So on my long arduous journey to discover whether atmospheric H2O can containment our samples in the Liquid Cell I have finally hit some pay dirt. It all started three weeks ago when I was trying to interpret IR spectra of some lithium solution I had specially made to test this exact question. Now since IR, at least how we use it, is such a qualitative tool (i.e. it can tell you if a –OH bond is present but can’t tell you how much), I was flustered trying to interpret whether 1 drop, 2 drops, or 5 drops of H2O was significantly different and if we could measure it. And even more flustered by the questions of what specifically qualifies as ‘significantly different’.
This thought brought me back to what my original question I had asked 6 months ago, “Does atmospheric H2O contaminate our samples?’, and more importantly I thought about why I had asked this question. It did not matter that a few molecules of H2O got into our samples, it mattered that a significant amount of H2O got into our samples, but why? Because if a ‘significantly’ amount of H2O got into our samples it would change our measured conductivity, viscosity, and density. This gave me a brilliant thought, why not use our physical property measurements as a way to gauge whether our samples become significantly contaminated while in the liquid cell.
The most sensitive and consistent instrument we have is the density meter, and the fact that the density meter was outside the glovebox allowed for me to measure samples with water contamination. This week I prepared 4 std samples, one at a 0.003 molar fraction H2O to methanol, one at 0.01, and the other at 0.02. These are small enough quantities that atmospheric absorption of H2O, if present, would not only be detectable but quantifiable. At this point I have some preliminary data indicating that indeed the liquid cell samples are being contaminated by atmospheric H2O and that this contamination is significantly altering our physical property measurements.
Now before I can say that statement with absolute certainty, there is one explanation that could explain my results. Although we tried to procure the purest methanol available it is very likely that it had a small amount of H2O was present in our starting material. This is validated by the fact that our pure methanol sample has a density of 1% higher than the literature value for methanol. Methanol’s bp is around 55C, I programed the liquid cell only went up to 45C, but it is likely that some methanol evaporated out of the liquid cell, if so this would increase the concentration of H2O in the liquid cell sample, mimicking the effects of atmospheric H2O contamination. I think we could attempt to validate or disprove this issue by measuring the volume of methanol that goes into the liquid cell, letting the liquid cell sit at RT, and then measuring the remaining methanol and the density.