IDENTIFICATION AND QUANTIFICATION OF OVERLAPPED PEAKS IN LIQUID-CHROMATOGRAPHY WITH UV DIODE-ARRAY DETECTION USING AN ADAPTIVE KALMAN FILTER

A method for the identification and quantification of overlapped peaks in liquid chromatography with UV diode array detection is reported. T he adaptive Kalman filter, which is optimized using simplex optimizati on, can detect low levels of isomeric impurities. The filter optimizat ion is based on the maximum information yield from the filtering proce dure, and is capable of compensating the model errors. This approach i s combined with the commercially available software to result in a com plete method for chromatographic peak analysis. A mixture of two isome rs, chrysene and benz(a)anthracene, has been evaluated. This method ca n detect as low as 1 mu M benz(a)anthracene in a 121 mu M chrysene sol ution, and 1 mu M chrysene in a 106 mu M benz(a)anthracene solution. B oth, identification and quantification of the major and minor componen ts have been achieved. For the overall method, the prediction errors a re within 2% for the major components and within 10% for the minor com ponents. If the reference spectrum of the minor component is unavailab le, this filter can also predict the concentration of the major compon ent within an error of 7%. The reliability of this method has been tes ted by the second isomeric system, a mixture of benzo(k)fluoranthene a nd benzo(b)fluoranthene. This method is valid for highly overlapped pe aks even when the chromatographic resolution is zero, but an overlap-f ree region is required in the spectra.