2.1. Materials

Fresh Artemia biomass was reared at the Vinh Chau experimental farm of Can Tho University, Vietnam. From 18 to 20 days after stocking, Artemia biomass was harvested. Egg albumin (EA) (India) and XG (Bob’s Red Mill, USA) were used.

2.2. Preparation of the Artemia Puree and Foam

The Artemia biomass were cleaned and then frozen at −10°C. Before starting each experiment, the samples were thawed at 3–5°C for 12 h. Thawed samples were steamed for 5 min to destroy microorganisms or enzymes. Next, the samples were cooled at ambient temperature about 10 min and was used for our experiments. To prepare the foam samples, the concentrations of EA (9.2–15.2%), XG (0.1–0.38%), and the WAR (2/1–6/1) were transferred to a 250 mL beaker, the mixture was mechanically whipped (Philips HR3705 300 W) at the highest speed for 6 min, fixing the batch volume at 100 mL. During whipping, the foam was developed by incorporating air in it to increase the surface area, which expanded in volume (according to the optimization conditions obtained in the previous study by Thuy et al. [10]). At the end of each treatment, the mixture was evaluated on the foam properties, including foam EV, FD, and volume drainage.

2.3. Optimization Process

The ranges and levels of variables investigated in the research are tabulated in Table 1. Three independent variables were studied namely EA (X₁), XG (X₂), and WAR (X₃). RSM was utilized to determine the numbers of experimental runs required for the Box–Behnken experimental design that produces reliable measurement of the response. The optimization was conducted for multiple responses of extraction yield (EV = Y₁, FD = Y_2, and volume drainage Y₃). Table 2 exhibits that the total of 18 runs were required for three variables employed in this experiment. The experimental plan based on Box–Behnken design consisting of six center points.

Table 1: Independent variables with range and factor level.

Table 2: Design of experiment of three variables.

RSM (using Statgraphics Centurion XV software) was applied to determine the optimal conditions for the foaming process from Artemia puree. A full second order reaction model (equation 1) was established.

Where: Y: Predicted responses (EV, FD, and volume drainage); β_o: Intercept coefficient; β₁, β₂, β₃: Linear terms; β₁₁, β₂₂, β₃₃: Quadratic terms; β₁₂, β₁₃, and β₂₃: Interaction terms and X₁, X₂, and X₃: Uncoded independent variables.

2.4. Analysis of Foam Properties

FD was determined by the method described by Abd Karim and Wai [14] (equation 2).

EV was analyzed to represent the amount of air introduced into the water during foaming and measured as a percentage increase in water volume, calculated using equation 3 [15].

Where V_o is initial volume (mL) and V₁ is final volume (mL).

Drainage volume (DV) was measured based on a method defined by Narender Raju and Pal [16]. 100 g of foam was poured into a Buchner filter (diameter 80 mm) and placed in a 100 mL cylinder. Volume of liquid released from the foam in 120 min, then uses a 5 mL pipette or a 100 mL measuring cylinder to determine the drain volume.

3.1. Effect of EA, XG, and Water/Artemia Ratio (WAR) on FD

One of the most difficult problems in foaming in preparation for drying is the stability of the foam during foaming and drying. If the foam is not stable, the porous structure will collapse, resulting in a deterioration in product quality after drying. Foaming ability can be assessed by the FD index, lower FD also means that more air is trapped in the foam [11,17]. The lower the FD, the more the foam expands and helps to transfer heat better for the subsequent drying process. All variables (EA, XG, and WAR) had a significant effect on FD. The values obtained ranged from 0.263 to 0.415 g/mL. Multiple range tests for FD by EA, XG, and WAR were performed. FD values were significantly reduced using EA from 9.2% (0.38 g/mL) to 12.2% (0.30 g/mL) and were not significantly different when increasing this concentration to 15.2% (0.31 g/mL). The low concentration of EA makes the bubble unstable because it cannot form the critical thickness required for the foam film. However, when the concentration of EA increases too much, it also reduces the overall foaming ability, possibly because the viscosity of the solution increases as the protein content increases, so it does not create favorable conditions for large gas escapes. Large amounts of gas are incorporated at the interface [18]. An increase in XG (0.1–0.38%) led to an increase in FD (values ranged from 0.31 to 0.34 g/mL). Azizpour et al. [12] reported that FD increased significantly with increasing concentration of XG, probably due to an increase in the thickness of the mixture. The high viscosity liquid phase will prevent air trapping during whipping or mechanical mixing. Salahi et al. [19] reported that when albumin and XG were added to cantaloupe pulp for foaming, an increase in albumin concentration from 1% to 3% resulted in a significant decrease in FD, while FD increased with increasing XG content from 0.05% to 0.2%. The increase in the ratio of water/Artemia caused the FD to decrease significantly, when the WAR changed from 2/1 to 4/1, the FD value decreased from 0.36 to 0.30 g/mL and could not be showed a significant difference between the 4/1 and 6/1 ratios. That may be because increasing dilution during foaming reduces the viscosity of the Artemia mixture and thus the FD decreased. The ANOVA table [Table 3] partitions the variability in FD into separate pieces for each of the effects, it then tests the statistical significance of each effect by comparing the mean square against an estimate of the experimental error. The results show that most of the independent variables, their first and second order interactions have a significant effect on FD. Specifically in this case, eight effects have P < 0.05, showing that they are statistical significance at the 95.0% confidence level, only the X₁X₃ interaction has no significant effect on FD.

Table 3: Analysis of variance for FD.

The equation of the fitted model is presented in equation 4, where the values of variables are specified in their original units. Since P-value for lack-of-fit is >0.05 (P = 0.65), the model appears to be adequate for the observed data at the 95.0% confidence level.

Where Y₁ is FD (mg/L), X₁ is EA concentration (%), X₂ is XG (%), and X₃ is WAR (v/w) (R² = 92.6%; R² adj. = 91.28%; SEE = 0.013)

3.2. Effect of EA, XG, and WAR on EV

The amount of air integrated into the Artemia puree during foaming was measured using the foam EV. The analysis of variance (ANOVA) results showed that eight effects have P < 0.05 (data are not fully presented here), indicating that they are significant different from zero at the 95.0% confidence level. It was observed that double interaction (X₁ X₃) did not affect the foam EV. This pane shows the regression equation that has been matched to the data, where the values of the variables are specified in their original units. The equation 5 of the fitted model is:

Where Y₂ is EV (%), X₁ is EA concentration (%), X₂ is XG (%), X₃ is WAR (v/w) (R² = 92.99%; Adjusted R² = 91.74%, SEE = 12.05)

Similar to FD, all three variables had significant influence on foam EV. The minimum and maximum analyzed EV values are 156.63 and 310.03%, respectively, at different levels of EA, XG, and WAR. The EV values were significantly increased from 183.12% to 262.8% in increasing of EA concentration from 9.2% to 12.2% and were not significantly different (269.27%) when increasing this concentration to 15.2% (269.03%). Foaming ability and foam stability are important functions of EA. Foam is a colloidal system consisting of small air bubbles that propagate in a continuous aqueous phase. The protein in EA acts as an amphoteric emulsifier between the air and water phases, resulting in good foam stability [20]. At the amount of XG was increased from 0.1% to 0.24%, the EV values increased from 240.14% to 258.34%. However, when increasing XG up to 0.38%, the EV value decreased (228.94%). Increasing the XG concentration affects the increase in the viscosity of the continuous phase. In addition, XG aids in the formation of durable films and stabilizes the interfacial films. However, when it is increased too much, the EV decreases. This is probably because when the viscosity of the mixture increases, the incorporation of air decreases, leading to a decrease in foam expansion and an increase in FD [21]. High viscosity liquids prevent air from being trapped during whipping or mechanical mixing. Similar trends were also reported for bael fruit pulp [22]. Increasing the ratio of water to raw material significantly increased the EV, when the WAR changed from 2/1 to 4/1, the EV value increased from 212.57% to 247.68% and showed no significant difference between the ratio 4/1 and 6/1 (59.57%). As explained above, it is possible that increasing dilution during foaming reduces the viscosity of the Artemia mixture and hence FD decreases and EV increases.

3.3. Effect of EA, XG, and WAR on DV

Drainage or liquid phase from the foam is measured as the factor that destabilizes the foam. Therefore, measuring DV is also one of the better methods to determine its stability. The measured DV values ranged from 0.2 to 7 mL. WAR and XG had significant effect on the DV (at 5% level of significance). It was observed that with higher ratios of WAR exhibited greater drainage. This is due to the increase of dilution rate and decrease of solid content of the mixture. In addition, increasing the XG concentration reduced the DV of Artemia foams (increased the stability). Pasban et al. [23] determined that XG at 0.3% gave good foam stability. Muthukumaran [24] also reported a strong effect of XG as a stabilizer. Some previous studies have also reported similar results [9,20,25,26]. In this study, XG and WARs affect DV clearly than EA, ANOVA for DV was performed. In this case, seven effects have P < 0.05 (data are not fully presented here), indicating that they are significantly different at the 95.0% confidence level. The interaction of X₁ and X₁X₂ is not significant (P > 0.05), so these interactions are omitted from the equation, the equation 6 of the fitted model is:

Where Y₃ is DV (%), X₁ is EA concentration (%), X₂ is XG (%), and X₃ is WAR (v/w) (R² = 88.96%; Adjusted R² = 87.29%, SEE = 0.68).

Since P-value for lack-of-fit in the ANOVA table is >0.05 (P-value of Lack-of-fit = 0.64), the model appears to be adequate for the observed data at the 95.0% confidence level.

3.4. Multiple Response Optimization

Optimizing multiple responses is more useful when it is necessary to evaluate the impact of multiple variables on responses. The numerical optimization results show that the desired maximum (0.91) can be achieved using the optimal concentrations of EA, XG, and WAR of 13.5%, 0.5% and 4.6:1. At these maximum levels, the predictive values of responses such as FD, EV and DV are 0.7 g/mL, 293.4%, and 0.27 mL, respectively, shown in asterisk (*) on Figure 1.

Figure 1: Overlay plot of foam density, foam expansion volume, and drainage volume at different levels of input variables. [Click here to view]

The validity of the model has also been confirmed through the experimental performance at the optimal level of the independent variables. The validation results gave the experimental values almost good agreement with the response values predicted from the selected models [Table 4]. In addition, the test data on the responses obtained at the optimum level have also shown that the foam has good properties and is ready for the next stage of drying.

Table 4: Predicted and experimented value of responses from optimal conditions.