Relation between characteristic temperature and elution temperature in temperature programmed gas chromatography -...

The method development process in gas chromatography can be accelerated by suitable computer simulation tools using knowledge about the solute-column interactions described by thermodynamic retention parameters. Since retention parameters usually are determined under isothermal conditions, the presented work offers a step to estimate one of the most important retention parameters, the characteristic temperature T_char by less laborious temperature programmed measurements. In the first part an empirical multivariate model was introduced describing the correlation between the elution temperature T_elu of a solute and its characteristic temperature T_char. Now in the second part a simulation model of GC and available retention data from a retention database was used to investigate the correlation between T_elu and T_char for an expanded range of heating rates and initial temperatures. In addition to part I, the simulation is used to investigate the influences of different properties of the separation column such as different phase ratios and column geometries like length and diameter or various stationary phases including SLB-5 ms, SPB-50, Stabilwax, Rtx-Dioxin2, Rxi-17Sil MS, Rxi-5Sil MS, ZB-PAH-CT, DB-5 ms, Rxi-5 ms, Rtx5 and FS5ms. The fit model is valid for all investigated stationary phases. The influence of the phase ratio to the correlation could be determined. Therefore, the model was expanded to this parameter. The expanded range of heating rates and the normalization for the system independent dimensionless heating rate required a further modification of the previously presented correlation model. The model now fits also under isothermal conditions. The results were used for estimation of the T_char of an analyte from the elution temperature in the temperature program. The prediction performance was investigated and evaluated for 20 different temperature program conditions and at two phase ratios (β=125 and β=250). Under best conditions the estimated and the measured T_char values show relative differences <0.5 %. With this novel model estimations for T_char are possible at 20 °C above the initial temperature, which expands the prediction range even for low and medium retained analytes compared to earlier approaches.