"Extremely dangerous" if done by amateurs or without a regard to safety. However, it's quite possible to work with flammable solvents in such a fashion that the risk of fire or explosion is greatly reduced (sealed-off systems, fume hoods, careful temperature control, general fire safety like no sparks or open flames, etc). Many labs do solvent extractions on a regular basis, and even amateur extractions can be done safely.
Also, solvent removal in general can be as easy as melting your extract and holding it at a certain temperature. The gold standard is to heat and melt it in a vacuum oven, at low pressure, to ensure any solvent turns to gas and is pumped away - and residual solvent usually has a noticeable smell as well as makes any extract much less solid.
THC extraction can also be done with supercritical CO
2 (used to decaffeinate coffee beans), which does have its own risks (high pressure gas/liquid handling, and CO
2 is in theory toxic if a massive leak occurs) and costs more, but also extracts cannabinoids well, evaporates instantly, and leaves no solvent residue. It can even be recycled.
Anyway, the reasons THC-A crystals are rarely if ever seen:
1. Most commercial/mature marijuana contains at least some THC in addition to THC-A. But the "desired product" is usually THC so growers prefer selling those strains.
2. THC-A is not super stable. Even at room temperature in the dark, some THC-A will decarboxylate, gradually. And as stated before, acid exposure, heat, and very likely base exposure and UV light all convert it to THC and CO
2. Thankfully there are no enzymes that decarboxylate it.
3. Separating THC-A from the other cannabinoids is not something that can be done super easily.
Normally if you take freshly dried buds and extract them with butane/hexane/heptane/naphtha, you will get an extract that can contain proportionally quite large amounts of THC-A (and if the plant produces CBD, CBD-A will also be present). Ordinarily this is hard to detect, typical bioassay techniques like "dabbing" or using a vape pen will, through heating, convert all THC-A to THC. And analysis on a GC by most methods will also decarboxylate the sample (as the sample on a GC first passes through an injection port heated to >300C or so), meaning analysis should ideally be done on a HPLC, but can be done on GC too, however derivitization is needed - reacting the sample with e.g.
BSTFA, a quick and easy reaction done by adding a premeasured amount of solution to a small sample, which masks the phenolic OH and the carboxylic acid COOH with protective trimethylsilyl groups
[Image: THC-A-2TMS - aka THC-A with a trimethylsilyl on the OH and COOH - structure and mass spectral 'fingerprint'], allowing them to survive a GC after all.
The two ways I have seen to seperate THC-A that would work are, adding an amine like triethylamine (or many others, even quinine(?)) to a solvent solution of cannabinoid extract, which precipitates crystals of triethylammonium tetrahydrocannabinol-carboxylate that can be carefully neutralized to isolate pure THC-A
[ref], or alternatively flash chromatography of the crude cannabinoid extract (which as a bonus can seperate individual cannabinoids as well, i.e. THC from CBD from CBN). However solvents are needed, and also some means of identifying the product.
Another method that might work but I am less confident about is simply dissolving the crude cannabis extract in e.g. pentane or hexane and cooling it to very low temperatures for a long time, in the hope that THC-A is less soluble and will precipitate (but also, other things could precip too like waxes/fats).
Another thing that has not been considered is that every cannabinoid the
Cannabis plant makes is actually produced as the carboxylic acid. So, there is also such a thing as CBD-A, CBG-A, and so on. Separating THC-A and CBD-A (or others) could only be done by chromatography unfortunately, and the easiest way to check if you have a mixture is by doing thin layer chromatography (unfortunately.)
Like its THC analog, CBD-A is much less bioactive, does not metabolize to CBD, converts to CBD when heated (slower and less completely than THCA->THC
[ref]), but converts instead to THC if exposed to acid (decarboxylation and the acidic resorcinol OH attacking the outlying isopropenyl group of the terpene "half" of CBD, forming an ether ring, i.e. THC is formed. It can happen either like CBD-A -> THC-A -> THC or CBD-A -> CBD -> THC.
The main reason this is not done is because there doesn't seem to really be a point. The cannabinolic acids are definitely not active at CB
1 or CB
2 - or eating raw weed would work way better. They don't have much research behind them but I would assume that they are mostly lacking in health benefits. Also, unless the THC-A is isolated properly, and kept in ideal conditions (preferably as cold as possible, in a sealed, dark container) it will degrade to THC, potentially quite quickly.
So you have a product that takes special effort to isolate, seemingly lacks psychoactivity, and is only
really useful when it readily decomposes into (possibly illegal) THC. There is also not a lot of demand.
Maybe some specialty stores might find a chemist willing to isolate some as a special order, but it would not be cheap.
If you want to experience THC-A, go eat an eighth of some random bud, especially if it hasn't specifically been cured/aged, and prepare to be underwhelmed. No need to isolate it.