Fast analytical procedures for biodiesel quality control

The present dissertation is focused on the development of analytical procedures for the determination of iodine value in biodiesel and vegetable oils as well as of ester content in biodiesel and biodiesel:diesel blends. The procedure for the determination of iodine value was based on discoloration of a triiodide solution measured by spectrophotometry. For biodiesel and vegetable oil samples, linear responses were obtained from 10-106 g I2/100 g and 20-144 g I2/100 g, respectively. The coefficient of variation (n = 10) and limit of detection (99.7% confidence level) were 5.0% and 2.5 g I2/100 g, whereas for vegetable oils, the values were 3.0% and 7 g I2/100 g, respectively. About 1.2 mL of sample, 0.365 mg of I2 and 40 mg of KI were consumed per determination, with generation of ca. 2.2 mL of waste. Matrix effects were circumvented by matrix matching. Alternatively, a spot test using digital images was developed, also based on consumption of iodine by unsaturated compounds in the samples, but relying on deposition of reagent aliquots on a filter paper followed by measurement of remaining iodine, through its complex formed with starch. The measurements were performed with a cell phone camera and treated as reflectance. A linear response was obtained within 10 and 106 g I2/100 g of biodiesel. The coefficient of variation (n = 10) and limit of detection were (99.7% confidence level) estimated at 4.9% and 8 g I2/100 g, respectively. About 40 ?L of sample and 50 ?g of I2 were consumed per determination, with generation of ca. 65 ?L of waste per determination. Ester determination was based on reaction of alkyl esters from biodiesel and hydroxylamine, yielding hydroxamate, which was measured by spectrophotometry as an iron(III) complex. The flow-based procedure exploited the lab-in-syringe approach, by using ethanol as mediator solvent to generate a single phase between the hydrophobic sample and the hydrophilic reagents. Linear responses were obtained from 4 to 99% (v/v) and 2 to 40% (v/v) for biodiesel and biodiesel:diesel blends, respectively. For biodiesel samples the coefficient of variation (n = 10), detection limit (99.7% confidence level) and sampling rate were estimated at 0.80%, 0.36%, and 15 h-1, respectively, whereas for biodiesel:diesel blends, the corresponding values were 0.20%, 0.03% (v/v), and 12 h-1, respectively. About 40 µL of biodiesel or 100 µL biodiesel:diesel blends, 0.860 mg of hydroxylamine, and 0.145 mg Fe were consumed per determination, with generation of ca. 3 mL of waste. The results obtained by all the proposed procedures agreed with the reference ones at the 95% confidence level. The proposed procedures are practical and feasible for on-line or at-line analysis, by consuming only low amounts of organic solvents and reagents, and generating low waste volumes (AU)