3.2. Fitting the Model
Analysis of variance (ANOVA) and multiple regression analysis were employed to evaluate the effects of individual and interaction factors using Design Expert 13. Box–Behnken design is the most frequently used RSM design, and the model equation is applied to predict the optimum CO2 gas absorption flow rate (Y1) and CO2 gas mass transfer coefficient (Y2).
The final equation in terms of coded factors is presented in Equation (13):
The final equation in terms of actual factors is presented in Equation (14):
The model equation was applied to predict the optimum CO2 gas mass transfer coefficient (Y2, kog).
The final equation in terms of coded factors is shown in Equation (15):
The final equation in terms of actual factors is shown in Equation (16):
where Y
1 is the predicted CO
2 gas absorption flow rate response, and Y
2 is the CO
2 gas mass transfer coefficient. Meanwhile, X
1, X
2, and X
3 are independent variables for absorption time, absorption flow rate, and CO
2 gas flow rate, respectively.
Based on the equation, it was shown that the influence of absorption times (X1), absorption flow rate (X2), and CO2 gas flow rate affected the CO2 absorption rate and CO2 gas mass transfer coefficient. These effects can be observed from the intercept and coefficients of the three optimization equations.
Figure 2 illustrates the influence of interactive variables.
Figure 2a displays the experimental and predicted data plot. The value predicted by the design response surface equation has a high degree of accuracy (R
2 = 0.97) and has an intercept of 1.00663. The distribution point spread above the prediction line demonstrates this. To obtain a respectable model, we must examine the normal probability value (%). As shown in
Figure 2b, the proposed equation model is appropriate for predicting the CO
2 gas absorption flow rate (Y
1) and CO
2 gas mass transfer coefficient (Y
2) when using absorbent NaOH 0.1 M. The value predicted by the design response surface equation is precise (R
2 = 0.94) and has an intercept of 1.25048. The correlation of the variable effects on variable responses can be seen in
Figure 2,
Figure 3,
Figure 4,
Figure 5 and
Figure 6.
In
Figure 2, NaOH is wholly neutralized at the phenolphthalein endpoint, Na
2CO
3 is half neutralized, and HCO
3− has not responded. From the phenolphthalein endpoint to the methyl symbol endpoint, the bicarbonate is neutralized. Therefore, only a few drops of titrant will be required for the NaOH to change from pH 8 to 4, which will be corrected with a blank indicator. As shown in
Table 4, v
1 is the volume of acid in millimeters used from the start of the titration to the phenolphthalein endpoint, and v
2 is the volume from the phenolphthalein endpoint to methyl orange.
During the absorption of CO
2 gas, a reaction occurs between CO
2 and OH
− because the pH value decreases, and the reaction proceeds according to the following criteria. If pH = 11.5, the reaction is:
If pH = 10, the reaction is:
If pH = 11, the reaction is:
After absorption of CO
2 gas, a mixture of OH
−, CO
3−, and HCO
3− ions occurs in the solution [
25,
34]. Sample titration using HCl whose molarity is known (HCl standard) will obtain two equivalence points. In the first step, OH
− reacts as a whole, whereas HCO
3− reacts only in the second step, and CO
32− splits in both. Therefore, direct calculation of the concentration is not possible but can be calculated if the
p-value and m-value have been determined. The
p-value (phenolphthalein price) is the volume of titration used to titrate 1 mol/L HCl for 1000 mL so that the color of the phenolphthalein indicator changes (or to a pH of 8.2 using a glass electrode). The M-value (methyl orange value) is the volume used to titrate 1000 mL of HCl mol/L (1 M) sample until the color of the methyl orange indicator changes (or at pH 4.3 by using a glass electrode). The formulas used in this regard are as follows:
p-value = V × M × 1000/aliquot, where V = volume of titration (mL), M = molarity of HCl used, aliquots = volume of sample titrated,
,
, and
CO
2 absorption is a function of pH; the reaction that occurs is a function of pH; the increasing absorption of CO
2, which continues to increase, is a function of pH, solubility, thermal, and longitudinal diffusion; the mechanism of mass contraction is the driving force in the form of differences in CO
2 concentrations and the partial pressure of each component and total pressure; the formula used is pH dependent, as it is determined by pH [
51,
52], and the reaction is according to the following criteria. If the pH value < 4.5, then
. If the pH value is between 4.5–8.3, then
. If the pH value is between 8.3–9.5, then
. Suppose the pH value is >9.5. Then
calculated from Equation (2).
Figure 3a,b shows the 3-D optimization of the effect of absorption time (minutes), the flow rate of 0.1 M NaOH absorbent (L/min), and the flow rate of CO
2 gas mixed with air (L/min) on the absorption rate of CO
2 gas in the Raschig packing column ring.
Figure 3a shows the effect of absorption time (X
1, min) and absorbent flow rate (X
2, L/min) on CO
2 gas absorption flow rate (Y
1, L/min). The effect of the two independent variables on the response variable (CO
2 gas absorption) increased, and optimal conditions were reached at 4.13 L/min.
Figure 3b shows the desired optimization of the effect of absorption time (minutes), the flow rate of 0.1 M NaOH absorbent (L/min), and air-mixed CO
2 flow rate (L/min) on the absorption rate of CO
2 gas in the Raschig ring packing column. The effects of absorption time (X
1, min) and CO
2 gas flow rate (X
2, L/min) on CO
2 gas absorption rate (Y
1, L/min) increased until it reached the optimum condition of 4.13 (L/min).
Figure 4a,b show the 3-D optimization of the effect of absorption time (minutes), the flow rate of 0.1 M NaOH absorbent (L/min), and CO
2 gas flow rate (L/min) on the mass transfer coefficient of CO
2 gas in the Raschig ring packing column.
Figure 4a shows the effect of absorbent flow rate (X
2, L/min) and CO
2 gas flow rate (X
3, L/min) on the CO
2 absorption flow rate. The absorption rate of CO
2 gas increases until it reaches optimum conditions, which are reached at a value of 4.1292.
Figure 4b shows the effect of absorbent flow rate (X
2, L/min) and CO
2 gas flow rate (X
3, L/min) on desirability. Desirability can be increased by choosing the proper criteria in planning the constraints to achieve the best optimal conditions. The best condition is achieved at the desired value of 0.99999.
Figure 5a shows the effect of absorbent flow rate (X
2, L/min) and CO
2 gas flow rate (X
3, L/min) on the CO
2 absorption flow rate. The CO
2 absorption flow rate continues to increase so that optimum conditions are reached at 4.13 L/min.
Figure 5b shows the desired optimization of the effects of absorption time (minutes), the flow rate of 0.1 M NaOH absorbent (L/min), and the CO
2 gas flow rate (L/min) on the mass transfer coefficient of CO
2 gas in the column packing Raschig ring. Based on the analysis and optimization of 3-D plots, the optimum conditions for the CO
2 uptake rate are 1.44656 L/min, and the mass transfer coefficient for CO
2 gas is 1.44656 mol/min m
2 atm.
Figure 6a shows the effect of the absorbent flow rate of 0.1 M NaOH (X
2, L/min) and CO
2 gas flow rate (X
3, L/min) on the mass transfer coefficient of CO
2 gas. The greater the flow rate of the absorbent and the flow rate of CO
2 gas, the higher the mass damping value, and the faster the mass locking occurs due to the driving force difference in CO
2 gas concentration in the absorbent and air. Optimal mass absorption conditions are achieved at 1.44656 mol/min m
2 or mol/min m
2 atm.
Figure 6b is the effect of 0.1 M NaOH absorbent flow rate (X
2, L/min) and CO
2 flow rate (X
2, L/min) on optimal desirability based on a design to determine constraints that produce the best-desired desire. The optimum desirability value is 0.9999 (≈1). Numerical optimization and desirability ramps are shown in
Figure 7 and
Figure 8, respectively.
Figure 9 shows the 3-D optimization of the effect of absorption time (min) and pH on the formation of HCO
3− and CO
32− as a result of the reactions between 2OH
− + CO
2 and CO
32− + CO
2 + H
2O
2 + HCO
3−, conditions under which the optimum was achieved at 154.994 mol/L HCO
3− and 105.664 mol/L CO
32−. This shows that absorption also occurs chemically, according to pH. This reaction occurs at a pH of 9.5–11, forming CO
32− and HCO
32− due to the reaction of OH
− and CO
2, whose products can be used to meet the needs of cosmetics and food ingredients.
Figure 10 shows the numerical optimization bar graph (Pareto graph) for desirability. The bar graph is a graphic display for each optimal solution. Optimal factor settings are shown with red bars, and optimal response predicted values are shown in blue. Optimum conditions for the desirability of each factor are, respectively, absorption time (X
1, min), absorbent flow rate (X
2, L/min), and CO
2 gas flow rate mixed with air = 0.99, CO
2 absorption desire rate (Y
1, L/min), and mass transfer coefficient CO
2 gas (Y
2, mol/min m
2 atm) = 0.999 each; and combined desire = 0.99.
Figure 11 is the desirability ramp for numerical optimization of the 3-D optimization of the effect of absorption time (minutes) and pH on the formation of HCO
3− and CO
32− as a result of the reaction between 2OH
− + CO
2 → CO
32− + H
2O and CO
32− + CO
2 + H
2O → 2 HCO
3−, optimal conditions achieved, i.e., absorption time = 15.3097 (X
1, min), pH = 11.52 (X
2), CO
32− = 105.664 mol/L, HCO
32− = 154.994 mol/L, a total of CO
2 = 76.9359 L, and desirability = 0.91.
Table 9 illustrates the numerical optimization. Constraints are designed according to appropriate criteria to produce reliable validity and high accuracy by looking at the influence and correlation between each independent variable and the response variable to produce the correct constraints and high desirability.
where, n is the number of responses in the measure; if all the important values are the same, the simultaneous objective function reduces to the normal form for desirability.
For the goal of maximum, the desirability will be defined by the following formulas:
For absorption time criteria, X1 (10–80 min) maximum goals: optimum X1 = 80 min absorption flow rate; X2 (2–5 L/min) maximum goals: optimum X2 = 5 L/min, CO2 gas flow rate, X3 (2–5 L/min): optimum X3 = 5 L/min on absorption flow rate Y1 = 3.967 L/min and desirability = 0.999999 ≈ 1. Desirability ramp for numerical optimization of three goals, i.e., the absorption time, X1 (10–80 min), absorbent flow rate, X2 (2–5 L/min) CO2 gas flow rate, X3 (2–5 L/min), on mass transfer coefficient (Y2, mol/min m2 atm) as the response variable. Optimum condition: X1 = 5 L/min, X2 = 5 L/min, X3 = 5 L/min, Y2 = 1.442 mol/min m2 atm and desirability = 0.999999 ≈ 1
Table 10 presents desirability function optimization. Analysis of variance (ANOVA) for responses to CO
2 gas absorption flow rate (Y1) indicates that effects of individual factors (absorption time, absorption flow rate, and CO
2 gas flow rate) are significant for degree of confidence ≥ 95% (
p-value ≤ 0.05). The RSM model selected is a quadratic model, R2 = 0.97, CV = 7.4, and the model is very significant.
Analysis of variance (ANOVA) for the response variable CO2 gas mass transfer coefficient (Y2) showed the influence of individual factors and significant interaction for the degree of confidence ≥ 95% (p-value ≤ 0.05), except for the interaction effect of X1 X 2 and X2 X 3, which was not significant. Therefore, the RSM model was chosen as a quadratic model with R2 = 0.94 and CV = 4, and the model is significant.