Response surface methodology to optimize the yield of alkyd resin from jatropha (jatropha curcas) and sesame (sesamum indicum) seed oils
Loading...
Date
2021-09-18
Journal Title
Journal ISSN
Volume Title
Publisher
Chemical Society of Nigeria (CSN)
Abstract
Utilizing the response surface methodology (RSM) of the Box Behken model to optimize the alkyd resin yield factors from low-cost feedstocks such as sesame (sesamum indicum) and jatropha (jatropha curcas) seed oils were investigated. The experimentally chosen input factor ranges of oil:glycerol (0.3–1), phythalic anhydride: glycerol (1–3), and catalyst (0.5–1.5 wt. % for optimization were used. For each of these distinct alkyd resins, optimal yield, viscosity, and specific gravity quality data were collected. The alkyd resin outcomes from the two sets of combination variable tests were utilized to validate the expected technique. The optimization solution utilizing CaCO3 catalysts and a combination of other process parameters assessed, as well as the related desirability functions, was discovered using analysis of variance (ANOVA) results for refined sesame alkyd resin (RSAR) and refined jatropha alkyd resin (RSAR) (RJAR). The optimization input variables for RSAR yield, used a CaCO3 concentration of 1.5 wt. % at molar ratios of oil:glycerol (1.0:1.0) and phythalic anhydride:glycerol (3.0:1.0), while the RJAR used a similar catalyst concentration of 1.5 wt. % at molar ratios of oil:glycerol (1.0:1.0) and phythalic anhydride:glycerol (2.8:1.0). Under these conditions, the anticipated and experimental biodiesel yields for RSAR were 48.26 % and 47.29 %, and 62.07 % and 61.61 % for RJAR, respectively which shows less than 0.5 % variations. The effects of calcium carbonate catalyst dosages on the yield, viscosity, and specific gravity of alkyd resin generated from the RSAR and RJAR were major characteristics that considerably influenced the yield, while the specific gravity and viscosity only varied moderately. The regression model provided the multivariate analysis coefficients (R2) as a variation of the mean, implying that the models were desirable.