Chromatography Forum

Hi,

I was wondering how precise electronically controlled split ratios usually are (I'm using an HP 7890) and whether it's appropriate to assume the split ratio is quantitatively that dilution of your sample?

On a more specific level - I have some biological extractions that I'm analyzing for derivatized fatty acid methyl esters, with these esters present at a very wide range of concentrations. I had been running all my samples and external standards at a 1:10 split ratio while turning off the detector for the abundant species, and making a 20X dilution of each sample into a new vial and also running that at a 1:10 split ratio to quantify just the abundant species. I decided to try running a 1:10 split and a 1:100 split on the same sample instead, which works great and saves on excessive use of vials - however I need to know whether the standards must be run at the same split ratio for accurate quantitation (so run each of six standards TWICE at 1:10 and 1:100), or whether I can run them just at the 1:10 split and assume the split ratio itself is accurate enough to quantify everything on that basis.

Should mention that the manual dilutions aren't terribly accurate either since the solvent is hexane, but I think the internal standards to which everything is normalized help account for evaporative concentration increases.

Thanks for any suggestions!

I decided to try running a 1:10 split and a 1:100 split on the same sample instead, which works great and saves on excessive use of vials - however I need to know whether the standards must be run at the same split ratio for accurate quantitation (so run each of six standards TWICE at 1:10 and 1:100), or whether I can run them just at the 1:10 split and assume the split ratio itself is accurate enough to quantify everything on that basis.

Do you want to be accurate or save effort? The answer depends on how accurate you need or want to be. You should assay the lower-level components at 10:1 v. their external standard, then assay the main component at 100:1 v. its external standard.

Personally, I'd assay both at 10:1 and dilute additionally to assay the main component, but what's 30 years experience mean these days....I guess because some of my units use manual pressure settings, not EPC....

I would calibrate at both split ratios - even if the EPC controls are working exactly to settings while the normal carrier gas flow is running, the actual split ratio depends on what happens at the moment of injection, and then the pressure pulse caused by evaporating solvent can change things dramatically - and non-repeatably. Molecular weight discrimination will also not be consistent with changed split ratios. You can check the performance of your system by injecting the same sample at 10:1 and 100:1 and checking to see if the peak areas are in 10:1 ratio.

What detector are you using ?, an FID will easily give linear responses over 4 - 5 orders of magnitude, and its response to a small peak is not affected by a large peak having just eluted, but is cannot be turned on and off in the middle of a run as you say that you do, . If you are using an FID, why (and how) are you turning it off for the large peaks ? If you are seeing the large peaks interfering with the small ones this is most likely to be an effect on the column, not the detector. What range of quantities per compound (i.e. how many ng) are you putting onto the column ?

Peter

I think you'd better run calibration in both split ratios. The split ratio is not only the dilution factor but it also controls the time the samples spend in the liner and it affects the way the components go in to the column.

On turning off the detector - if you are using a mass spec and are having issues with the larger peaks being solved with a tenfold dilution, look at using a weaker ion in the analytes for quanntitation of the larger peaks. As long as you are not saturating the ion source you can quantitate on any ion you want. It does not have to be a strong ion in the spectrum.

You need to be careful as the EPC won't control as well at very low total flows (10 ml/min? - can't remember). Depending on your column flow, you may not be too low. You may get better precision at a higher split, if you can get the LOD that you need.