Baseline rise with temperature

[Chris Jones]

Well it seems to have calmed down a fair bit and the baseline noise is around 20 uV. Baseline rise over the course of a 100'C program is about 900 uV, which is acceptable. So I've started my sample runs. I'd say we've got gases which are right at the bottom of acceptable purity limits, hence the garbage we see on the baseline. Air Liquide will be put on notice

Ron, the baseline rise and spikes have nothing to do with the column. All of the pictures posted above were screenshot while the column was disconnected, so it's entirely FID noise. Anyway, while the signal might look fine to some, I have got far, far better chromatography out of the old 17A right next door.

Will let you all know how we get on.

Cheers,
CHRIS

[Chris Jones]

Hi Again everybody,
Well the mysterious baseline rise saga continues! We have installed a brand new column and set it up for a conditioning run. After filling the column with He we took the overn from 80 up to 200'C, whcih is 20 above the maximum in my program.

Sure enough, the baseline rose 40,000 uV before coming back again. Once it had settled down, we did an injection of ethylene gas (as a way of checking the column flow rate). Same deal, except it was only 10,000 uV rise. The third and forth programs (80 to 200 at 10'C per min, hold 20 min) resulted in familiar rises, but each being progressively lower than the last. At present we're sitting at an average of 4000 uV rise.

As previously posted, we have done chromatography with a baseline rise of 800-1000 uV but, it can be much better. I will see if we can bypass the traps, but they are deep inside the instrument somewhere.

The FID seems to be heating up enough to drive some crap off, resulting in the massive signal. So what sorts of compounds tend to do this? Heavy grease?? I have no idea where it could have come from.

Cheers,
CHRIS

PS I have a dilution series of sandalwood oil to run tomorrow. We'll see how it looks then.

[Peter Apps]

Hi Chris

Earlier you mention a bad batch of hydrogen, this might have contaminated all the plumbing to the FID.

Do you have other GCs in the same lab using the same gasses, did they also get the bad hydrogen, and do they show the same symptoms. ?

Is the hydrgen feed line accessible enough that you could heat it up independently of the FID tower and GC oven, say with a small hot air blower or a soldering iron (I'm assuming the lines are stainless steel) ?

If a brand new instrument (of course it isn't any longer) gives trouble it is the supplier's problem. Did it ever pass the installation checks ?

Peter

[Ron]

Are you using compressed air from a cylinder or from a compressor? The sine wave makes sense if you are on a compressor and the incoming pressure is fluctuating. If there are not adequate traps on a compressor it is very easy to contaminate lines and the detector. If you are on a compressor or a central supply system I would try running the instrument from a cylinder of zero grade air and see if that helps.

[Chris Jones]

Hi folks.

It looks like we have resolved some of the issues with an FID cleanout. We took the jet out and cleaned it as best we could and popped it back in. The baseline rise with temperature still occurs but not nearly as bad.

So here is the diagnosis: The GC17A right next door uses the same H2 and air cylinders and it gives perfect chromatography. There are no scrubbers on the H2 lines; they go from the cylinder straight to the detector with no filters. In the past, the 17A gave bad chromatography but we'd determined it was due to some very dirty H2. All I can think is the last of the contaminated H2 was burned up on the brand new instrument, coating the detector with crap.

Still, it proves to all chromatographers out there you need to go through a process of elimination to ascertain where the problem lies.

Now I have a new problem - the brand new column actually gives worse separation then the old one!! The peaks are very nice and not tailing, but where we used to get nice separation, we now have double peaks. Bugger.


Will post some chromatograms in a minute.

Cheers,
CHRIS

[Chris Jones]

Okay, a little admission here...

So, I did several bakes of the FID and arranged new gases. The signal vastly improved, but was still giving a 3000-5000 uV baseline rise. So I pulled the jet out and cleaned it - it was rather dirty and greasy.

That would do it. But I now realise why my chromatography was both ugly and the detector constantly getting filthy...

I started analysing a different species of sandalwood oil on the machine - Santalum spicatum has many more compounds in it than S. album, which is what I'd mostly analysed in the past. FOr some reason I'd get these large, scalene triangles of peaks appearing, and never at the same retention time.


...It was peaks that should have come out from the previous run, but never got a chance because the temp program only went up to 180 for 20 mins, then returned to 80. It simply wasn't enough to get it off the column before the next sample came on... bit of an oversight...

So there it is folks - for a new method, always run a hot program with a final bake-off to make sure all the peaks are off. Make sure the detector is running at least 250'C and bake it out regularly.

Always learning...

CHRIS