Retention modeling and simultaneous optimization of pH value and gradient steepness in RP‐HPLC using feed‐forward neural networks

Abstract

Retention modeling and simultaneous optimization of pH value and gradient steepness in RP-HPLC using feed-forward neural networks A novel approach is proposed for the simultaneous optimization of mobile phase pH and gradient steepness in RP-HPLC using artificial neural networks. By presetting the initial and final concentration of the organic solvent, a limited number of experiments with different gradient time and pH value of mobile phase are arranged in the two-dimensional space of mobile phase parameters. The retention behavior of each solute is modeled using an individual artificial neural network. An "early stopping" strategy is adopted to ensure the predicting capability of neural networks. The trained neural networks can be used to predict the retention time of solutes under arbitrary mobile phase conditions in the optimization region. Finally, the optimal separation conditions can be found according to a global resolution function. The effectiveness of this method is validated by optimization of separation conditions for amino acids derivatised by a new fluorescent reagent.

Optimisation of HPLC gradient separations using artificial neural networks (ANNs): Application to benzodiazepines in post-mortem samples

“…In addition, having flexible starting and final concentrations avoids the use of long analysis times that are required to implement slow gradients when concentrations are fixed. Another advantage of this method is that separate ANNs do not need to be constructed for each analyte, unlike the method presented by Shan et al [13].…”

“…Firstly, in contrast to the methods presented by Madden et al [11] and Shan et al [13], the initial and final compositions of the organic modifier do not need to be constrained. Instead, the initial composition of organic modifier and the slope of the gradient can be used as input variables for an ANN, and the final composition can be calculated from the gradient slope after substituting in an appropriate time.…”

“…For the true optimum to be located, retention times should be predicted over all possible gradient conditions within the experimental space. In contrast, Shan et al [13] used an ANN to predict analyte retention times under all possible pH and gradient conditions, and found the optimum conditions by calculating resolution function values at each point. However, the retention time of each analyte was modelled using a separate ANN which, for the separation of the seventeen amino acids in this study, required the construction of seventeen individual ANNs.…”

Analysis of amino acids using serially coupled columns

“…The main problem is the inability to get sufficiently good resolution in practical analysis times, using conventional columns and instrumentation for HPLC. It is thus not surprising that a number of authors have been interested in developing optimisation protocols to improve the chromatographic performance in amino acid analysis . The most successful optimisation strategy is the in silico scanning of the performance of a large number of arbitrary conditions by inspecting their computer‐predicted chromatograms .…”