Anyone who has ever worked with me has heard me say this, “When troubleshooting an instrument, any instrument, the problem will lie in one of 3 places, sample introduction, chemistry, electronics.” In gas chromatography, the clues lie in the chromatogram. I remember starting out as a technician years ago and being so impressed with senior techs who could look at a chromatogram and tell what sample it was based off the peaks. When I started, Chemstation was not as prevalent, some of our GC’s still had strip chart recorders, and you may not be aware, but there was a time that you would have to physically cut the peak out and weigh the paper. I never had to do this, but I suppose you weighed each peak and then determined the percentage of each, thankfully that is no longer the case, eventually I too memorized the peaks. I don’t run samples anymore so when I am looking at a chromatogram it is to troubleshoot.
When I say electronics, I am primarily thinking of detectors. There are other electronics to consider, such as an electronic pressure controller, remote start cables, communication cables, etc. But the detector signal is what I am looking at, every instrument has a detector. Have you ever had an injection where some peaks come out, and then the signal drops? Or maybe the baseline has lots of noise in it, these are the things I am looking at. To me problems with electronics are not very likely. Most manufacturers have stringent quality control protocols, they must, they can’t make any money if there are defects, people will stop buying their products. There are some bad batches sometimes, but this is rare.
Chemistry in a GC may not be obvious, but there is chemistry going on in the GC column. We leverage chemical properties of the different analytes to separate them from each other, be it boiling point or polarity. When looking at a chromatogram, look at peak resolution, the distance between peaks. This tells you if the chemistry on the column has gone bad, many times there is contamination on the column that can be baked off. The great thing about GC is that you can literally see the problem in the chromatogram, you can even go back and see when the problem started. Titration is a more obvious chemistry problem, sometimes the titration solvent goes bad, so there is no color change, or as you get to the bottom of the bottle, it becomes more or less concentrated, resulting in erroneous results.
Sample Introduction is always my number one suspect, I would say about 90% of analytical instrumentation problems lie here. This presents as some change in signal. Consider that we are always looking for the concentration of an analyte, and most of the time the concentration is based on the amount injected. If you inject less sample, your signal goes down. In a previous article, “Why the Sample Point Matters”, I mentioned that the sample analysis begins at the sample point. In “Why Sample Quality Matters”, I mentioned how contamination will affect the results. Next in line is sample introduction.
I once worked somewhere where someone wanted to test my troubleshooting skills, (this is the nicest way to say this). After a vacation I came back to several instruments that were having problems.
One of the instruments had a backflush for removing the heavies to prevent contaminating the column, for some reason, the back flush was not working. Upon investigation of this new instrument, I opened the valve cover and discovered that the airline to the valve actuator had been cut, preventing the backflush from happening, problem, sample introduction.
The next instrument was a super critical fluid analyzer, it presented with lower broad peaks. As it turns out bolts on the pump had been loosened, a pump I had just worked on before vacation, again the problem was sample introduction.
The last problem was more challenging. This was a DHA type of analysis, and the problem presented with the peaks being bigger than normal. I made sure the syringe size was correct, I checked the method to be sure that the injection volume was correct, but no explanation as to why my peaks were bigger. The problem with bigger peaks is that you lose resolution, in GC, less is more, so we are always looking for that sweet spot that gives us the best resolution as well as the most sensitivity. So puzzled by this problem, I was discussing it with my manager, and as I was explaining the problem, I figured it out. Remember my favorite quote “A problem well stated, is a problem half solved.”? So, the problem was the peaks were bigger, which can only happen if more sample is injected. I checked for increased injection amount and found nothing in the method or the syringe. But this was a 5890, and the technology is less advanced. There is one more piece of the sample introduction system that I had not considered, the split ratio! We may inject 1 uL, but we do not necessarily analyze 1 uL. It is very common to have a 100:1 split ratio, meaning 1/100th of a microliter actually gets injected, less sample is more resolution. On the old 5890’s, there was no EPC, electronic pressure controller, so you must turn a knob manually and measure the flow from the split vent with a flow meter. Sure enough, the split vent flow had been changed! I changed it back and we were back in business.
My examples all revolve around GC, but I have used this strategy on any instrument, if you are having problems with your analysis, split the analysis up into each of its components, sample introduction, chemistry and electronics, then check your sample introduction.
Just in case you have someone who is trying to help you, like they helped me (for some reason there is always someone wanting to help me), remember, the obstacle is the way, these challenges really boosted my confidence, and like the AJR song “100 Bad Days” says:
“A hundred bad days made a hundred good stories.
A hundred good stories made me interesting at parties.
No, I ain’t scared of you, no, I ain’t scared of you no more!”