2.2.2. Design of experiment (DoE)

The statistical analytical technique was performed using MODDE software version 12.1 (Umetrics Inc., Sweden). D-optimal design for the DoE with quadratic model was selected, which was further fitted using partial least squares (PLS) method. Whereas response surface modelling (RSM) was employed to investigate and optimize the non-linear multidimensional relationship between factors and responses. A total of 24 experiments were produced, that include triplicate runs to evaluate the repeatability and error estimation. To fit the quadratic model. The experiments were conducted depending on proposed run order which was given by the software; so, the randomness of the process could be assessed. Table 2 specifies the factors and responses which were used in the DoE, respectively.

List of factors and responses with their details that were employed in the DoE study.

*Polymer concentration %w/v: −1 = 0.4%, 0 = 0.8%, +1 = 1.2%.

**Plasticizer concentration %w/v: −1 = 0.1%, 0 = 0.2%, +1 = 0.3%.

In the design, none of the factors or responses underwent transformation of values and hence the type was regular. However, various proportions of the polymer concentration and plasticizer concentration were encoded in the D-optimal design as −1, 0 or 1 that stand for the lowest value, intermediate value, and the highest value respectively. Table 3 summarizes the D-optimal design worksheet with the proportions of factors, the total number of runs as well as the run order.

The D-optimal design worksheet with factors, responses the total number of runs as well as the run order (Inc/Ecl: inclusion and exclusions; Gly: glycerine; Sor: Sorbitol), *are the used responses in the DoE study.