total petroleum hydrocarbons method linearity problem

[Widebore]

I need to validate a TPH method in our lab using GC-FID. Now, I'm normally the LC person and so I may have missed something. I got a method set up and the standard curves are good for C10-C22 (average response factor RSD ~2%), barely passable for C24and C26 (RSD = 17-20%) and terrible for C28-C40 (RSD = 30). Oh and it is the high end of the curve that is deviating from linearity but it can not be going ouside the FID linear range as earlier peaks are much larger.

I'm using splitless injection mode with the inlet set to 280 degrees celcius. I have an inlet purge (40 mL/min) set to start at 2 minutes before the temperature ramp starts. The oven temperature ramp starts at 40 degrees and ends at 320 degrees. GC flow at 1 mL. FID temp set to 340 degrees C (20 above the top of the oven temperature ramp). FID gases are set to 400 for air, 40 for hydrogen and 40 for makeup.

If anyone has any suggestions as to what the problem could be then that would be appreciated.

[chromatographer1]

It sounds like a classic case of discrimination from extremes in vapor pressure (boiling points) at injector temperature being used.

On column might give you a lot better results, especially if you have an injector that is PVT.

You can try increasing the INJ TEMP, but if there are cold spots in the injector that may not help, especially if you have larger than appropriate sample volumes injected.

best wishes,

Rod

[Widebore]

Thanks Rod

That's the conclusion I was leaning towards. I will try a higher inlet temperature (320 degrees) on Monday. Also, you say that if there are cold spots in the inlet then that could be a problem as well. Are these cold spots caused by poor manufacturing of the instrument or some other factor?

kind regards

[chromatographer1]

"Are these cold spots caused by poor manufacturing of the instrument or some other factor? "

Poor design manufactured precisely would better describe it, and limits of physical science. Remember you are evaporating a solvent which at the same time cools a part of the injector and then reheats. This evaporation is 'explosive' and happens quickly, it 'flashes', so material can leave the hot areas into places that are not as hot, despositing part of the matrix with low volatility.

Keep the entire sample inside the column, allow the solvent to evaporate at a controlled rate and in a controlled direction and you will measure more accurately.

Rod

[AICMM]

Widebore,

Are you using wool? I would for this analysis. Not a wisp of wool (like pesticides) but rather a decent bit of wool right in the middle of the liner.

Best regards,

AICMM

[Widebore]

[quote="AICMM"]Widebore,

Are you using wool? I would for this analysis. Not a wisp of wool (like pesticides) but rather a decent bit of wool right in the middle of the liner.

Best regards,

AICMM[/quote]

Hi AICMM

Yes, I'm using wool in the liner. I've done some further work. The best I could get with splitless was 1.5% RSD for C10-22, 10% C24-26 and between 22-26% for C28-40. So I would say that The lower mass compounds perform very well but the higher mass compounds still cause me problems with analytical performance.

I also tried split mode (2:1) with the inlet at 320 degrees Celsius and the RSDs were 6% for C10 - C30 and 8%-15% for C32-C40. So although the low end got slightly worse the rest of the compounds have gotten better. The only problem now is that there seems to be some fronting on the peaks but I may be able to live with that. If only I could get all of the RSDs to be less than 10% then I would be satisfied.

[Peter Apps]

Hi Widebore

A split of 2:1 is not really split at all - the gas flow through the inlet is still not high eneough to flush it quickly. Try a minimum split ratio of 10:1. All the peaks will be smaller of course, which may or may not be a problem depending on the LOQ that you need.

I suspect that the improvement that you saw going from splitless to 2:1 split was due to the increased inlet temperature. Law of troubleshooting: only change one thing at a time. Try splitless with the inlet at 320C.

Do you see any changes in area of any of the peaks under the different conditions - especially do you see the late peaks getting larger (or larger in proportion to the early ones) with a hotter inlet ?

You might also need to play with autosampler settings - what hardware do you have ?

Peter

[Peter Apps]

Hi Widebore

Another question: for how long is the inlet kept splitless ?

Peter

[AICMM]

Peter,

Earlier Widebore states 2 minutes splitless. On the long side but I don't think it should adversely affect the heavies in the manner described...??

Widebore, another thing I would suggest looking at. Do you have a manual thermometer (especially one of those pencil types) that you can install in the inlet with the liner removed? Maybe your heater is out of position and your are not heating the injector block to the temperature that you think you are.

Best regards,

AICMM

[Peter Apps]

Peter,

Earlier Widebore states 2 minutes splitless. On the long side but I don't think it should adversely affect the heavies in the manner described...??

AICMM

Hi AICMM

That's what I thought he said - but actually he says that the split opens two minutes before the temperature programme starts, which could be quite soon after the injection. A short splitless time could be the cause of the poor repeatability at the heavy end.

Peter

[chromatographer1]

Peter is right on.

If the injector temperature has not stablilized after the cooling from the injector then a longer wait may improve the recovery of the high end components. It sounds like that could be the problem.

But his RSD values are awful. Since his early components are less than 5%, <2% I think he said, he should do better if he really had good volatilization of the sample.

He should also try to inject less into the injector and see if his numbers improve. I bet they will.

best wishes,

Rod

[Widebore]

Thanks for the replies.

Peter Apps (26 Oct) - I'll definitely give a large split ratio a go. I apologize by my lack of clarity because I had compared splitless to split at 320 degrees. The 2:1 split did see a slight overall performance increase, especially for the heavier compounds. The splitless injections were kept splitless for 2 minutes and then purged. Do you think that if I wait until three or 4 minutes with the splitless injections and then start the temperature ramp/purge that things may get better?

AICMM (27 Oct) - Thanks I think I'll pull the inlet apart and have a look.

Chromatographer1 (27 Oct) - I'll definitely wait longer to see if there are improvements with splitless and see if I can inject say 0.5 uL instead of 1.

The only other anomaly with the 2:1 split is that the early eluting compounds are fronting quite a bit. But I assume that this might be due to the pulsed injection. Since I have more time to work on the problem in the next few days, I'll work out if split or splitless is the answer.

[Peter Apps]

Hi Widebore

You mention pulsed injection - does that mean that you have a pressure pulse programmed into the GC inlet gas flow controls ? If so, see what happens with no pulse.

A 2 min splitless time should be long enough, as long as the inlet is hot enough, and is clean.

Peter

[cjm]

I run TPH and CT ETPH daily, calibrating C9-C36 and have almost zero discrimination for all compounds. It took a while to perfect it. I could describe my setup if you're still having trouble.

[Spuzzin]

Widebore,

What exactly are you calibrating with ? An alkane mix such as Florida Mix or something like a diesel/lube oil solution ?

Rich