A three factor central composite design (CCD) was applied to optimize the drying parameters of papaya slices using Response Surface Methodology [16][20]. Drying time, X₁ (8 h, 9 h, 10 h), temperature, X₂ (60˚C, 70˚C, 80˚C), and slice thickness, X₃(5 mm, 7.5 mm, 10 mm) were selected as independent variables based on the author’s previous work [9]. TFC and TAC were considered as response variables. The rotatable central composite design was constructed with an axial distance α = 1.68179 . Each independent variable was coded as (−1, 0, +1) and the axial points as ( ± α ). Considering the axial points the complete experimental ranges were 5.32 - 8.68 h, 53.18 - 86.82˚C, and 3.30 - 11.70 mm. The design consists of 8 factorial points, 6 axial points, and 6 replicated center points, resulting in a total of 20 experimental runs. The experimental design and data analysis were performed using the trial version of Minitab software. Each design point was evaluated using three independent experiments. Details of experimental runs are provided in Table 1.

Each response variable in RSM was modeled using a second-order polynomial equation, including linear, quadratic, and interaction terms, as described by Montgomery (2017) [21]. The equation can be expressed as follows:

where Y is the response variable, γ 0 is the constant; γ i , γ i i and γ i j are the coefficients of linear terms, quadratic terms and interaction terms respectively; X_i andX_j are the independent variables; m is the number of variables; ε represents the random error of the model. The two-way and quadratic interactions are shown by the variables X i X j and X i 2 , respectively.

Analysis of variance (ANOVA) was performed to evaluate the significance of model terms and estimate regression coefficients. Optimal drying conditions were identified and validated through laboratory experiments by comparing predicted and actual response values.

Fresh papaya fruits (Carica papayaL.) were collected from a commercial farm (cultivar Mr. Sumon Chakma) located in Ghagra, Rangamati, Chattogram, Bangladesh. Fruits of uniform size, shape, and maturity stage approximately 75% yellow coloration and free from visible defects, were selected. A total of 10 papaya fruits were collected from one batch for the experiments.

The fruits were allowed to ripen at ambient temperature (25 ± 2˚C). Ripened papayas were washed thoroughly with distilled water, peeled manually, and cut into slices of predetermined thickness. Drying experiments were carried out using a laboratory hot air oven (Raypa, Model DOD-150, Spain) at three temperature levels (60˚C, 70˚C, and 80˚C). The air velocity in the convective dryer was 1.5 m/s throughout all experiments. After drying, samples were cooled in a desiccator for 30 minutes to prevent moisture absorption. The dried samples were then ground and stored in airtight containers at 4˚C until further analysis.

All chemicals used in the study, including methanol, ethanol, and 2,2-diphenyl-1-picrylhydrazyl (DPPH), were of analytical grade and purchased from Merck (Darmstadt, Germany).

2.3.1. Preparation of the Extract

Papaya extracts were prepared following a modified method of Unal et al. (2014) [22]. Each dried sample (2 g) was soaked in 100% ethanol (20 mL) and shaken at room temperature for 72 h where solvent-to-sample ratio was 10:1. The filtrates were then collected, and used as sample extract. All filtrates were combined and evaporated at 60˚C using a rotary evaporator (Heidolph Hei-VAP Digital). Crude extracts were weighed and stored at 4˚C for further analysis.

2.3.2. Determination of Total Flavonoid Content

Total flavonoid content (TFC) was determined following Chang et al. (2002) [23]. Extracts (1 mg/mL) and quercetin standards (0.02 - 0.10 mg/mL) were mixed with ethanol, aluminium chloride, potassium acetate, and distilled water. After 30 min incubation at room temperature, absorbance was measured at 415 nm using a UV spectrophotometer (UV-2600). TFC was expressed as mg quercetin equivalent per 100 g extract (mg QE/100 g).

2.3.3. Determination of Antioxidant Capacity

Antioxidant activity was evaluated using the DPPH assay as described by Azlim Almey et al. (2010) [24]. Extract (1 mL) was mixed with ethanol (6 mL) and DPPH solution (2 mL, 6 mg/100mL methanol), then incubated in the dark for 30 min. Absorbance was measured at 517 nm, and results were expressed as mg Trolox equivalent per 100 g extract (mg TE/100 g).

The experimental results on the response variables (Total flavonoid content and Total antioxidant capacity) under different drying conditions are displayed in Table 1. The initial values were 1.45 mg QE/100g and 2.45 mg TE/100g. All experiments were done in triplicate.

Table 1. Central composite design of three factors and results of responses.

The statistical data was evaluated using Minitab software. An analysis of variance (ANOVA) needs to be conducted to assess the reliability and suitability of the model [20]. A mean value from each trial was utilized to establish a second-order polynomial model, resulting in the regression equations. The predictive performance of a model can be assessed by evaluating its predicted R². In order to assess the performance of the model, various statistical measures were computed, including the mean square, sum of squares, degree of freedom (DoF), P-value, and F-value. Based on the results of the variance analysis, an F-value greater than 2.6 indicates a more precise estimation of the parameters. Furthermore, if the P-value is below 5%, the statistical model is considered valid.

3.2.1. Effects of Drying Parameters on TFC

Table 2 displays the outcomes of the analysis of variance (ANOVA) conducted on the total flavonoid content (TFC). Since the P-value for the model is less than 0.05, it indicates that there is a statistically significant relationship between the terms in the model. With an F-value of 13.51, the model can be estimated statistically significant. Terms that have P-values below 0.05 are deemed significant.

Table 2. Analysis of variance of TFC and TAC.

The predicted R² value of 0.8394 and the adjusted R² value of 0.8329 are highly similar. The coefficient of determination (R²) for the response variable was 0.9237, signifying that 92.37% of the overall variability was effectively accounted for by the model. The precision value of 9.43 exceeds 4, so, we can infer that this response is more precise and reliable [20]. Consequently, it is feasible to enhance this dependent variable by implementing the model.

Equation (2) represents the regression model fitted with the total flavonoid content.

In the regression equation, positive linear and interaction factors imply that TFC levels in papaya rise with increases in relevant variables, while negative ones present declines.

(a) (b)

(c)

Figure 1. (a)-(c) 3D plots of the variation of TFC.

Figure 1 depicts the overall quantity of TFC present in papaya. The response surface plots were generated to demonstrate the cumulative effects of the factors on the responses in the fitted model. The plots were constructed by varying two independent variables, while keeping the third variable fixed at the center of the plot. A three-dimensional graph is employed to illustrate the links between the experiment’s outcomes and its many components.

Table 2shows that the effect of drying time on the TFC is insignificant. However, drying temperature and sample thickness have a significant effect on TFC. Figure 1 shows that the increase in temperature and drying time increased total flavonoid content in papaya. On the other hand, the flavonoid content of papaya slices was significantly reduced with the increase of sample thickness. Besides, it is also clear that temperature had a more significant effect on the total flavonoid content than time and sample thickness.

The findings are consistent with those of a few other studies that have been published [25][26]. Furthermore, the flavonoid content was found to be higher when the drying process was carried out at medium and high temperatures. As the sample is dried out, its dry matter content increases, for which the total flavonoid content tends to rise. On the other hand, the negative trend of TFC might be attributed to the heat sensitivity of the variable once it has reached its maximum value [14].

3.2.2. Effect of Drying Parameters on TAC

Table 2 presents the ANOVA findings for the total antioxidant capacity (TAC). A statistically significant correlation between model variables is denoted by a P-value below 0.05. A 5.92 F-value signifies a statistically significant model. Terms with P-values less than 0.05 are considered significant. The predicted and adjusted R² values of 0.8310 and 0.8349 are comparable. The R² of the response variable was 0.8419, signifying that the model accounted for 84.19% of the variability. The dependent variable can be enhanced by the implementation of the model. The regression model incorporating total flavonoid content is represented by Equation (3).

The data presented in Table 2 demonstrates that the drying time has no significant impact on the TAC. On the other hand, the TAC is significantly impacted by the drying temperature as well as the sample thickness. Figure 2 shows that the total antioxidant capacity of papaya increased as a consequence of the rise in temperature as well as the increase in drying time. However, when the thickness of the sample was increased, there was a significant decrease in the amount of TAC. On top of that, temperature exerted the greatest influence on total flavonoid content, followed by sample thickness, and then drying time. The reduction of moisture from the sample has enhanced the dry matter content, which is the cause of the increasing trends of TAC. There are a few more research on quince fruit and black rosehip fruit that have been published, and the findings are in agreement with those of the studies [11][27].

3.2.3. Assessment of Model Adequacy

Figure 3 presents the residual analysis for both TFC and TAC, including the normal probability plot, the random distribution of residuals, and residuals versus observation order. For TFC, the values are close to a straight line (Figure 3(a)), ranging from −5 to 4 (Figure 3(b)), with the largest values observed at orders 5, 12, and 19 (Figure 3(c)). It indicates that the model accurately represents the data.

(a) (b)

(c)

Figure 2. (a)-(c) 3D plots of the variation of TAC.

(a) (b)

(c) (d)

(e) (f)

Figure 3. Residual plots for assessing the adequacy of the developed models: Normal probability versus residuals (a, d), residuals versus fitted values (b, e), and residuals versus observation order (c, f) for the total flavonoid content (TFC; a-c) and total antioxidant capacity (TAC; d-f).

Similarly, for TAC, the values approximate a linear trend (Figure 3(d)), ranging from −4 to 4 (Figure 3(e)), with the highest residuals at orders 9, 10, and 19 (Figure 3(f)). These mean the residuals are approximately normally distributed. It confirms the precision of the model. Overall, these plots demonstrate the suitability and reliability of the model in explaining the observed responses.

The optimum drying conditions for papaya, which resulted in the highest flavonoid content and antioxidant content are shown in Table 3. These optimal conditions yield the highest possible value for the desired function, which was achieved by numerical optimization. A composite desirability value close to 1 indicates that the chosen conditions are nearly ideal for achieving all objectives. The optimal drying conditions were found with a composite desirability of 0.796, which confirms that the optimized drying parameters provided an acceptable balance for retaining both flavonoid content and antioxidant capacity in dried papaya. The predicted optimum variables were time 7.64 h, temperature 80˚C, and thickness 5 mm. The corresponding response variables were TFC 18.93 mg QE/100g and TAC 27.37 mg TE/100g. After that, a drying experiment was conducted using the optimum drying conditions 7.64 h, 80˚C, and 5 mm. The desired response values yielded 19.17 ± 0.14 mg QE/100 g for TFC and 26.51 ± 0.25 mg TE/100 g for TAC, corresponding to errors of 1.25% and 3.2%, respectively. It was found that the experimental results are in good agreement with the predicted response value.

Table 3.Optimum drying conditions for the oven drying of papaya.