Sample concentration for 250µ Capillary columns and FID

[WillyOne]

I'm analyzing pesticides formulations by GC or HPLC
When using GC the typical situation is a 30 m x 0.25 mm Capillary column (0.25µ thinn film) of 5% Phenyl/Methyl silicone.
I tryed to inject 1 µl of a 5 mg/ml solution in split mode 50:1 but it overloads the column. (fronting peaks) The same result with 100:1 split.
I can increase the split ratio but the Fid sensitivity can become a problem for linearity check.
I can continue testing but perhaps some of you have experienced this problem.
Any suggestions?

[Peter Apps]

You are injecting 5 ug to the inlet, of which 100 ng goes onto the column at 50:1 split. An FID will give a cleanly detectable peak from 1 ng, you will be getting into signal:noise issues with 0.1 ng. How much you can put onto the column depends on what the pesticies are.

The problem with an unselective detector like an FID is that any real sample will generate a forest of huge peaks that will swamp any response to the target compounds - no matter how small an amount you can see in a clean standard, and no matter how linear the calibration is.

Peter

[WillyOne]

I did continued testing this morning.
It seems that from a split ratio of 200:1, means 25 ng on the column, I can obtain a USP asymmetry > 0.9 that seems enough.
Then I need check linearity from 2.5 to 50 ng. and repeatbility at 25 ng.
I work with internal standar, but anyway I'm not sure I will be able to handle so little peaks with automatic integration.
I use a Varian 430 and Galaxie for data treatment. And sometimes I loose the peak.

[Peter Apps]

If the smallest quantity that you need to detect is 2.5 ng then the FID should give you a very easily detectable peak that automatic integration will find and integrate repeatedly and reliably. If this is not so, there is something wrong with your instrument.

What kind of samples are you planning to analyse ?

Peter

[WillyOne]

I normally work with pesticides formulations. That I look for is quantify the active ingredient content. It means that I look for a highly concentrated component ( as high as 60 % w/w) That's quality control and I have no interest on impurities, only to separate them.
I usually try, (given that there is a label) to work in a very close concentration range. Means I weight 100 mg of analytical standar and 200 mg of a 50% formulation in order to match the concentration in the injected solutions. but I need investigate the linear range, repeatability, and accuracy to be sure of my results when a formulation presents a deviation.
I'm not sure that I'm answering your question

[Peter Apps]

That answers the question very nicely thank you - at least the indiscriminate response of the FID is not likely to present any problems.

Given that you know how much pesticide there should be in the sample, do you really need to demonstrate linearity over a wide range ? Regulated QC is not my area (thank goodness) but in pharmaceuticals I understand that a range needs to be from 80% to 120% of the expected concentration of an active ingredient in a formulation.

A move to a 0.32 mm diameter column with the same phase ratio will give much greater loadability, is there a reason to use 0.25 mm ?

Peter

[WillyOne]

Thanks a lot per your help.
Talking about molecules, I've "validated": Bifentrin, Buprofezin, Clomazone, Clorpirifos, Dimethoate, Fenitrothion, Methalaxyl, Methyl Chlorpiryfos, Oxyfluorfen, Procimidone, Tebuconazol and Triadimenol.

About 0.32 mm i.d. column. That was the solution I was thinking on. I use 0.25 mm because that was the column that our Residues Department gave to me. And I begun working with the materials I had.
Really I'm not sure if that is only an estetic problem because the accuracy test I use gives me good results.
This is a Official Laboratory dependent of the Spanish administration. Means that not being manufacturers, I don't have formulations blanks. So I check Accuracy by Internal Standar Addition.
And it works, obtaining deviations of less than 2% between results by ISA and the proposed GC method. But the Chromatogram is horrible...

[Peter Apps]

Your first step needs to be to use a 0.32 mm column - since you are in an official regulatory lab your method needs to be robust.

I have a concern over how you can be sure that you do not have impurities, or components of the diluent co-eluting with the pesticide peak. A straightforward approach would be MS. If you only have FID then I would have thought that you would need to run the sample on two different columns, and get the same peak area on both.

How are you doing your sample prep ? - I would have expected "dilute and shoot".

Peter

[WillyOne]

Well, not exactly the same Area but the same Area ratio between a.i. and Internal standar added.
I feel that Analytical Standar addition technique can be more accurate that the simple comparation of area ratios in two different columns.
As in Atomic absortion Spectroscopy the idea is obtain a result free of matrix effects. This is done by analyzing solutions that keeps constant the internal standar concentration but where I spike different quantities of analytical standar. the representation of the corrected areas front the additions gives us to a straight line and with extrapolation to the concentration of the sample.

About sample prep is as you told dissolve and shoot. Only a few points.
The election of solvent should warrant the complete extraction of the active ingredient. Sometimes mixture of solvents to break the formulation.
I add the internal standar by volume. Means I prepare a volume of the IS choosen in a concentration that gives an area similar to the 5 mg/ml solution of active ing. Then I add the same volume ( by dosifier) to the Standar and samples. Sonnication 5 min and magnetic stirrer 15 min. Decante or filter and shoot...

[Peter Apps]

Known additions is certainly a sound technique for quantitation - my question was how do you know that the analyte peak contains only analyte ? (I am assuming that you have no co-elutions with the internal standard of course). If there is another component of the formulation that elutes at the same time as the pesticide, the pesticide peak will be bigger than it should be, and the amount of pesticide will be over-estimated. Since you have a non-selective detector you cannot detect co-elutions. If the pesticide peak is pure it will have the same area (or relative area vs the standard) when run on two columns with different phases (i.e. polar vs non-polar) since it is very unlikely that two compounds (the pesticide and the interference) will co-elute on two columns.

Peter

[WillyOne]

Thank you by your suggestions. The use of two different columns can be a great idea. I will try to put it at work.