Well, let me tell you several things.
First, DMF is indeed quickly decomposed by conc. aqueous alkali at RT! ([confused] can anyone approve or disprove this fact? - SWIM is still unsure if his DMF might have some (Me2NH2)HCOO in it). On contrast, when a solid alkali, such as KOH is added, no smell is observed (doesn't dissolve much, either).
Second, a trial was conducted on vanillin using DMS in DMF/K2CO3. Everything was done a la Hest (
https://www.thevespiary.org/rhodium/Rhodium/chemistry/345-tmb.html
), xcept the mixtr was incubated for 5 hrs instead of 3 and an occasional shaking was employed instead of stirring. Result -
nada. Got all vanillin back unchanged. SWIM also noted that when he added DMS to the mixtr it heated considerably - something that normally shouldn't happen. So the probable cause of failure seemed to bee the presence of dimethylammonium formate in DMF.
So SWIM took his DMF and boiled it for an hour, hoping to convert the formate to formamide. Do you guys think it'd work OK or something else is needed? In any case, both boiled and unboiled DMF showed absence of ammoniac smell when treated with granular KOH - so maybee SWIM was wrong about dimethylammonium salt... Could something else ruin the reaction? Like water?
SWIM has no doubts about the validity of Hest's procedures - and it drives him insane that he can't really figure why it didn't work. It must bee smth REALLY obvious - but SWIM's DMF seems to bee devoid of both significant qtties of water and dimethylammonium salts - otherwise there'd bee smell upon treatment with KOH, no?
Fortunately, SWIM beegan to consider the variation with acetone as solvent. Drying acetone is easy - it was refluxed for 5hrs over CaCl2 and distilled.
Initially he was reluctant to take this approach since the patent, quote from which can bee read at Rh's (
https://www.thevespiary.org/rhodium/Rhodium/chemistry/mmda.mescaline.html
) stated a
24hr reflux was needed. Yikes (SWIM feeds his condenser from the kitchen sink which is normally used as such
![Smiley :)](https://thevespiary.org/talk/Smileys/default/smiley.gif)
).
But then he re-read
Patent US3867458
, mentioned in the beginning of this thread - and to his great surprize found out that, according to its authors,
only 3hrs was needed! Upon careful analysis the following reasons for that discrepancy were found:
1. Solvent to reactant ratio is twice as high in the 24hr variation.
2. Na2CO3 + a little KOH is used as a base in the 1st patent, whereas the 2nd patent clearly says that 'in presence of Na2CO3 instead of K2CO3 the reaction does not proceed'.
Well, SWIM tried it and it worked! Unfortunately, he made a fatal mistake during workup (DO NOT add NaOH to a soln containing both acetone and aldehyde - aldol condensation) so the proper writeup is yet to come - but for now it's clear that this modification works well, gives no byproducts and does not require any stirring (after an hour or so the reaction begins to bump, which efficiently mixes the flask's contents).
SWIM would really like to master that DMF variation though... Really confused about what happened and how to prevent that (or, for that matter, how to efficiently test for presence/remove an ammonia salt from DMF). Any ideas, bees?
Antoncho