minitab® software, which aims to identify a combination of conditions that jointly optimizes a set of responses (using maximizing, minimizing, or targeting goals for each response), as required by the user, who establishes lower or upper as well as target values The optimization was based on the following combination: maximizing all tensile properties (i.e., finding a proper balance among them), while minimizing WVP The definition of lower/upper and target values for each response were based on the ranges obtained for each one of them The conditions set as optimum were experimentally verified (with three repetitions), and the mean relative deviation modulus (G) was calculated for each response 211 Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al granules presented oval or spherical shapes (Fig 1), and the SNCs were round like those reported for waxy maize SNCs (García et al., 2011) The major X-ray diffraction peaks of both MKS and SNC (Fig 1) were observed at 2θ around 12°, 15°, a doublet one at 17–18°, and another near 23°, a typical A-type starch pattern (Angellier et al., 2006; Mukurumbira, Mariano et al., 2017) Although Le Corre et al (2011) has described that A-type starches typically produce square-like nanocrystals, the A-type mango kernel starch of this study produced roundlike nanocrystals, corroborating our previous study (Oliveira et al., 2018) This may be due to the sonication applied on SNC isolation Indeed, the morphology of starch nanostructures was demonstrated to depend on the method used to produce them, ultrasound generating ellipsoidal nanoparticles instead of the platelet-like nanocrystals produced by acid hydrolysis (Haaj, Thielemans, Magnin, & Boufi, 2016) The combination of acid hydrolysis and ultrasound used in this study may explain the production of round-like nanocrystals Some peaks of SNC were sharpened when compared to those of MKS, which probably results from the selective acid hydrolysis of the starch (Mukurumbira, Mellem, & Amonsou, 2017) The CrI of the SNC was about 50% higher than that of MKS, confirming the effectiveness of the hydrolysis in partially removing the amorphous parts of starch The FTIR spectra of the MKS and SNC are almost identical (Fig 2A), except that the height ratio of the bands at 1045 and 1018 cm−1 increased in SNC, indicating that the ratio of crystalline to amorphous starch has increased (Warren, Gidley, & Flanagan, 2016), as expected Also, the band at 2930 cm−1 (CeH region), which varies with the amylose/amylopectin ratio (Kizil, Irudayaraj, & Seetharaman, 2002), was lower in SNC, which is ascribed to decreased amylose content, which was also expected, since amylopectin is the crystalline fraction Some bands are present in MKS and both kinds of nanocrystals, such as those ascribed to antisymmetric CeOeC stretching of glycosidic links at around 1160 cm−1 (Robert, Marquis, Barron, Guillon, & Saulnier, 2005), CeO stretch at 1034 cm−1 (Mohebby, 2008; Peng et al., 2012), CeH and OeH bending at 1420–1430 cm−1 and 1317–1337 cm−1, respectively (Peng et al., 2012) The CNC yield (Table 1) was similar to the one reported from sugarcane bagasse pith (Oliveira, Bras, Pimenta, Curvelo, & Belgacem, 2016) The CNCs were needle-like (Fig 1), with an average aspect ratio (L/d) slightly higher than those reported for CNCs from cotton fibers (Sapkota, Natterodt, Shirole, Foster, & Weder, 2017; Teixeira et al., 2010), and good dispersability, as represented by their high zeta potential (Table 1) The XRD pattern of CNC (Fig 1) exhibited two main crystalline domains at 2Ɵ about 12° (plane 110) and 22° (plane 020) (French, 2014; Santmartí & Lee, 2018) The yield of lignin powder (on MSP) was 15%, while its purity was 89%, indicating that about 88% of the lignin in MSP (Table 1) was recovered Table Determinations on components from mango kernels and mango seed shell Determination Average ( ± standard deviation) MKS 38.8 ± 3.5 0.12 ± 0.02 0.02 ± 0.001 25.1 ± 1.95 10.2 ± 2.1 39.5 24.4 ± 2.7 67.1 ± 21.7 −24 ± 62.6 15.8 ± 2.1 270 ± 71 18 ± 15 −54 ± 80.3 Yield (wt%, on dry kernels) Ashes (wt%) Proteins (wt%) Amylose (wt%) Granule size (μm) Crystallinity index (%) Yield (wt% on starch) Particle size (nm) Zeta potential (mV) Crystallinity index (%) Yield (wt% on MSP) Length (nm) Diameter (nm) Aspect ratio (L/d) Zeta potential (mV) Crystallinity index (%) SNC CNC MKS: mango kernel starch; SNC: starch nanocrystals (from MKS); CNC: cellulose nanocrystals (from MSP) Results and discussion 3.1 Characterization of film components The amylose content of MKS (Table 1) may be considered as medium, when compared to those classified by Colussi et al (2017) Medium- to high-amylose starches are more suitable to form films, since the mainly linear chains of amylose allow it to form more hydrogen bonds when compared to amylopectin (Colussi et al., 2017) Also, amylose tends to re-associate easier than amylopectin by forming double-helix on retrogradation due to its linear structure (Zhu et al., 2017) Starches with higher amylose contents result in films with better tensile strength and barrier properties (Fu et al., 2018; Wang et al., 2017), although more rigid and brittle (Biduski et al., 2018; Wang et al., 2017) On the other hand, high-amylose starches should be avoided to obtain starch nanocrystals, since nanocrystals are mostly derived from amylopectin (Le Corre, Bras, & Dufresne, 2011) Thus, a medium amylose may be regarded as suitable to obtain both starch and starch nanocrystals for films The SNC yield from MKS was similar to those reported from amadumbe starch (Mukurumbira, Mariano, Dufresne, Mellem, & Amonsou, 2017) and waxy maize (Sanchez de la Concha et al., 2018) The MKS Table Experimental conditions and responses for film properties Film Control 10 11 Replacement of SNC with CNC (wt%) Total nanofiller content (wt%) C U C U −1.41 −1 −1 −1.41 1.41 0 0 0 14.5 85.5 14.5 85.5 100 50 50 50 50 50 −1.41 −1 −1 1 0 −1.41 1.41 0 0 3.16 3.16 8.84 8.84 6 10 6 σ ε E WVP 3.2 Characterization of films 13.05 14.37 16.25 16.72 19.77 15.50 18.04 13.47 25.10 25.41 23.50 25.66 17.72 10.71 8.78 10.66 5.43 11.92 7.35 15.31 2.53 3.50 3.44 3.28 1158 1181 1420 1311 1751 1247 1577 1214 1967 1902 1848 1898 The tensile properties and water vapor permeability of the films are shown in Table When compared to an MKS film containing just wt % SNC (from our previous study, Oliveira et al., 2018), the film #8 from this study (containing wt% SNC and wt% CNC) presented higher tensile strength and modulus, corroborating the higher effectiveness of combined SNC and CNC on tensile properties when compared to SNC alone Fig presents the contour plots of the models, while Table presents the regression coefficients and statistical evaluation, indicating that all the regressions were significant The replacement of SNC with CNC resulted in decreasing WVP (reflected by a significant and negative regression coefficient, Table 3) and increasing modulus, indicating that CNC were more effective than SNC to increase stiffness and water vapor barrier of the films The higher effect of CNC on modulus was also observed by González et al (2015), and is ascribed to the higher aspect ratio of CNCs, which 1.579 1.483 1.298 1.341 1.248 1.386 1.303 1.334 1.133 1.311 1.263 1.269 C: coded values, according to the central composite design, varying from −1.41 to 1.41 (corresponding to to 100 wt% replacement of SNC with CNC, and from to 10 wt% total nanofiller (SNC + CNC) content on films U: uncoded (experimental) values; σ: tensile strength (MPa); ε: elongation at break (%); E: elastic modulus (MPa); WVP: water vapor permeability (g mm h−1 kPa−1 m-2) 212 Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al Fig SEM micrographs and XRD diffractograms of mango kernel starch (MKS), starch nanocrystals (SNC), and cellulose nanocrystals (CNC) Magnification of micrographs: MKS: 2000×; SNC: 28,000×; CNC: 41,000× elongation has been impaired, as observed in previous studies with SNC alone (Kristo & Biliaderis, 2007; Oliveira et al., 2018), CNC alone (El Achaby et al., 2017; Pereira et al., 2017; Sukyai et al., 2018), and combined CNC and SNC (González et al., 2015) The FTIR spectrum of the film containing wt% of each kind of nanocrystals (film #8) exhibited some changes when compared to that of the control film (Fig 4) The 1043 cm−1 to 1022 cm−1 height ratio increased with the addition of nanofillers, which is ascribed to the increasing crystalline to amorphous starch ratio (Htoon et al., 2009; Warren et al., 2016) The band at 1155 cm−1 (antisymmetric CeOeC stretching of glycosidic links, according to Robert et al., 2005) had its intensity increased with the addition of nanofillers, as previously reported from CNC addition to alginate films (Huq et al., 2012); the presence of charged sulfate ester bonds on CNC resulting from sulfuric acid hydrolysis may have contributed to this band (Zhao, Zhang, Lindström, & Li, 2015) Other bands were increased with the addition of nanocrystals, such as those at 1462 cm−1 (CH2 bending), 1423 cm−1 (CH2 scissoring), 1202 cm−1 (OH bending, probably increased by the S]O vibration on CNC), 1111 cm−1 (ring breathing of CeOeC), and 931 cm−1 (out-of-plane bending of OeH) The unfilled MKS film (Fig 5, A1 and A2) presented a relatively smooth surface, with a good transparency On the other hand, the film with wt% CNC and wt% SNC (Fig 5, B1 and B2) was rougher and more opaque, which is ascribed to aggregations of nanocrystals, which was also noticed previously on starch films added with SNC (Jiang, Liu, Wang, Xiong, & Sun, 2016; Li et al., 2015; Oliveira et al., 2018) or CNC (Johar & Ahmad, 2012) None of the films exhibited cracks or discontinuities, and the signs of nanocrystal aggregation were not reflected in impaired tensile and barrier properties; on the contrary, those properties were improved by those nanocrystals (Table 2, Fig 3) Fig FTIR spectra of: MKS: mango kernel starch; SNC: starch nanocrystals; CNC: cellulose nanocrystals results in higher mechanical reinforcement (González et al., 2015; Wang, Tian, & Zhang, 2010) Although the effects of replacing SNC with CNC on strength and elongation were not significant, their values may be interpreted as a tendency of CNC to be more effective than SNC on enhancing film strength, but also less favorable to elongation The better effectiveness of CNC (when compared to SNC) on barrier properties would be surprising if the SNC had presented platelet-like morphology, such as those from waxy maize starch (González et al., 2015; Putaux, Molina-Boisseau, Momaur, & Dufresne, 2003), which would be expected to create a more tortuous diffusion path for the permeants (González et al., 2015; Le Corre, Bras, & Dufresne, 2010; Le Corre, Bras, & Dufresne, 2012) On the other hand, the round-like morphology presented by the SNC in this study (Fig 1) partly explains why this was not the case Moreover, the barrier properties not depend only on the diffusion, but also on polarity and structural features of the polymer (González et al., 2015), and the lower crystallinity of SNC when compared to CNC (Table 1) could be a major reason for their lower effectiveness on decreasing WVP Increasing the total nanofiller (SNC + CNC) content resulted in increased strength and modulus and decreased WVP, although the 3.3 Optimization The optimization was based on the following criteria: (a) tensile strength: goal, maximizing; lower, 14 MPa; target, 16 MPa; (b) elastic modulus: goal, maximizing; lower, 1200 MPa; target, 1400 MPa; (c) elongation at break: goal, maximizing; lower, 9%; target, 12%; (d) WVP: goal, minimizing; upper, 1.5 g mm kPa−1 h−1 m−2; target, 1.3 g mm kPa−1 h−1 m−2 The optimized conditions (with a composite 213 Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al Fig Contour plots for tensile properties and water vapor permeability of films desirability of 0.96) were thus established as 10 wt% total nanocrystals, with 15% replacement of SNC with CNC, that is to say, 1.5 wt% CNC and 8.5 wt% SNC The predicted responses at those conditions, as well as the corresponding experimental responses and mean relative deviation modulus (G), are presented at Table The strength of the film obtained at those conditions was well above the lowest acceptable value for food packaging (4 MPa, according to Tajeddin, Rahman, & Abdulah, 2010), and its elongation was about 10%, which has been considered by Khwaldia, Banon, Desobry, and Hardy (2004) as the minimum value required for most industrial applications of edible films The mean relative deviation modulus (G) values of all models (Table 4) were between and 10, indicating reasonably good predictions (Roy, Gennadios, Weller, & Testin, 2000) Table Regression coefficients (in coded values) for film properties Term WVP σ ε E Constant x y x2 y2 x.y R2 (%) F p 1.281 −0.050 −0.059 0.045 −0.010 0.023 85.19 5.75 0.04 24.85 1.066 2.789 −4.352 −3.098 0.292 90.82 9.90 0.01 3.408 −1.703 −2.684 3.019 2.661 −0.825 84.99 5.66 0.04 1882 143.3 190.6 −256.8 −167.6 50.37 92.93 13.14 < 0.01 WVP: water vapor permeability; σ: tensile strength; ε: elongation at break; E: elastic modulus x: replacement of SNC with CNC; y: total nanofiller (SNC + CNC) content on films Regression terms in bold are significant (p < 0.05) Conclusions Starch and starch nanocrystals (SNC) were obtained from mango kernels, and cellulose nanocrystals (CNC), from mango seed shells Mango kernel starch films were prepared with different combinations of SNC and CNC When compared to SNC, CNC were more effective to enhance overall tensile properties and water vapor barrier of the films, whereas SNC tended to exhibit a lower effect on decreasing elongation The optimized conditions (1.5 wt% CNC and 8.5 wt% SNC on a starch basis) resulted in a film with enhanced strength and modulus (30% and 17% higher respectively) and water vapor barrier (22% lower permeability) than the unfilled film, although the elongation has been reduced by 40% (but still with a value suitable for edible film applications) The nanocomposite mango kernel starch films may be used for a variety of food applications as renewable and biodegradable, even edible, films Acknowledgements Fig FTIR spectra of mango kernel starch films: unfilled (control) and with wt% CNC and wt% SNC (film #8) The authors gratefully acknowledge the financial support of the Brazilian Agricultural Research Corporation (Embrapa, 214 Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al Fig SEM micrographs and pictures of mango kernel starch films (A1) and (A2): unfilled (control) film; (B1) and (B2): film #8 (with wt% CNC and wt% SNC) waxy maize starch nanocrystals nanocomposites Biomacromolecules, 7, 531–539 Arora, A., Banerjee, J., Vijayaraghavan, R., MacFarlane, D., & Patti, A F (2018) Process design and techno-economic analysis of an integrated mango processing waste biorefinery Industrial Crops and Products, 116, 24–34 ASTM D882-12 (2012) Standard test method for tensile properties of thin plastic sheeting West Conshohocken: ASTM International ASTM E1755-01 (2015) Standard test method for ash in biomass West Conshohocken: ASTM International ASTM E96/E96M-16 (2016) Standard test methods for water vapor transmission of materials West Conshohocken: ASTM International Banerjee, J., Singh, R., Vijayaraghavan, R., MacFarlane, D., Patti, A F., & Arora, A (2018) A hydrocolloid based biorefinery approach to the valorisation of mango peel waste Food Hydrocolloids, 77, 142–151 Biduski, B., Silva, W M F., Colussi, R., El Halal, S L M., Lim, L.-T., Dias, A R G., et al (2018) Starch hydrogels: The influence of the amylose content and gelatinization method International Journal of Biological Macromolecules, 113, 443–449 Bradford, M M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding Analytical Biochemistry, 72, 248–254 Chen, Y., Liu, C., Chang, P R., Cao, X., & Anderson, D P (2009) Bionanocomposites based on pea starch and cellulose nanowhiskers hydrolyzed from pea hull fibre: Effect of hydrolysis time Carbohydrate Polymers, 76, 607–615 Colussi, R., Pinto, V Z., El Halal, S L M., Biduski, B., Prietto, L., Castilhos, D D., et al (2017) Acetylated rice starches films with different levels of amylose: Mechanical, water vapor barrier, thermal, and biodegradability properties Food Chemistry, 221, 1614–1620 El Achaby, M., El Miri, N., Aboulkas, A., Zahouily, M., Bilal, E., Barakat, A., et al (2017) Processing and properties of eco-friendly bio-nanocomposite films filled with cellulose nanocrystals from sugarcane bagasse International Journal of Biological Macromolecules, 96, 340–352 French, A (2014) Idealized powder diffraction patterns for cellulose polymorphs Cellulose, 21, 885–896 Fu, L., Zhu, J., Zhang, S., Li, X., Zhang, B., Pu, H., et al (2018) Hierarchical structure and thermal behavior of hydrophobic starch-based films with different amylose contents Carbohydrate Polymers, 181, 528–535 García, N L., Ribba, L., Dufresne, A., Aranguren, M., & Goyanes, S (2011) Effect of glycerol on the morphology of nanocomposites made from thermoplastic starch and starch nanocrystals Carbohydrate Polymers, 84, 203–210 González, K., Retegi, A., González, A., Eceiza, A., & Gabilondo, N (2015) Starch and cellulose nanocrystals together into thermoplastic starch bionanocomposites Carbohydrate Polymers, 117, 83–90 GTZ (Deutsche Gesellschaft für Technische Zusammenarbeit) (1992) Manual de Table Predicted and experimental responses obtained from the optimized conditions (1.5 wt% CNC and 8.5 wt% SNC) Responses Predicted Experimentala G σ (MPa) ε (%) E (MPa) WVP (g mm kPa−1 h−1 m−2) 16.87 10.73 1353 1.238 16.68 ± 1.400 10.35 ± 0.593 1299 ± 57.55 1.265 ± 0.091 8.52 7.18 6.33 7.16 G: relative deviation modulus a Averages from at least four replicates of repetitions 02.14.04.002.00.00) They also thank the Central Analítica-UFC/CTINFRA/MCTI-SIS-NANO/Pró-Equipamentos CAPES for the use of their facilities for the SEM analyses, the X-Ray Laboratory at Federal University of Ceara for the XRD analyses (National Council for Scientific and Technological Development, CNPq, 402561/2007-4), as well as Celli R Muniz (Embrapa Agroindústria Tropical) for the SEM micrographs of starch nanocrystals We thank the Coordination for the Improvement of Higher Education Personnel (CAPES, 1376677 and 2017SLR-17925) and Fundaỗóo Cearense de Apoio ao Desenvolvimento Cientớco e Tecnológico (FUNCAP, BMD-0008-00640.01.11/15) for the scholarships granted to A.V Oliveira, A.L.S Pereira, and A.P.M Silva respectively The authors also acknowledge CNPq for the Research Productivity Fellowships granted to M.F Rosa and H.M.C Azeredo (305504/2016-9 and 302381/2016-3 respectively) References Angellier, H., Choisnard, L., Molina-Boisseau, S., Ozil, P., & Dufresne, A (2004) Optimization of the preparation of aqueous suspensions of waxy maize starch nanocrystals using a response surface methodology Biomacromolecules, 5, 1545–1551 Angellier, H., Molina-Boisseau, S., Dole, P., & Dufresne, A (2006) Thermoplastic starch- 215 Carbohydrate Polymers 211 (2019) 209–216 A.P.M Silva et al 4744–4759 Pereira, P H F., Waldron, K W., Wilson, D R., Cunha, A P., Brito, E S., Rodrigues, T H S., et al (2017) Wheat straw hemicelluloses added with cellulose nanocrystals and citric acid Effect on film physical properties Carbohydrate Polymers, 164, 317–324 Persin, Z., Stana-Kleinschek, K., Foster, T J., van Dam, J E G., Boeriu, C G., & Navard, P (2011) Challenges and opportunities in polysaccharides research and technology: The EPNOE views for the next decade in the areas of materials, food and health care Carbohydrate Polymers, 84, 22–32 Piyada, K., Waranyou, S., & Thawien, W (2013) Mechanical, thermal and structural properties of rice starch films reinforced with rice starch nanocrystals International Food Research Journal, 20, 439–449 Plant-O-Gram (2015) Mango tree Tommy Atkins variety grafted Available at:https:// plantogram.com/product/mango_tommy_atkins/ Putaux, J L., Molina-Boisseau, S., Momaur, T., & Dufresne, A (2003) Platelet nanocrystals resulting from the disruption of waxy maize starch granules by acid hydrolysis Biomacromolecules, 4, 1198–1202 Robert, P., Marquis, M., Barron, C., Guillon, F., & Saulnier, L (2005) FT-IR investigation of cell wall polysaccharides from cereal grains Arabinoxylan infrared assignment Journal of Agricultural and Food Chemistry, 53, 7014–7018 Roy, S., Gennadios, A., Weller, C L., & Testin, R F (2000) Water vapor transport parameters of a cast wheat gluten film Industrial Crops and Products, 11, 43–50 Sanchez de la Concha, B B., Agama-Acevedo, E., Nuñez-Santiago, M C., Bello-Perez, L A., Garcia, H S., & Alvarez-Ramirez, J (2018) Acid hydrolysis of waxy starches with different granule size for nanocrystal production Journal of Cereal Science, 79, 193–200 Santmartí, A., & Lee, K (2018) Crystallinity and thermal stability of nanocellulose In K Lee (Ed.) Nanocellulose and sustainability production, properties, applications, and case studies (pp 67–86) Boca Raton: CRC Press Sapkota, J., Natterodt, J C., Shirole, A., Foster, E J., & Weder, C (2017) Fabrication and properties of polyethylene/cellulose nanocrystal composites Macromolecular Materials and Engineering, 17, 595–606 Slavutsky, A M., & Bertuzzi, M A (2014) Water barrier properties of starch films reinforced with cellulose nanocrystals obtained from sugarcane bagasse Carbohydrate Polymers, 110, 53–61 Statista (2018) Mango production worldwide from 2000 to 2016 (in million metric tons) Available at:https://www.statista.com/statistics/577951/world-mango-production/ Sukyai, P., Anongjanya, P., Bunyahwuthakul, N., Kongsin, K., Harnkarnsujarit, N., Sukatta, U., et al (2018) Effect of cellulose nanocrystals from sugarcane bagasse on whey protein isolate-based films Food Research International, 107, 528–535 Tajeddin, B., Rahman, R A., & Abdulah, L C (2010) The effect of polyethylene glycol on the characteristics of kenaf cellulose/low-density polyethylene biocomposites International Journal of Biological Macromolecules, 47, 292–297 TAPPI T222 om-02 (2000) Acid-insoluble lignin in wood and pulp Teixeira, E M., Corrêa, A C., Manzoli, A., Leite, F L., Oliveira, C R., & Mattoso, L H C (2010) Cellulose nanofibers from white and naturally colored cotton fibers Cellulose, 17, 595–606 Terrazas-Hernandez, J A., Berrios, J J., Glenn, G M., Imam, S H., Wood, D., Bello-Pérez, L A., et al (2015) Properties of cast films made of chayote (Sechium edule Sw.) tuber starch reinforced with cellulose nanocrystals Journal of Polymers and the Environment, 23, 30–37 UNCTAD – United Nations Conference on Trade and Development (2016) Mango—An INFOCOMM commodity profile Available at:http://unctad.org/en/ PublicationsLibrary/INFOCOMM_cp07_Mango_en.pdf Wang, Y X., Tian, H., & Zhang, L (2010) Role of starch nanocrystals and cellulose whiskers in synergistic reinforcement of waterborne polyurethane Carbohydrate Polymers, 80, 665–671 Wang, K., Wang, W., Ye, R., Xiao, J., Liu, Y., Ding, J., et al (2017) Mechanical and barrier properties of maize starch-gelatin composite films: Eff ;ects of amylose content Journal of the Science of Food and Agriculture, 97, 3613–3622 Warren, F J., Gidley, M J., & Flanagan, B M (2016) Infrared spectroscopy as a tool to characterise starch ordered structure—A joint FTIR-ATR, NMR, XRD and DSC study Carbohydrate Polymers, 139, 35–42 Zhao, Y., Zhang, Y., Lindström, M E., & Li, J (2015) Tunicate cellulose nanocrystals: Preparation, neat films and nanocomposite films with glucomannans Carbohydrate Polymers, 117, 286–296 Zhu, J., Zhang, S., Zhang, B., Qiao, D., Pu, H., Liu, S., et al (2017) Structural features and thermal property of propionylated starches with different amylose/amylopectin ratio International Journal of Biological Macromolecules, 97, 123–130 exportacion: frutas tropicales y hortalizas Eschborn34p Haaj, S B., Thielemans, W., Magnin, A., & Boufi, S (2016) Starch nanocrystals and starch nanoparticles from waxy maize as nanoreinforcement: A comparative study Carbohydrate Polymers, 143, 310–317 Henrique, M A., Silvério, H A., Flauzino Neto, W P., & Pasquini, D (2013) Valorization of agro-industrial waste, mango seed, by the extraction and characterization of its cellulose nanocrystals Journal of Environmental Management, 121, 202–209 Htoon, A., Shrestha, A K., Flanagan, B M., Lopez-Rubio, A., Bird, A R., Gilbert, E P., et al (2009) Effects of processing high amylose maize starches under controlled conditions on structural organisation and amylase digestibility Carbohydrate Polymers, 75, 236–245 Huq, T., Salmieri, S., Khan, A., Khan, R A., Le Tien, C., Riedl, B., et al (2012) Nanocrystalline cellulose (NCC) reinforced alginate based biodegradable nanocomposite film Carbohydrate Polymers, 90, 1757–1763 ISO 6647-1 (2015) Rice – Determination of amylose content – Part Reference method Geneva: International Standardization Organization Jiang, S., Liu, C., Wang, X., Xiong, L., & Sun, Q (2016) Physicochemical properties of starch nanocomposite films enhanced by self-assembled potato starch nanoparticles LWT – Food Science and Technology, 69, 251–257 Johar, N., & Ahmad, I (2012) Morphological, thermal, and mechanical properties of starch biocomposite films reinforced by cellulose nanocrystals from rice husks BioResources, 7, 5469–5477 Kaur, M., Singh, N., Sandhu, K S., & Guraya, H (2004) Physicochemical, morphological, thermal and rheological properties of starches separated from kernels of some Indian mango cultivars (Mangifera indica L.) Food Chemistry, 85, 131–140 Khwaldia, K., Banon, S., Desobry, S., & Hardy, J (2004) Mechanical and barrier properties of sodium caseinate–anhydrous milk fat edible films International Journal of Food Science and Technology, 39, 403–411 Kizil, R., Irudayaraj, J., & Seetharaman, K (2002) Characterization of irradiated starches by using FT-Raman and FTIR spectroscopy Journal of Agricultural and Food Chemistry, 50, 3912–3918 Kristo, E., & Biliaderis, C G (2007) Physical properties of starch nanocrystal-reinforced pullulan films Carbohydrate Polymers, 68, 146–158 Le Corre, D., Bras, J., & Dufresne, A (2010) Starch nanoparticles: A review Biomacromolecules, 11, 1139–1153 Le Corre, D., Bras, J., & Dufresne, A (2011) Influence of botanic origin and amylose content on the morphology of starch nanocrystals Journal of Nanoparticle Research, 13, 7193–7208 Le Corre, D., Bras, J., & Dufresne, A (2012) Influence of native starch’s properties on starch nanocrystals thermal properties Carbohydrate Polymers, 87, 658–666 Li, X., Qiu, C., Ji, N., Sun, C., Xiong, L., & Sun, Q (2015) Mechanical, barrier and morphological properties of starch nanocrystals-reinforced pea starch films Carbohydrate Polymers, 121, 155–162 Matharu, A S., Houghton, J A., Lucas-Torres, C., & Moreno, A (2016) Acid-free microwave-assisted hydrothermal extraction of pectin and porous cellulose from mango peel waste—Towards a zero waste mango biorefinery Green Chemistry, 18, 5280–5287 Mohebby, B (2008) Application of ATR infrared spectroscopy in wood acetylation Journal of Agricultural Science and Technology, 10, 253–259 Mukurumbira, A., Mariano, M., Dufresne, A., Mellem, J J., & Amonsou, E O (2017) Microstructure, termal properties and crystallinity of amadumbe starch nanocrystals International Journal of Biological Macromolecules, 102, 241–247 Mukurumbira, A R., Mellem, J J., & Amonsou, E O (2017) Effects of amadumbe starch nanocrystals on the physicochemical properties of starch biocomposite films Carbohydrate Polymers, 165, 142–148 Nawab, A., Alam, F., Haq, M A., & Hasnain, A (2016) Biodegradable film from mango kernel starch: Effect of plasticizers on physical, barrier, and mechanical properties Starch – Stärke, 68, 919–928 Nawab, A., Alam, F., Haq, M A., Haider, M S., Lufti, Z., Kamaluddin, S., et al (2018) Innovative edible packaging from mango kernel starch for the shelf life extension of red chili powder International Journal of Biological Macromolecules, 114, 626–631 Oliveira, F B., Bras, J., Pimenta, M T B., Curvelo, A A S., & Belgacem, M N (2016) Production of cellulose nanocrystals from sugarcane bagasse fibers and pith Industrial Crops and Products, 93, 48–57 Oliveira, A V., Silva, A P M., Barros, M O., Souza Filho, M S M., Rosa, M F., & Azeredo, H M C (2018) Nanocomposite films from mango kernel or corn starch with starch nanocrystals Starch – Stärke, 70, 1800028 Peng, F., Bian, J., Peng, P., Guan, Y., Xu, F., & Sun, R C (2012) Fractional separation and structural features of hemicelluloses from sweet sorghum leaves Bioresources, 7, 216 ... Germany), and stored in sealed glass flasks The starch yield was expressed as extracted starch to initial raw kernels (wt%, on a dry basis), as an average value of the three batches The starch was analyzed... physical properties of mango kernel starch films, as well as to compare the effects of SNC and CNC (both from mango seeds) on those films The production of starch nanocrystals (SNC) was based on an acid... et al 2.3 Isolation of starch nanocrystals (SNC) Hasnain, 2016, 2018) Our group have studied the properties of mango kernel starch films as affected by SNC from mango kernel starch (Oliveira et al.,