Ayahuasca: alkaloids, plants & analogs
Section 1 :
The Hasenfratz Method of Harmine and Harmaline Isolation
With additional comments by Keeper of the Trout
The green seeds of Peganum harmala, if available, will give a higher yield than the ripe seeds. Roots of Peganum harmala may also be used and, while somewhat lower in total alkaloid, will have a much cleaner alkaloid profile.
Banisteriopsis caapi may be used as well but gives a substantially lower return and thus requires a far larger amount of dried plant material. For example, an extraction of 95 grams (dry weight) of cultivated B. caapi leaf yielded a total of 0.6 grams (600 milligrams) of alkaloid. (i.e. 0.63% total alkaloid. This is nearly 4.5 kilograms of leaf per ounce of total alkaloid recovered as the hydrochloride salts. Recovery would be higher if stem bark or roots were used but it would still fall short of Peganum harmala's yield.)
In the case of B. caapi the principle alkaloid present will be harmine, with lesser amounts of tetrahydroharmine; if harmaline is present at all it will usually be as a minor component (but the odd exception has been reported). It also lacks peganine etc. found in P. harmala as well as the fixed oils so it will start out as a greenish white to white precipitate and will not need as many recrystallizations to purify.
A few other plants are potential sources of harmine or harmaline such as Tribulus terrestris, and Passiflora incarnata. These have tetrahydroharmine , harmine or harmaline in even lower amounts and require far larger quantities of plant material to be processed. [Hundreds of grams of dried leaf and stem per dose in the case of Passiflora incarnata or an even larger quantity of fruit with Passiflora edulis  These also have other alkaloids present in greater quantities than the harmine and require more complex purification.
Tribulus is used medicinally due to the presence of some of these other compounds.
Passiflora is also used medicinally due to the presence of some of these other compounds. Its most abundant beta-carboline alkaloid should always be expected to be harman, never harmine.
The following approach came from Marion's article in Manske & Holmes (eds.) The Alkaloids Volume Two (1952), from page 393 and Hasenfratz (1927) Annales de Chimie 10 (7) 151-226 [pages 154-155]). [Marion cited Hasenfratz.]
The seeds of Peganum harmala are crushed or ground  and covered with 3 times their weight of 3% acetic acid . A thick dough is formed which is pressed  after 2 or 3 days. Then, twice the weight of seeds worth of acetic acid is added and after maceration they are pressed again. The two liquids are combined and sodium chloride (100 grams per liter of liquid; i.e. 1 gram per 10 ml) is added. This is then refrigerated until cold and the bulk of the liquid is siphoned off leaving the crystalline residue layer on the bottom. The crystalline residue is filtered with suction (i.e. with a Buchner Funnel) and then redissolved in hot water. The subsequent addition of sodium chloride and chilling will cause the hydrochlorides (harmine and harmaline) to precipitate as a "crystalline mush". Recrystallization is to be repeated (i.e. redissolving in hot water, adding salt and then chilling to precipitate the crystals) until the hydrochlorides acquire a yellow color (they start out reddish). Usually 3 to 5 times are sufficient. Occasionally two such recrystallizations are enough. At this point the material can be dried by gentle heat . The paper and material will dry rapidly and the yellowish or greenish material thus obtained will fall off easily and should then be packed in a tightly sealed lightproof container and kept refrigerated.A crucial point is the removal of fine particulates from the first step (acetic acid extracts). Careful and thorough filtration or use of a centrifuge is recommended. Otherwise these particles will remain with the product. Another potential source of (a trivial) adulterant is trying to work the powder loose from the dried paper. If a person merely removes what falls off easily as chunks this is not a issue .
Excess salt is not desirable. Measure all liquids and weigh the salt carefully. (Any excess salt will become a part of the product .
Salt is the most likely, and single largest, potential adulterant when using this process. Careful measurement of salt can help minimize it.)
Vacuum filtration with a Buchner funnel and Whatman filters will make this easy. Otherwise, the tendency of all of the above solutions to clog filter paper will cause it to be tedious and time consuming .
The above separation takes advantage of the fact that when sodium chloride is added to a solution of harmine and harmaline acetate it transforms them into their hydrochloride salts.
The hydrochloride salts of harmine and harmaline are insoluble in cold sodium chloride solutions and thus precipitate during the chilling.
Further purification is possible by separating the alkaloids (see below) or the mixed alkaloids can readily and simply be sublimated to further purify them (Sublimation will not separate them but will highly purify them.)
Similarly, simply dissolving the hydrochlorides into warm water, neutralizing them with ammonia, and recovering the precipitated free base by filtering, will purify the product and remove any excess salt.
By almost all accounts harmaline is a stronger MAO inhibitor than harmine and when measuring dosages this should be taken into account.