Comprehensive 2D Gas Chromatography Making GC Separations Work Harder

By Dr. Philip Marriott, Professor of Chemistry, RMIT University, Melbourne Australia

We are entering a period in its development where the expectations of comprehensive two-dimensional gas chromatography (GC×GC) should — justifiably — match the rhetoric. Since its inception about 15 years ago, researchers who have made it their (life) goal to develop and promote GC×GC have waxed lyrical about the advantages of GC×GC to the GC community. If we were to list the three primary contributions that are often ascribed to GC×GC, these would be: (i) greater separation capacity; (ii) greater sensitivity; and (iii) retention structure in the 2D data presentation that permits the analyst to identify, or predict the identity of, related compounds based on the molecular properties that control retention.

At this point, I should admit that I count myself guilty of being amongst those who have promulgated these advantages! Further, I also strongly support the position of GC×GC, and the benefits it holds for volatile and semi-volatile chemical analysis. And if these benefits are indeed general outcomes of GC×GC, then it is only logical that, sooner or later, this coupled column technique will supplant the single-column method that has served us so well for many years. But we might query whether single column GC has really served us so well. Admittedly, it has been just about all we had, so we have had to learn to live with its inherent limitations. Just as we might have recognised, and been frustrated by, the limited separation capacity of single column GC (i.e., as we searched for more complete understanding of the molecular composition of complex samples), analysts turned their attention to GC/MS which became routinely available. Considerable effort was devoted to implement solutions based on mass-detection to provide the necessary unique identification of individual compounds in (grossly) overlapping chromatograms. The mantra that MS can solve (all) our overlap problems probably became a crutch that somewhat numbed our realisation, according to my Research Group’s philosophy, that often “the only Solution is better Resolution.”

So, now that we have this new tool, what does it mean to the analyst? Well, in a simple answer — everything! With extra separation, the rationale for having to rely on MS for compound measurement (as opposed to identification) might now be negotiable. This is a considerable conceptual departure from the classical reliance on GC/MS. Extra sensitivity is a useful property to analysts, but this may be a lesser advantage of GC×GC. The ability to remove column bleed from solute elution does have benefits (when doing GC×GC/MS). The most significant advantage is separation power. To be able to resolve many more compounds immediately enables a much more complete ‘picture’ of the composition of a sample. Picture is used deliberately here, since the 2D GC presentation is very much akin to a picture. The comparison of 1D GC results is via a conventional GC trace — a one-dimensional time-response plot. The comparison is limited by the extent to which peaks coincide, or give multiple compound responses at one point. In GC×GC, the greater separation and picture-style GC plot means that we can simply compare two 2D pictures. Each compound now resides in its own 2D location which is determined by, or depends upon, the specific chemical-physical properties of a molecule which generate the peak position though specific interactions with the column stationary phases. The 2D plot has been called a chemical property retention map, which has axes controlled by retention mechanisms on each of the two columns. Choice of column phases is crucial to the effectiveness with which compounds are located within the available 2D space. Here, we will not consider how we generate the GC×GC experiment (i.e., the modulation methods used), however a few comments on the column selection are warranted in this text.

In GC×GC we usually couple a long 1D column directly to a short 2D column (or a regular elution column to a fast elution column). The second column has to work hard! We ask it to resolve peaks that are overlapping on the 1D column. Being about 1 m in length, with a need to complete continual, on-the-fly analyses of effluent from the 1D column within about 4-5 s, performance is everything. We use high carrier flow and narrow bore columns, but actual conditions are flexible. We now commonly find some regions of 1D GC analyses where up to 5 — 10 or more compounds co-elute. This is clearly beyond the scope of MS deconvolution. GC deconvolution through real separation is a more rugged and desirable outcome — and we can still combine GC×GC with MS for further identification. It is certainly true that GC×GC demands improved performance capabilities of GC instruments, new software, and better column quality control (e.g., improved batch-to-batch column reproducibility for the 1 m lengths of 2D columns that we use). These cannot be realised without the compliance of instrument and column manufacturers. As examples of generic GC×GC applications, a low polarity (5% phenyl) 1D column coupled to a short polar (wax phase) 2D column is useful for essential oils, but recently wax-low polarity column combinations have proved equally valuable. For petrochemicals, where higher temperature operation is needed, a low polarity (5% phenyl) 1D column coupled to a short polar (50% phenyl) 2D column is often used. For environmental analysis of PCBs, a carborane phase 2D column has been reported, where selectivity towards the extent of compound planarity is sought. In this short commentary, there is no space to engage in specifics of certain GC×GC methods, but obviously there is considerable opportunity to optimise methods, and use sound principles of GC and phase selection to get the best out of GC×GC.

We are at the threshold of a new era in GC, and getting the best out of our GC×GC methods is a task that an increasing number of analysts will be striving for. The comments of Professor Walt Jennings (Restek Advantage, 2006.01) also ring true for GC×GC. When a method has been developed for GC×GC, and one of the columns has to be replaced, to what extent will the 2D plot faithfully reproduce our archived or master analytical result? This must be addressed in the two-dimensional experiment, to prove that the analyst can have confidence that their data interpretation protocols survive column change and routine maintenance of the system. But with the impressive capabilities of GC×GC, it is important that analytical methods and the greater information content it offers are supported by validated and reliable operation.

Editor’s note: Dr. Marriott is one of the world's leading experts in 2D-GC.

 

An Example of 2D Gas Chromatography

GCxGC analysis of organochlorine pesticides combines primary column and confirmation column results.

Organochlorine pesticides separated from interferences in tomato extract.

Columns:	Rtx®-5 9m, 0.18mm ID, 0.20µm (10m column, cat.# 40201, with 1m removed) Rtx®-200 1m, 0.18mm ID, 0.20µl;m (1m of 10m column, cat.# 45001)
Inj.:	1µL, split, 250°C, split ratio 50:1
Oven:	Primary: 50°C (0.2 min.), 30°C/min. to 140° (no hold), 5°C/min. to 250°C (no hold) Secondary: 50°C offset from primary oven
Instrument:	LECO GCxGC/ECD
Modulator:	Temperature offset: 30° Modulation time: 6 sec
Det.:	ECD, 325°C, 150mL/min. nitrogen makeup gas, 50Hz