Methylsulfonylmethane, a nontoxic useless dietary supplement the FDA deems “generally recognized as safe”, melts at 109° and can reportedly dissolve aluminum trichloride, from which a new paper reports that aluminum metal can be electrolytically deposited under an anhydrous argon atmosphere in a glove box (I think the paper said at 140°). (Aluminum trichloride alone reportedly sublimates at 182.7°, so some kind of solvent is needed.)
Aluminum trichloride is itself relatively safe; PubChem says it’s in antiperspirants at up to 15% concentration, though it “may cause severe irritation to eyes” and is “toxic by ingestion” (hexahydrate LD₅₀: 3311 mg/kg rat (oral); anhydrous 380 mg/kg). A tricky thing is that it’s maybe best to consider the anhydrous form as a separate compound from the hexahydrate — if you heat the hexahydrate ((Al(H₂O)₆)Cl₃)), it doesn’t dehydrate to the anhydrous form, but rather releases muriatic acid and forms alumina at 100°! The Merck Index says, “Fumes in air, strong odor of HCl; when heated in small quantities volatilizes without melting. Combines with water with explosive violence and liberation of much heat.”
It’s normally produced from aluminum metal, as explained in US patent 3,343,911, though US patent 3,760,066 documents a process for producing it from chlorine, carbon, and alumina at 800°–1000°, part of the patent office’s classification “C01F7/60 - Preparation of anhydrous aluminium chloride from oxygen-containing aluminium compounds”.
At the risk of making a fool of myself by trivializing a patent classification with 16 patents in it (all expired: 01919, 01923, 01925, 01926, 01927, 01928, 01932, 01964, 01971, 01973, 01975, 01977, 01978, 01982, 01990, and 02004), this reaction seems like it would be pretty straightforward:
2 Al(OH)₃ + 6 HCl → (Al(H₂O)₆)Cl₃ + AlCl₃
That is, if you react dry aluminum hydroxide with anhydrous muriatic acid, you should get a 1:1 molar ratio of the hexahydrate and the anhydrous, which you then must separate. Without using water! Worse, you probably don’t really have this neat separation of water-complexed aluminum ions from non-complexed ones; I think you’ll get polymeric aluminum chlorohydrate.
But I bet that if you heat up the mixture to drive off the HCl you’ll end up with a mix of anhydrous aluminum chloride and alumina, and I bet the alumina is insoluble in methylsulfonylmethane.
This is particularly interesting because purified Al(OH)₃ is the result of the Bayer process by which aluminum is electrowon today — but by dissolving alumina in cryolite, rather than aluminum chloride in methylsulfonylmethane. Crudely we should expect this process to be cheaper because it runs at such a low temperature. Could this be a new economic route to aluminum refining?
It would be even more exciting if it turned out to be possible to do the same trick with, for example, magnesium chlorate, magnesium formate, magnesium perchlorate, or titanium tetrachloride.
(Thanks to Mina for help with the calculations!)