A comparison of model fitting using different runs in Box-Behnken design for alginate film formulation

Abstract

Box-Behnken design (BBD) is a response surface method used for experimental design in several applications such as pharmaceutical and food. BBD experiment can be run using different central points for analyzing optimal model fitting for responses. This study aimed to compare the model fitting between 3-central points (15 runs) and 5-central points (17 runs) BBD in preparation of alginate film. The concentrations of trehalose, sodium alginate, and CaCl₂ were dependent variables and film properties, thickness, opacity, and water vapor permeability (WVP), were responses. The results showed the model fitting was quadratic. Although values of adjust R² of each response in 17-run BBD were higher than 15-run BBD, ANOVA of model fitting indicated the “lack of fit” in thickness and WVP of 17-run BBD were significant (p = 0.0084, p = 0.0148, in orderly). It seemed that the 17-run BBD might have higher number of erroneously observed model prediction. Therefore, the 3-central point BBD was more suitable to be used as a model for evaluation of response and prediction of further optimized film formulation.