GC-HS problem with aceton in water

[GT]

Dear All,
I'm analyzing aceton and ethanol in water soluble Active Pharmaceutical Ingredients by GC-HS plus FID.
The problem is poor reproducibility with aceton (RSD >>10%) and not with ethanol (RSD< 3%) obtained by injecting the same standard solution containing both solvents.
Here it is GC conditions (Agilent G1888+6890):
HS, Oven 80 °C, time 20 min., loop 90°C, T.line 100°C, shake high,
GC, Injector 200|C, DB-WAX 30m, 0.53, 1um.
Standard solution is 1 ml of water in GC-HS vial with about 1 ppm of each solvent.
Why problems only with aceton??

Thanks in advance for your help

[Peter Apps]

I am pretty sure this problem or something very like it has already been covered on this forum. Try a search of the archives.

Peter

[krickos]

Hi

The most obvious and likely reason is that you have exceeded the boiling point for acetone with a gross margin. This is not good when using water as sample solvent (DMF etc retains solvents more) because you get a very unstable partion coefficient between solvent and gas phase(more or less completly forced to the gas phase). The high water content in the gasphase at 80°C may also cause troubles, at 90°C the water vapour is known to actually decrease sensitivity in at least some cases (trace halocarbon analysis for example).

Ethanol has a higher boiling point and somewhat different physical properties.

So a good "rule of thumb" is to try not to exceed the boiling point of the solvent of intrest when using water as sample solvent.

Additionally: If you only have 1pmm of solvent in the vial you may expect a higher RSD % than usual, compare to the SST for class 1 solvents (limits 1-10ppm typically) where a RSD of 15% is allowed if recalling correctly.

[sdegrace]

I developed a very similar kind of method to analyze acetone in a (kind of) water soluble product - well, the product isn't really soluble in anything, but I was able to dissolve it destructively using 44% w/w phosphoric acid, so it's still basically an aqueous system. I'm using THF as an internal standard, and I got excellent results in my validation study. The use of an internal standard was extremely helpful.

I found that you really need to lower the oven temperature. Try 70 or 65 degrees, maybe. If you get too close to the boiling point of water you are getting too much water vapour and your reproducibility suffers. I found that definitely 80 C is too high. I finally went with 70 C oven, 105 C loop and 115 C Tx line.

Stephen