Fast parallel column multi-dimensional GC

Jiří Ševčík
Department of Analytical chemistry, Charles University, Albertov 2030, CZ-128 40 Prague 2

Development of separation systems toward a low ultimate uncertainty (ultimate uncertainty increases with increasing number of compounds in a sample and decreasing efficiency of a separation system) could be realized by a couple of different approaches. Either the number of compounds simultaneously presented in the sample should be limited (serial column multi-dimensional systems with a heard-cut) or the system efficiency should be very high. There has been developed different MDGC systems, among them comprehensive GC and recycle GC are most promising.

The other approach of ultimate uncertainty reduction is combinatorial system based on an additivity of probabilities of retention characteristics, evaluated in the logic IF-THEN-AND-AND-...-AND. This system is realized by means of parallel columns with a different stationary phases analysing simultaneously the same sample.

The system with three parallel capillary columns, conected to the front column spliter and micro TCD detectors has been investigated. An attention has been paid to the optimization of:

It has been found, that for 42 test substances, there has not been found selective compound disproportination until the split ratio upto 1:500, thus the sample injector was inactive (average coefficient of variation was 2,3%), there has not been sample amount discrimination between different columns, thus front column splitter has been properly designed.

It has been found, that all the columns combinations (from 5 to 15 meter lenght, 0,32 mm to 0,20 mm ID, five different stationary phases from apolar OV-101 to polar Carbowax 20M) led to reproducable retention characteristics (average coefficient of variation of retention indices was 0,8%).

It has been found, that micro TCDs with make-up gas could be operated under similiar sensitivities and they are mass-flow dependent (average coefficient of variation was 2,3%). Under this conditions peak area becomes additional input for identification algorithm.

It has been found that compounds identification based on a match of measured data with library data sets suffere from missing correction for actual stationary phase characterization. Some correction algorithms have been developed.

Fast parallel column multi-dimensional gas chromatography is the approach to achieve high information content, thus a low ultimate uncertainty, high information flow under low cost of information.