The LCGC Blog: Column Overload in Gas Chromatography with Vacuum Ultraviolet Detection

Photo Credit: Patrick Strattner/Getty Images

Kevin A. Schug, University of Texas Arlington, Texas, USA

Column overload is a very commonly encountered issue in gas chromatography (GC) for beginners. Changes in peak symmetry, generally observed as peak fronting, can be subtle in the sharp peaks generated by GC, but the result can be significant shifts in retention times, loss of resolution, and error in peak integration. LCGC Blogger Kevin Schug explains more.

Split injectors were invented to ensure that wall-coated open-tubular capillary gas chromatography (GC) columns are not overloaded. Because it is not practical to reduce actual injection volumes much lower than tenths of microlitres and the capacity of thin-film stationary phases coated on the capillary wall surface should not be exceeded, the ratio of gas flows in the injection port directed through the column versus to waste (through the split vent) is adjusted to set an appropriate split ratio. Split ratios are generally reliable between 10:1 and 400:1 (where the majority of analyte is split to waste) to ensure that the target analyte of interest exhibits good peak symmetry. In fact, column overload is a very commonly encountered issue in GC for beginners. Changes in peak symmetry, generally observed as peak fronting, can be subtle in the sharp peaks generated by GC, but the result can be significant shifts in retention times, loss of resolution, and error in peak integration. Traditionally, it is preached that column overloading conditions should be avoided for those reasons; however, in cases where the greatest sensitivity is desired to perform ultratrace analysis, a splitless injection, where all of the analyte is transferred onto the column (the split vent is closed for a minute or two during the injection phase of the analysis), can be performed. One can imagine that the choice between the use of split or splitless injection would be driven by monitoring peak area and peak shape, and then choosing the conditions that give a nice symmetric peak and significant signal for the target analyte over the desired concentration ranges for analysis. In our experience with the new vacuum ultraviolet (VUV) absorption detector for GC, these considerations may not be so critical.

A VUV detector was introduced into the market in 2014.¹ I have written previously about the concept and the general operational advantages it provides.^2,3 Since then, we have applied GC–VUV for the investigation of permanent gases,⁴ pesticides,⁵ and fatty acids.⁶ We, as well as other groups around the world, are currently looking at several other application areas. The majority of VUV detectors to date primarily have been sold in the petroleum and petrochemical industries.

In short, the principle of VUV spectroscopic detection is full-range absorption measurements between 120 nm and 240 nm, where all chemical entities absorb and have unique gas–phase absorption signatures. VUV detection is quite complementary to mass spectrometry (MS), because it can well discern isomeric (isobaric) species that are often indistinguishable based on electron-ionization MS spectra. Further, standard concepts from Beer’s law apply - the magnitude of absorption is directly proportional to the amount of analyte present and its absorptivity (cross-section), and absorption for overlapping signals is additive. This latter point means that unresolved components can be easily deconvolved if the reference spectra for each are present in the VUV spectral library. A ton of effort need not be placed in fully resolving all peaks of interest. The strength of this capability is impressive. For example, we are currently drafting a manuscript on our effort to fully speciate Arachlor (previously manufactured by Monsanto) samples, which are complex mixtures of polychlorinated biphenyl compounds (PCBs). Each of the 209 PCBs has a unique spectrum - they can be well differentiated from one another, even if they chromatographically overlap.

Returning to the question at hand, the deleterious effects of GC column overload should be quite well handled for GC–VUV analyses. Significant peak fronting can compromise resolution; it can cause overlap of neighbouring peaks that would be very difficult to deconvolve using any other GC detector. Yet, we have shown in most of our applications that two or three (and probably more are possible) distinct analytes can be completely resolved from coeluted peaks into their respective contributions to the overall signal. Assigning the retention time to the peak apex will cause the retention time to shift more and more as overloading is increased. This is not really a major problem, except for the context of analyzing a number of samples with a wide range of analyte concentrations. In such cases, where some peaks might front (high analyte concentration) and some might be symmetrical (low analyte concentration), it would be necessary to understand that there is potential for the retention time to shift.

At some point, when larger and larger amounts of analyte are eluted through the column into the VUV flow cell, the absorption signal might saturate the detector, especially in regions of the absorption spectrum where absorptivity of the molecule is very high. However, because quantitative analysis is typically performed through averaging the signal across a wide wavelength range (in contrast to typical UV quantitation, where quantitation is often performed at a predefined peak maximum), when one spectral feature goes off scale (that is, it becomes saturated), the quantitation performed by the VUV detector can be defined based on absorption across less–absorbing wavelengths in the spectrum. Because we know the shape of the absorption spectrum for an analyte across the full range of 120–240 nm, and because the ratios of intensities for electronic transitions across that range remain constant, an off-scale response in one region of the absorption spectrum does not matter. It is a simple matter to model the shape of the response of that off-scale region based on the magnitude of response in other regions of the spectrum. This treatment could effectively increase the dynamic range of the detector.

These concepts have not been shown empirically using GC–VUV; these are simply my thoughts on the subject. However, there is nothing to prevent methods from being conceptualized where column overload is used to achieve lower detection limits in GC–VUV analysis. Currently, VUV detection is about as sensitive as MS in a scan mode (50–200 pg on-column LODs, depending on the analyte chromophore). In truth, the main limitation might actually be the volume of the injection port liner and the thermal expansion of the various sample solvents used. In other words, you can only increase injection volume so high (unless you become quite creative and experienced in performing methodically slow splitless injections) before the capacity of the liner will be reached and vaporized sample solutions will overflow into other parts of the injection port - an undesirable situation. Of course, there are large–volume injectors commercially available. Some more obscure standard methods (I can think of one for disinfection by-products by GC with negative chemical ionization MS) require this type of hardware for ultratrace detection limits. Further, the dynamic range of the response for the method can be extremely large, considering that the strategy above can be used when detector saturation is reached.

So, I need to get my students to try this approach (or perhaps one of you readers with access to a VUV system could try it): Prepare a series of solutions ranging from absurdly dilute analyte concentration up to some concentrations a few orders of magnitude higher. Use large volume splitless injection to see how low you can go and how wide the dynamic range of response could be. I purposely use “dynamic range” here, because nothing says that the response has to be linear all the way up to high concentrations. As long as the response changes in a predictable, even if nonlinear, fashion across the concentration range, then a calibration curve could be constructed. The progression of peak shape and size should go from small and symmetrical up to very large and fronting. At some point, the detector will saturate in part of the wavelength range, but that problem can be mitigated with some modelling of the response using intensities in less–responsive regions of the spectrum. There is some precedent for this technique in the literature using lower abundance isotope species in MS,^7,8 but it is far from commonplace. Dare I say, it would be fun to try it and see what performance would be like.

References

1. K.A. Schug, I. Sawicki, D.D. Carlton Jr., H. Fan, H.M. McNair, J.P. Nimmo, P. Kroll, J. Smuts, P. Walsh, and D. Harrison, Anal. Chem.86, 8329–8335 (2014).
2. K.A. Schug, The LCGC Blog 11 September 2014. http://www.chromatographyonline.com/lcgc/Blog/The-LCGC-Blog-My-New-Obsession-Gas-Chromatography-/ArticleStandard/Article/detail/853093?contextCategoryId=50130
3. K.A. Schug and H.M. McNair, LCGC North Am. 33(1), 24–33 (2015). http://www.chromatographyonline.com/gc-detectors-thermal-conductivity-vacuum-ultraviolet-absorption
4. L. Bai, J. Smuts, P. Walsh, H. Fan, Z.L. Hildenbrand, D. Wong, D. Wetz, and K.A. Schug, J. Chromatogr. A1388, 244–250 (2015).
5. H. Fan, J. Smuts, P. Walsh, and K.A. Schug, J. Chromatogr. A1389, 120–127 (2015).
6. H. Fan, J. Smuts, L. Bai, P. Walsh, D.W. Armstrong, and K.A. Schug, Food Chem.194, 265–271 (2016).
7. H. Liu, L. Lam, L. Yan, B. Chi, and P.K. Dasgupta, Anal. Chim. Acta850, 65–70 (2014).
8. H. Liu, L. Lam, and P.K. Dasgupta, Talanta87, 307–310 (2011).

Kevin A. Schug is a Full Professor and Shimadzu Distinguished Professor of Analytical Chemistry in the Department of Chemistry & Biochemistry at The University of Texas (UT) at Arlington. He joined the faculty at UT Arlington in 2005 after completing a Ph.D. in Chemistry at Virginia Tech under the direction of Prof. Harold M. McNair and a post-doctoral fellowship at the University of Vienna under Prof. Wolfgang Lindner. Research in the Schug group spans fundamental and applied areas of separation science and mass spectrometry. Schug was named the LCGCEmerging Leader in Chromatography in 2009 and the 2012 American Chemical Society Division of Analytical Chemistry Young Investigator in Separation Science. He is a fellow of both the U.T. Arlington and U.T. System-Wide Academies of Distinguished Teachers.

E-mail: kschug@uta.edu

Website: www.chromatographyonline.com