| In Chapter 1, the information power of experimental designs is evaluated before experiments are carried out. The experimental designs investigated (full-factorial designs, central composite designs, and random designs) were reported in the literature for fitting, in a least-squares sense, full second-order polynomial models in two continuous factors. The relations between each experimental design and the precision of both parameter estimates and response predictions were evaluated.;In Chapter 2, near-infrared (NIR) spectroscopy and principal component analysis (PCA) were used to monitor the progress of three esterification reactions between glacial acetic acid and methanol. The initial molar ratios of acetic acid and methanol were 2:1, 1:1, and 1:2. The loading and score plots resulting from the PCA of the NIR spectra were interpreted. The equilibrium constant for the esterification reaction was estimated from the score plots. The mixture space in which the three esterification reactions are constrained was reconstructed from the NIR spectra of binary mixtures of acetic acid-methanol, acetic acid-methyl acetate, acetic acid-water, methanol-methyl acetate, methanol-water, and methyl acetate-water. PCA of the NIR spectra for the binary mixtures acetic acid-methanol and acetic acid-water suggest that four major NIR detectable chemical species are successively present in solution when the concentration of acetic acid is increased. |
