15 Mar 2023
GC Column Killers
Our technical department is often involved in helping clients with ‘GC column autopsies’ and over the years we have come to know the most common ways that you can effectively despatch a column.
However, like anything to do with health, there are cures and preventative measures which can help to revive an ailing column or avoid a premature demise. Everyone fears the signs and symptoms; unusually broad peaks, peak tailing, peak splitting, changes in selectivity and loss of resolution.
But fear not, by following some simple guidelines, we can ensure your GC columns live a long and healthy life, and even when worrying signs appear, there may well be something we can offer to extend column lifetime and healthier-looking GC results.
Cost isn’t always the primary driver for prolonging column lifetime. Relatively, the cost of a GC column per number of samples analysed has reduced dramatically, primarily due to the quality of the GC columns now available and the high-quality materials from which instruments and consumables are manufactured. However, the nuisance value, of having to change the GC column or to delay analytical results is more often the motivating factor, and often the reason for the worried call or e-mail that lands into the in-box of our technical support department.
So here, I offer some words of wisdom to avoid that panic situation and whilst some of this may seem a little bothersome or esoteric, I can guarantee that a little time spent with ‘housekeeping’ and ‘going the extra mile’ with your system and column preventative measures, will pay dividends in reducing your chromatographic angst.
Stationary phase contaminants can be both semi-volatile and non-volatile and can arise from a variety of sources including the sample matrix and anything which comes into contact with the sample such as vials, pipettes, glassware etc. It may also arise from the system itself, from released contamination from inlet septa, column ferrules or gas traps (filters).
Clearly the most effective way to reduce matrix contaminants is to undertake thorough sample preparation and the more selective the sample preparation technique, the more effective it will be at removing matrix contaminants. So, for example, mixed mode solid phase extraction which targets analyte extraction based on specific functional chemistry, will be much more selective at removing sample contaminants than a simple liquid / liquid extraction. However, even with highly targeted analyte extraction matrices such as biological fluids, soils, waste waters and foods may still contain a relatively high amount of non- and semi-volatile materials.
The symptoms of column contamination can be difficult to spot, as they can mimic other problems. Non-volatile contamination can cause broad peaks due to the adsorbed contaminants interfering with the partitioning of the analyte into and out of the stationary phase. Any active contaminants (polar and charged substances) can interact with polar analytes (acids, amines, alcohols, phenols, aldehydes and thiols) and cause peak tailing. Semi-volatile analytes will move slowly through the GC column, that is, they will elute over extended periods of time and eventually appear on your baseline or within a chromatogram as a peak which is perhaps broader that the analytes eluting before and after. All contamination can result in peak area reduction and baseline instability and wander.
The second most effective way to prevent column contamination is to use a guard column, a 1-3 meter length of deactivated silica placed in between the GC inlet and the analytical column and connected with a zero dead volume union. The guard column is designed to ‘trap’ the matrix contaminants and in this respect, it is a sacrificial consumable. Typically, the most pronounced contamination effects are seen when the head (or start) of the analytical column becomes fouled and therefore, on first signs of a degradation in chromatographic performance, the guard column can be swapped.
If the column performance degrades suddenly, this can more often be the result of contamination from the system, rather than catastrophic contamination from a single sample injection. Ensure that all vials, pipette tips and glassware used for sample preparation are of the highest quality affordable, new (i.e. don’t wash and re-use vials or pipette tips) and free from visual contamination. If ‘coring’ of septa is suspected, introducing shards of silicone materials into the GC inlet, check the injection needle to ensure it is free from burs and use septa with a protective PTFE backing to reduce the risk of coring and silicone breakthrough.
One needs to carefully monitor the condition of any gas traps connected to the GC carrier gas line. Self-indicating traps are most convenient as the visual colour change of the ‘tell-tale’ within the trap indicates when a change is necessary. Failure to change an exhausted trap risks a major deposition of contaminant into the system and this can be catastrophic over a very short space of time. Non-indicating traps should be on a regular replacement schedule, based on the type of samples you are analysing.
Avoid carrier gas bottles running out of gas. Even with newer dip tube design cylinders, the final volumes of gas within the cylinder are susceptible to pulling water, oxygen (air) and particulates into the GC system or saturating the gas line filters.
Be aware that splitless or low split flow analytical methods will result in the deposition of more sample matrix components into the GC column. Samples with complex matrices using splitless injection almost always require some form of sample pre-treatment in order to preserve column lifetime.
If column contamination is suspected, one can take several courses of action, including;
- Thermal bakeout at the isothermal maximum temperature of the column, for one to two hours. Any longer than this will render any non-volatile contaminants insoluble and will not be able to be rinsed from the column.
- Remove 0.5 – 1 meters of stationary phase from the inlet end of the column and re-install.
- Solvent rinse the column. This is often a measure of last resort and manufacturers will all have their recommendations for this operation. Rinsing is often the most effective method for removing non-volatile residues.
Of course, expected column lifetimes based on contamination issues very much depend on the nature of your samples and the analytical method, but following the guidelines above will help to give your GC column the best fighting chance for a long and healthy life.
Oxygen is the enemy of all GC stationary phases as it catalyses the ‘back-biting’ reaction which results in the evolution of cyclic siloxane species via oxidative degradation of the stationary phase – often know as ‘column bleed’. This reaction proceeds at increasing rates with;
- Increased column temperature
- Increased stationary phase polarity
- Increased stationary phase film thickness
Column bleed can lead to an increased baseline response, pronounced peak tailing of polar analytes and loss of column efficiency represented as peak broadening. Unfortunately, there are no good remedies for extensive oxygen damage of the stationary phase. Removal of the first 0.5 – 1 meters of the column may temporarily restore performance, but this will always be short lived and the original column performance may never be restored.
Instantaneous exposure to oxygen during column change, inlet septum change or injection of small amounts of air is not usually a problem, the issue arises when the stationary phase is constantly exposed to oxygen due to an exhausted carrier gas filter, poor quality carrier gas or a leak within the GC system.
Use the highest quality carrier gas available and regularly monitor the state of oxygen filters and changed at the earliest opportunity. The GC system (especially the GC inlet) should be regularly checked for leaks. Leaks can occur at the column inlet connection, the septum nut or fastening and the inlet body nut which is used to access to liner. Pressurising the GC system, setting the carrier gas (column) flow to zero, turning off the split-line flow and turning off the incoming carrier gas supply will cause the inlet pressure to decay over time and most instrument manufacturers will have a specification for this decay – i.e. the amount of pressure loss over time. A rapidly decaying gas pressure will indicate a leak within the carrier gas supply line or around the GC inlet connections and this should be addressed prior to any further analysis.
Over-used (cored) inlet septa or septa which split because they are under too high a torque due to an overtightened septum nut, can cause air to bleed into the system and regular septa changes are also highly recommended.
At lower temperatures, oxidative degradation of the stationary phase is much reduced but as soon as one begins to increase the oven temperature, the reaction rate increases significantly. Freshly installed columns may have a significant amount of dissolved oxygen within the stationary phase and one should avoid immediately increasing the oven temperature, to thermally condition the column, after installation. Column lifetime can be extended significantly by purging the dissolved oxygen with flowing carrier gas for 30 minutes at lower temperature (40oC is typical), prior to increasing the column oven temperature for conditioning of the stationary phase. If this process is followed, conditioning times over 1-2 hours are rarely required. For more detailed information on column conditioning – see the following reference;
Thermal damage (degradation) of the stationary phase results in all the same symptoms of those mentioned above for Oxygen damage. Fortunately, thermal damage occurs at a slower rate than oxygen damage as long as we can be sure there is no oxygen within the GC column. Usually, prolonged exposure of the GC column above the thermal limits is required to produce noticeable degradation.
All of this being said, thermal damage is usually the way in which most GC columns ‘die’ over a prolonged lifetime and we need to be mindful to reduce the column to exposure to elevated temperatures. Setting the oven maximum just a few degrees over the column gradient maximum temperature is a good idea to prevent accidental exposure to higher temperatures during thermal gradient GC analysis. Do not set the oven limit to exactly match the top temperature of your method as sometimes the GC oven will ‘overshoot’ the top temperature by a few degrees prior to settling back to the desired temperature. This may cause the GC system to enter a ‘fault’ state and interrupt your analysis.
If thermal damage is suspected, disconnect the column from the detector and heat at the isothermal temperature maximum overnight and trim 1—15 cm from the detector end of the column prior to re-installation. This may not restore the column to an ‘as-new’ state, but it may buy you some time to order a replacement column without disrupting your analytical productivity!
Strangely, this is perhaps the topic that I get asked about the most when teaching GC troubleshooting and maintenance courses. There are very few solvents which cause significant damage to GC stationary phases, even if GC column performance may appear degraded. Let’s cover a few of the more dangerous solvents and deal with the ever-popular topic of using water as an injection solvent in GC.
Mineral acids or bases are not good for stationary phases and can cause significant damage, primarily due to their low volatility which will see them accumulating at the head of the analytical column and doing their worst to degrade the stationary phase in this important region of the column which is largely responsible for determining the peak shape of the analytes, especially in splitless injection methods and gradient thermal programmed methods which start at a low oven temperature. The more water-soluble acids and bases can be carried through the system with any water in the samples and providing that the method is suitable for the rapid elution of water from the GC column (i.e. methods where the column is at least 115oC) then residence time will be low and column damage minimised.
One needs to take care if using samples which have been dissolved in typical HPLC eluents, especially those containing reagents which are popular for LC-MS, especially trifluoroacetic acid (TFA), pentafluoropropianoic acid (PFPA) and heptafluorobutyric acid (HFBA). These reagents are less volatile and tend to be deposited at the head of the analytical column where they can cause significant stationary phase degradation. This is especially true where splitless injection is used.
Chemical damage usually results in the exposure of active sites within the silicone backbone of the GC stationary phase and as such severe peak tailing or splitting may occur for polar analytes. One useful tip to discriminate between GC stationary phase damage and system issues with regards to peak tailing is to review the type and nature of the peaks which are tailing. If all peaks within the chromatogram tail, then this is much more likely to be a system issue. If only polar analytes show tailing behaviour, then is issues is more likely to be related to GC stationary phase degradation.
Chemical degradation of the stationary phase usually affects the first meter or so of the GC column and as such removing (trimming) the first meter or so of the column can often restore satisfactory performance. In more severe cases, up to 5 meters of column may need to be removed and this will only result in satisfactory performance if the GC column is 30 meters or longer. The use of a sacrificial guard column is recommended if aggressive chemicals such as those mentioned above are likely to be present within the injected sample.
The use of water and acetonitrile are often discussed, and many urban myths exist around the use of these solvents as sample diluents in gas chromatography.
Let’s start by saying that with care, each of these solvents is acceptable for use with chemically bonded GC columns, however, careful is very much the operative word here. Most modern GC columns are now chemically bonded to the silica which forms the stationary phase support, however with older GC columns where the stationary phase was merely immobilised onto the silica, then water could indeed dissolve and wash away the stationary phase, causing fairly catastrophic damage.
Transient exposure of chemically bonded GC stationary phases to water is acceptable, but this requires that column temperatures above 115oC are used to prevent the water from de-wetting and travelling through the column as liquid water rather than water vapour. If this occurs, then more the more water soluble (polar) analytes will display severe peak shape degradation.
The main issue with both water and acetonitrile is that they are ‘difficult’ sample solvents. Their boiling points are relatively high and therefore initial oven temperatures need to be high to ensure that solvent peaks are not tail badly. In splitless injection, these solvents are incompatible with non-polar stationary phases and therefore split speaks are highly likely, as the solvent ‘pools’ on the stationary phase surface during the initial ‘focussing’ of the analyte onto the stationary phase surface. Both solvents also have a high expansion coefficient and therefore only small amounts of them (<<1mL) can be introduced into the inlet to avoid backflash of the solvent vapours in the GC inlet which can result in quantitative irreproducibility and carryover.
Two special groups of GC columns with which water should not be used is the PLOT or porous layer open tubular columns and the highly polar Wax based columns which are based on a polyethylene glycol backbone. In the former group of columns, the water will be included into the pores of the molecular sieve materials used for their construction and will render the column unusable. In the latter, the water forms strong hydrogen bond associations with the stationary phase and will lead to major changes in the selectivity of the phase.
The final point to make here is the use of sample derivatising reagents such as BSTFA, MTBSTFA etc. used to reduce the polarity of analytes and make them more ‘GC’able’ by reducing peak tailing through the chemical exchange of analyte hydroxyl functional groups for trimethyl silyl groups. These reagents require a non-aqueous solvent and many recommend the use of acetonitrile or acetone as a good solvent. This is perhaps where acetonitrile get’s it’s bad reputation. However, the issue is not the solvent but the derivatising reagent, which when present in vast excess can damage the stationary phase of the GC column and even derivatise the phase, causing changes to the selectivity of the GC column.
Gc columns are coated with polyimide to given them thermal and mechanical stability, and these rarely spontaneously break within the GC column oven, with one notable exception. If the GC column is allowed to contact the walls of the GC oven, then the polyimide may ‘bake’ at the contact spot and become more brittle than usual. When the GC oven fan activates, the resulting vibration of the column may cause the column itself to break.
In this regard, always make sure that the column is secure on it’s hangar and that the column is hanging clear of the side walls of the oven shroud. Also make sure that the column leading from the inlet and to the detector is not touching the oven walls, is not longer than it needs to be and is not user stress.
Column Cutting & Positioning
Whilst column cutting and positioning doesn’t really fit the criteria of ‘column killer’ they can certainly contribute to chromatographic problems which resemble those described above.
If the column is not cut using the correct technique, and is cut at an angle or has rough shards or is jagged on the end, peak splitting and tailing can result. If the column is not properly positioned in the inlet or detector outlet, then peak area response can reduce and peaks will broaden. Therefore, instrument manufacturers guidelines must be closely followed to avoid these issues. For more information on correct column preparation and installation see;
As you can see, there are many ways to damage and reduce the performance of your GC column and many of the issues caused by column degradation are similar, often leading to confusion over the underlying cause and confounding our troubleshooting efforts. For this reason, it is imperative that we ensure proper preventative measures are in place. If issues do arise, then I hope that the above guide serves as a useful reference when troubleshooting and problem solving.