The optimized reversed-phase HPLC separation of 14 different samples is reported, based on simultaneous changes in temperature and gradient steepness. Four experimental runs are required for each sample, following which preferred conditions can be predicted using computer simulation software (DryLab). The overall accuracy and effectiveness of this method development approach is discussed, with particular attention to the use of resolution maps provided by the software. These maps are useful for maximizing resolution for the total sample, for optimizing the separation of a smaller number of selected sample compounds, and as an initial step in the separation of more demanding samples.

The preceding paper (Part I) suggests that simply optimizing temperature and gradient steepness will often provide an adequate reversed-phase HPLC separation. In some cases, however, this procedure will prove unsuccessful, and then further method-development experiments (involving change in other separation conditions) will be required. One strategy is to change a variable other than temperature or gradient steepness, followed by re-optimization of the latter two variables. The present paper examines the application of this approach with the aid of computer simulation to several samples.

In gradient elution separations, it may be required to change either column length (to increase resolution or shorten run time) or column diameter (for an increase in sensitivity or for preparative separations). In either of these changes of column dimensions, it is usually desired to maintain the same relative band spacing (selectivity), so as to increase resolution in proportion to (column plate number)1/2 when increasing column length, or to maintain constant resolution when changing column diameter. Changes in band spacing as a result of change in column size are of special concern when developing procedures for preparative chromatography under gradient conditions.