Polar solvents and non-polar column (eg. HP5)

[rambiochem]

Is it true that polar solvents such as water, methanol, acetone are not compatible with the GC columns such as DB5/HP5? What will happen to these column if water/methanol/acetone is directly injected into them? Will the column get damaged? If yes, in what way?

[Ron]

These solvents wil not damage a modern bonded phase column, but there can be some other issues. Water has a very large expansion volume, and is not suitable for splitless injections. Methanol has the same issues, but not as bad since the expansion volume is smaller. These are not usually the best solvents for GC use, but can work well as long as the method is designed properly.

[Don_Hilton]

I've used acetone with a DB-5 type column and it has worked well enough for me. Methanol is a bit tricky and I have learned that you need to keep the column above the boiling point of methanol or you will have some ugly shaped peaks in the chromatogram. Keep it hot and you get good results there as well.

[rambiochem]

Thanks Ron for the comment
What do you mean by 'water is not suitable for splitless injections'? What exactly will happen if sample in water is injected splitless? Will it result in ugly peaks as Don said? How is it exactly related to the expansion volumes of the solvents as you said?

[krickos]

Hi

First, a 1ul injection of water in a hot inlet will give a vapour volume of about 1,2ml, which is larger than many standard liners (~1ml), consequently you typically inject like 0,5ul of water solutions if you not can avoid using water at all.


uhh lets see if I recall this right
When doing splitless the property of sample solvent becomes even more important. As you likely know you want a narrow starting band of your analytes. Polar solvents such as water and methanol may form droplets on the column surface rather than a "flooded" area that evapourates into a narrow start band. So if you get droplet formation you get a poor/streched starting band of your analytes and consequently poor peak shapes.

Also if remembering right there are some column types that does not like too much water/metanol condensating on the phase surface (certain WAX columns for instance if recalling right).

[Don_Hilton]

I've heard some argument on water and methanol on wax columns. And, I've heard it proposed that you are OK if you elute water below 100 degrees C. I dont' think I've ever tried this. And, give the nature of wax, would probably avoid it.

The flip side is that a column is a consumabe item and if the data has sufficient value - and an injection of water onto the column is the only way to get it, frequent replacement of the column factors into the cost per result. (But this can be a hard sell. Some believe a $300 column should last for 20 years.)

And there is nothing like a relaxing afternoon, cleaning an MS source...

[Ron]

The expansion volume is one of the major problems with injecting water in a splitless injection. The volume has to be 0.5 uLor less as pointed out above, or you will have backflash, which is vaporized solvent and analyte spilling out of the liner and into areas where it should not go. This is a great source of ghost peaks in chromatograms.

Starting temperature of the oven program is also critical for using water. Most column surfaces are hydrophobic, so if the column is at 35C you can actually form droplets of water that roll through the column. This can cause real problems if you get droplets of liquid water in the detector. I don't like to use a starting oven temperature of less than 60C for water injections, the vapor pressure is high enough at that temperature to avoid formation of drops of liquid on the surface of the column.

The formation of droplets of water can also affect the width of the injection band, and thus affect the peak shape.