zick-zack baseline

[sb.monkey.sb]

and here comes another newbie with a hopefully not too stupid question:

today i installed a new column (HP-5) on our GC/FID 6890.

i checked with butane gas (1 uL) for the peak shape of an unretained peak and it looked wonderful (inlet 200C, split 1:40, 1.3 mL/min, oven 40C, detector 290C). the baseline was already slightly drifting down. inlet liner was replaced, as well as new ferrules. i also replaced the FID jet with a "cleaner" version. next i tried to condition the column at 300C.

the signal from the FID looks like a zick-zack line (or a distored saw pattern), which repeats itself about every 5-7 minutes.

I want to attach a picture, but don't know how to add it in right now - so I might post it right after...

this also happens at lower temperatures (50C for example)

therefore i thought maybe the "cleaner looking" jet was contaminated, so i replaced it with the old one: it seems that the up and down pattern of the FID is less now, but still dominantly present in the signal.

i observed a similar pattern almost 2 years ago on another 6890 with a Beta-Dex column - i was busy, couldn't fix it and the column was replaced with another one later... therefore I couldn't figure out what the reason for this strange looking FID signal was.

Any suggestions? My assumption is that something with the detector is not right, but I might be wrong.

[sb.monkey.sb]

[Peter Apps]

First guess is that it is a gas flow control problem, so you need to measure the actual flows of the FID hydrogen, air and nitrogen, using an independent flow meter, not the on board gas controls of the GC.

Are you using gas generators or cylinders ?

Peter

[sb.monkey.sb]

we use house air, N2 cylinder and H2 generator to feed the FID

how do I measure the individual gas flows on the FID?

but that brings me to another idea to test the gases: we have several other GC's attached to the same supply gases and they should therefore show a similar behavior...
i will test that tomorrow...

thanks

[Peter Apps]

To measure gas flows to the FID you turn them on one at a time, plug a gas flow meter (electronic or bubble) into the outlet tube at the top of the FID tower, and measure. If you have an electronc meter the claibration might not be correct for all three gasses, but what you are looking for is fluctuations over a time scale that matches the baseline fluctuations. Subtract the carrier gas flow if you want to be really accurate.

House air ? How stable is the pressure ?, and how are you cleaning it ? When was the last time the compressor tank was drained ?

Peter

[cjm]

I use house air as well, but run it through a zero air generator. When first installed I saw a similar pattern but on a much longer timescale and we found it perfectly timed with the compressor turning on and off. Problem was the air pressure fluctuated. I would recommend a regulator just before the instrument set to a pressure just below the minimum pressure coming from the compressor.

[aldehyde]

I ran into something like this recently, sinusoidal baseline due to compressor used for FID air.

I set the air to 200 mL/min and the severity of the fluctuation decreased, check and see if you can influence the amplitude similarly.

Set your carrier and makeup to 0 to make sure helium isn't influencing it.

[sb.monkey.sb]

alright, first of all thank you all for the helpful hints:

it turned out that our house air system was not running properly. nobody knew about it, until the mechanics pointed it out today

when i carefully watched the pressure on the air regulator, i could see it clearly going up and down between 7 and 8 bars. the cycle time was exactly constant to what i observed on the chromatogram.
i followed the advice and reduced the regulator slightly below and i'm trying to condition the column again: it seems better now, although there is still some up and down, now in the form of a sinusoidal baseline.
let's see what happens when the house air problem is fixed...

i also checked on another 6890 which is fed by the same gas sources: clean baseline. most likely because it has its own pneumatic controls of H2, air and N2 on the detector side, whereas the other 6890 has only on/off. am i right about this?

just for your info: we use house air which goes through a manifold of regulators, then into zero air generators followed by a high purity filter. we can also backup the air with cylinders that are meant to feed the GC's if the house vacuum breaks down completely.

thanks again for all the help and i will keep you updated...

[Peter Apps]

Thanks for the feedback.

6890 GCs have internal controls for their detector gasses, but they can only do so much when the input pressure is fluctuating. If you cannot rely on the house air it would be a good idea to have a supplementary pressure regulator, set about half a bar lower than the low fluctuation of the house air immediatley upstream of where the air supply enters the GC.

Peter

[aldehyde]

Good job identifying the problem, your air system sounds great. When you said the supply was broken I was worried, until you said you have good filters on the line. Make sure to keep an eye on indicator in case of water getting into the system--removing water from EPC is a giant pain and will most likely be very costly so being vigilant is important .

Keep up the good work.

[sb.monkey.sb]

alright,

it seems that everything is fine now. see attached a measurement of a coworker. baseline doesn't fluctuate much anymore and peaks are well separated.



the problem was definitely the pressure change in our house air system. since this 6890 has no additional EPC's on the detector (as far as I can tell, everything is manually set and one can only switch it on or off), the baseline started to go up and down with the fluctuations in pressure of the house air.

another 6890 which is setup at the same gas supply lines, did not show this abnormal behavior, since it has full EPC of N2, H2 and air, respectively. therefore, it smoothened the baseline through carefully providing a constant air pressure by regulating internally.

I tried the trick with the regulator set slightly below the minimal fluctuation of the house air: it seemed that the zick-zack behavior of the baseline was drastically reduced, but still ended up in a sinus waveform.

thanks again for all your help!

now it's time for me to find good conditions on a chiral Restek betaDEXcst column, to finally separate my enantiomers on another 5890