Determination of low levels of water

[HbJ]

This topic spawned from http://www.sepsci.com/chromforum/viewtopic.php?t=5906

My setup for the determination of low concentrations of water in solvents is:

- DANI 85.10 with macro TCD (~ 10 ml/min flow required) and 651.1 electrometer
- Carrier He 12 ml/min (5.0 purified with molsieve and SUPELCO high capacity purifier)
- Porapak R 100/120 1/8" SSx6 ft.
- Injector: 210 °C
- Detector: 280 mAx1; x10; Cell: 220°C
- Oven: 215 °C isothermal
- Peak of water at around 1 min.
- Injection volume: 10 yl

It even allows for the detection of water in ethanol (both peaks are nicely separated).

The determination of water in a solid matrix is a classical application for headspace GC.

Hope this helps you

[AICMM]

HBJ,

Thank you for the post. I am curious, did you have decent peak shape or lots of tailing. My work involves a thick film CP-Sil8 column and peak shape is pretty decent so I am curious about your luck with the Porapak R.

Best regards.

[HbJ]

Porapak R gives a very nice peak shape so I'm happy (you can even use peak height to determine the concentration).

What's your matrix and what are your method parameters?

[Davide]

Why dont you perfom a Karl Fischer analysis, for water determination, it's sounds simplier


Regards

Davide

[HbJ]

Why dont you perfom a Karl Fischer analysis, for water determination, it's sounds simplier

There are two reasons for this:
- I don't have the equipment (most important fact)
- Maybe my matrix interacts with KF so GC is the most reliable way to get the data.

But in a perfect world I'd surely use KF.

[HbJ]

Just as a reminder: How does your method look like AICMM?

[AICMM]

HBJ,

Agilent 6890 with helium ionization detector. Low ppm detection limits in monomer using thick film CP-Sil 8 with a little bit of tailing. Tailing is slight enough to not affect reproducability. Can't say too much more at the moment.

To speak to Davide's point, matrix is toxic so we don't want to bubble it into Karl Fischer if we can avoid it.

Best regards.

[Serpico]

- Maybe my matrix interacts with KF so GC is the most reliable way to get the data.

Of course, you can get KF cells with an oven compartment...this removes the matrix effects as well.

However, reason 1 is still valid - if you don't have any KF kit then your GC-TCD method looks good.

[Serpico]

OK...I actually might have a need for water content by GC.

Sample reacts with iodine, its also thermally unstable - liberating water upon heating.

I was thinking about trying GC-TCD, hope the water liberation (in GC inlet) is constant and use a standard addition method to determine the 'free' water in the solid sample.

However, one more stumbling block is that its also water sensitive in solution! So ideally I'd want to keep it in the solid form. So possibly headpace GC-TCD.

However, I still need to deal with the water liberated on heating so is there anyway to do a standard addition method using a head space but without dissolving the sample?

I guess I was pondering if anyone makes dual compartment headspace vials...could put solid in one side and a high boiling solvent with x amount of water in on the other.

Am I talking crazy? Anyone have any other ideas?