Osmium's variation on the Al/Hg reductive amination of MDP2P

By Osmium

[ Back to the Chemistry Archive ]


My first experiments were performed according to Shulgin's procedure with P2P and MDP2P. Yields were a bit on the low side, probably because of lack of purity of the ketones, but mainly due to the presence of too much water and vigorous reaction conditions causing too much loss of MeNH2. Even with strong cooling of the reaction to avoid MeNH2 losses, yields didn't improve. Lit. search revealed several other procedures, e.g. different solvents, amalgamating in the reaction solution, using big excess of Hg (1g HgCl2 and 7g Al wire! Dunno where I found that shit!), using MeNo2 as the MeNH2 source (MeNO2 is added to the reaction, together with the ketone) etc. The best variations however were found in 2 old American patents. I don't have them anymore, but I gave Strike some hints where to find them. They will hopefully find their way into TS2. In these patents 2-MeO-P2P (and I think some others, too) was reacted with several amines in more than 80% (i guess even 90%) yields. hey gave me confidence that Al/Hg really works. So I experimented and found out the following:

1. Solvent:

In Europe denatured EtOH is dirt cheap and found everywhere. At least in all the countries I've been so far. This alcohol is usually ~94% EtOH, the rest water and very little denaturant. It is usually impossible to remove them, so I didn't even try it. This EtOH worked perfectly. 99% IPA is OTC, too, but it costs 3 times as much.

2. Al:

Since the amalgamation and the reduction of the intermediate imine takes place at the solid-liquid interface, reducing the surface area of the Al both cuts amount of Hg necessary to effect complete amalgamation and slows down reaction rate to a usable amount. In reduction using thin Al foil (huge surface), a lot of the Al is lost because it is consumed without reducing anything but liberating H2 which is useless in this reaction. Of course Al is cheap, but the Al eventually forms Al(OH)3 sludge, which must be removed by filtration and which retains a lot of valuable product. It is thereore preferable to use as little Al as possible. Al turnings or sheet have a much smaller surface area than foil for a given weight. This means that the reduction will take much longer for completion. But because of a better Al surface to imine concentration ratio, much less Al will be lost reacting with the solvent. I don't remember for sure , but I guess I used about 50-60g of Al per mole of ketone. This still is a healthy excess (>100%), but much less than Shulgin uses. This means there will be less Al sludge to filter. Reduce this amount if you think there's still too much of the Al left when the reduction is finished.

Don't bother etching or cleaning the Al. It is unnecessary.

3. Hg:

Very little Hg is needed for complete amalgamation of the sheet's surface. Remember, surface area might be several dozen times smaller compared to foil! Usually I didn't even weight it. I added a spatula or two, and achieved good, reproducible yields without using the same amount each time.
BTW, you can't "over-amalgamate" the Al. Excess Hg salts will be reduced to the element. No Hg was ever found in the final product. In my opinion, no sizable amounts of Hg will make it through the sludge filtration and acid/base purification, but of course I can't guarantee it ;-). Some of you might ask: Can I omit the Hg? What function has it anyway? There are several answers: Al is a very reactive metal. The only reason it does not decompose spontaneously like Na is a thin protective layer of Al2O3 on its surface. This effect is called passivation. Hg destroys this protective layer and avoids its formation. When a reactive metal reacts with acids (or bases or water), H2 is formed. The Hg sucks the liberated electrons out of the dissolving metal as every other more electropositive metal would do. When you connect two different metal strips and put them into an acid, the more electronegative metal will dissolve, while the more electropositive will show the H2 evolution. This is used for corrosion protection of pipelines and marine equipment like boat engines. The same happens here. But H2 evolution does not work well on a Hg surface. I don't know the English term for this effect, but I can give you another example: When a NaCl soln. is electrolysed with a Hg cathode, almost no H2 is formed. This is used in industrial chemistry to produce NaOH which isn't contaminated with Cl(-) or ClO(-) by decomposing the formed Na amalgam with water in a different reactor, producing NaOH and Hg. The Hg is now "enriched" with electrons. They are transferred to the imine, reducing it. Electron transfer to H(+) or H2O, which would produce H2, is taking place at a much lower rate. This can be observed in the Al/Hg amination: when all of the imine is reduced, the remaining Al sheet will dissolve at a much slower rate, now with H2 evolution.

4. H2O:

The EtOH I used already contained maybe 5% water. The MeNH2 used was usually 40% aq. soln, so there's more from this source. This is all the water the reaction needs. Don't add extra water like Shulgin. It is not necessary, even when you use higher percentage alcohols, like 99% IPA. H2O is needed in this reaction. Although Al can also dissolve in alcohols, the formation of Al2O3/Al(OH)3 seems to be much more exothermic, which probably (?) helps the reaction (don't ask). An experiment using abs. EtOH and NH3 gas in huge excess yielded absolutely nothing. Too much H2O of course is bad, because imine formation is reversible.

5. Amine:

40% commercial aq. MeNH2 was usually used in my experiments. Ketone:amine molar ratio was usually about 1: 2.5-3.5. Using NH3, even in big excess, gives lousy yields of MDA, I was told by some other researchers. EtNH2 should work even better, because of lesser volatility and better imine formation.

6. Stirring:

Strong overhead stirring is necessary, especially for big aminations, because the Al sludge will cover the Al avoiding contact with solvent/imine. Magnetic stirring will be insufficient for bigger aminations.

7. Temp:

Depending on the Al surface area, it will probably be necessary to heat the reaction. It usually was necessary in my experiments. A slight reflux was usually used. With an efficient reflux condenser and ice-cold water, MeNH2 losses through the condenser were very small.

General procedure:

  1. Cut Al sheet into small squares.
  2. Dissolve Hg salt in EtOH, add squares. About 800ml EtOH/mole ketone (? forgot it, sorry)
  3. Let amalgamate for several minutes, until bubbling starts/sludge formation is evident/silvery appeareance.
  4. Add MeNH2 soln. and a few grams of NaOH. This will speed up reaction a little. Heat up the flask to a temp below reflux.
  5. Slowly, dropwise, add the ketone dissolved in some EtOH. The idea is that the ketone reacts with a big excess of amine, favoring imine formation. Ideally, ketone should be added as fast as it is consumed. Just add it over the course of several hours. Or add it all at once. It probably does not make that much difference ;-). Let reflux slightly.
  6. Let react until you see reaction slows down. Internal temp will drop, asexothermic amination is finished. Or let react until all Al has dissolved. At this point 15-20% NaOH can be added to convert the Al sludge into a more filterable granular consistency. Again heat it up to reflux to speed this conversion.
  7. Switch off stirrer and let the sludge settle for some hours or overnight. Decant the solvent, add fresh or rotovaped EtOH to sludge. Boil again, let settle, decant. Repeat this as often as you want or until you think you have extracted all the goodies.
  8. The decantations are filtered to remove residual carried-over sludge, and the EtOH is removed by distillation. This EtOH contains MeNH2. Acidify with HCl and again evaporate to obtain MeNH2.HCl. Or reuse EtOH in the next reduction. Or use it to extract the Al sludge, or discard if you want to.
  9. The crude product/H2O mixture is dissolved in dil. HCl, washed several times with DCM, made basic and the freebase extracted several times with DCM. This base extract is washed with water, the DCM removed by distillation and the freebase distilled. More than 80% yield when done correctly.
  10. Dispose of the Al sludge properly. It contains elemental Hg!.

This scheme looks very time consuming. The reduction will take a day or longer, the extraction/decantation several hours. But the reaction actually needs little attendance, you can do other things in your lab. Like cleaning everything ;-), or doing other reactions. Having a rotovap (or
two) will speed up the evaporation of solvents considerably. I usually left the reaction unattended for many hours. Nothing bad ever happened.

And the best thing is: this reaction can be scaled up almost indefinitely. It was designed for several hundred gram reductions, and I was told that some evil people used it repeatedly for 300-400g ketone reductions. When done properly, you will obtain more weight of MDMA.HCl than ketone used (this sounds better than giving an abstract yield number. Shulgin reports 42g X from 53g ketone...).

Nobody using this procedure reported an upper limit yet. I guess it can even be done with kg quantities. Try that with Al foil! Sorry I don't give exact numbers. All written records are lost. I remember some of the details, but I'm not absolutely sure about them. Use your immagination and common sense to make it work. Combine this with the MeOH/O2 Wacker, which gives >80% yield, and the doors to massive and easy X production really are wide open.

Cool. The War is Over.

MDMA from MDP2P using Osmium's Al/Hg variation

By Modesty

55g of thick aluminium sheet (about 0.2-0.3mm) was cut in about 1cm wide strips, about 4cm long. These were folded around the finger to a ring-shape.

MeNH2-HCl was added to a cooled beaker containing 360ml. of 25% NaOH, until pH was neutral. NaCl settled to the bottom.

800ml of 99.5% IPA was poured in a 4-litre flatbottom, and a knives-edge (a bit less than a gram) of Mercuric Chloride was thrown in. After dissolving with occational swirling for about ten minutes the aluminium rings were thrown in, and a little IPA was added to cover the rings completely. The ring shape makes sure the foil is evenly distributed throughout the solvent, because we do not have the means to stir this reaction. If flat squares of foil is used, they will sink to the bottom, and quickly be covered in the sludge generated.

After what seemed to be a longer time than usual, silvery spots appeared, and the reaction was started:

The beaker containing the methylamine was poured into the flask, the NaCl was left in the beaker.

120g of MD-P2P mixed with 100ml IPA was poured into the flask, and the beaker was rinsed with a small amount of IPA - all poured into the flask. The flask was swirled around to mix the reagents and it was placed on a hotplate. Gentle heat was applied to speed up the reaction. Aiming for about 50°C.

Every 15-30 minutes the reaction was checked up on, swirled, temperature controlled. After a couple of hours, the heat was turned off as things were getting too hot. Some cooling was needed. The effervescence created by this reactions bubbeling was considerable, and after 5-6 hours the reaction was left to its own devices. It was judged that it would not heat up to unreasonable levels. Sleep....

The next morning the reaction had finished, and the whole mess was poured into two litres of 40% NaOH, premade and at room temp. Now something happened that has never happened with methanol as solvent. It separated into two layers. The top alcoholic layer was separated off and the bottom layer was extracted once with 500ml. of Xylene.

About half the alcohol was evaporated off and the remainder poured into 1.5 litres of distilled water. This solution was then extracted twice with 500ml. portions of Xylene. The second extraction created a nasty emulsion which cleared up when about 100g of non-iodized table salt was poured in and mixed.

All the Xylene extracts was combined, washed three times with distilled water, and once with a saturated NaCl solution, dried with Magnesium Sulphate and gassed with HCl.

Yield of a post-vacuum filtered xylene-wet white crystalline powder weighs in at 273g! How long will this take to dry?

After recrystallization in IPA the yield was 94g of the hydrochloride. Not as good as it looked when wet, but it's the best Al/Hg I've seen. Stirring it would bring it up more I guess. The amount of MeNH2-HCl used would be somewhere around 150g. It was wet from EtOH recrystallization, so enough to make the pH of the NaOH solution neutral was thrown in. (and another half spoon thrown in for good

To me it looks as if this bee has been shopping around.

  • Osmium's Al-Hg with a little more Al to compensate for the lack of stirring. Chose to use IPA instead of 95% EtOH because of less water involved.
  • The cleanup was lifted from Ritters writeup in TS2, no toluene was availiable but Xylene apparently worked well. Apparently this type of cleanup leaves very pure crystals, but recrystallization from IPA is probably a good thing anyway. Another thing to remove possible Hg salts