2.5. Experimental design
A four-factor and two-level CCD consisting of 27 experimental
runs with three replications at central point (Table 1) was used
for optimization of independent variables. The range for cellulase,
pectinase, temperature and time was selected as 3–9 FPU/g-cellulose
in the pretreated BP, 25–75 IU/g-pectin in the pretreated BP,
30–40 C and 8–20 h, respectively. Because, cellulase concentration
was based on the cellulose content in the pretreated BP, relatively
higher concentrations for cellulase were selected in
comparison with the previously reported studies (Grohmann
et al., 1996; Wilkins et al., 2007). However, it is not clear from
the previous reports, whether the cellulase addition was done on
the dry or wet substrate basis. Pectinase concentration range was
selected based on the results of previous research (Wilkins et al.,
2007). In a SSF process, both enzymes and yeast cells are accommodated
in the same vessel at the same time, therefore, a compromise
was made in selecting the temperature range. The aim of this
study was to improve process productivity, thus the maximum fermentation
time was selected as 20 h. Ethanol production was fixed
as a response for process optimization using RSM. The residual sugar
concentration was not considered as a response as sugars produced
are simultaneously fermented in a SSF process. Initial pH of
the fermentation medium was set at 5.5, since formation of galacturonic
acid (GA) by hydrolysis of pectin results in decline in pH
during fermentation, thereby adversely affecting the fermentative
ability of S. cerevisiae cells. The experiment was performed in 27
capped baffled fermentation flasks, each flask containing 75 ml of
the pretreated banana peel medium and inoculated with a freshly
prepared inoculum of yeast cells at 10% (v/v) having cell concentration
of 3 109 cells/ml. Experimental data from the CCD was analyzed
using RSM algorithm Design Expert 7.1 and fitted according
to Eq. (1) as a second-order polynomial equation including main
effects and interaction effects of each variable:
Y