A membrane reactor process for the production of biodiesel

Abstract

Bench and pilot scale membrane reactors were designed and assembled to carry out the transesterification of different lipids with methanol. The membrane reactors integrate many procedures and reach several process integration objectives such as combining reaction and separation in a single unit, continuous mixing of raw materials, maintaining high mass transfer between the immiscible phases and maintaining two phases during the reaction. Several factors affecting the biodiesel production process and the biodiesel purity such as membrane pore size, volume ratio of feedstocks, recycling of the polar phase, catalyst (NaOH) concentration, and feedstock type and quality were investigated. Thus, a series of experiments using canola, palm, soybean and waste oils were performed at various alcohols to oil ratios, recycling proportions, catalyst (NaOH) concentrations, and residence times. Carbon membranes of different pore sizes (namely, 0.05, 0.2, 0.5 and 1.4 micron) were shown to retain the unreacted lipid feedstock in the reactor, which indicated that the oil droplets present in the reactor were larger than all the pore sizes; no triglycerides were found in the permeate. The effects of initial methanol/canola oil loading were also studied in semi-continuous mode. Permeate was observed at the 0.38, 0.47, and 0.64 initial methanol volume fractions (&phis;1) whereas it was not observed at &phis; 1 = 0.29, which indicated that initial volume fractions of methanol up to 0.38 could be treated in the membrane reactor. A semi-empirical model was developed and validated as a predictive equation to determine phase inversion in the reactor. The membrane reactor produces a permeate stream which readily phase separates at room temperature into a FAME (fatty acid methyl ester)-rich non-polar phase and a methanol- and glycerol-rich polar phase. To decrease the overall methanol:oil molar ratio in the reaction system, the polar phase was recycled. Three recycle ratios were tested: 100, 75 and 50%, at the same residence time and operating conditions. The permeate consistently separated to yield a FAME-rich non-polar phase containing a minimum of 85 wt.% FAME (the remainder being methanol) as well as a methanol/glycerol polar phase. At the highest recycle ratio, the FAME concentration ranged from 85.7 to 92.4 wt.% in the FAME-rich non-polar phase. The total glycerine and free glycerine contents of the FAME produced were below the ASTM D6751 standard after only a single reaction step. Under essentially the same reaction conditions, a conventional batch reaction was not able to achieve the same degree of FAME purity. A kinetic study was conducted to seek the rate constants of transesterification in the membrane reactor at various catalyst (NaOH) concentrations and RTs. The proposed mathematical kinetic model fitted the experimental results well. It was found that the catalyst (NaOH) concentration increased the reaction rates, and the RT had no significant influence on the reaction rates. The forward rate constants were over 450% higher than previous work using a batch process. This was attributed to the excellent mixing in the membrane reactor loop, the higher methanol:oil molar ratio used in the study and the continuous removal of product from the reaction medium. The work shows the advantages of product removal in enhancing the reaction rate in transesterifying canola oil in a membrane reactor. In conclusion, this work presents the first systematic study of biodiesel production using a membrane reactor. Among the most significant factors investigated were the completion of high purity biodiesel production process based on the permeate de-phasing discovery. Empirical models were selected for prediction of phase inversion in the reactor. Lower catalyst (NaOH) concentration than conventional process can be used in membrane reactor, and a very broad range of feedstocks can be used in the membrane reactor to produce biodiesel. (Abstract shortened by UMI.)