Two types of separations are described in the literature: isocratic procedures, in which the mobile phase does not change during the run, and gradient meth- ods, in which the mobile phase is changed in a regular and reproducible manner.
Isocratic methods are usually preferred for separating simple mixtures of com- pounds because column reequilibration is not necessary and the effluent reaching the evaporative interface is consistent except for changes in peak components.
However, the column may not be able to separate the target compounds in a mixture under isocratic conditions, or the compounds that are retained longest may take too long to elute off the column. Gradient procedures often allow sep- arations that cannot be achieved with a simple mobile phase and allow washing out of late-running peaks, but columns must be reequilibrated to the starting mobile phase before the next sample can be injected. Gradients most run slowly enough to be reproducible and generally will test your pumping system’s per- formance. An inefficient pumping system will give poorly reproducing gradient performance and can be checked by running sample step gradients into a UV detector at 254 nm with one of the solvent reservoirs spiked with acetone as described in Chapter 4. Detection of the acetone will provide a method of visu- alizing the baseline changes. Bad gradient formation will not show perpendicular
ISOCRATIC AND GRADIENT METHODS DEVELOPMENT 47 slope steps and will show overshooting of the plateaus. This is covered in detail inHPLC: A Practical User’s Guide.
The resolution of a separation is controlled by three factors: retention (K), separation (α), and efficiency (N). The resolution equation ties all of these factors together in a single equation:
R= 1 4
α−1 α
√
N K 1+K where (see Figure 5.1)
K= VA−V0
V0
α= VA−V0
VB−V0
N =16 VA
WA
2
=5.42 VA
W0.5
2
Polar components of the sample come off first on these columns, with non- polar compounds being retained the longest. Remember: “Like attracts like.” By comparison, polar normal-phase columns retain polar components while releas- ing nonpolar compounds first. The longer a compound is retained on a column, the better the chance that it will be separated. The retention factor is a measure of the retention of a compound on a column relative to the void volume of the column (Figure 5.1).
Increasing the amount of the stronger, nonpolar solvent used with a octylde- cyl–silica column can decrease retention times by decreasing the mobile-phase polarity. The relative positions of the eluting peaks will remain the same, but compounds will not stick as tightly and will wash out faster. This will help
VA V0
WA B
A
FIGURE 5.1 HPLC separation control factors.
48 SAMPLE PREPARATION AND SEPARATIONS DEVELOPMENT
get later runners off faster, but you may lose the resolution of the early-running peaks. The retention factor,K, portion of the resolution equation is a convergent term meaning that after you increase the retention past a certain point, diffusion overcomes the gain you get from making things stick longer on the column.
Selectivity,α, is measured by differences in the retention factors of two adja- cent peaks. These selectivity changes cause changes in the relative position of chromatographic peaks. Altering the chemical nature of the column bonded phase, the solvent, or the sample usually changes the selectivity. The easiest change is usually to change the more nonpolar solvent (i.e., methanol, acetonitrile, or tetrahydrofuran) in the aqueous mobile phase. You can also switch to a column with a different organic bonded phase, but this will require buying a second column, downtime to replace the column, and reequilibration time. Although most separations are carried out on C18–silica columns, other columns pro- vide different specificities for different uses. Octyl–silica columns hold nonpolar compounds less tightly and may show peak position switching compared to octyldecyl–silica. Phenyl–silica columns are selective for aromatics and com- pounds with double bonds, such as fatty acids. Amino–silica columns will separate among monosacchrides, such as glucose, galactose, and fructose, as well as among polycaccharides.
Temperature changes and sample derivatization can also alter selectivity but not in any predictable way useful in methods development. Temperature increases generally will shorten run times while producing separation changes. Temperature as a control variable on silica-based columns always comes with many compli- cations. Bonded-phase hydrolysis and column packing solubility increase with increased temperature, and thermally labile compounds will decompose.
The efficiency factor, N, is the final factor influencing resolution. The effi- ciency factor measures the sharpness of peaks and the decreasing overlap between adjacent peak bottoms and is measured by dividing the retention volume by the peak width. It is controlled by column packing diameter and shape, packing technique, flow rates, extracolumn tubing volumes, and column length. It is a square-root factor in the resolution equation, so large changes in efficiency do not produce a proportional change in resolution. Columns are often sold on effi- ciency comparisons, but this can be very deceptive. Efficiency values can change dramatically with the instrument the columns are run on, the methods used for calculation, and the standards selected. Efficiency measurements are valuable for studying changes in column performance over time and definitely should be mea- sured under a standard set of conditions when new columns are being introduced into an accepted method. Following the nature of changes in the separation under these standard conditions over time provides information on the rate of column aging and cleanliness.
Solvent gradients use retention changes to improve separations that cannot be accomplished with a constant solvent. They can be run as a series of step gradient changes to speed elution of later-running peaks. Or they can be run as a continuous solvent change with a final plateau to ensure that all peaks are eluted, followed by a reequilibration step to prepare for the next injection. Timed
AUTOMATED METHODS DEVELOPMENT 49 changes in the gradient slope can be incorporated in the gradient change to improve separation of compressed or expanded areas in the separation. Finally, scouting and washing gradients with other solvents can be incorporated in a multipump gradient system. Generally, only two solvents will be used in an analytical gradient, although you may see a procedure with a constant level of a third solvent added. Developing three- and four-solvent gradients is very complicated, and results are difficult to predict.