This paper introduces a new approach for the quantitative resolution of overlapped chromatographic peaks. In this new approach, sigmoid models are used to resolve area versus time data for overlapped chromatographic responses. The capability of the approach to resolve peak areas was evaluated for artificially overlapped chromatographic peaks and real unresolved multicomponent samples with degrees of resolution from 1.53 to 0.39 and peak-height ratios (leading/trailing) from 0.15 to 5.94. Results demonstrated the feasibility of using area data rather than peak data to resolve components that are not fully resolved in chromatographic processes. The degree of peak overlap has less impact on the sigmoid-based options than on a conventional peak-shaped option using the exponentially modified Gaussian (EMG) model. For small degrees of overlap (R-s > 0.7), performance characteristics for the sigmoid-based models are similar to those for the EMG model. However, for larger degrees of overlap (R-s < 0.7), performance characteristics for the sigmoid-based models are superior to those for the EMG model. For example, relative errors for sigmoid-based resolution results for overlapped responses for 2- and 3-nitrotoluene (R-s = 0.6) were mostly within about 2.5% whereas errors for EMG resolution were mostly in the range of 20-30%. (C) 2000 Elsevier Science B.V. All rights reserved.
