I am developing a method where I determine Isovaleric acid in the headspace over an aqueous solution.
The basic method is that isovaleric acid is added to 5mls of a 5% NaCl solution, and the headspace sampled using a 100um PDMS SPME fiber exposed 5minutes at 22.0C. I am trying to calibrate from 0 to 100uls added to each vial, and am having linearity/repeatability problems trying to generate the calibration curve.
Now, I don't think my difficulties are due to outside variables such as temperature control, sample prep error etc. There are two isovalerate ester impurities in my isovaleric acid stock solution. If I plot a linear regression of the isovalerate esters, I get good linearity, and good repeatability.
I expect that isovaleric acid is more volatile than the two ester species. And of the two esters, the higher molecular weight one has the best linearity and reproducibility. So I think my difficulty is related to the volatility and partitioning of the isovaleric acid, more than sample prep error, etc.
Any SPME experts out there?
By Anonymous on Friday, March 21, 2003 - 12:18 am:
Reduce the amount of ionised acid.
The variation in ion concentration influences the partition solution-headspace as can be seen with the esters which are not ionised.
I would recommend lowering the pH of the solution with a buffer so that more acid stays UN-ionised thus more volatile.
[OT: Not taking the UN into account leads to unpredictable results ;) ]
By Anonymous on Friday, March 21, 2003 - 06:27 am:
Just a note from another analytical chemist who is a very poor organic chemist, you might not want to make the solution too acidic if you are interested in the esters. I think they might start to hydrolyze in the aqueous acid media. Just my two cents worth.
By Anonymous on Friday, March 21, 2003 - 10:44 am:
I would expect the ester to be more volatile than the free acid so there's no surprise that your ester curves are better. The acid is probably more water soluble. In addition to adjusting your pH, as suggested by the others, I would try using a higher concentration of NaCl, to coax the acid into the headspace. I use a saturated NaCl solution for some of my analyses.
By Mike on Friday, March 21, 2003 - 12:55 pm:
Actually, I can't adjust pH. The idea is that I add a known quantity of isovaleric acid to a vial, then add a product designed to absorb/neutralize the acid. I then measure the residual acid remaining. pH adjustments may interfere.
I'll try higher salt concentration
Thanks for the suggestions,
By Anonymous on Monday, March 24, 2003 - 01:18 am:
How about HPLC instead then?
By Anonymous on Monday, March 24, 2003 - 06:07 am:
Have you tried longer equilibration times? It sounds to me like you may not have reached an equilibrium condition in 5 minutes.
By Anonymous on Wednesday, April 9, 2003 - 12:34 pm:
In case anyone is still interested, it seems that it is an equilibrium issue.
First I tried increasing the sampling time to 20minutes, this increased my peak areas, but did not improve repeatability for Isovaleric acid. Isovaleric RSDs are all over the place while the two ester species are around 5%.
Next I compared standards made individually with aliquots taken from a master batch. %RSDs from the master batch are around 5% for all three species, while isovaleric %RSD from the individual samples is around 26%.
Next, I repeated the comparison using an automated headspace sampling system, heating at 80C for ten minutes. In this case all three analytes show good %RSD for the aliquots and individual standards.
Finally, I prepared samples individually, and aliquots from a single batch as described on a friday and allowed them to sit over the weekend. Monday morning - smooth sailing. %RSDs are good for all three analytes prepped both ways. So, now I prep my samples a day or two ahead of when I plan to analyze, and all is well.
Thanks for the help.