*Columns Week meets #RealTimeChem*
This is an issue that crops up quite a bit whenever methanol-based columns are mentioned (ex: here, here and in the comments thread of yesterday’s post). Moving above 10% MeOH in DCM is said to dissolve standard flash grade silica, causing large quantities of amorphous solid to form in the test tubes. Against doom-saying comes a relatively convoluted theory, put forward by Milkshake and others, which suggests that the relatively high heat of solubilization for silica/methanol leads to a wave of expansion in the column, forcing smaller silica particles through standard coarse frits.
In the last few years the majority of my columns have used methanol in one form or another, either as MeOH/DCM mixtures or as part of reverse phase columns. I hadn’t noticed any issues with the standard flash silica (perhaps because I use sand instead of glass frits), but flushing my glass-fritted reverse phase column with pure methanol often–but not always–lead to the buildup of silica mentioned by others.
What I haven’t found online was anyone actually testing the solubility of silica gel in methanol. So, I did the legwork. 1.0040 g of Silicycle P60 grade silica gel was suspended in 100 mL of pure methanol, and left stirring for 48 hours. The solvent was then filtered through Celite [1], concentrated, and the residue weighed.
Against expectations, the RBF did contain a measurable amount of silica (see above). However, the mass clocked in a just under 4 mg. The amount of silica observed following some reverse phase columns was closer to several grams, and given the conditions of this experiment I think we can safely discount dissolution as a significant factor to test tube/silica contamination.
——————————————————————————————————-
[1] Celite was added to the filter as a slurry in water, rinsed 5x with distilled water, and then rinsed at least 3x with methanol. The collected solids were washed 3x with methanol as well.
Most of my columns are about 30% methanol and I’ve never detected any silica in my fractions. Based on your test, it may be that there is a negligible quantity in each fraction…
Where do you think this myth originates? I have heard it time and time again from reputable sources.
Well, under the right conditions you can get silica in the test tubes (glass frit, extreme MeOH gradient). I think silica/MeOH solubility was just a plausible, if incorrect, explanation.
The real issue is the lack of followup. This wasn’t a difficult test to do, and pretty clearly disproves the solubility hypothesis.
if you’re worried ab silica coming through the column, first wash the silica it with a large quantity of the methanolic mobile phase.
My opinion is …Silica is not soluble in methanol but silicagel( another name is silicic acid) is soluble in methanol columm process. when i was in master’s course, i experienced same thing but it observed in high portion of methanol eluent.
i agree with anonymous…if you pre-wash the column with the methanolic eluent, you don’t get any important contamination…at least that’s what I routinely observe getting up to 40%MeOH…
Many silica nanoparticle syntheses are performed in MeOH… never heard this mentioned.
I think that depending on the quality/ age/ previous handling of the silica there will be some grinding of it’s particles, as will with stirring with a stir-bar for 48 hours. So I think it’s possible there’s some insanely fine silica, too fine to be filtered even by the celite, and that’s what you find!
Would you see peaks in NMR or MS if you have trace amount of it ? Any references ?
If there were silica in an NMR tube it’s possible that you’d see a really broad peak (ie. 1 ppm or more) from surface Si-OH bonds.
But as large polymer the silica won’t dissolve in most solvents, so generally I’d expect silica particles to settle at the bottom of the tube, out region excited during an NMR scan. Similarly, I wouldn’t expect much from EI or ESI MS, though a broad MALDI peak is theoretically possible (always filter your samples!).
Is fumed silica soluble in toluene ??