Troubleshooting Real GC Problems

There are only a few topics in gas chromatography (GC) that are responsible for a large number of problems posed to the "Ask the Expert" function in ChromAcademy (www.ChromAcademy.com). This month's instalment presents a summary of some of these topics relating to poor peak shape and rising baselines, to act as a guide when setting up or troubleshooting a method.

Problems with split and shouldered peaks are typically caused by a disruption to the way the sample band is introduced into the GC column. Check that the column is correctly installed into the instrument — usually the depth of insertion into the inlet is critical in this respect. The column cut is absolutely critical in determining peak shape, and one should ensure that the cut is at 90° to the column wall and that it is clean (not jagged or rough). The exposure of high numbers of silanol groups and the formation of turbulent eddies at the head of a roughly cut column can cause major peak shape issues. Always inspect the column cut with a magnifier or lower power microscope — we cannot overemphasize the importance of good quality column cuts at both the inlet and detector ends of the column. The homogeneity of the analyte band can also be disturbed by problems with the internal column surface at the head of the GC column. If stationary phase has been stripped, exposing silanol groups, or if nonvolatile sample matrix has been deposited on the surface, the analyte band will interact differently with these areas than with the bonded phase, causing peak splitting or shouldering. Typically, these issues can be solved by trimming a few centimetres from the head of the column. Occasionally it may be necessary to trim the column by up to 10% of the total column length to solve the problem; however, note that peak retention times will decrease and peak identification windows may need to be altered in your data system.

Further peak shape issues are more specific to splitless injection modes, but again all relate to the homogeneity (contiguous nature) of the sample band as it enters the GC column. One should ensure, especially in splitless injection, that the polarity of the sample diluent solvent matches that of the stationary-phase chemistry. Further, the initial oven temperature should be at least 10 °C (preferably 20 °C) below the boiling point of the sample solvent, which will act to condense the analyte as it slowly evolves from the inlet and focus each analyte band to give sharp peaks within the chromatogram. These factors combined, ensure that the sample vapours condense as contiguous bands within the GC column before they revolatilize as the oven temperature is raised.

We are often asked about the causes of rising baselines within GC separations — and these typically fall into three categories. If operating a temperature programmed separation, with constant carrier gas head pressure, the flow rate (and linear velocity) of the carrier gas will decrease because gas viscosity increases as a function of temperature. If one is using a mass- or flow-sensitive detector (a flame ionization detector, for example), which responds not only to the amount of analyte but also the rate at which analyte passes through the detector, then the baseline position will naturally rise. The solution to this issue is to operate in a constant flow mode in which the instrument increases the carrier gas head pressure to maintain a constant flow (or linear velocity with some instruments) during the whole of the temperature programme. Note that in switching operating modes, retention times, especially of later eluted compounds, will change.

Baseline rise can also be caused by an increase in column bleed with temperature. Ensure that columns are properly conditioned before use, which will involve a short time at room temperature with carrier gas flowing (this step is very important), followed by no more than 30 min at 10 °C higher than the upper operating temperature of the analytical method. Remember that more-polar and thicker-film GC columns will show greater bleed and to set a bleed specification beyond which the column will not be used.

The third common cause of rising baselines is an improperly optimized splitless injection. Although the initial phase of a splitless injection should be carried out with the split valve closed, the split should then be initiated to remove excess solvent and sample vapours from the inlet. This "splitless" or "purge" time needs to be carefully optimized; too short and sample will be lost resulting in poor quantitation, too long and a large tailing solvent peak with rising baseline will result. Typically, the purge time is optimized by choosing a time value (usually in seconds) at which repeated injection of the sample gives reproducible analyte peak areas, but which results in the narrowest solvent peak width.

Finally, we are very often asked about peaks that tail badly in capillary GC, that is, peaks whose asymmetry or United States Pharmacopeia (USP) tailing factor is greater than one. Most often peak tailing occurs because a certain proportion of the analyte molecules are being subjected to a secondary mechanism of retention compared to the rest and this is usually some type of silanol interaction with analyte polar functional groups. The silanol groups are present on the surface of your quartz glass inlet sleeve or liner, glass wool used for liner packing, and in the silica from which the wall coated capillary columns are manufactured. To avoid peak tailing one should use only professionally deactivated inlet liners and glass wool packing, ensure good column cuts, trim the inlet end of the column to remove exposed silanol groups because of phase stripping, and, lastly, consider derivatizing analytes to "cap" or "mask" polar functional groups.