Software is described that allows the rapid development of separations by means of isocratic reversed-phase liquid chromatography (RP-LC) based on the optimization of column temperature (T) and mobile phase strength (%B). For a given sample, four initial experiments are carried out at two different temperatures, using either isocratic or (better) gradient elution. If isocratic experiments are chosen for computer simulation, it is necessary to select appropriate values of %B for these initial runs. Literature data for solute retention as a function of T are reviewed as a basis for estimating values of %B at the two values of T selected. The paper describes use of a newly introduced version of DryLab to optimize reversed-phase isocratic separations by varying temperature and %B.

The choice of T and t_G as variables for controlling selectivity and resolution during reversed-phase liquid chromatography (RPLC) method development can be used to minimize problems caused by column batch-to-batch irreproducibility. When a new column fails to provide adequate separation of the sample, altered values of T and t_G can be predicted that will restore the separation obtained with the previous column. Alternatively, columns from different manufacturers can be tested during method development, in order to find a common set of conditions (T and t_G) that provide acceptable separation with two or more of these columns. In this way, any of several columns from different sources become usable for the method. Examples are shown of these different computer-assisted procedures for minimizing problems due to column variability.

The difficulty in separating two compounds generally increases as the molecular structures of the two compounds become more similar. Isomers represent a "worst case" scenario, which can serve as a test of the efficacy of a given method development approach. We have advocated the use of DryLab for method development, with particular stress on simultaneous changes in temperature T and either isocratic %B or gradient time tG for the purpose of optimizing selectivity and band spacing. The application of the latter procedure to 137 different isomer pairs resulted in the separation of 90% of these pairs with a resolution of at least Rs = 1.0. It is concluded that optimizing temperature and gradient time is a good first step in method development.

When separations by reversed-phase liquid chromatography (RP-LC) are carried out at temperatures other than ambient, resulting retention times and bandwidths can depend on the equipment used. As a result, an RP-LC separation that is adequate when carried out on one LC system may prove inadequate when the separation is repeated on a second system. In the present study, various temperature-related problems that can result in a failure of method transfer for non-ambient RP-LC methods were examined using DryLab. Means for correcting for such effects, and thereby ensuring method transferability, are described. Using temperature to optimize HPLC separation, care must be taken to ensure that the column is at the correct temperature. An experimental study is described that leads to simple rules for ensuring good method transfer for methods run at temperatures > ambient.

Optimization of the HPLC separation conditions for the determination of albendazol and its main metabolites by gradient elution using a 2-dimensional tG-T-DryLab-model is demonstrated. The optimal separation of the compounds of interest from endogenous plasma constituens was obtained by simultaneously optimizing gradient range, temperature and gradient time. DryLab sufficiently resolved the peaks of interest from the endogenous plasma components. The results show excellent comparisons of the DryLab models with the real experiments.

Different C18 columns were used with DryLab for the optimization of temperature and gradient steepness for the separation of impurities from a pharmaceutical product. For this application, each of nine different columns gave similar results (a resolution Rs equal to 2.1-2.7), while a column with an embedded polar group gave somewhat better separation (Rs = 3.2).

Four experimental runs where temperature T and gradient time t_G are varied allow the computer-prediction of reversed-phase liquid chromatographic (RPLC) separation for different combinations of temperature and gradient time. This in turn can provide significant changes in selectivity and a resulting optimization of separation. If this procedure is repeated for different columns, additional control over selectivity and resolution becomes possible. The simultaneous variation of T and t_G for columns from different sources was studied for two samples, as a means of evaluating the general advantage of this approach for RPLC method development. Changes in relative retention with T were found to be approximately constant for different values of t_G and for different RPLC columns; similarly, changes in relative retention with tG were roughly independent of changes in temperature or the column. The latter relationships can be useful in matching ("tracking") peaks between runs during method development based on the present approach, as well as for other applications discussed in here and in Part II.