Selectivity in Gas Chromatography

Colin Poole is a polychromatographer with broad interests in the separation and detection of small molecules in biological, environmental, and food samples. He is the coauthor of over 400 papers and eight books, an editor of Journal of Chromatography A and three scientific encyclopedias, and a member of the editorial boards of five other analytical chemistry journals. Dr. Poole is a professor at Wayne State University and has served in several capacities as a Science Advisor to the U.S. Food and Drug Administration.

All chromatographers need a working knowledge of selectivity to facilitate the development of separation methods with the desirable properties of adequate resolution in a reasonable time. Selectivity is usually measured by the separation factor (the ratio of the retention factors for any two peaks in the chromatogram). The separation factor is an observation. Why two peaks are separated to different extents on different columns requires some basic understanding of the interactions between the two compounds and the separation system. This knowledge is not widespread, and many practicing chromatographers pride themselves on their ability to achieve a separation without ever unearthing how it was achieved. This is possible because resolution is the observed result from the contribution of kinetic and thermodynamic parameters on a separation. In gas chromatography the smallest useful separation factor is about 1.005 and requires a column with a plate count of slightly more than one million. A difficult task but just achievable with existing columns. With a separation factor of 1.05 a plate count of about 12,000 is all that would be required and this represents a very easy separation. Thus large separation factors are never needed for gas chromatography and the range of separation factors required for any peak pair in the chromatogram is in the range 1.005 to 1.05. So long as mixtures are not too challenging, success seems almost guaranteed because of the high kinetic performance of the columns in gas chromatography. Or put another way, you don’t have to be very lucky to take any column from the shelf and find that it can separate moderately complex mixtures by identifying a suitable temperature. If success and expertise are seen as related, then even the newest member of staff is suddenly an expert.

The how of separations, to me a true measure of expertise, requires an intimate knowledge of stationary phase chemistry and an understanding of how compounds interact with and become distinguished by the stationary phase. Over quite a lengthy career in gas chromatography, I have participated in the development and abandonment of many one-time exciting and ultimately disappointing approaches to describing and modeling selectivity in chromatography. The selectivity of a stationary phase can be defined as its relative capacity to enter into specific intermolecular interactions represented by dispersion, induction, orientation, and hydrogen bonding. The transfer of a compound in the gas to the stationary phase requires that initially a cavity is formed in the stationary phase of the same size as the solute; this is accompanied by reorganization of the stationary phase molecules to minimize the disruption of cavity formation and to establish a more favorable orientation for intermolecular interactions. Lastly the solute is inserted in the cavity and establishes solute-stationary phase interactions that vary in magnitude depending on the complementary capabilities of the solute and stationary phase to interact with each other. Cavity formation is the penalty extorted to gain entry to the stationary phase and only depends on the properties of the stationary phase. Polar stationary phases have stronger stationary phase internal interactions and cavity formation is more costly in free energy currency. Thus, the spacing between adjacent members of a homologous series is smaller on polar phases than on low polarity phases because of the higher cost of cavity formation extorted for an increase in cavity size. To make the payment for cavity formation, the solute-stationary phase interactions set up when the solute fills the cavity need to exceed the cost of cavity formation. The complementary nature of these interactions is where selectivity comes in. All solutes will establish dispersion interactions that scale approximately with size, but interactions of a dipole-type (induction and orientation) and hydrogen bonding depend on the dipolarity, polarizability, and proton donor/acceptor properties of the solute and stationary phase. The different combinations of these interactions are why stationary phase chemistry is important in separations and why different columns are required for optimized separations of mixtures of varied compounds.

The roles of temperature in optimizing separations in gas chromatography are, first, that a minimum temperature is required to create a gas phase mixture (achieve an adequate vapor pressure to have a presence in the gas phase) and, second, that cavity formation and solute-stationary phase interactions are temperature dependent. For the latter, there are two temperature contributions: one that affects only the stationary phase (cavity formation) and the other that involves both the stationary phase and the solute (the intermolecular interactions). An important recent observation is that for moderately polar stationary phases that are stable to high temperatures (in this case, 300 °C) polar interactions endure, although they are weaker than at lower temperatures. In addition, selectivity differences between stationary phases also endure, but are generally not as great as at lower temperatures. This means that selectivity endures to high temperatures and that stationary phase selection remains important. Thus, the myth that all columns behave the same at high enough temperatures is not true for the range of typical temperatures used in gas chromatography (at least up to 300 °C). It is also possible to demonstrate that individual differences in selectivity between stationary phases are not preserved as temperature is varied, and therefore, single temperature scales of selectivity are of limited use for column selection.

There are now quite large databases for open-tubular columns that include quantitative information on selectivity and its variation with temperature. For polysiloxane stationary phases it is clear that what might be referred to as the selectivity space is not uniformly populated. There are no stationary phases that are hydrogen-bond acids and so this interaction is not exploited, although many compounds are hydrogen-bond bases. The reliance on a limited number of monomers for the synthesis of stationary phases results in islands in the selectivity space surrounded by empty space inaccessible with current stationary phases. In reality, therefore, there are plenty of opportunities for stationary phase development, and we should not become complacent with the status quo, or try to solve all problems as the new staff member alluded to earlier, by overexploitation of column efficiency.