GC-FID problems

[vincentyin]

We are trying to set up a GC/FID but noticed a number of problems (drifting retention time, ghost peaks, FID areas were not reproducible). These chromatograms were made yesterday after cleaning the injection port, liner and insert and conditioning at 275C overnight. Also we added glass wool to the liner. The chromatograms, from top downward, are n-Octane, Ethylbenzene and o-Xylene, each sample has been injected for two runs to verify its repeatability, therefore the top two are n-Octane, middle two are Ethylbenzene and the lower two are o-Xylene. They are 1 microL injections, without solvent.

The machine is a HP 5890 series II GC/FID with HP autosampler 7673A, running with a Petrocol DH Capillary Column by Supelco with 100m*0.25mm*0.5micron film thickness. The method being used is injection port 220C, FID detector 220C, temperature ramped at 5C/min from 50C to 250C then hold for 10 minutes, with programmed syringe twice pre-washing, twice post-washing and 3 times pump. Carrier gas is Helium, FID is using Air and H2.

We measured the split vent and purge vent flow rate to be 104ml/min and 3ml/min, so we are assuming the split ratio to be 100:1 because of the description in the manual, or what other ways are there for us to judge the split ratio? This machine goes with a manual gas controller.

So the major problem is what you can see in the chromatograms, no effective signals(first one), drifting the retention time(3 and 4), suspicious remnant peak from the previous run(5 and 6),and unsteady peak area between runs.
Please help us, any comments and suggestions will be greatly appreciated!

[Peter Apps]

Assuming that the 100:1 split is more or less correct then you are putting aboout 6 ug of compound onto the column and through the detector. This is about 100 times as much as you need.

The first step is to inject some methane or the gas from a cigarette lighter and measure how long it takes to elute. This will tell you the average linerar gas volcity through the column. It needs to be between 35 and 50 cm/s.

Then try injecting 1 ul of a 1% solution of octane in hexane with a 100:1 split. Do not worry about the huge hexane peak, look for the octane peak and see how it looks for shape and area repeatbaility.

Peter

[larkl]

Did you use silanized glass wool? Not doing so could create some funky peaks.

[vincentyin]

Did you use silanized glass wool? Not doing so could create some funky peaks.

No, it was just normal glass wool without silanization, would it be worse than without it? I just simply compared the chromatagrams after and before putting the glass wool, looks no obvious changes.
Thanks.

[vincentyin]

Assuming that the 100:1 split is more or less correct then you are putting aboout 6 ug of compound onto the column and through the detector. This is about 100 times as much as you need.

Why would you say 6 ug?

The first step is to inject some methane or the gas from a cigarette lighter and measure how long it takes to elute. This will tell you the average linerar gas volcity through the column. It needs to be between 35 and 50 cm/s.

How to achieve this? by the retention time?

Then try injecting 1 ul of a 1% solution of octane in hexane with a 100:1 split. Do not worry about the huge hexane peak, look for the octane peak and see how it looks for shape and area repeatbaility.

Then for the following samples, so I need to dilute each of them using the same manner?

Thanks,
Vincent

[DSP007]

1) Probable cause you problem - the carrier gas where a siphon, or "gushes like water from a tap, ".
HP5890 machine is old, the small leak is not responding.
2) Possible leaks- hole in septa, destroyed column, the valve is split.
However, it seems you understand inzhekor to change the liner - because this too can be a problem.
Recommendation -This procedure is spelled out in the service manuals, the page can not remember.
Switch off the injector nut with a blank (no holes ferrulа or graphite ball
), close the valve and relief valve blowing gaskets, give the maximum pressurein the the injector and turn off the supply valve . Check how long the appliance keeps the pressure in the injector.Pressure must be kept for at least 15 minutes.

If the pressure is not kept - here it is cause.
Next, disassemble the injector, put the new rings, the bottom strip up to 250 C is better not to gold and Teflon. And collect the injector again.
As we say. "If being tortured for a long time - something to be "

Respected Peter Apps recommended you the correct procedure for checking the device. When correct the problem, spend the two procedures.
However, for Petrokol as a unilateral matter, I recommend not use lighters gas , but use formalin, acetaldehyde, or methanol or acetone (as very polar liquid on nonpolar column) at 250 C.

[Peter Apps]

Assuming that the 100:1 split is more or less correct then you are putting aboout 6 ug of compound onto the column and through the detector. This is about 100 times as much as you need.

Why would you say 6 ug?
density of hydrocarbons is about 0.6 g/ml, 1 ul is then 600 ug, split by 1/100 is 6 ug

The first step is to inject some methane or the gas from a cigarette lighter and measure how long it takes to elute. This will tell you the average linerar gas volcity through the column. It needs to be between 35 and 50 cm/s.

How to achieve this? by the retention time?
Divide the length of the column by the retention time - that gives velocity.

Then try injecting 1 ul of a 1% solution of octane in hexane with a 100:1 split. Do not worry about the huge hexane peak, look for the octane peak and see how it looks for shape and area repeatbaility.

Then for the following samples, so I need to dilute each of them using the same manner?
You have a lot of trouble shooting to do before you worry about diluting samples

Thanks,
Vincent

Formalin and acetaldehyde give very poor signals by FID

Peter

[DSP007]

[quote="Peter Apps]
Formalin and acetaldehyde give very poor signals by FID

Peter[/quote]
Yes of course. But there a very poor sorbtion in dimethylpolisiloxane column. Its unsorbtion subcstancies for "unilateral peak". Also may use ethanol, but need most high temperature - 300 C.