Optimization of textural characteristics of restructured pimiento strips by response surface methodology

Abstract In this study, the effect of guar gum (0.5%–1% w/w), sodium alginate (1%–2% w/w), and calcium chloride (2%–8% w/w) on textural properties of restructuring pimiento strips (RPS) was investigated. The gums were added to the pimiento strip formula, and different quality attributes including rupture force, energy to fracture, hardness, adhesiveness, cohesiveness, springiness, and chewiness were determined. Based on the textural properties of RPS, it was optimized by response surface methodology. All the textural properties of RPS were found to be significantly affected by alteration in guar gum, sodium alginate, and calcium chloride. The regression models for product’s response like rupture force and energy to fracture were highly significant. Results showed that restructured pimiento strip formula containing guar gum 1% w/w along with sodium alginate 2% w/w and 8% calcium chloride improved the textural and tensile properties. According to the RSM results on the textural properties of RPS, it is feasible to achieve the high elasticity and rigidity of pimiento strips as well as obtain the ability to tolerate thermal and mechanical stresses with appreciable textural integrity during processing such as pasteurization that would be investigated in another work.


| INTRODUC TI ON
In the recent years, restructuring process is a key aspect of the food science, including complex combination of raw materials, components, and number of texturizing and construction processes (Laaman, 2011;Raharjo et al., 1995). In restructured products, a key material which is usually a natural raw material (e.g., vegetable, meat) is re-formulated and the product is further processed. The aim of developing these products is to take natural food apart and then rebuilds them again in order to achieve better properties while the appearance, texture, and flavor are maintained (Laaman, 2011). In other words, by the process of restructuring, food can be produced with novel characteristics that did not exist in the main food (Laaman, 2011). For example, it can be pointed out that restructured food is able to tolerate thermal processes such as pasteurization, sterilization, while the raw material of the product does not have such properties. Moreover, the restructured food can be designed in such a way to stable at different conditions of preservation or storage. As a case, reconstructed pimiento strip can be pasteurized in jars and stable in the acidic calcium solution and keeps its texture characteristics up to time required for consumption (Laaman, 2011).
In preparation of stuffed olive pimiento, the process requires the elimination of stones and the creation of a cavity inside olive, and then, the cavity is filled with a piece of restructuring 2 | MATERIAL S AND ME THODS

| Pimiento strip preparation
The pimiento strips were prepared according to our previous work . In brief, pimiento puree was thoroughly mixed with deionized water, potassium sorbate (0.1 g), sodium alginate, and guar for 5 min at 21-48 g over a juicer blender (SHARP -COUNTERTOP BLENDER SBTI172G) in order to achieve no lumps and the mixture became homogenous.
Then, the mixture was rest for 30 min to remove the entrapped air bubbles. The product, viscous pimiento paste, was extruded by a head cut laboratory syringe  under gravity into the setting bath containing calcium chloride at concentrations 2.0%, 5.0%, and 8.0% w/w to develop a wide continuous sheet with 2 mm thickness. The setting sheet moved along the bath propelled by more paste was extruded. At the end of setting bath, the strip was left for 20 min to complete the gelation. In order to obtain the proper gel strength, the strips were stored in the aging bath containing NaCl 8% w/w and CaCl 2 2% w/w in the presence of citric acid 1% w/w at room temperature to complete evenly the gelation for 1 week. Essentially, this is a rearrangement of the gel network (Skjåk-Braek, Grasdalen, & Smidsrød, 1989). The effect of sodium alginate, guar gum, and calcium chloride levels on the tensile and textural properties of pimiento strips was investigated ( were cut and clamped between tensile grips. The initial distance between grips was 30 mm, and the crosshead speed was 0.5 mm/s (Karki, 2011). From force-time curves, rupture force (RF) and energy to fracture (EF) were determined. RF was taken as the maximum force peak height (N) required to break the sample, and EF was calculated as the area under the deformation curve (Herrero et al., 2008).

| Texture profile analysis and fractural properties of gels
Pimiento strips were taken out of the aging bath and cut into uni-

| Experimental designs and statistical analysis
Response surface methodology (RSM) was adopted in the experimental design. The central composite rotatable design for three independent variables of sodium alginate, guar gum, and calcium chloride concentration was selected. Each independent variable was used according to the literature study. The complete design is given in Table 3. Modified quadratic model was used for three-factor design which is given in Equation (1): where Y is the predicted response, β 0 , β 1 , β 2 , β 3 , β 11 , β 22 , β 33 , β 12 , and β 13 are the coefficients for linear, quadratic, and interaction terms, respectively, and ε is the residual error. x 1 , x 2 , and x 3 are the real values of independent variables, that is, sodium alginate, guar gum, and calcium chloride concentrations. In the current work, hardness (H), adhesiveness (A), cohesiveness (Co), springiness (S), chewiness (C), rupture force (RF), and energy to fracture (EF) were determined. The statistical analysis was carried out using Design-Expert 10 (version 10 by STAT-EASE inc., USA). To check the adequacy of regression model, R 2 , adjusted R 2 , adequate precision, and Fisher's F test were used (Montgomery, 2001).
The confidence level of the experiments was selected at 95%.

| RE SULTS AND D ISCUSS I ON
Variation of response (hardness, adhesiveness cohesiveness, springiness, chewiness, rupture force, and energy to fracture) of pimiento strips with independent variables (sodium alginate, guar gum, and calcium chloride concentration) was analyzed and shown in Table 3.
The result of regression analysis is shown in Table 4.

| Hardness
Hardness of strips was determined by measuring the force in gram required to break the product, and it varied from 1,385 to 2,956 g.
The hardness of the product is a sensory perception of the humans and is correlated with development. The effect of sodium alginate, guar gum, and calcium chloride concentrations on the hardness is represented in Table 3. Regression analysis showed that hardness of strips was significantly increased by the amount of calcium ion, although no significant effect was found by the guar gum and sodium alginate (Table 4). Second-order nonlinear regression model showed a good relationship. The lack of fit of the model was achieved at the 0.01% level and estimated coefficient of variation was 57.76%, which indicates that a quadratic model can be used to express the relation.
Increasing in hardness can be related to the more cross-linking of the alginate by calcium, which improves the gel strength. Similar results were also reported in our previous work .
The results showed that the model is significant in the F value of 3.83. The effect of calcium ion and guar gum interactions with calcium was significant on the hardness of pimiento gel (Figure 1). In other words, this interaction showed a positive effect on the hardness of the gel. The coefficient of variation was 13.88% and model accuracy was more than 7.165 which is higher than 4 are desirable, and both coefficeints showed the model is suitable. The hardness prediction model for the pimiento gel is recommended as the following equation (Equation (2)):

| Adhesiveness
Adhesiveness of strip gels was determined by measuring work necessary to overcome the attractive forces between the surface of (1) the pimiento strips and the surface of the plate of Texture Analyzer which comes in contact with it, and it varied from 11 to 40 g/mm.
The effect of sodium alginate, guar gum, and calcium chloride concentrations on the adhesiveness is represented in Table 3. All adhesiveness values were negative, and the least adhesiveness was obtained at the lowest concentration of calcium. Thus, less work is required to pull the compressing cylinder probe away from the gel sample at 2% level of calcium chloride. Nonlinear polynomial curve relationship between the variables and the response was achieved, and the analysis of variance showed a significant effect of variables on the adhesiveness of the pimiento gel at p ≤ 0.05 (Table 4). However, adhesion was decreased by increasing alginate levels and increased by increasing guar and calcium chloride ( Figure 2). The analysis of variance of adhesiveness of the pimiento strip is presented in Table 4. Second-order nonlinear regression model does not show a desirable relationship. The lack of fit of the model is not achieved at 0.01% level, which can be due to error, and the coefficient of variance is 33%. The results showed that the model is not significant in the F value of 0.55. A negative "Pred R 2 " implies that the overall mean may be a better predictor of this response than the current model. The coefficient of variation was 333%, and model accuracy was 3. The adhesiveness prediction model for the pimiento gel is recommended as the following Equation (3):

| Cohesiveness
Cohesiveness can be explained as a measurement of how well the structure of a product withstands compression, the work needed to break internal bonds. It can be defined as extent to which pimiento strip can be deformed before it ruptures, and it varied from 0.70 to 0.89. It is anticipated that the more the structure is deformed, the more the internal bonds might be broken. The effect of sodium alginate, guar gum, and calcium chloride concentrations on the cohesiveness is presented in Table 3. Nonlinear polynomial curve showed relationship between the variables and the response, and the analysis of variance showed a significant effect of variables on the cohesiveness of the pimiento gel at p ≤ 0.05 (Table 4). Increasing the concentration of alginate has a positive effect on gel cohesiveness. However, guar gum did not have a significant effect, and to some extent, it was increased and then decreased, which had the opposite effect response for calcium chloride. Thus, it was slightly reduced then ineffective and it was finally increased (Figure 3). Second-order nonlinear regression model showed a good relationship. The lack of fit of the model was achieved at 0.01% level. The results showed that the model is not significant in the F value of 1.93. A negative "Pred R 2 " implies that the overall mean may be a better predictor of this response than the current model. The coefficient of variation was 5.04% and model accuracy was more than five which is higher than four and it is desirable, which shows that both models are suitable.
The cohesiveness prediction model for the pimiento gel is recommended as the following Equation (4):

| Springiness
Springiness is rate at which a deformed pimiento strip gel goes back to its undeformed condition after the deforming force is removed, and it varied from 1.94 to 2.90 (Table 3). As it can be found from   Table 3, springiness was initially decreased at 5% of calcium chloride and then increased at 8%. Moreover, the springiness decreased with increasing alginate levels in the calcium chloride-alginate gels (Wang, 2015). The initial reduction in springiness may be related to the increasing of the gel hardness. As the alginate and calcium levels increased, a stronger gel structure formed and provided more resistance to compression, which resulted in less deformation and therefore, after compression, the gels would recover less.

Parameters H (g) A (g/mm) Co (-) S (mm) C (g/mm) RF(N) EF (g/s)
However, increase in springiness was associated with gel hardness F I G U R E 1 Response surface plot for hardness as a function of calcium, guar, and sodium alginate levels F I G U R E 2 Response surface plot for adhesiveness as a function of calcium, guar, and sodium alginate levels at 8% calcium level. Springiness of pimiento strip was also much lower than the carrot pulp restructured alginate gels (Manjunatha & Das Gupta, 2006), which can be attributed to the effect of guar gum in the alginate gel systems.
Second-order nonlinear regression model showed a good relationship, and the resulted analysis of variance showed a significant effect of variables on the springiness of the pimiento gel at p ≤ 0.05 (Table 4). Increase in the concentration of alginate and guar gum has a positive effect on the gel springiness. However, guar's effect is more moderate than that of alginate. Gel springiness decreases with increasing amount of calcium ion, but the binary interaction of alginate and guar, as well as guar and calcium chloride, has a significant effect on gel resistance (Figure 4). Second-order nonlinear regression model showed a desirable relationship. The lack of fit of the model is not achieved at 0.01% level. The results showed that the model is significant in the F value of 15.26. Calcium ions and alginate were significant parameters of the model. The "Pred R 2 " of 0.55 is not as close to the "Adj R 2 " of 0.84 as one might normally expect; that is, the difference is more than 0.2. This may be representative of a large block effect. The coefficient of variation was 4.57%, and model accuracy was more than 14 which is higher than four are desirable, which both show that model is suitable. The springiness prediction model for the pimiento gel is recommended as the following Equation (5):

| Chewiness
Chewiness is defined as energy required masticating pimiento strip gel to a state ready for swallowing: a product of hardness, cohesiveness, and springiness, and it varied from 1.94 to 2.90 (Table 3). Second-order nonlinear regression model showed a good relationship, and the resulted analysis of variance showed a significant effect of alginate on the chewiness of the pimiento gel at p ≤ 0.05. However, calcium's effect is more moderate than alginate, and the guar gum has no significant effect on the chewiness.
At the same time, the interaction between alginate and guar had a significant effect ( Figure 5). Second-order nonlinear regression model showed a good relationship. The lack of fit of the model was not achieved at 0.01% level. The results showed that the model is significant in the F value of 3.8. The coefficient of variation was 10.31% and model accuracy was more than 7.5 which was higher than 4 and showed the model is suitable (Table 4). The chewiness prediction model for the pimiento gel is recommended as the following Equation (6):

| Rupture force
Rupture force is maximum peak height resisted by pimiento gel, and it varied from 189 to 838 (N) ( Table 3). Second-order nonlinear regression model showed a good relationship, and the resulted F I G U R E 3 Response surface plot for cohesiveness as a function of calcium, guar, and sodium alginate levels analysis of variance showed a significant effect of variable on the rupture force of the pimiento gel at p ≤ 0.05. Increasing the concentration of alginate resulted in increased fracture force (Figure 6).
In addition, guar and calcium chloride concentration did not influence on the rupture force. However, the binary interaction of alginate and calcium chloride and C 2 resulted in a reduction in the gelling strength of the gel. However, A 2 and B 2 resulted in increasing the gel fracture. The analysis of variance of rupture force of the pimiento strip is presented in Table 4. Second-order nonlinear F I G U R E 4 Response surface plot for springiness as a function of calcium, guar, and sodium alginate levels F I G U R E 5 Response surface plot for chewiness as a function of calcium, guar, and sodium alginate concentrations regression model showed a good relationship. The lack of fit of the model was achieved at 0.01% level and estimated coefficient of variation was 7.99%, which indicates that a quadratic model can be used to express the relation. The results showed that the model is significant in the F value of 7.99, and model accuracy was more than 27, which showed the model was suitable (Table 4). Rupture force prediction model for the pimiento gel is recommended as the following Equation (7)