Behaviour of the isothermal retention indices of n-alkylbenzenes on stationary phases of different polarity

Anomalous temperature dependence of gas chromatographic retention indices of polar compounds on non-polar stationary phases

Increasing the reliability of both GC and GC-MS identification requires appropriate interlaboratory reproducibility of gas chromatographic retention indices (I). Known temperature dependence, I(T), is the main source of non-reproducibility of these parameters. It can be approximated with a simple linear function I(T). However, since mid-1990s-beginning of 2000s some examples of anomalous temperature dependence, I(T), preferably for polar analytes on non-polar stationary phases were revealed independently by different authors. The effect implies the variations in the sign of the temperature coefficients β=dI/dT for selected compounds and, hence, the appearance of the I-extrema (usually, minima). The current work provides evidence that the character of the anomalous I(Т) dependences (ascending, descending, or with extrema) is strongly influenced by the amounts of analytes injected into the chromatographic column, but these anomalies appeared not to be connected directly with the mass overloading of separation systems. The physicochemical model is proposed to describe the observed anomalies of I(T) dependence. This model is based on three previously known principles of chromatography, namely: The superposition of these objectives allows understanding both the unusual temperature dependence of retention indices, and the influence of the amounts of polar analytes injected into GC column on the parameters of this dependence.

A new accurate quadratic equation model for isothermal gas chromatography and its comparison with the linear model

The gas holdup time (tM) is a dominant parameter in gas chromatographic retention models. The difference equation (DE) model proposed by Wu et al. (J. Chromatogr. A 2012, http://dx.doi.org/10.1016/j.chroma.2012.07.077) excluded t(M). In the present paper, we propose that the relationship between the adjusted retention time t'RZ and carbon number z of n-alkanes follows a quadratic equation (QE) when an accurate tM is obtained. This QE model is the same as or better than the DE model for an accurate expression of the retention behavior of n-alkanes and model applications. The QE model covers a larger range of n-alkanes with better curve fittings than the linear model. The accuracy of the QE model was approximately 2-6 times better than the DE model and 18-540 times better than the LE model. Standard deviations of the QE model were approximately 2-3 times smaller than those of the DE model.