Deciding appropriate oil concnetrations to shoot in GC-FID

[chroma11]

I have neat oil. I prepared concentrations ranging from 1000 ppm down to 500 ppb of this oil in DCM solvent. I ran sample with 1 ppm on GC-FID but there are very weak signals

Can anyone tell me how can i decide appropriate concentrations for the injection? Any theoretical background would be helpful.

Thanks alot!

[Don_Hilton]

If everything in the oil will pass through a GC column, you had plenty of material in solution. With a splitelss injection, you should get a signal for individual compunds even at the lower part per billion levels for individual compounds. And a mixture is made up of many individual compounds.

But not only do you have to get enough material into the GC, but you have to get the materal through the column. And, the way that neat oil is made would lead me to expect some/many high boilers in the mixture.

A few other questions: What was your split ratio - or was this splitless. What inlet conditions were used - how hot was the inlet?
What column was used and carrier? Did you have conditions to give high enough temperature to elute the compunds present?

(And, pull the inlet liner and look to see if it has darkened from nonvolatile materials decomposing.)

[chroma11]

Thanks you for your detailed reply. Oil was light crude. I used splitless injection. Inlet temperature was 280 degree C. Column uses was Rtx-1, 20 m length, 100 micrometer ID and 0.4 film thickness. Carrier gas used was Helium. Flow rate was 0.5 ml/min. Temperature program was 35 oC for 5 min and then 35-307 oC @ 0.66 oC/min. I checked the inlet liner, it appears to be fine; no stain was there.

[AICMM]

chroma11,

At 1 ppm you are shooting about 1 ng/uL. Let's say that you have 500 constituents in your crude (conservatively) so each constituent is getting on column at about 2 pg's each. Then your temperature ramp is slow, I assume to improve resolution, so you will widen your peaks even further. The amount you inject and the way you ramp are at the extreme limits of the FID. If speciation is not so critical, you could try a much faster ramp rate but I doubt you will be able to get to 1 ppm even so.

A technical aside. Make sure you are using make-up and preferably nitrogen make-up with this column configuration.

Best regards.

[Bruce Hamilton]

I suspect that the GC may not be at optimum separation conditions for helium carrier gas. The flow rate seems fast ( linear velocity faster than optimum for helium ), and the oven ramp rate a bit slow - 6.9 hour run time?.

That's quite a thick film for a narrow bore column - will permit higher initial temperatures ( no noisy sub-ambient cooling required ) , but peaks will be broader than for thin film.

My vague recollection is somebody ( Ettre? ), in early 1980s, showed optimum oven ramp rate for volatile HCs on non-polar fused silica capillary phase was about 3 - 4C/min with hydrogen carrier. Helium might be a bit lower but 2 - 3C/min would be my guess.

I'd also increase sample concentration, for reasons noted by others above.

[Don_Hilton]

A couple of things:

1) my comments about getting things to chromatograph was based on the term "neat oil." My mind wandered off to neatsfoot oil or oil from the neat tree... (This is what I get for logging on the computer before breakfast)

2) With a flow of 0.5 mL/min going through your column, you should be able to see peaks from your mixture, if they are there. While optimization should help, this is not the big sensitivity problem

As the others have suggested a faster ramp rate may help out. I would try more material on column and faster chromatograpy and then back off to separate the components that are of particular interest.

The next question is what is your injection volume? While you have some film thickness to help - it is more than easy to overload a 0.1 mm ID column. And, the volume of solvent is part of the loading. Is it a problem? I would suggest ramp rate, quantity of analytes, and detector optimization first.

I would suggest that you dilute the crude in carbon disulfide (which has no signal in the FID) at enough dilution that you do not have a viscosity problem in the syringe and then use a 5 microliter syringe - or smaller and inject as small a volume as you can with your autosampler. Perhaps 0.1 uL.

And use a very small volume liner. At the low flow rates, it takes a long time to get everything on the column (and too short a purge off time will really hurt!). A smaller liner with a smaller solvent volume - to avoid overfilling the liner with vapor takes less time to transfer to the column - and you avoid band broadening on the front end.

Using a 0.2 mm ID colum, I've injected small volumes (0.2 or so uL) of naphthas (neat) to profile the mixtures. I wanted solvent free injections because I was using a mass spec...