Computer-simulation with commercially available software (DryLab GC) allows the prediction of isothermal or temperature-programmed gas chromatographic (GC) separation as a function of experimental conditions. In either case, two experimental runs are carried out initially, using a linear temperature program (heating rate different, all other conditions the same). Data from these two runs are entered into the computer, and separation can then be predicted for other conditions: different temperatures in the case of isothermal runs, or any kind of temperature program for programmed runs.

The reliability of resulting predictions was evaluated in the present study for several samples and a wide ranger in separation conditions. Retention time predictions were usually accurate within a few percent, and sample resolution was predicted within about ± 10%. The use of computer simulation should be a considerable help for the rapid development of superior GC methods.

Several different samples and three stationary phases of varying polarity have been examined for changes in band spacing as a function of temperature. The results of these studies have been expressed as a relative variability in the temperature coefficients of retention (S) for adjacent bands. A theoretical analysis suggests that useful changes in band spacing vs. temperature (or heating rate) can be expected when the difference in S values (ΔS) for two bands is larger than 1 to 2%. The samples studied exhibited variations in average values of ΔS of 0.6–6%. This suggests that optimizing the (isothermal) temperature or (programmed) heating rate of a gas chromatographic separation will often be advantageous.

Based on reversed-phase isocratic experiments and gradient optimization modeling, acetonitrile was found to provide optimum separation of nitrated polycyclic aromatic hydrocarbons (nitro-PAHs). A 31-min linear gradient between 24% and 80% acetonitrile in water at 35°C and 0.5 ml/min flow-rate was established. Nitro-PAH retention indices, I, were measured under these conditions. It was found that retention index values varied with changing column temperature and/or mobile phase compositions.

Diode-array, fluorescence (FD) and chemiluminescence (CD) detection were studied for nitro-PAHs. Diode-array detection responded linearly with detection limits between 2 and 12 ng/compound injected. In addition, dual-wavelength UV absorbance ratio (A230/A254, A330/A254 and A230/A330) measurements at these wavelength pairs were reported. Fluorescence and chemiluminescence provided increased selectivity and sensitivity. Four orders of magnitude linear range were found for both detection methods with detection limits between 10 and 15 pg and 50 pg (on an NO2/compound mole basis), respectively

Examples of DryLab use in pharmaceutical method development.

Description of commercial program and application to a complex sample (spearmint oil).

A review of the use of gradient elution and computer simulation (DryLab G/plus) for the separation of large biomolecules.