Preparation and characterization of nanoemulsion based β-carotene hydrogels

Abstract

The aim of this study was to develop β-carotene hydrogels using nanoemulsions, with increased β-carotene aqueous solubility, bioavailability and improved physical and chemical stabilities. The nanoemulsion of β-carotene was prepared using a solvent-displacement technique and converted into hydrogels using sodium alginate as stabilizer and calcium chloride as cross-linker. The effects of formulation parameters, mainly, the effects sodium alginate and calcium chloride concentrations on the physicochemical properties of hydrogels were evaluated using a surface response methodology. The second order polynomial equations, subsequently, were suggested to predict the changes of studied physicochemical characteristics of hydrogels, with relatively high regression of coefficients values. Based on numerical multiple optimization, it was concluded that using 4.1 g/l sodium alginate and 5.7 g/l calcium chloride, resulted in a hydrogel with the most desired physicochemical characteristics. No significant differences between the measured and predicted data, reconfirmed the accuracy of the models.

Introduction

In recent years, bioactive compounds have drawn great attention because of their various health-promoting activities such as anti-oxidative, anti-microbial, anti-inflammatory and anti-cancer actions (Akinosho and Wicker 2015; Anarjan et al. 2010). β-Carotene is one of the most important lipophilic nutrients, composed of a polyene chain with 11 conjugated double bonds and two β-rings at the ends (Silva et al. 2011). Besides its several useful pharma-nutritional pro-vitamin A, antioxidant and anticancer effects, it can be used in the food formulations as a natural colorant and antioxidant compound (Mehmood et al. 2018). However, its uses are limited due to its low water-solubility and chemical instability, especially in water-based food and beverage constructions (Anarjan and Tan 2013; Tan and Nakajima 2005). Therefore, attempts to overcome these problems have led to incorporate these compounds into various nano-sized delivery systems or their encapsulation into natural biopolymer based hydrogel beads (Huang et al. 2017). In general, encapsulation at nano-level is recommended to increase their solubility, and consequently bioavailabilities, as well as their stabilities (Horn and Rieger 2001).

Solvent displacement, sometimes known as nanoprecipitation, is perhaps the simplest method for preparation of the nanoemulsions of water-insoluble bioactive compounds. This technique generally consists of mixing an aqueous phase containing a hydrophilic emulsifier or polymer with a dissolved lipophilic bioactive compound in a water-miscible organic solvent such as ethanol or acetone. Beside active compound, the organic solvent can optionally contain the dissolved polymer, oil, and lipophilic emulsifiers, as well. Nanoparticles are formed instantaneously by rapid diffusion of the organic solvent in the aqueous phase, which induces interfacial nano-deposition of the bioactive compound at the interface between the organic and aqueous phases (Anarjan and Jouyban 2017; Horn and Rieger 2001).

Alginate is among the most used polysaccharides based biopolymer for encapsulation of lipophilic bioactive compounds in nanodispersion systems, for increasing their chemical and physical stabilities or controlling their release to the site of actions and changing the nanoemulsions into hydrogels (Wang et al. 2018; Zhang et al. 2016). However, there are limited works on the preparation of hydrogels from nanoemulsions of lipophilic compounds such as carotenoids. Therefore, the aim of this study is to develop alginate-based hydrogels of β-carotene from their nanoemulsions via low cost, simple and controllable solvent-displacement technique. The effects of polysaccharide and calcium chloride concentration as a cross-linker were evaluated on the characteristics of achieved hydrogels. The selected independent variables were also optimized to give the hydrogels the most desired characteristics such as minimum particle size, PDI as well as maximum zeta-potential, gel fraction and swelling ratio.

Materials and methods

Materials

β-Carotene (purity ≥ 98%) was purchased from Akras (Austria, Biedermannsdorf). Tween 20 [Polyoxyethylene (20) sorbitan monolaurate], Calcium chloride (Calcium chloride) and sodium alginate were acquired from Merck (Germany, Darmstadt). Acetone was provided from Neutron (Iran, Tehran).

Preparation of β-carotene nanoemulsions

For the preparation of β-carotene nanoemulsion briefly 0.05 g of β-carotene was dissolved in 5 ml acetone under magnetic stirrer at 25 °C and coded as the organic phase. Tween 20 (5 ml) were also dissolved in 95 ml double distilled water and coded as the aqueous phase. The organic phase was then added to the aqueous phase slightly, under magnetic stirrer. The acetone was removed from the system using a vacuum rotary evaporator at 0.5 atm., 45 °C, 50 rpm and 15 min. The mixture would be converted to nanoemulsion after solvent removing. The nanoemulsion was volume up to 100 ml by addition of double distilled water.

Preparation of alginate based hydrogels

For encapsulation of β-carotene into alginate hydrogels, sodium alginate was added in various concentrations (according to Table 1) to nanoemulsion. Calcium chloride at designed concentrations (as Table 1) was also added to the system. All samples were volume up to equal sizes.

Table 1.

The independent and response parameters in preparation of β-carotene nanoemulsion hydrogels

Sample no.	Independent factors		Dependent or response variables (experimental data)
			Nanoparticle characteristics			Color characteristics
	Sodium alginate (g/l)	Calcium chloride (g/l)	Mean particle size (nm)	PDI	Zeta- potential (mV)	${\text{L}}^{\ast }$	${\text{a}}^{\ast }$	${\text{b}}^{\ast }$
1	133	23	210.3	0.4760	− 6.520	18.58	14.73	14.45
2	100	50	148.7	0.2849	− 4.150	19.67	14.59	14.49
3	133	77	452.6	0.3954	− 0.392	19.56	16.19	12.90
4	100	50	184.4	0.2807	− 4.100	19.02	14.95	14.19
5	100	50	148.3	0.2828	− 4.200	19.04	14.60	14.40
6	67	23	655.0	0.4739	− 4.250	23.25	18.35	17.16
7	100	10	620.0	0.6685	− 7.630	23.01	18.17	16.73
8	100	50	149.0	0.2821	− 4.130	19.67	14.59	14.49
9	100	90	493.3	0.4774	− 4.000	20.13	15.99	13.31
10	150	50	347.0	0.5097	− 3.700	18.56	16.46	11.92
11	50	50	598.9	0.4998	− 5.480	20.67	16.50	16.66
12	100	50	148.0	0.2842	− 4.120	19.03	14.63	14.10
13	67	77	395.7	0.3514	− 5.330	18.27	14.13	14.90

Analysis

Mean particle size, polydispersity index (PDI) and zeta-potential measurements

The mean particle size, PDI, and zeta-potential of samples were measured based on dynamic light scattering (DLS) using a Zetasizer Nano (Malvern Instruments Ltd., UK). PDI representing the size distribution of nanoparticles in gained hydrogels. Samples were diluted with deionized water prior to measurements to avoid multiple scattering effects. This analysis was performed 1 day after the preparation of the samples and the measurements were carried out at 25 °C. The average of three recorded results for each sample was reported as the mean particle size (Anarjan and Jouyban 2017).

Color measurements

Color parameters ($L^{\ast }\times a^{\ast }\times b^{\ast }$) for samples were estimated using a Hunter Lab colorimeter (Labscan^® XE, Hunter Associates Laboratory Inc., Reston, US) coupled with Easy Match QC software. Color values were recorded in triplicates for each sample.

Gel fraction and swelling percentage

The gel fraction and swelling percentage were calculated using the following equations:

$$\text{Gel}\text{fraction}(\%)=\left(\frac{w_1}{w_0}\right)\times 100\%$$ 1

where w₁ and w₀ refer to the weights of the dried sample after extraction of water, and original specimen, respectively.

$$\text{Percentage}\text{Swelling}(\%)=\frac{\left(w_t-w_d\right)}{w_d}\times 100\%$$ 2

where $w_t$ is the weight of swollen gel after passing time t, and $w_d$ is the weight of the initial dried gel.

For this purpose, the three pre-weighed dry hydrogel discs (10 mm diameter) from each sample were immersed in distilled water for 72 h at 25 °C. The samples were dried in a vacuum oven until obtaining constant weight values.

Scanning electron microscopy (SEM)

The optimum hydrogel was cast dried on an aluminum pan, mounted on a conductive carbon tape, coated with gold/platinum using a sputter coater (MC1000, Japan co., LTD), and its morphology was evaluated using SEM (Hitachi SU-70, Pleasanton, CA, USA).

Experimental design

Response surface methodology was performed to evaluate the significant effects of process variables on each of the responses. For encapsulation process, significant variables, such as sodium alginate ($x_1$) and calcium chloride ($x_2$) concentrations, were chosen as the independent variables. To estimate the full model, a total of 13 experiments were designed by central composite design (CCD). The independent variables levels for each experiment were shown in Table 1.

Minitab v. 17 software was used for experimental design, data analysis, model presentation and variable optimizations. Experimental data were fitted to second-order polynomial models, and regression coefficients were obtained. The analysis of variance (ANOVA) was performed to justify the significance of the developed regression model. The used general equation for this purpose was as:

$$y_i=a_0+a_1{x}_1+a_2{x}_2+a_{11}{x}_1^2+a_{22}{x}_2^2+a_{12}{x}_1{x}_2$$ 3

where $y_i$ is the response (particle size, PDI, zeta-potential, $L^{\ast }$, $a^{\ast }$ and $b^{\ast }$), 1 and 2 subscripts were used for sodium alginate and calcium chloride concentrations, respectively. $a_0$ is the constant coefficient, $a_1$ and $a_2$ are the coefficients of the linear effects of parameters. While the $a_{11}$ and $a_{22}$ are the coefficients of the quadratic impacts and $a_{12}$ is the interaction influence coefficient after data collection, a mathematical model is built for each response. The goodness of fit was evaluated by calculation of the determination coefficient ($R^2$) and $R^2$ adjusted. The significance of linear, quadratic and interaction terms in the confident level of 95 or 90% in each model was evaluated by using variance analysis (ANOVA). The terms with p > 0.05, (> 0.1) can be considered as insignificant in 95% (90%) confident level, which can be removed from the model without any considerable changes in its quality. Moreover, the terms with less p and higher F-ratio are more significant as compared to the other significant terms (Anarjan et al. 2010).

Results and discussions

General aspects

The physicochemical characteristics of the prepared nanoemulsion-based β-carotene hydrogels in different concentrations of sodium alginate and calcium chloride, such as mean particle size, PDI, zeta-potential of β-carotene nanoparticles, gel fraction, swelling percentage and color characteristics ($a^{\ast }$, $b^{\ast }$ and $L^{\ast }$) of hydrogels were measured for all prepared samples. Based on the results there was no significant difference between the β-carotene load, gel fraction and swelling percentage of gained hydrogels. Thus, the regression analysis of these three responses were ignored. However, the regression analysis was carried out for the rest of characteristics. All the size and color characteristics of samples could be fitted with second-order polynomial functions of independent variables. The high coefficients of variables for each model ($R^2$ > 90) indicated the suitability of models. The mean particle size, PDI, zeta-potential and color characteristics of all experiments (measured data) were shown in Table 1.

In addition, the quality of the regression models was checked by evaluation of their coefficients of significance (p value of regression), in which the found p value more than 0.05, also re-confirmed the effectiveness of models. The regression coefficients of suggested models for each response were shown in Table 2. Regression significance coefficients for each term are indicated in Table 3.

Table 2.

The significance probability (p value, F-ratio) and R² and R²-adj of the regression coefficients of suggested polynomial second order models for studied the physicochemical characteristics of β-carotene nanoemulsion hydrogels

Parameter	Mean particle size (nm)	PDI	Zeta-potential (mV)	${\text{L}}^{\ast }$	${\text{a}}^{\ast }$	${\text{b}}^{\ast }$
Linear terms
X1
p Value	0.002	0.359	0.239	0.002	0.018	0.000
F-ratio	22.28	0.99	1.71	26.91	10.48	160.2
X2
p Value	0.001	0.016	0.007	0.000	0.001	0.000
F-ratio	39.35	8.78	15.94	56.27	32.54	179.7
Quadratic terms
x11 (x1 × X1)
p Value	0.000	0.005	0.015	0.197	0.004	0.375
F-ratio	56.29	19.08	11.26	2.10	20.68	0.95
x22 (x2 × X2)
p Value	0.000	0.003	0.019	0.005	0.001	0.466
F-ratio	86.38	21.87	10.03	19.15	32.48	0.62
Interaction term
x12 (x1 × X2)
p Value	0.00	0.012	0.001	0.011	0.001	0.482
F-ratio	65.91	12.53	32.78	13.33	33.39	0.58
R2 (%)	96.22	91.67	95.06	92.59	91.77	98.16
R2 adj (%)	92.43	83.35	90.13	85.19	83.55	95.95

p value > 0.05, 1: sodium alginate, 2: sodium chloride

Table 3.

Regression coefficients of polynomial regression models suggested for each response variables

Coefficient	z-average	PDI	Zeta-potential	${\text{L}}^{\ast }$	${\text{a}}^{\ast }$	${\text{b}}^{\ast }$
a0	+ 12,426	+ 5.78	− 49.8	+ 73.6	+ 97.7	+ 23.03
a1	− 4015	− 1.908	+ 19.12	− 13.51	− 26.44	+ 0.42
a2	− 1468	− 0.631	+ 7.24	− 8.09	− 10.45	− 0.64
a11	+ 341.9	+ 0.1836	− 1.484	+ 0.741	+ 2.157	− 0.284
a22	+ 52.12	+ 0.02419	− 0.1724	+ 0.2752	+ 0.3327	+ 0.0287
a21	+ 218.8	+ 0.0880	− 1.500	+ 1.103	+ 1.621	− 0.128

p value > 0.05, 1: sodium alginate, 2: sodium chloride, a₀: constant term, a₁, a₂: linear terms. a₁₁, a₂₂: quadratic terms, a₁₂: interaction term

The surface and contour plots for each response were also provided in order to visualize of the predicted model equations. This graphical figure can be utilized to achieve the optimal condition, as well.

Mean particle size

The mean particle size of β-carotene in prepared hydrogels was varied from 148 to 655 nm. Thus it can be concluded that the nanoparticles of β-carotene could be produced in selected independent variables’ ranges. However, the size of β-carotene particles was relatively large, especially at some used concentrations for sodium alginate and calcium chloride.

ANOVA analysis indicated that the mean particle size of β-carotene in hydrogels is significantly (p < 0.05) affected by linear and quadratic terms of sodium alginate and calcium chloride concentrations (Table 2). Moreover, the interaction term of the variables also influenced the particle size variation of β-carotene in hydrogels considerably (p < 0.05). According to the results of Table 3, with respect to the negative sign of linear term ($a_1$ and $a_2$) positive sign of quadratic term ($a_{11}$ and $a_{22}$) and coefficients of both sodium alginate and calcium chloride concentrations indicates that the average particle size decreased with increase in polymers and cross-linker concentrations. However, further increase of these variables increased the particles size. Thus, there was an optimum level for both sodium alginate and calcium chloride concentrations leading to production the β-carotene nanoparticles with the smallest particle size in hydrogels. The positive sign of the interaction term also indicate that ncreasing or decreasing the alginate and calcium chloride concentration would increase the mean particle size of nanoparticles. The effects of the studied independent parameters are shown in Fig. 1a. The obtained coefficient of determination of the suggested model indicated that it can predict more than 96% of the mean particle size variation of the produced nanoparticles.

Fig. 1.

The size characteristics. a Mean particle size, b PDI and c zeta-potential of nanoemulsion based β-carotene hydrogel

Decreasing the mean particle size at high concentrations of calcium chloride can be due to the positively charged calcium ion interaction with negatively charged sodium alginate; calcium condensates the nanoparticle, reduces the length of the bond and decreases the size of the nanoparticles (Lin et al. 2017; McClements 2017).

Increasing the mean particle size with alginate and calcium chloride concentration can be related to the formation of thickening layer around the active compound to provide steric stabilization. However, at less alginate and calcium chloride concentrations, the gel matrix would be very weak, thus the entrapped nanoparticles can move easily toward each other, join together and cause a considerable growth in particle size of the systems. Moreover, high concentrations of calcium chloride decreased the electrostatic stabilization by neutralization of the particles charge (Anarjan and Tan 2013).

PDI

The PDI of gained β-carotene nanoparticles is a good index for heterogeneity of the system in term of their size distribution. This parameter is very important since it affects the physical stability of the produced nanoparticles. Since the Oswald ripening which is most important physical instability mechanism in nanoparticle systems, are significantly promoted by increasing the heterogeneity of system. It should be noted that the system having higher PDI, is more heterogeneous and consist of the nanoparticles with wide size distribution (Anarjan and Jouyban 2017; Anarjan et al. 2012). The system with less PDI is preferred due to its higher stability. The ANOVA analysis for PDI of produced hydrogels revealed that the except for linear term of sodium alginate, all linear, quadratic and interaction term of suggested second order polynomial model for this response are significant terms (Table 2). Among these significant terms, the quadratic effect of calcium chloride was the most effective term (highest F-ratio). The sign evaluation of the effective coefficients (negative sign of linear and positive sign of quadratic terms) also shows that increasing either calcium chloride or sodium alginate led to a decrease in PDI of system and creation more homogeneous hydrogel. While, further increase of them influenced inversely, and caused the PDI to increase (Table 2). This trend is very similar to particle size variation which is evaluated in the previous section. These observations can also be confirmed in Fig. 1b. It can also be seen that the calcium chloride concentration is the most effective parameter in PDI changes as compared to sodium alginate. The high coefficient of determination $\left(R^2=90.31\%\right)$ of the presented model indicated its suitability and ability to predict the PDI variations by selected formulation parameters. Increasing the homogeneity of the system (decreasing the PDI) by increasing the alginate concentrations can be related to a compact network of biopolymer chains. However, further increase in stabilizer concentration, acted inversely, in which the higher degree of ionic interaction between alginate and calcium chloride in the structure caused a significant increase in PDI by creation the particles with different sizes. The observed similar pattern in PDI by increasing the calcium chloride, also was due to the positive charged ${\text{Ca}}^{2+}$ causes a strong electrostatic repulsion effect (Lin et al. 2017; McClements 2017).

Zeta-potential

Since the sodium alginate, a large molecular biopolymer, was used as the stabilizer in preparation of nanoemulsions, the zeta-potential of the nanoparticles is of significant important. Thus, this characteristic was also evaluated for the products. The nanoemulsion-based hydrogels owned less net zeta-potential values despite prefiguration, ranged from 0.39 to 7.63 mV (with negative sign). Generally, the stabilized organic nanoparticles with large molecular biopolymers, namely polysaccharides and proteins should have higher net zeta-potential values due to their electrostatic stabilization effects. However, it seems that in the present system the nanoparticles were stabilized through the steric mechanism and the used calcium chloride probably created various cross-linking bonds between the stabilizers of the particles and neutralized the produced nanoparticles. The nanoparticles with less zeta-potential values mostly are prone to physical destabilization. In other words, the systems with higher net zeta-potential values are usually desired. Therefore, the formulation parameters were optimized to produce the system with highest net zeta-potential values. Moreover, it should be considerable that the zeta-potential sign is just related to the partial polarity of functional groups on the stabilizing compounds, in which the alginate and Tween 20 containing various –OH groups with negative sign, leading to negative zeta-potential of nanoparticles (Anarjan and Jouyban 2017; Anarjan et al. 2012; Horn and Rieger 2001). The results of ANOVA and regression analysis were shown in Tables 2 and 3. As indicated in Table 2, the linear terms of sodium alginate are only the insignificant term of the suggested model for zeta-potential changes of β-carotene nanoparticles. Among the significant terms of the model, the interaction term has the highest influence on the response.

The sign evaluation of regression coefficients of significant terms, especially interaction term (Fig. 1c), revealed that the effects of alginate and calcium chloride concentration on zeta-potential were different at various levels. Increasing calcium chloride caused an increase in zeta-potential of nanoemulsions, while this trend was inverse at higher alginate concentrations. Similar behavior was observed for the effect of alginate concentration on zeta-potential of the system at different levels of calcium chloride. The results can be explain by neutralizing effect of calcium chloride; increasing the alginate caused frequent coupling of its -OH groups with those for Tween 20 and increase of negative zeta-potential values of the system (Anarjan et al. 2012; Cornacchia and Roos 2011; Hu and McClements 2015). While this influence gets negligible at high concentrations of calcium chloride due to its cross-linking, viscosity enhancer and ion counteracting effects (Horn and Rieger 2001; Lin et al. 2017).

Lightness (L^*)

This parameter represents the samples’ turbidity. Generally, the lower $L^{\ast }$ shows the less capacity and the smaller particles. The size, dispersed phase ratio and refractive index are some influencing factors on the capacity of the nanoemulsions, and the increase of these parameters can increase the opacity $\left(L^{\ast }\right)$ (Anarjan et al. 2012; Silva et al. 2011). Thus, $L^{\ast }$ of nanoemulsions can be corresponded to the mean particle size changes of the system. According to the results of ANOVA analysis, the quadratic term of sodium alginate is insignificant in the suggested model for $L^{\ast }$ variations of the system. However, the linear term of calcium chloride is the most significant term in this model due to its highest F-ratio (Table 2). The regression analysis (Table 3), as well as Fig. 2a, shows that at simultaneously less (or more) uses of both sodium alginate and calcium chloride uses, the produced hydrogel is more turbid, like mean particle size changes pattern.

Fig. 2.

The color characteristics. a $L^{\ast },$ b $a^{\ast }$ and c $b^{\ast }$ of nanoemulsions based β-carotene hydrogels

It can be concluded that in these areas the particle size of the system was the most critical characteristic affected the system turbidity or $L^{\ast }$. However, with fewer uses of the biopolymer, increasing the calcium chloride would decrease the turbidity of the system resulting in production of a clearer hydrogels. The similar trend was also observed for increasing the biopolymer concentrations with less calcium chloride concentrations. Lower $L^{\ast }$ of the system at high calcium chloride and lower alginate can be related to creation of less emulsifier micelles due to presence of various positive and negative ions $\left({\text{Ca}}^{2+},{\text{Cl}}^{-},{\text{Na}}^{+}\right)$ and considerable increase in continuous phase dissolving capacity of colloidal particles (Anarjan et al. 2010; Anarjan and Tan 2013). Furthermore, at reducing sodium alginate uses, whole added alginate is consumed for surrounding and stabilizing the nanoparticles and no alginate would be remained freely to create the clarity-reducing semi-soluble micelles (Anarjan et al. 2012; Mun et al. 2016; Silva et al. 2011). According to the $R^2$ of the suggested model for this characteristics, it is claimed that the suggested model can predict more than 92% changes hydrogel turbidity $\left(L^{\ast }\right).$

Redness (a^*) and yellowness (b^*)

The appearance and color of food hydrogels can affect the consumer selection considerably. Moreover, β-Carotene is an intense orange color, thus the redness $\left(a^{\ast }\right)$ and yellowness $\left(b^{\ast }\right)$ of hydrogels, can directly match to their β-carotene content, in which the systems with higher positive $a^{\ast }$ and $b^{\ast }$ can correspond to the more intense orange color, higher β-content, and more chemical stability.

ANOVA indicated that all model terms had significant (p < 0.05) effects on $a^{\ast }$ of samples. The interaction term with the most F-ratio was the most effective term on $a^{\ast }$ changes (Table 2). The relatively high $R^2$ of the suggested model (91%) was reconfirmed the suitability model for describing $a^{\ast }$ variations of hydrogels. The sign evaluation of significant terms’ coefficients (Table 3) and Fig. 2a, shows that the effects of alginate concentrations on $a^{\ast }$ of samples were different at various levels of calcium chloride. Then, at reducing calcium chloride uses, increasing the alginate led to the considerable decrease in redness of hydrogels, while at higher concentrations of calcium chloride, the redness was augmented by increasing the alginate concentration. Analogous effect was also seen for the influence of calcium chloride concentrations at various alginate concentration levels. The positive sign of the interaction term coefficient $\left(a_{12}\right)$ also gave rise to complex variations of the response variable with independent variables. Therefore, an increase or decrease in the ratio of alginate and calcium chloride, simultaneously, increases $a^{\ast }$ of products.

The ANOVA results of $b^{\ast }$ illustrated that despite $a^{\ast }$, only the linear terms of both independent factors are just the significant (p < 0.05) terms on color characteristic of the hydrogels (Table 2). As shown in Table 3 and Fig. 2c, increasing either sodium alginate or calcium chloride decreases the $b^{\ast }$, yellowness and consequently, the intensity of yellow in samples. The high coefficient of determinations ($R^2>91\%$) of suggested models could also be reconfirmed their acceptable ability in prediction of the redness and yellowness changes of samples, as other studied responses.

The binary numerical optimization to get the hydrogels with the most color intensity ($a^{\ast }$ and $b^{\ast }$), and accordingly the highest β-carotene content, proposed the least concentrations of both alginate and calcium chloride, as optimum conditions in preparation of β-carotene nanoemulsion hydrogels.

The observed increase in $a^{\ast }$ at high concentrations of alginate and calcium chloride, can be due to considerable increase of the transverse bonds between the active particles (β-carotene) and sodium alginate. This results in making the hydrogel system stronger, and blocking the oxygen from the environment and suppress the oxidation and loss of β-carotene resulting in a subsequent reduction of the redness (Shao et al. 2017; Tan and Nakajima 2005). However, reducing either $a^{\ast }$ or $b^{\ast }$ by increasing the alginate and calcium chloride can be related to possible chemical bonding between the β-carotene and other stabilizer molecules or existed ions. This declines the bonds responsible for light absorption at red-yellow wavelengths (Lin et al. 2017; Shao et al. 2017; Silva et al. 2011).

Gel fraction and swelling percentage

The range of the gel fraction and the percentage of swelling for samples were quite less ranged 31.5 ± 8.26 (% w/w) and 40.1 ± 6.80 (% w/w), respectively. All samples had almost similar gel fractions and the swelling percentage, maybe because of using fewer concentrations of polysaccharide base biopolymer as the stabilizer and gelling agent, as well as calcium chloride as cross linker. Thus, these data were not included in the regression analysis of RSM. It seems that more biopolymer and crosslinker would be needed for creating the most strong and inflatable hydrogels. However, stronger hydrogels may lead to more collision of organic nanoparticles with each other and rejoining them together to make bigger micro-sized particles (McClements 2017).

Optimization and validation procedures for the optimum processing condition

In the current research, a β-carotene nanoemulsion hydrogel was considered as the optimum product if it has the smallest mean particle size, least PDI, highest zeta-potential, smallest $L^{\ast }$ and greatest $a^{\ast }$ and $b^{\ast }$. Thus, multiple-response optimizations were carried out to determine the best alginate and calcium chloride concentrations leading to the best response goals using either graphical or numerical optimization procedures.

The graphical optimization was performed by overlapping all response, known as the overlaid contour plot, which was shown in Fig. 3a. The white area shown in Fig. 3a indicates the range of best sodium alginate and calcium chloride concentrations to achieve the optimal nanoemulsion hydrogel product. It can be seen that using alginate ranging from 3 to 4.5 g/l and calcium chloride from 4 to 8 g/l would give the desired products.

Fig. 3.

a The graphical optimization of formulation parameters to get the most desirable nanoemulsion based β-carotene hydrogels, b Scanning electron microscope (SEM) image of optimum nanoemulsion based β-carotene hydrogels prepared at optimum concentrations of independent variables (4.1 g/l sodium alginate and 5.7 g/l calcium chloride)

For determination the exact optimum concentrations of sodium alginate and calcium chloride, the multiple numerical optimizations was also applied by using the desirability function. The desirability function provides the optimum conditions by converting the multiple the response into a single one. In general, the ultimate goal of optimizing conditions for the production of nanoparticles of bioactive lipid compounds bases nanoemulsion hydrogels. The results of numerical optimization suggested that using 4.1 g/l sodium alginate and 5.7 g/l calcium chloride would give the hydrogels with the smallest mean particle size ($d_{\text{min}}=151.1\text{nm}$), PDI ($PD{I}_{\text{min}}=0.294$), ($L^{\ast }{L}_{\text{min}}^{\ast }=19.2$) and the highest zeta-potential ($z{p}_{\text{max}}=4.18\text{mV}$), $a^{\ast }$ ($a_{\text{max}}^{\ast }=14.7$) and $b^{\ast }$ ($b_{\text{max}}^{\ast }=14.2$).

The accuracy of the models was checked by comparison of the experimental data to the predicted ones. The p value and T-value of applied comparison t test were for all responses were 1 and 0.00, respectively. Therefore, due to obtained p values = 1 (more than 0.05) and T-value = 1 for all responses, it can be concluded that there were no significant differences between the experimental and calculated data for each response. Thus, nothing outstanding was seem differences between them confirmed the suitability of presented models.

Moreover, the optimum hydrogel was also prepared in triplicate, at offering optimum concentrations of sodium alginate and calcium chloride, and subjected to study characteristics. Their averages of mean particle size, PDI, zeta-potential, $L^{\ast },$ $a^{\ast },$ and $b^{\ast }$ were measured as $146.2\pm 5.59,$ $0.301\pm 0.041,$ $4.32\pm 0.810,$ $20.3\pm 1.05,$ $15.01\pm 1.55$ and $13.69\pm 0.03,$ respectively. The performed comparison Tukey-test between the measured characteristics and predicted ones of the optimum sample, representing no significant differences and re-confirmed the suitability of models.

Morphology of nanoemulsion-based β-carotene hydrogels

The morphology of cross-linked hydrogels was evaluated by using SEM images (Fig. 3b). Figure 3b shows that the optimum prepared hydrogel owns a uniform and homogeneous phase, with spherical nanoparticles with less than 100 nm in size. According to the DLS results, the mean particle size of this sample is around 146 nm. This inconsistency can be related to the different measurement mechanism between them, in which the SEM images usually excludes the surrounded stabilizer molecules around the β-carotene nanoparticles. While the reported diameter size by DLS, known as Z-average, is the diameter of both β-carotene and surrounded stabilizing molecules. These were calculated by using Stokes–Einstein theory based on the particles Brownian motion.

Conclusion

Nanoemulsion based functional lipid hydrogels, such as β-carotene nanoemulsion hydrogels can be successfully prepared through preparation of their nanoemulsions and then converting them into hydrogels. The hydrogels’ characteristics could be improved by adjusting the formulation parameters, especially the concentrations of biopolymer and cross-linking agent. RSM could be used effectively to present the mathematical models in order to predict the changes of hydrogels’ characteristics by formulating the parameters as well as to optimize them to achieve the best product with the most desirable characteristics. The produced optimized nanoemulsion based hydrogels could be used in various water-based food, beverage and pharmaceutical formulations and wound-healing applications.