whoa, wack, i didn't see there was a second page... =)
well then..
okay, first off, yeah, water won't work. as you experienced. methylene chloride is a possibility, but it's halogenated and that's bad; ethanol, in which THC should be quite soluble, is nicer for trees and birds and flowers and such. (i am so indoctrinated.)
now, it's difficult to predict exactly what solvent is the best here, especially considering this whole sodium ethoxide thing. take advantage of the huge K_ow for THC, 6,000:1 octanol:water, and have a solvent you can handle...
all the stuff i've read seems to indicate that you will at no point get sparkling white crystals. this is unfortunate.. but to optimize your results there are several things you can do. keeping in mind that under perfect conditions you are going to get an oil.
- use ethanol as your solvent. reportedly acetone is a better solvent, but it'll be a problem later when you need to use NaOCH2CH3.. which I should say I'm not entirely in favor of.. but.. oh well. this ethanol will be as dry as possible. you will need a large supply of anhydrous MgSO4, but i think you should know how to deal with that. you will need to dry the solvent at intermediate steps.
- if at all possible, use either a proper Soxhlet extractor or try to devise some way to extract the material into your solvent in the vapor phase. just throwing plant material into a solvent will doubtless result in a significantly lower ratio of desirable to undesirable extracted solute (e.g. proteins and other biomolecules, lipids.. plant stuff you don't want), and possibly, ironically, a lower yield. exposing the very loosely packed but finely powdered plant matter (held in a paper 'thimble') to solvent vapor should work better. more importantly, this method results in no solid plant matter messing up your solvent, which should prove to be a very good thing. unfortunately a Soxhlet extractor is an extremely complicated piece of glassware, far more so than a mere Graham condenser, but there should be some way you can force vapor to travel through your sample then drip back into the solvent flask (the advantage of a proper soxhlet is that it allows the returning solvent to take a much easier path back to the flask as it condenses)
- have you considered using decolorizing charcoal? yeah, there's tons of problems with using it and it's easily possible to adsorb your product by using too much, but provided that what you want to extract isn't going to be absorbed (you'll have to look into that), charcoal could prove to be very useful.
- in theory the method of preparing NaOCH2CH3 isn't all that bad (although my book only has it on microscale), though of course sodium metal cannot touch water. absolute ethanol and a perfectly clean, dry (incubated overnight, in the real world=) flask suffices. the solution would be prepared first, then the crude extract would be added. theoretically, you would have the salt; but you won't have Sparkling White Crystals precipitating out of the ethanol. sodium cannabinols may be soluble in much more than you want them to be from an extraction point of view. you'll want to test the pH at this point - THC will be 50% ionized at pH 10.6 which is probably too high for your needs. if in fact there are IMable oils (which is something I don't know much about), then you could add a volume of water and boil off the ethanol (monitoring temperature), hoping that very little solute ends up in the receiving flask. if by chance the salt is soluble in water, then there you go. but the pharmas don't formulate it like that, and there may be good reasons for it.
really, you don't need to show off here. you could just do the first stage, evaporate the solvent, and vaporize the residue. this would be far easier than getting totally pure aqueous THC (if that's even possible), and far more pleasant than smoking the devil's-weed.