NCJ Number
203122
Date Published
September 2003
Length
9 pages
Annotation
This paper reports on the optimization of the separation of 12 common explosives by capillary electrophoresis in the micellar electrokinetic chromatographic (MEKC) mode with the aid of an artificial neural network (ANN).
Abstract
After an explosion, the majority of the products formed from the explosive materials are generally simple gaseous oxides, which are not useful for forensic analysis; therefore, the detection and identification are most often limited to any unreacted original explosive material or possibly condensed products. The techniques most often used for the analyses of explosives include thin layer chromatography (TLC), high performance liquid chromatography (HPLC), gas chromatography (GC), and infra-red spectroscopy (IRS). HPLC is currently considered the method of choice, primarily because the analysis can be conducted at room temperature, thus resolving the problem of thermal instability encountered in GC. Capillary electrophoresis in the MEKC mode can be used as an alternative technique for the analysis of explosives. MEKC involves the addition of a surfactant to the electrolyte, which forms micelles that have a hydrophobic interior; this allows separation of neutral molecules. Some advantages of MEKC are high efficiency separations with relatively short analysis times, less reagent use, and very small sample sizes; however, MEKC has been found to offer less reproducible migration times and peak areas than HPLC. In the current study, the separation of 12 explosives was optimized by using MEKC with the aid of ANN's. The selectivity of the separation was manipulated by varying the concentration of sodium dodecyl sulfate (SDS) and the pH of the electrolyte, while maintaining the buffer concentration at 10 mM borate. The concentration of SDS and the electrolyte pH were used as input variables, and the mobility of the explosives was the output variable for the ANN. Eight experiments were performed based on a factorial design to train a variety of ANN architectures. An additional three experiments were required to train ANN architectures to model the experimental space. A product resolution response surface was constructed based on the predicted mobilities of the best performing ANN. This response surface indicated two optima: pH 9.0-9.1 and 60-65 mM SDS, and pH 8.4-8.6 and 50-60 mM SDS. Separation of all 12 explosives was achieved at the second optimum. The separation was further improved by changing the capillary to an extended cell detection window and reducing the diameter of the capillary from 75 mm to 50 mm. This provided a more efficient separation without compromising detection sensitivity. 5 tables, 5 figures, and 12 references