Abstract:
An eco-friendly bio-ethanol from biomass waste is an alternative to petroleum
products which are becoming scarce due to increased demand and depletion of source
besides their environmental pollution effects. Most studies on optimization of process
variables using Response Surface Methodology exploit Central Composite Design yet
other designs exist. Optimal designs have fewer trials employed with the aim of
obtaining efficient designs for fitting reduced quadratic or higher order models.
Efficiency as discriminating criteria allows comparison between any design and the
best design. This study sought to apply a second order optimal rotatable design in
four dimensions to the ethanol production through fermentation of pineapples peels
using yeast. The D-, A-, E- and T-Optimal values of the general design with their
corresponding relative efficiencies were obtained. The design was used to carry out
an experiment in order to determine the effects of time, pH, temperature and
concentration of substrate on ethanol yield during fermentation of pineapple peels.
Finally the optimum settings of time, temperature, initial pH and substrate
concentration that led to optimal yield of ethanol were determined. The coded values
of a design constructed by Das and Narasimham were used to obtain a design matrix
X which was then used to construct a moment matrix X / X. The moment matrix was
then utilized to determine the optimal values and the required efficiencies. The ratio
of the optimal value of the general design to the corresponding optimal variance of
the optimal design was used as a measure of relative efficiency of the design. A
second order model was fit into experimental data in order to study the effects of the
process variables, and the optimization through response surface plots and analytical
method. The D-, A-, E- and T-Optimal values were found to be 0.6796529,
0.04104631, 0.002856958 and 1.135448 respectively, and the corresponding relative
efficiencies were 98%,71.7%, 87.5% and 1% respectively. Normal probability plots
and R-squared of 0.95 and Adjusted R-squared of 0.911 for E-optimal (most efficient
with only 32 runs) design indicated the model fit the data well. An optimal yield of
12.35g/L of ethanol realized after 54.35 hours at 4.96 levels of pH, 34.67 0 C
temperature and 28.03g/L of substrate concentration translated to 0.441g of ethanol
per gram of substrate (roughly 86% of the theoretical yield of 0.511g ethanol per g of
substrate) which compares well with findings from similar studies. The yield by E-
optimal design was slightly lower than that of general design by 0.040g/L. The design
was found reliable in modeling, optimizing and studying effects of the factors to the
processes of fermentation of pineapples peels for ethanol production. A comparison
of these results and the result of rotatable design with four factors constructed using
balanced incomplete block design when replications are more than three the number
of times pairs of treatments occur together in the design is suggested in studying the
effects of the four process variables on ethanol yield and optimization of the process
using pineapple peels as feedstock. Study established the factor settings for optimal
ethanol yield from pineapple peels. These wastes if not properly disposed can be a
major source of pollution. A cheaper fuel than fossil fuel is provided while managing
wastes.