Page 104 - Multidimensional Chromatography

96 Multidimensional Chromatography

be transferred from the mobile to the stationary phase. This then alters the distribu-
tion constant and also the retention factor (i.e. retention time) of the compound. If
two compounds have the same chemical potential on one column, then in order to
separate them in a multidimensional experiment we have to alter their chemical
potentials. This can be done by using a different temperature on the two dimensions,
although a more useful approach is readily achieved by choosing different stationary
phase types. Figure 4.11(b) shows that the first dimension separates by the property
of volatility, and the second by polarity. The response axis indicates the detector
response to the solute. Temperature variation on the two columns is a less practical
solution, since we require two independently controllable temperature regions–such
as a two-oven system. Most studies of multidimensional gas chromatography
employ different column phases, and as a typical example we can consider a hypo-
thetical experiment of a processed petroleum (e.g. kerosene) sample separation.
To a first approximation, the analysis of the petroleum sample on one stationary
phase column versus another column will to a large extent appear very similar. The
chromatogram will be dominated by the saturated alkanes, with the normal hydro-
carbon suite providing the recognizable dome shape. Between the major components
will be a range of branched and aromatic compounds. We cannot distinguish these
minor components due to the large number of overlapping components, and one
column is unlikely to be very much better than another. Admittedly, with mass spec-
tral detection, we might be able to say that one column gives a better result than the
other, but with flame-ionization detection the complexity is overwhelming. Different
phases will shift the peak relative retentions, but for the kerosene trace components
this is a little like changing one scrambled chromatogram for another. Each column
separates or retains (primarily) by the component boiling point, and then imposes a
selectivity or relative peak position adjustment based on what we might call polarity,
but is better referred to as specific solute–stationary phase interactions.
With comprehensive GC, we can now choose a rational set of columns that should
be able to ‘tune’ the separation. If we accept that each column has an approximate
isovolatility property at the time when solutes are transferred from one column to the
other, then separation on the second column will largely arise due to the selective
phase interactions. We need only then select a second column that is able to resolve
the compound classes of interest, such as a phase that separates aromatic from
aliphatic compounds. If it can also separate normal and isoalkanes from cyclic alka-
nes, then we should be able to achieve second-dimension resolution of all major
classes of compounds in petroleum samples. A useful column set is a low polarity
5 % phenyl polysiloxane first column, coupled to a higher phenyl-substituted
polysiloxane, such as a 50 % phenyl-type phase. The latter column has the ability to
selectively retain aromatic components.
The concept of tuning a separation was succinctly summarized by Venkatramani
et al. who stated (32):
A properly tuned comprehensive 2-dimensional gas chromatograph distributes
substances in the first dimension according to the strength of their dispersive interac-
tions . . . and in the second dimension according to their specific non-dispersive