D J Salvat | PubMed Health

PubMedDate not supplied

Liquid chromatography solvent optimization: potential pitfalls when using a black box for developing a quality separation.

ICOS and DIAMOND are two commercially available, semi-automated HPLC solvent optimization software packages. The resultant optimized chromatographic separation is dependent on a combination of the operator's objective, the capability of the software system and the appropriateness of the data input. The latter contains components that represent the match between the requirements of the algorithms used and the information content of the data on which those algorithms operated. Knowledge about the sample content, stability and potential sample-solvent interactions can have a significant effect on the quality of the optimal solvent composition that is calculated. The results generated during the optimization of the separation of a mixture of U-83,757 and a variety of related compounds illustrate the need to consider the significance of the contribution to the calculated optimal separation of each of these potential pitfalls, both individually and in combination with the mode of operation of the relevant algorithms. Our results indicate that the quality of the final result is highly dependent on the intelligence content of the data used.

Acetonitriles↗

PubMedDate not supplied

Automated HPLC optimization--not all systems are the same.

ICOS and DIAMOND are two LC solvent optimization software packages commercially available from Hewlett-Packard and Unicam, respectively. U-83,757 and various related compounds were chosen as the test mixture for the comparison of these systems. Chromatographic data were collected on both systems using the same 10 mobile phase compositions, equally spaced across an iso-eluotropic plane. The comparison focused on determining the performance of both packages with respect to the prediction of the mobile phase composition required to achieve an optimal separation. Both software systems are semi-automatic, with differing amounts of operator involvement required and employing slightly different approaches to interpolating peak movements between the 10 sets of data. The predicted optimal solvent compositions are evaluated in terms of the extent the information collected across the iso-eluotropic plane was used by the various algorithms in the two systems. Our results demonstrate the importance of comprehending the component operations involved in and the limitations of any software package that is used in analytical development. The operator should always remember that both systems are simply tools and design experiments appropriately, since the quality of the final result is highly dependent on a combination of the operator's objective, the capability of the system and the appropriateness of the data input.

Algorithms↗