Modification of Oxalis tuberosa starch with OSA, characterization and application in food-grade Pickering emulsions

Abstract

The emulsifying properties of Oxalis tuberosa starch (native and chemically modified) were evaluated in Pickering emulsions based on the emulsification index, emulsion stability over time and emulsion morphology. The best conditions of chemical modification were found by esterification of starch with octenyl succinic anhydride (OSA) at a concentration of 3% and a reaction time of 2 h, achieving a degree of substitution of 0.033 ± 0.001. The results obtained using Fourier-transform infrared spectroscopy, a Rapid Visco Analyzer, and differential scanning calorimetry, indicated that the starch underwent a change in its structure and that the insertion of the OSA groups was achieved. The amphipathic characteristics of OSA starch were evaluated by forming oil-in-water emulsions. Various concentrations of OSA-starch granules (1, 2.5 and 5 wt%) were used. A higher concentration of particles produced a smaller droplet size of emulsions (76.5 ± 0.9 μm) compared to those formed at a lower concentration of 1% (92.5 ± 1.0 μm). Therefore, the starch modified with OSA displayed the necessary characteristics to be adsorbed at the oil–water interface, achieving Pickering emulsion stabilization.

Introduction

An emulsion is a dispersed system in which the phases include immiscible liquids and an emulsifying agent such as surfactant or solid particles. Solid-stabilized emulsions are known as Pickering emulsions. It has been proposed that low molecular weight emulsifiers and surfactants have a negative impact on gut health and inflammation when used in foods, being Pickering stabilizers from natural sources a good option to replace them (Marefati et al. 2017). Recently, the use of starch granules to stabilize Pickering emulsions has attracted the attention of researchers (Li et al. 2013; Marefati et al. 2017; Fonseca-Florido et al. 2018). The different types of starches differ from each other due to their granule size, microscopic appearance, and chemical constitution because the amylose-amylopectin ratio is not always the same and will depend on the species from which it is extracted (Palabiyik et al. 2017). Native starches have limitations for their use in the stabilization of Pickering emulsions; therefore, they must be subjected to modifications. Modification can be achieved by different methods to obtain desirable surface properties. Among the compounds used for the modification of starches, the FDA has allowed octenyl succinic anhydride (OSA) to be used in foods at concentrations not higher than 3.0% (Bhosale and Singhal 2007). Octenyl succinic anhydride (OSA) starch is derived from the hydrophobic esterification of the native starch molecules, which introduces OSA compound groups in replacement of the OH groups. Therefore, the esterified starch contains hydrophilic and hydrophobic groups within its structure, which allow its adsorption at the oil–water interface. So, OSA starch has great advantages as an emulsifier and its nontoxic characteristics allow it to be used in the food industry (Li et al. 2013).

Oxalis tuberosa (Oca) is considered to have strong potential in the production of starch. Oca is a tuber considered a good source of carbohydrates (13.8-85.2 g/100 g dry matter), and its protein, fat and fiber content vary from 3.0 to 8.4 g, 0.5 to 1.0 g and 1.4 to 5.1 g/100 g of dry matter, respectively, with a starch content of 56.8 g/100g of dry matter (Chirinos et al. 2009).

The scientific research carried out on Oca has been related to its nutritional and functional value, its qualitative and quantitative characterization of the main phenolic compounds and their antioxidant capacity (Chirinos et al. 2009), the evaluation of flour obtained from it for the preparation of gluten-free breads (Güemes et al. 2018), the isolation and partial characterization of Oca starch (Valcárcel-Yamani et al. 2013) and its esterification and physicochemical characterization (Velásquez-Barreto et al. 2019). Therefore, there is a need to evaluate the ability to form Pickering emulsions with modified starch from Oca, since the characteristics of this unconventional source starch have not been thoroughly elucidated. Therefore, the aim of this work was to determine the ability of the starch obtained from unconventional sources, such as the tuber of Oca, to be chemically modified with OSA and to adopt the necessary surface properties to be used as a new source of solid particles for the stabilization of food-grade Pickering emulsions.

Materials and methods

Chemicals

Oxalis tuberosa tuber was purchased from a local market in Tulancingo, Hidalgo State, Mexico. 2-Octen-1-yl succinic anhydride (OSA, 97% purity) was obtained from Sigma-Aldrich Co. (St. Louis, MO, USA). Analytical-grade hydrochloric acid (HCl), sodium hydroxide (NaOH), acetone and isopropyl alcohol were purchased from Química Laitz, S.A. de C.V. (Mexico City). Canola oil (Capullo^®, Unilever de Mexico, S.A. de C.V., Tultitlan, Mexico) was purchased at a local supermarket. All the water used in the experiments was double distilled and deionized (DDW).

Isolation of soluble starch from Oxalis tuberosa

The isolation of the soluble starch of Oxalis tuberosa was carried out following the method described by Miao et al. (2014). The tubers were washed to remove the impurities and cut into pieces of 2 to 3 cm to be crush subsequently after the addition of DDW. The obtained mixture was filtered through 100-mesh sieves and centrifuged at 10,000 g for 10 min. The supernatant was collected and rinsed twice in DDW. The resulting sediment was dried in an oven at 40 °C for 24 h. The dried solid was milled to obtain a powder, which was used for further experiments.

Modification of starch granules by OSA

The Oxalis tuberosa starch was chemically modified with OSA using the method of Zhang et al. (2017) with some modifications. The starch (20 g, dry weight) was suspended in DDW (30% wt/wt) with continuous stirring. 2-Octen-1-yl succinic anhydride (OSA) was diluted 3 times in isopropyl alcohol. The amount of OSA varied from 0.6, 1.2, 2.4 and 3% wt in relation to the dry matter of the starch over a time period of 2, 4, and 6 h. The pH was kept constant at 8.5 with a 0.1 N NaOH solution during the reaction. When the pH is constant at 8.5, no further addition of NaOH is necessary (Marefati et al. 2017). After the reaction, it was neutralized to pH 6.5 with 0.5 N HCl. The modified starch (MS) was coded as MS_X,Y, where the subscript X denotes the concentration of OSA, and the subscript Y denotes the reaction time. The native starch was designated NS.

Isolation of starch particles

A total of 190 g of DDW was added to the suspension of the reaction obtained in the previous section and centrifuged at 5000 g for 7 min. The pellet was suspended again in 350 g of distilled water and centrifuged. This second sediment was suspended in 300 mL of acetone, stirred for 5 min and centrifuged again. The third sediment was dried at room temperature for 24 h and then in a convective dryer at 45 °C for 4 h. Under these conditions, the acetone evaporated, and the starch was dried below its equilibrium water content. Finally, the starch was conditioned at room temperature for 2 days to reach its equilibrium moisture content (Marefati et al. 2017).

Determination of the degree of substitution

The degree of substitution (DS) is the average number of hydroxyl groups substituted per glucose unit. The DS of OSA starch was determined by titration (Song et al. 2006). To this end, the OSA starch sample (5 g, dry weight) was accurately weighed and dispersed by stirring for 30 min in a 25 mL 2.5 M HCl-isopropyl alcohol solution. Then, 100 mL of 90% (v/v) aqueous isopropyl alcohol solution was added, and the slurry was stirred for an additional 10 min. The suspension was filtered through a glass filter, and the residue was washed with 90% isopropyl alcohol solution. The starch was redispersed in 300 mL distilled water, and then the dispersion was cooked in a boiling water bath for 20 min. The starch solution was titrated with 0.1 M standard NaOH solution using phenolphthalein as an indicator. The DS was calculated by the following equation:

$$DS=\frac{0.162\times \left(A\times M\right)/W}{1-\left[0.210\times \frac{A\times W}{W}\right]}$$

where A is the titration volume of NaOH solution (mL), M is the molarity of NaOH solution, and W is the dry weight (g) of the OSA starch.

Fourier transform infrared spectroscopy (FTIR)

The determination of the chemical structure of the native and modified starch was carried out by means of FTIR (Perkin Elmer, Waltman, MA, USA) to determine the DS of the hydroxyl groups by the carbonyl groups of the OSA. For each sample, the region was scanned from 400 to 4000 cm⁻¹ at a resolution of 4 cm⁻¹ over 48 scans. The correction of the baseline was made automatically using the software that accompanies the equipment.

Pasting properties

The pasting properties were analyzed using a Rapid Visco Analyzer (RVA-4500, Warriewood, Australia). A suspension (28.0 g) containing 3 g of starch and with a correction of the moisture content at 14% was equilibrated at 50 °C for 1 min, then heated to 95 °C at a rate of 6 °C/min; this temperature was maintained for 5 min followed by cooling to 50 °C at the same speed (6 °C/min). The speed of rotation of the paddle was 160 rpm. The parameters evaluated were pasting temperature (PT), peak viscosity (PV), breakdown (BD), and trough viscosity (TV). The analysis was carried out in duplicate, and the results were expressed in °C for the pasting temperature and in cP for the other parameters (Zabot et al. 2018).

Solubility

The solubility of the native and modified starch was determined following the method proposed by Leach et al. (1959) with slight modifications. In this experiment, 0.4 g of starch was placed in a 50-mL centrifuge tube containing 20 mL of DDW. The tubes were vortexed for 15 min followed by heating at 95 °C for 30 min. The gelatinized starch was cooled to room temperature and centrifuged at 8000 rpm for 15 min. The supernatant was dried at 105 °C to constant weight and weighed. The solubilities of both the modified and the native starch were calculated according to the following equation:

$$Solubility\left(\%\right)=\frac{weightofdrysupernatant}{weightofdrystarch}\times 100.$$

Thermal properties

The thermal properties of the native and modified starch were analyzed using a differential scanning calorimeter (DSC). A sample of starch was weighed on an aluminum tray, and DDW was added in a 1:3 ratio (dry starch: water). The sample tray was hermetically sealed, and an empty tray was used as a reference. The samples were heated from 30 to 120 °C at a heating rate of 10 °C/min. The gelatinization enthalpy (ΔH), onset (T₀), peak (T_P) and conclusion temperature (T_C) were determined (Marefati et al. 2017).

Particle size measurements of starch granules

The Beckman Coulter Z1 (Beckman Coulter, U.K.) double-threshold equipment was used to measure the particle size of the starches. This measurement is based on the method of the electrical detection zone of the Coulter principle and is used for the counting of cells and particles within a range of 1 to 120 microns. A small volume of starch granules was suspended and added to the flow system that pumped the samples through the optical chamber for measurements (Timgren et al. 2013).

Formulation and characterization of Pickering emulsions

Formulation of emulsions with different starch contents

Pickering emulsions (O/W) were prepared containing a dispersed phase mass fraction of 0.3 dispersing oil in the continuous phase (DDW + starch granules) with the aid of an Ultra Turrax homogenizer (model T25, IKA Works, Inc., Wilmington) at 16,000 rpm for 5 min. The concentration of starch granules used for the preparation of the Pickering emulsions were 1, 2.5 and 5 wt% of the total emulsion, thereby obtaining 3 formulations coded as O/W_X, where the subscript X indicates the concentration of starch granules (Estrada-Fernández et al. 2018).

Emulsification index

The emulsification index (EI) of the samples is expressed as a relation between the height of the cream layer and the total height of the emulsion. Each sample was made in triplicate (Fonseca-Florido et al. 2018).

$$EI=\frac{Heightofthecreamlayer}{Totalheightoftheemulsion}$$

Morphology and droplet size of Pickering emulsions O/W

The mean droplet size (d_1,0) and the size distribution of the freshly prepared emulsions and emulsions stored for 40 days at 5 ± 1 °C were determined by analyzing micrographs taken with an Olympus BX45 phase contrast microscope (Olympus Optical Co., Ltd., Tokyo, Japan) coupled with a Camedia C-3030 digital camera (Olympus Optical Co., Ltd., Tokyo, Japan) and analyzed with Image-Pro Plus software version 7.0. Selected optical micrographs obtained using the 40× objective are shown.

Statistical analysis

Data were analyzed using one-way analysis of variance (ANOVA) and the Tukey test with a statistical significance of p ≤ 0.05 using the software NCSS 2000 (Kaysville, Utah, USA). All experiments were performed in triplicate.

Results and discussion

Effect of OSA concentration and reaction time on the esterification of Oxalis tuberosa starch

The DS of the modified starches with different concentrations of OSA was determined (Fig. 1a). A maximum concentration of 3% wt/wt based on dry starch was used, since it is the maximum concentration allowed by the FDA for foods (Miao et al. 2014). An increase in DS of the modified Oxalis tuberosa starch was observed as the OSA concentration increased. This increase in DS is due to a greater availability of OSA molecules near the starch molecules. The immobility of the hydroxyl groups in the starch limits the substitution of OSA such that the esterification depends on the availability of OSA around the starch molecules to achieve a high substitution rate (Hui et al. 2009). The reaction time also plays an important role in the esterification reaction (Fig. 1a). Esterification was more likely to occur in the short time (2 h) of reaction, since the DS began to decrease with increasing reaction time (4 and 6 h). It has been found that a short reaction time contributes to a DS higher than the prolonged reaction time. The esterification and hydrolysis process deplete OSA to a point at which the concentration of OSA begins to decrease, and eventually, all OSA will be involved in the reaction (Song et al. 2006). At this point, extending the reaction time will only cause the side reactions to become dominant. NaOH played an important role in these unwanted side reactions, but it is also important to improve the nucleophilicity of the hydroxyl group and the swelling of the starch particles (Hui et al. 2009). Therefore, the optimal reaction time was 2 h when the OSA concentration was 3% w/w, since it had the highest DS = 0.033 ± 0.001, indicating that more hydroxyl groups in the starch molecules have been substituted. This behavior can be observed in greater detail in Fig. 1b, where a 3% w/w of OSA was used to determine the DS every 30 min, which verified that the reaction time of 2 h and 3% OSA presented the highest DS for Oxalis tuberosa starch.

Fig. 1.

a Effect of concentration and reaction time on the degree of substitution; and b effect of reaction time on the degree of substitution using a constant concentration of 3% w/w OSA

Fourier transform infrared (FT-IR) spectroscopy

FT-IR spectroscopy was used to identify changes in the structure of the starch after modification. The structures of the native starch (NS) and modified starches (MS_3,2, MS_3,4 and MS_3,6) presented similar profiles, and their spectra are presented in Fig. 2a, b, respectively. The peak present at 3293 cm⁻¹ indicates the presence of hydroxyl groups (OH), and the peak at 2929 cm⁻¹ corresponds to the vibrations of the stretch CH–, of the glucose unit (Lin et al. 2017). Adsorption at 1628 cm⁻¹ is due to water bound to the starch granule. In the region of the fingerprints of the starch spectrum, 4 peaks appear between 800 and 1200 cm⁻¹, attributed to the stretching of the C–O bond (Simsek et al. 2015). Peaks close to 1073 and 1157 cm⁻¹ are characteristic of the anhydrous glucose ring of the C–O stretch; the peak near 990 cm⁻¹ was assigned to the skeletal mode of vibration of the glycosidic bond α (1–4), while the peak near 860 cm⁻¹ corresponds to the deformations C–H and CH₂ (Palabiyik et al. 2017). Comparing the native starch with the OSA modified Oxalis tuberosa starch spectrum, in the region of 1800 to 1500 cm⁻¹ (Fig. 2b), two additional peaks are presented in 1725 and 1575 cm⁻¹ that appeared after the esterification with OSA. The peak at 1725 cm⁻¹ is attributed to the stretching vibration of C=O, suggesting the formation of carbonyl ester groups (C=O) (Bharti et al. 2019). The peak at 1575 cm⁻¹ is attributed to the vibration of the asymmetric stretch of the carboxylate (RCOO–) (Hui et al. 2009). In addition, the intensities of these two absorption peaks increased in the following order: MS₃,₂ > MS_3,4 > MS_3,6. A higher intensity of these two new absorption peaks suggested that more carboxyl groups were bound to starch (Lin et al. 2017), suggesting that the higher intensity is related to the increase in the DS displayed (Simsek et al. 2015; Zhang et al. 2017). For this reason, starch modified with OSA (MS_3,2) was selected for subsequent analyses.

Fig. 2.

FT-IR spectra of: a Oxalis tuberosa native starch (NS) from 4000 to 500 cm⁻¹; and b bands between 1800 and 1500 cm⁻¹ of NS, modified starch with 3% OSA and 2 h of reaction time (MS_3,2), modified starch with 3% OSA and 4 h of reaction time (MS_3,4), and modified starch with 3% OSA and 6 h of reaction time (MS_3,6)

Pasting properties

The pasting profile is presented in Fig. 3, while the pasting properties of NS and MS_3,2 are shown in Table 1. The pasting profile represents the fingerprint of a particular starch (Bharti et al. 2019). In Fig. 3, NS showed a rapid increase in peak viscosity (PV) within a narrow temperature range with a clear maximum viscosity because the starch of a tuber is considered to be highly swelling, while MS_3,2 showed a significant decrease in peak viscosity over a longer time. The insertion of some compounds into the native starch granules has an inverse relation with the PV and their presence restricts the starch granules to reach their maximum swelling capacity prior to the leaching of the starch components thus reducing the viscosity, and made them more resistant to rupture under shear forces and high temperature (Galkowska and Juszczak 2019).The peak viscosity (PV) indicates the maximum size that the starch granules can swell and is therefore a measure of the water retention capacity of the starch. This parameter indicates the resistance of the swollen granules to the shear and the swelling performance of the granules (Singh et al. 2015). The decrease of the peak viscosity (PV) was the result of several effects on the starch granules during the modification process. According to Xiao et al. (2011), there is a strengthening between starch chains through cross-linking groups. Previous research has shown that starches modified with OSA are affected by bulky octenyl succinate groups and that they contribute to a lower water retention capacity and granule swelling capacity due to the formation of amylose-OSA complexes (Ovando-Martinez et al. 2017) and result in a lower PV.

Fig. 3.

RVA curves of native starch (NS) and modified of Oxalis tuberosa with 3% OSA and 2 h of reaction time (MS_3,2)

Table 1.

Pasting properties and differential scanning calorimetry of native and OSA-modified starch of Oxalis tuberosa

Sample	Pasting properties				Gelatinization parameters
	PV	TV	BD	PT	T0	TP	TC	ΔH
	(cP)	(cP)	(cP)	(°C)	(°C)	(°C)	(°C)	(J g−1)
NS	8478 ± 3.2ª	2650 ± 3.4ª	5828 ± 1.5ª	66.49 ± 0.3a	56.935 ± 0.01a	60.53 ± 0.01a	70.56 ± 0.72a	10.44 ± 0.08a
MS3,2	5675 ± 1.2b	2877 ± 4.2b	2798 ± 5.2b	65.91 ± 0.8b	53.475 ± 0.04b	58.39 ± 0.03b	66.15 ± 0.27b	7.69 ± 0.14b

Results are expressed as mean values ± standard deviations. Superscripts with different letters in the same column indicate significant difference (p ≤ 0.05)

PV Peak viscosity, TV Trough viscosity, BD Breakdown, P_T Pasting temperature, correspond to pasting properties. Meanwhile T₀ Onset temperature, T_P Peak temperature, T_C Conclusion temperature, ΔH Gelatinization enthalpy correspond to gelatinization parameters

With respect to the breakdown stage (BD), a lower value was presented in MS_3,2 than in NS. The higher the BD is, the lower the tendency of the starch to resist the shear force during heating is, and the lower value for MS_3,2 is associated with stability of the starch granules versus the strength of cutting and heat treatment (Bharti et al. 2019), probably facilitating the interactions and the formation of a network. These interactions results in a lower decrease in trough viscosity (TV) after the temperature is maintained at 95 °C (Wang et al. 2019).

The pasting temperature (PT) is the minimum temperature at which the starch granules begin to swell due to the absorption of water; therefore, a detectable increase in viscosity is recorded (Galkowska and Juszczak 2019). The MS_3,2 of Oxalis tuberosa presented a lower PT compared to the native starch (Table 1). The insertion of the octenyl succinate groups causes a disorder of the components of the starch, which allows the granules to begin to swell at lower temperatures than those of their native counterpart (Bello-Flores et al. 2014). Modified starches can swell and absorb water earlier than native starch, promoting the gelatinization of starch and increasing viscosity at lower temperatures (González et al. 2018).

Thermal properties

The results regarding the thermal properties of NS and MS_3,2 are summarized in Table 1. All the samples presented thermograms with a single endothermic peak (data not shown), characteristic of the gelatinization processes with a high water content (González et al. 2018). Many DSC studies have shown that modification of starch alters thermal transition temperatures and enthalpy associated with gelatinization (Tao et al. 2019). Starch MS_3,2 exhibited lower gelatinization temperatures than its native counterpart (NS) due to the breakdown of the crystal structure after modification (Galkowska and Juszczak 2019). The decrease in transition temperatures during modification with OSA may be due to the cleavage of the glycosidic bonds induced by heating (Ovando-Martinez et al. 2017) and by the introduction of a bulky group such, as the OSA, changing the degree of hydrogen bonding and weakening the bonds, which causes the starch to begin to swell at a lower temperature, and gelatinize more easily.

The ΔH was also reduced for MS_3,2, which means that less energy was required for the gelatinization. These results show that the crystal structure and the double helix structure of the starch were damaged, since ΔH is a general measure of crystallinity and is an indicator of the molecular order loss within the granule (Singh et al. 2009).

Solubility

The solubility of MS_3,2 (8.7 ± 0.031%) was significantly higher (p < 0.05) than that of NS starch (4.8 ± 0.023%). The increase in the solubility of MS_3,2 may be due to a weakening of the intermolecular hydrogen bond due to the introduction of the OSA group. Polymers with high branching are highly soluble in water due to the lower amylose content (Bharti et al. 2019). The introduction of branching points in the amylose (octenyl succinate groups) restricts the recrystallization of the starch and therefore contributes to the increase in solubility. In addition, the increase of new branch points in amylopectin inhibits the formation of the highly ordered structure and inhibits the realignment of the glucan chain (Timgren et al. 2013). The increase of the amorphous phase in MS_3,2 (greater disorder) facilitates the water intake and its interaction with the polymer chains, favoring the solubility of the starch granules in the water (González et al. 2018).

Starch particle size

It is well-known that the size of the starch granules depends on the source from which they are obtained. Other authors (Agama-Acevedo and Bello-Perez 2017) have reported sizes from 15 to 30 μm for Oxalis tuberosa starches, which are close to that found in this work [20 ± 0.53 μm for native starch (NS) and 25 ± 0.25 μm for modified starch (MS_3,2)], observing a slight increase in the size of the granules of MS_3,2 in comparison with NS. According to Simsek et al. (2015), this increase may be due to aggregation, since the more hydrophobic OSA starch granules are dispersed in an aqueous solution, making it more energy-efficient for them to aggregate. This slight increase may also be due to the swelling of the starch granule during chemical modification (Timgren et al. 2013).

Formulation and characterization of Pickering emulsions

Emulsification index

The modification of the Oxalis tuberosa starch with OSA allowed the starch granules to remain in the continuous phase, whereas the native starch was not able to do so. Figure 4 shows the behavior of the emulsification index (EI) at different evaluation times. The values presented a range of 0.8 for O/W_1% to 0.9 for O/W_5% on the first day of storage. A decrease in EI can be observed in the O/W_2.5% and O/W_1% emulsions, reaching a minimum value during the 40 days of storage, 0.6 and 0.5, respectively. In contrast, the EI decreased during the first 6 days in the O/W_5% emulsion but subsequently remained constant at 0.8 during the 40 days of storage. This finding indicates that the concentration of starch used for the preparation of the emulsions affected the EI. A higher concentration of starch granules allows a greater volume of emulsified phase after 40 days of storage (Timgren et al. 2013). Also, in Fig. 4, it was observed that in some of the treatments (O/W_1%), the granules of starch sedimented, with three regions in the tubes with samples being observed: (I) a first region containing the emulsion; (II) a second phase containing the continuous phase; (III) and a third region containing sedimented starch. As the evaluation time elapsed, the volume of the region corresponding to the emulsion showed that different volumes were higher on day 1 and lower on day 40, which indicates a decrease in stability. In general, the values of EI decreased, similar behavior has been reported in literature for Pickering emulsions stabilized with different starches (Fonseca-Florido et al. 2018; Leal-Castañeda et al. 2018). When the initial oil and water interface and the migration rate of starch granules were fixed, a higher concentration of particles in the aqueous phase caused more granules to adsorb at the interface, thereby contributing to the stabilization of a higher interfacial area (Fonseca-Florido, et al. 2018).

Fig. 4.

Emulsification index of Pickering emulsions stabilized with Oxalis tuberosa starch modified with 3% OSA and 2 h of reaction time (MS_3,2) during storage. The tubes with samples show three regions: (I) The emulsion; (II) The continuous phase; (III) The sedimented starch

Droplet size and morphology of Pickering emulsions

Oil-in-water emulsions were prepared using NS and OSA-modified starch as emulsion stabilizers, and the formation and stability of emulsion droplets were analyzed. NS was not able to stabilize emulsions. A phase separation occurred immediately after the homogenization process, and the starch particles sedimented due to the gravitational force because the native starch was not able to adsorb at the interface and to stabilize an emulsion (Simsek et al. 2015). This behavior was similar to that reported by Lu et al. (2018) when using native corn starch.

Chemical modification is a way to expand the functional properties of starch. The hydroxyl group (–OH) of the glucose unit present in the starch can be modified by OSA in an aqueous dispersion. The introduction of OSA groups in the chains of starches produces esters that have hydrophilic and hydrophobic groups. This amphipathic nature makes the modified starch a good emulsifier (Galkowska and Juszczak 2019). It has been reported that starch modified by OSA, in its granular form, can be used as a stabilizer of Pickering emulsions (Agama-Acevedo and Bello-Perez 2017). Therefore, the starch MS_3,2 was able to stabilize the oil drops. In Fig. 5a, the stability over time of the Pickering emulsions formed with concentrations of 1, 2.5 and 5% can be observed. The diameter of the O/W_X emulsions decreased as the concentration of starch particles increased. The newly formed emulsions exhibited the following values d_1,0, from highest to lowest: O/W_1% = 99.99 ± 1.4 μm; O/W_2.5% = 92.5 ± 1.0 μm and O/W_5% = 76.5 ± 0.9 μm. When the concentration of particles was low, they were not adsorbed on the surface of the oil in an amount sufficient to stabilize the emulsion. Larger droplets were formed that had an insufficient amount of particles to stabilize them compared to the stabilized drops with a larger amount of particles in which the formation of smaller droplets occurred (Lu et al. 2018). A higher concentration of particles can stabilize a larger total interfacial area which, in turn, means that a smaller average droplet size can be achieved (Li et al. 2013). According to Timgren et al. (2013), the stabilization of Pickering emulsions does not require complete coverage of very compact particles at the water–oil interface; however, they did not manage to stabilize emulsions with concentrations lower than 10% by weight, only achieving it with high concentrations (20 and 30% by weight), in comparison to those obtained in this study for Oca, for which only 5% was used. Simsek et al. (2015) reported stable emulsions with starch particles even when large droplet diameters are present (< 100 μm) and determined that Pickering emulsion drops are more stable when covered uniformly with starch granules, which may be the reason for the high stability obtained in emulsions with larger droplet diameters in this work.

Fig. 5.

a Coalescence rate of O/W Pickering emulsions stabilized with Oxalis tuberosa starch modified with 3% OSA and 2 h of reaction time (MS_3,2) at different concentrations (1, 2.5 and 5%). b Optical micrographs of stabilized emulsions with different concentrations of starch granules: (I) O/W_5 %, (II) O/W_2.5 % and (III) O/W_1 %. Scale bar = 20 µm

The droplet size of the emulsions fluctuated during the first 2 to 6 days of storage. For the O/W_1% and O/W_2% emulsions after day 12 of storage, they showed a remarkable increase in the diameter of the emulsions, reaching 192.3 ± 2.8 μm and 172.4 ± 2.5 μm, respectively. This behavior is related to what was observed in the emulsification index, where creaming of the emulsions was observed, which is a typical characteristic of Pickering emulsions due to the formation of large droplets (Lu et al. 2018). After 40 days of storage, the volume of the emulsified phase decreased, as shown in the EI results, and more starch was settled on the bottom, which indicates that the oil droplets may have suffered limited coalescence during storage due to low surface coverage of oil droplets (Lu et al. 2018; Timgren et al. 2013).

On the other hand, as the concentration of particles increased, the creaming in the O/W_5 % emulsion decreased, which could be due to the increase in the surface coverage of the oil droplets. This emulsion had a diameter of 104.3 ± 2.0 μm on day 12 with a slight increase at day 40 (111.2 ± 2.1 μm). The average size of the droplets increases after the preparation of the emulsion, quickly reaching a limit value that depends on the quantity of solid particles (Ge et al. 2017). The initial increase in the diameter of the emulsions may have been observed because coalescence was limited until the coverage of the starch particles was sufficient to stabilize the drops (Lu et al. 2018).

The oil droplets formed in the emulsion stabilized by the OSA starch during the first day are shown in Fig. 5b. The starch particles can be observed on the surfaces of the drops of the emulsions, which is the characteristic feature of the Pickering emulsions (Marefati et al. 2017), corroborating the results of the particle size of the starch. The starch granules formed a thick layer at the water–oil interface, which indicates the stabilization of the emulsions by the starch MS_3,2 because the OSA groups had an affinity for the dispersed phase (oil), which allows the starch granules to form a dense film around the oil droplets (Fonseca-Florido et al. 2018). The O/W_1% emulsion presented the larger droplets due to the limited granule-oil interaction, while the O/W_5% emulsion had a smaller droplet size. A higher concentration of particles in the aqueous phase will also facilitate the adsorption of more particles in the oil–water interface, contributing to the stabilization of a larger interfacial area and limiting the coalescence process (Lu et al. 2018).

Conclusion

The octenyl succinic anhydride (OSA) modification of Oxalis tuberosa starch with the highest degree of substitution was achieved at a starch concentration of 30 wt% and 3 wt% OSA for 2 h. The resultant OSA starch displayed suitable amphiphilic surface activities that allow forming stable O/W Pickering emulsions. Despite different sources of modified starches that have been used as Pickering emulsions stabilizers, the required amount of OSA starch from Oxalis tuberosa was significantly lower (5 wt% vs 30 wt% in literature). Therefore, this work contributes to the body of knowledge regarding obtaining new solid food-grade particles as O/W Pickering emulsions stabilizers.