화학공학소재연구정보센터
Fuel, Vol.97, 531-535, 2012
Determination of the short-chain fatty acid pattern in biodiesel using high throughput syringe solvent extraction and ion exclusion chromatography
A simple, robust, and sensitive technique to determine the distribution of short-chain fatty acids (SCFA) in biodiesel and their relative contribution to the total acid number (TAN) was developed. The method is based on miniaturized volume extractions in order to (1) remove the diesel background and transfer the acids to an aqueous phase for robust ion exclusion chromatography with conductivity detection, and (2) regulate the sensitivity of the method so that it fits the concentrations for both aged and un aged biodiesel down to trace levels. The quick extraction (10 s) resulted in extraction recoveries from 90% to 100% for the six acids evaluated and the overall method precision from extraction to detection was at most 2.3% (relative standard deviation) for the tested acids (lactic-valeric). The suggested standard method covers SCFA concentrations from 0.1 mg/kg to similar to 985 mg/kg but can be extended to 2 mu g/kg by alteration of the extraction properties and sample loop size (injection volume). In order to demonstrate the usefulness of the method, unaged and aged rapeseed (B100) biodiesel were analyzed. The inherent characteristic short-chain fatty acid pattern for unaged biodiesel include lactic, formic acetic and propionic acid that all together initially constituted 33 mol% of the TAN in the specific rape seed biodiesel tested in this study. After 14 days of aging in 80 degrees C freely exposed to dry air, lactic acid concentration in the biodiesel was unaltered while formic, acetic and propionic acids increased altogether to reach over 50 mol% of the TAN. Formic acid alone contributed 42 mol% of the TAN which corresponded to an increase in absolute terms of 15 mg/kg to >300 mg/kg after aging. Longer acids such as butyric and valeric acids as well as branched isomers were not found in the tested biodiesel sample but will be detected by the proposed method if present. (C) 2012 Elsevier Ltd. All rights reserved.