1. Introduction
Products containing water-in-oil (W/O) emulsions are widely used in the pharmaceutical, cosmetic, petroleum and food industries. In food, almost all water-in-vegetable oil (W/VO) emulsions (e.g., margarine) are used in a semi-solid state (Ushikubo & Cunha, 2014). The emulsion stabilization mechanism in margarin is mainly due to fat crystal-based Pickering and network formation (Ghosh & Rousseau, 2011). Emulsifiers are also critical in the formation of W/O emulsion. Most emulsifiers used to form W/O emulsions are non-ionic lipophilic molecules (e.g., polyglycerol polyricinoleate (PGPR), Spans, monoglycerides). The selection of emulsifiers is based on their hydrophilic-lipophilic balance (HLB), which must be low (3 – 6) for W/O emulsion. Various research groups have worked with different combinations of oil, emulsifiers and aqueous phase ingredients to obtain stable liquid W/O emulsions. Opawale and Burgess (1998) investigated concentrated 50 wt% water-in-light mineral oil emulsions with Span 20, 80, 83 and 85 (1 – 5 %w/v) in the oil and sodium chloride (0 – 1 M) in the aqueous phase. The addition of the salt (S) within an optimum range enhanced the stability of the emulsion. Scherze et al. (2006) investigated stabilization of 30 wt% water-in-medium-chain triglyceride (MCT) oil emulsion with PGPR (4 wt%) and soy lecithin (2.5 wt%) in the presence of 0.6 % S in the aqueous phase. For lecithin, S favoured water droplet coalescence; however, for PGPR, the incorporation of S was necessary to obtain a stable W/O emulsion (Scherze et al., 2006). Lindenstruth and Muller (2004) investigated 30 wt% W/VO emulsions prepared with a 1:1 mixture of olive oil and MCT with 20% soya lecithin as an emulsifier. It was proposed that the cis double bond of oleic acid needed more space at the W-O interface, which prevented efficient interfacial packing of lecithin (Lindenstruth & Muller, 2004). In another research, Márquez et al. (2010) prepared 20 – 40 wt% W/VO emulsion using sunflower oil with 0.2 – 1 wt% PGPR and an aqueous phase containing various concentrations of calcium chloride and other salts (calcium lactate or carbonate; sodium, magnesium or potassium chloride). The incorporation of any salt allowed emulsion stabilization with higher water content. The stabilization effect of calcium salts was associated with a reduced attractive force between the water droplets and a higher PGPR adsorption density at the droplet surface. Prichapan et al. (2017) prepared 20 wt% W/VO emulsions with rice bran oil and stabilized them with rice bran stearin and PGPR. Emulsions without rice bran stearin (a source of fat crystals) were highly unstable to phase separation after one day; however, the stability improved by increasing the PGPR concentration from 2 to 4 wt% due to the reduction in water droplet size. However, the most stable emulsion was obtained with an oil phase containing 45 wt% rice bran stearin, which indicates that incorporating fat crystals as a stabilizing factor was more important than increasing the PGPR concentration.
PGPR is one of the most commonly used emulsifiers to develop food-grade W/O emulsions for its ability to stabilize water droplets by a polymeric steric barrier (Goubran & Garti, 1988; Mettu et al., 2018). Although it is classified as GRAS (generally recognized as safe), in 1979, the Scientific Committee for Food of the European Community determined an acceptable daily intake of 7.5 mg/kg body weight. In 1980, the Canadian Regulations reduced the maximum concentration of PGPR to 0.25% in chocolate (Wilson et al., 1998). However, in 2017, PGPR was limited just for chocolate products with a maximum of 0.5 wt% (Canada, 2017). For this reason, different groups investigated various combinations of PGPR with other ingredients to lower its concentration in W/VO emulsion (Ghosh & Rousseau, 2009; Rafanan & Rousseau, 2017). Nevertheless, there is an urgent need to completely replace PGPR from W/VO emulsions with a more natural small molecule emulsifier. Hence, one of the aims of the present study was to develop stable liquid food-grade W/VO emulsions by completely replacing PGPR with glycerol monooleate (GMO) without any stabilizing fat crystals. GMO is a widely studied low HLB (3 – 4) emulsifier that is non-toxic, biodegradable, and classified as GRAS. It is commonly used to prepare margarine-type semi-solid emulsions; however, the presence of saturated fat crystals is essential for emulsion stability (Ghosh & Rousseau, 2011). To our knowledge, no research so far has investigated stabilizing liquid W/VO emulsions without using PGPR and saturated fat crystals. The challenge here is that liquid W/VO emulsions prepared solely with GMO are impossible to stabilize due to its desorption from the water droplet interface towards VO. Ghosh et al. (2011) observed that GMO’s desorption from the W-VO interface strongly relies on its stronger hydrogen bonding interactions with the VO triacylglycerol (TAG) fatty acid carboxylic groups than the weaker hydrogen bonding interaction with water hydroxyl groups. The authors also showed that, in the presence of a hydrocarbon oil (such as mineral oil), GMO could very well remain at the water droplet surface due to the lack of interaction with the oil, thereby improving emulsion stability (Ghosh et al., 2011). Therefore, we hypothesized that the presence of hydrogen bond-forming molecules in the dispersed aqueous phase would form strong hydrogen bonds with GMO at the W-O interface, thereby improving the stability of liquid W/VO emulsions. Three different hydrogen bond-forming agents (two small molecules with multiple carboxylic groups, ascorbic acid (AA), and citric acid (CA) and one biopolymer, low methoxyl pectin (LMP)) with or without the presence of salts (sodium chloride (S) or calcium chloride (Ca)) were investigated to improve the stability of water-in-canola oil (W/CO) emulsions. As a control, light mineral oil (MO, made of paraffin hydrocarbons) was also used to form water-in-mineral oil (W/MO) emulsions, and its influence on emulsion stability and rheology was compared to W/CO emulsions.
2. Materials and methods
2.1 Materials
Canola oil (CO) was purchased from a local grocery store and stored at 4 °C. Distilled monoglycerides DMG 0298 (90 – 95% monoglycerides, containing mainly 75 – 91% oleic, 2 – 17% linoleic, 3 – 6% stearic, and 3 – 5% palmitic) was donated by Palsgaard (Juelsminde, Denmark). L-ascorbic acid (AA) and Nile red were purchased from Sigma Aldrich (Oakville, ON, Canada). Citric acid anhydrous (CA), light mineral oil (MO, code O121-4), and hydrochloric acid (HCl) (1 N) were purchased from Fisher Scientific (Toronto, ON, Canada). Sodium chloride (S) and calcium chloride (Ca) were obtained from BDH (VWR International, USA). Low methoxyl pectin (LMP) amidated, Genu® pectin type LM-101 (degree of esterification 36% and degree of amidation 14%) was provided by CP Kelco (Lille Skensved, Denmark). Purified water by Milli-QTM (Millipore Corporation, MA, USA) was used for the dispersed phase.
2.2 Preparation of solutions
AA and CA aqueous solutions were prepared with 0.125 and 0.5 wt% acids with 0 and 0.125, wt% S at room temperature with stirring (400 rpm) and stored at 4 °C. LMP solutions at pH 3.0 with 1N HCl were prepared with 1.5 and 2 wt% LMP with 0 and 0.045 wt% Ca at 90 ± 0.5 °C, stirred at 200 rpm for 20 min. The final LMP solutions were cooled down to room temperature before W/O emulsions preparation. The concentrations of the ingredients were selected based on preliminary experiments with a wider range; however, only the ones showing significant effect at minimum concentrations are reported. The zeta potential of LMP, measured at pH 3.0 using a zeta potential analyzer (Litesizer500, Anton Paar, Montreal, QC, Canada), was -18.68 ± 0.2 mV. Control solutions with just S at 0.125, 0.5 and 1 wt% and Ca at 0.045 wt% were also prepared using similar methods. All the solutions were prepared the same day as emulsions. The oil phase was prepared by dissolving 4 wt% GMO in canola oil (CO) or mineral oil (MO) under stirring (400 rpm) at 60°C for 25 min.
2.3 Preparation of W/O emulsions
The aqueous and oil solutions were warmed at 40 ± 2 °C before weighing. The oil phase (70 g, 80 wt%) was weighed in a glass beaker (100 mL), and the aqueous phase (17.5 g, 20 wt%) was added dropwise while stirring at 400 rpm. The same weight of each phase and similar size glass beakers were used for all emulsion batches. Coarse emulsification was done with a rotor-stator blender (Polytron, Brinkman, ON, Canada) at level 10 for 1 minute. The coarse emulsions were then further homogenized in a high-pressure homogenizer (Emulsiflex C3, Avestin Inc., Ottawa, ON, Canada) at 20,000 psi for three cycles. Preliminary experiments showed that three cycles were enough to reduce the droplet size, and after that, no further decrease was observed. The final emulsions were collected in 250 mL beakers and cooled down at room temperature while stirring with a magnetic stirrer (size 3.5 x 0.7 cm) at 290 rpm for 20 minutes.
2.4 Emulsion storage stability
Freshly-made emulsions (25 mL) were transferred to 40 mL clear glass vials (VWR, Edmonton, AB, Canada) and stored at room temperature for seven days. The images were taken with an iPhone 8 camera (12 MP). The images of the vials were used to measure the height of the emulsion layer on day 0 and day 7 with ImageJ (v1.5 2i – Fiji project) to calculate the fraction of emulsion height (Eq.1).
\(Fraction\ of\ emulsion\ height=\frac{\text{Emulsion\ height\ }on\ day\ 7}{Total\ emulsion\ height\ on\ day\ 0}\)Eq. 1
2.5 Microstructure analysis