MDMA via direct aminomercuration of safrole

This file is a part of the Rhodium site archive. This Aug 2004 static snapshot is hosted by Erowid
as of May 2005 and is not being updated. > > Back to Rhodium Archive Index > >

[www.rhodium.ws] [] [Chemistry Archive]

 

 

Discussions on the Aminomercuration of Safrole

Synthesis of MDMA from Safrole by Mercury(II)
Mediated Addition of Methylamine to Allylbenzenes

HTML by Rhodium

A Synthesis of MDMA from Safrole by Aminomercuration

by Zbyszek Szczesny (ziggi@polbox.com) - July, 1997
According to my experience I would suggest a different method of MDMA synthesis - in my opinion simpler than what I see on the web. The biggest advantage of my "developement" is that all processes are conducted in aqueous solution under mild conditions instead of in organic solvents, reducing the risk of fire or explosion. In fact, you need some toluene at the very end - to extract the final amine - but this is normal and brings no difficulties. Reduction with sodium borohydride is very elegant but pH control is a must with the buffer used for it.
Yield is very good - total over 90%. Because aminomercuration is a very simple and mild reaction it should present no difficulties for an experienced chemist, but since mercury is very poisonous I will not give all the details to the broad public. I believe collegues with a chemical background will know how to perform these reactions correctly. I strongly recommend against anyone without training in chemistry performing these reactions. It could result in serious injury or poisoning.

Scheme:

Step 1: A Markovnikov-order aminomercuration.

Safrole → Intermediate 2-Methylamino 1-mercuric salt

Step 2: Reduction with sodium borohydride, producing elemental mercury which is easily removed.

Intermediate salt → MDMA + Hg

References:

1. Lattes, Perie; Compt. Rend. Acad. Sci., Ser. C, 262, 1591 (1966)
2. Tetrahedron Lett. 5165 (1967)
3. Bäckvall, Åkermark; J. Organomet. Chem. 78, 177 (1974)
4. Gasc, Perie, Lattes; Tetrahedron 34, 1943 (1978)
5. Barluenga, Villanana, Yus; Synthesis 375 (1981)
6. Koziara, Olejniczak, Osowska, Zwierzak; Synthesis 918 (1982)

Safrole is a reactive unsaturated compound which goes mercuroamination when trated with methylamine in a presence of mercuric acetate. The resulting 3-(3,4-methylenedioxyphenyl)-2-methylamino-1-mercuroacetate may be easily reduced with sodium borohydride (in slightly acidic buffer). The mercury is sedimenting as a free metal. MDMA is obtained with a total yield over 90%. The allover picture is very much more convinient than any other synthetic path.
Further Notes by Ziggi (Jan 1998)

1. As most of you probably know oxidation of alkenes with mercury(II) salts is one of the most important pathways to alcohols. Please notice that this reaction called 'oxymercuration' has the following properties:


• High yield.
• 100% Markovnikov orientation.
• In some cases some organomercurial byproducts are formed.


2. Most popular oxymercurations are in fact 'hydroxymercurations' what means that the reactants are an alkene, a mercury(II) salt (usually mercuric acetate) and water (which is also the solvent).


3. But if water is replaced with alcohol the resulting product is corresponding ether. The reaction goes well with primary or secondary alcohols but when a tertiary alcohol is used the mercure salt must be highly reactive mercuric trifluoroacetate.


4. If the oxymercuration is carried out in a presence of a hydroperoxide instead of water (or alcohol), the product (after demercuration with NaBH₄) is an alkyl peroxide (→ peroxymercuration).


5. The latest example is the best to explain aminomercuration idea. Those interested may look into:
   
   • Ballard & Bloodworth, J. Chem. Soc. C 945 (1971)
   • Bloodworth & Loveitt, J. Chem. Soc., Perkin Trans. 1 1031 (1977)
   • Bloodworth & Courtneidge, J. Chem. Soc., Perkin Trans. 1 3258 (1981)
   • Bloodworth & Courtneidge, J. Chem. Soc., Perkin Trans. 1 1807 (1982)
   • Sokolov & Reutov, J. Org. Chem. USSR 5, 168 (1969)
   • Schmitz, Reiche, Brede, J. Prakt. Chem. 312, 30 (1970)
   
   I said it is a good idea because the reagent is not the solvent in this case!


6. It is necessary to comment the problem of mercuric organocompounds which may occure when the aminomercuration is carried out with aromatic compounds (like safrole undoubtly is...). I would say there is not many things to worry about in this point. Typical way of cleaning the final product (the amine) is a measure good enough to avoid any significant mercuric pollution: extraction with ether, steam distillation, fractional vacum distillation, cristallisation of the sulfate from isopropanole.


7. It is important to remember that THF is the necessary co-solvent used in amino/oxy-mercuration!

Aminomercuration-demercuration Discourse at the Hive:

ReFlux

So the subject of hypothetical Aminomercuration-demercuration of our favourite olefin safrole has been widely discusssed on this board. I'd now like to get to some specifics, so here are a few questions that are on my mind:

1. What is the ideal solvent for this reaction? Most journal refs call for THF, but the problem with THF is its perpensity for formation of explosive peroxides upon distilling that can make it go BOOM! Also, it is relativley expensive. How about 1,2-Dichloroethane (DCE) or methylene chloride (DCM)? Factors to consider are solubility of Safrole in the solvent, solubility of the Amine source in the solvent, and non-reactivity of the solvent.
2. For creation of secondary aminated product (MDMA) what form should the amine source (MeNH₂) be in? The journal refs call for it to be used in aq. solution, but can a solution of the hydrochloride salt (MeNH₂·HCl) be used somehow?
3. For creation of primary aminated product (MDA) what amine source can be used? Ammonum Acetate is not very soluble in THF or DCE, however, this should only slow down the speed of the creation of the organomercurial complex. More important to consider is will the formed complex dissolve in these solvents? What other sources for NH₃ can be used? Perhaps aq. ammonium?
4. What species of Mercury(II) salt should be used? Mercuric Acetate is widely used in oxymercuration reactions, but is less reactive than Mercuric Nitrate or Chlorate. It appears that speed of reactivity with amine source to form organomercurial complex is the only main issue, and as such, can the acetate salt be used with MeNH₂?
5. What is the best procedure for issolation of the resulting amines (MDA/MDMA)? It is important to carefully issolate the product from the rxn. mix so as to avoid any possible mercury contamination. Are standard solvent extractions, acid-base washings enough?
6. How to issolate pontential anti-Markownikoff aminated product? According to some journal refs, though most of the product from these rxns will proceed with Markownikoff addition of the amine, some anti-addition products can occur. How can these be seperated/issolated from the product ?

The refs that I have looked into for researching this method are:
One-step synthesis of N-substituted α-methylphenethylamines via aminomercuration-demercuration
Ronald C. Griffith, Robert J. Gentile, Thomas A. Davidson, and Francis L. Scott
J. Org. Chem. 44, 3580-83 (1979)
The question of the reversibility in the aminomercuration of olefins
José Barluenga, Julia Pérez-Prieto, Ana M. Bayón and Gregorio Asensio
Tetrahedron 40(7), 1199-1204 (1984)

Labrat:

A while ago I got interested in that reaction too. I did a little reading and would like to share my thoughts:

1. THF is probably the best solvent you can get. If you want to substitute, why don't you use diethyl ether? That's an ether too. I wouldn't use chlorinated hydrocarbons like DCM or DCE, cuz they might react with the Hg-salt. Distilling THF is not a problem if you throw in some reducing agent to take care of the peroxides, and then distilling.


2. I've asked the same questions a while back. Ziggy told me that you have to use methylamine freebase.


3. For primary amines, why don't you use benzylamine? You'll make the N-benzylamine, which can be split off by hydrogenolysis. This aminomercuration is actually suited better to make secondary amines, like MDMA.


4. After the aminomercuration, you'll pour in some NaOH solution to make the mixture basic, then reduction with NaBH₄. This'll reduce the Hg²⁺ to Hg(l). Mercury isn't soluble in water, so you can separate the layers. I guess acid-base extraction are sufficient to prevent mercury-contamination.


5. This bothered me too, until I read in PIHKAL, that the anti-Markovnivkov product, 1-(3,4-methylenedioxyphenyl)-3-propylmethylamine, is a known under the name GAMMA and is an active compound. So why bother removing it?

I hope I've helped you a little with this very interesting reaction, if you find anything good, keep me posted OK?

Reflux:

I really like your benzyl-substituted amine idea, but I think you meant Aniline for MDA as Benzylamine gives MDMA.
As for the anti-markownikoff product, GAMMA is the MDA anti-addition product (gamma-3,4-methylendioxy-phenylproplyamine), and though it is active, it didn't seem to have a desirable effect (see PIHKAL). Any info on the MDMA anti-product? In any case I think we should think of ways to remove them.

From Pihkal #100:
The n-propylamine counterpart (as if one were to move the amine function the other direction, from the beta-carbon to the gamma-carbon of the three carbon chain of the amphetamine molecule) is gamma-3,4-methylenedioxyphenylpropylamine or 1-amino-3-(3,4-methylenedioxyphenyl)propane, GAMMA. The hydrochloride salt has a mp of 204-205°C. At oral levels of 200 milligrams there was some physical ill-at-ease, possible time distortion, and a feeling of being keenly aware of one's surroundings. The duration of effects was 4 hrs.

Labrat:

Reflux: no, I didn't mean aniline, I did mean benzylamine. Hydrogenation with Pd/C is known to split off benzyl groups, just check out the literature to remove benzyl groups from benzyl alcohols. What GAMMA is gamma-3,4-methylendioxy-phenylproplyamine, not the methylamine? Damn, my mistake. But I still think it's active. And I would just use it as is. Otherwise you'll be forced to use chromatography to separate the anti-Markovnikoff from the Markovnikoff compound. Ziggy is indeed the guy who posted the aminomercuration method for MDMA.
I've done some reading too. I've read the JOC 44, 3580 (1979) again and found some interesting leads.
It's stated that: "In all cases, rapid formation of amine-mercury complexes was observed prior to addition of olefin". Hmm, what would happen if one would try aqueous MeNH₂ here? Will there be formation of amine-mercury complex? If there is, one could use aqueous methylamine. You probably have to use Hg(ClO₄)₂ instead of Hg(NO₃)₂. They also state: "We have observed that the aminomercuration can be performed in aqueous or anhydrous THF...this means hydroxide ion does not compete successfully with ethylamine as a nucleophile for the intermediate mercurinium ion or a direct addition of amine-mercury complex is involved". Hmm, I'm guessing that you could probably use aqeous methylamine solution.
Then I also read another good article: Chem. Het. Comp. 11, 4 (1975)
This is actually about intramolecular not intermolecular aminomercuration, but they give some properties of the reaction:
→ Nature of the anion on Hg is important
HgCl₂ > Hg(OAc)₂ > Hg(NO₃)₂ = Hg(ClO₄)₂
→ The reaction has an ionic character:
first a pi-complex between the olefin and the mercury salt is formed, then the amine adds to give the Markovnikoff addition product
→ reduction with NaBH₄ has an radical OR an ionic mechanism. If the mechanism is ionic, then during the reduction aziridinium ion is formed, which explains the presence of the two isomeric amines.
Remember when reading this article, that intramolecular aminomercuration is a lot easier to accomplish then intermolecular aminomercuration and that with intermolecular aminomercuration more hydroxymercuration occurs.
Well, that's it for today. I'm gonna find some more refs on this very interesting reaction and would like to communicate with you on it.

Reflux:

Labrat: I'm confused now. Aniline has a benzyl group doesn't it? (just one less methyl group before the amine when compared to Benzylamine) Or is a Benzyl group the benzene ring plus a single methyl (essentially a toluene less one H)? Please clarify.
As to rxns. with aq. MeNH₂, I too believe that it would work fine. In J. Org. Chem. 44, 3581 (1979) they used aq. EtNH₂ with m-methoxyallylbenzene & Hg(ClO₄)₂ and produced then ethylaminated product in high yields (87%).
Can one make an aq. MeNH₂ sol'n from MeNH₂·HCl? If it were to be disolved in H₂O, and then the sol'n basified with NaOH, would that not result in MeNH₂(g) which would stay disolved in the H₂O plus NaCl?

Labrat:

Reflux, aniline is phenylamine, benzylamine is like toluene with an -NH₂group substituted for a -H! Basic chemistry me boy!
Using aqueous methylamine is indeed a very viable option, although you have to heat the mixture of mercury salt and aqueous MeNH₂ to 60°C to expedite formation of the amine-mercury complexes. If the Labrat remembers correct, the bp of an 40% aqueous solution is about 50°C (correct?). I wonder if refluxing the amine-mercury mixture will take care of the complex formation. Have any thoughts on this?
Making aqueous methylamine is easy! Just put your MeNH₂·HCl in a flask, attach a distillation setup to it and fill a dropping funnel with 50% aqueous NaOH. Now slowly drip the alkalic solution on the salt to create MeNH₂ gas. Now bubble the gas into water until it has gained the proper amount of weight to get aqueous MeNH₂! You could neutralise it as you described, but you'll have NaCl contamination. I don't know if this will affect the yield.

007:

I previously posted a lot of ref and notes from Ziggy on the rxn on this board. elf has posted attempts with aminomercuration but with the MeNH₂·HCl salt which is definitely bad. I noticed that in one of the major ref for this rxn in JOC uses ethylamine as one of their variable amines and in methods they mention that it is used as a 70% solution (but i dont remember if it was aqueous of alcoholic) but this example made me think that aq methylamine would work. question is: is 40% enough and will much be lost during reflux. ziggys suggestions to over come this problem was doing rxn in a presurized container of sorts.
i think this rxn has a lot of potential, someone just has to ave the balls to dream about it a few times to perfect a procedure. my guess is that Hg(OAc)₂and HgNO₃ salts will work, and that HgNO₃ and Hg(ClO₄)₂ would be best (based on the articles i have read-they seem to use these the most).

Reflux:

I don't think that the conc. of the aqueous amine (ethyl-methyl) solution is that important. 70% aq. EtNH₂ & 40% aq. MeNH₂ are standard concentrations and that is probably why the JOC paper used the 70% soln.
As for the heating of reaction mix to 60-65°C to "speed up the formation of the intermediate organomercurials", I believe that here again the researchers in the JOC paper chose this temp as it just below the bp of THF (66°C). The paper did not call for refluxing the mix, but simple stirred heating.
The paper also indicates that "In all cases, rapid formation of amine-mercurcy complexes was observed prior to the addition of the olefin substrate. The rate of aminomercuration proved to be a sensitive function of the mercuric salt employed."
This indicates to me that there is no need for any refluxing or pressurized/sealed reaction vessels. The amine source is converted almost immediately into an amine-mercury complex which is dissolved in the THF. Simply add the 40% aq. methylamine to the stirring Mercuric salt in THF soln. Then add the Safrole, heat to say 60°C, and leave stirring for prolonged period of time (72-96 hrs.) until all of the safrole is converted to an organomercurial intermediate.
As for the species of Mercuric salt used, only dreaming will tell! As far as I can tell, Hg(NO₃)₂ would be ideal, but Hg(OAc)₂ is easier to aquire. Also, no need to worry about the hyrdated state of your Hg salt (mono-hydrate, tri-hydrate etc.) Use whatever type is easiest/cheapest to aquire and then adjust the amount used in reaction so that it is used in the correct molar ratios.
P.S. I looked up the entry for 40% MeNH₂(aq) in the Acros catalog, and they gave bp 48°C for it.

Labrat:

ReFlux, thanx for the comments. It's good to discuss this with a fellow chemist, cuz it seems we agree on a lot of things here. So we should definitely use 40% aqueous methylamine and we don't have to heat it to reflux (but I think it's better to speed up the complex formation).
I think that when using aqueous methylamine it's better to use Hg(ClO₄)₂ than Hg(NO₃)₂ and you can definitely forget Hg(OAc)₂. That anion is very important for the course of the reaction. That anion is the nucleophile that has to compete with hydroxide and methylamine nucleophiles, so you'd better choose the right anion. It's easy to prepare those salts from HgO, just by dissolving this mercuryoxide in e.g. perchloric acid you'll end up with Hg(ClO₄)₂. I'll try to find the intricate details for this.

Reflux:

In Tetrahedron 40(7), 1199-1204 (1984) they discuss reversibility in aminomercuration. They found that aminomercuration with Mercury(II) Acetate is an irreversible process while with Mercury(II) salts derived from strong acids (Hydrochloric, Nitric, Perchlorate etc) it is reversible. Could this pose a problem? The only main differnce is the rate of the rxn. which I think would be fine with acetate salt. Nucleophilicity is not an issue as hydroxy mercuration never occurs even with very weak nucleophilic amines and mercuric acetate. However, yeilds can be affected by weaker salts, though I wonder if that is because the reactions just aren't given enough time to complete?
In Tet. 34, 1943-1950 (1978) they discuss using a catalytic amount of concentrated (70%) perchloric acid sol'n with mercury(II) acetate which results in greater yields and much accelerated reaction time. My french really sucks, so If yours is any better maybe you could look up this ref. as it has a lot of info on the process and mechanisms of aminomercuration.
Also, won't using a strong acid based salt will result in highly acidic conditions in the reaction mix after aminomercuration is complete? The anion will turn the aq. sol. acidic will it not, which could damage the methylenedioxy ring structure, no?

Reflux:

Just one more point, I don't see the point in having a 2.67 X molar excess of amine to mercury salt as suggested by the JOC ref. Since in the first part of the reaction, there is a near immediate formation of organomercurial complex (before addition of the olefin) then what is the point in having all that highly nucleophilic (as in the case of MeNH₂) excess amine floating around in the reaction mix.
I would suggest equimolar or if not then just a slight excess of amine (1.1 X maybe? relative to the mercury salt), while still maintaining the excess of mercury salt relative to olefin ratio (1.5 X). What are your thoughts?
On a related note, how about using an aq. Ammonia sol'n for hypothetically creating MDA in this reaction. Any thoughts?

Labrat:

ReFlux - sorry it took a while to answer this, but my computer got fucked. I tried to post the answers Friday, but no good. I'll try it again today:

> Tet. Vol 34 (1978) pg. 1943-1950

This is a real goldmine! I'll translate the most important things for you:
They used 15 molar excess of water compared to the amine and got NO oxymercuration products!
They added the olefin to a aniline/Hg(OAc)₂ mixture in dry THF, then added water. They obtained uniquely aminomercurations compounds. When they added water to Hg(OAc)₂ in THF, then added the amine, the oxymercuration products were obtained. This means it's important to have the amine in the mix before the alkene is added.
Mercury has a greater affinity for nitrogen (N) than for oxygen (O). In the reaction conditions used, there is competition between two nucleophiles: the amine and water. Since no oxymercuration product was found, they concluded amines are better and stronger nucleophiles then water.
Aminomercuration is a compromise between the nucleophilicity of the amine and the weak stability of the amino-mercury complex. There's always complex formation, that's why you have to use at least 4 moles of amine to 1 moles of Hg-salt. In the case of intermolecular aminomercuration the stability of the complex formed between the amine and the mercury salt plays a dominant role in determining the rate of the reaction. In reactions with strongly basic amines (like ammonia, methylamine) a reaction time of several days is necessary to get a decent yield.
The absence of a reversible binding of the complex mercury salt-olefin makes nucleophilic attack on the carbon atom possible even for weak nucleophiles.
Now onto the questions:

> Also, won't using a strong acid based salt will result in highly acidic
> conditions in the reaction mix after aminomercuration is complete?
> The anion will turn the aqueous solution acidic will it not,
> which could damage the methylenedioxy ring structure, no?

No! The salt of a strong acid is by definition a very weak base! Theory me boy!

> I would suggest equimolar or if not then just a slight excess of amine
> (1.1 X maybe? relative to the mercury salt), while still maintaining the
> excess of mercury salt relative to olefin ratio (1.5 X). What are your thoughts?

In Tet. 34, 1943 (1978) they're talking about complex formation between the amine and the Hg-salt is a competing reaction with complex formation of the Hg-salt with the alkene. With aniline, two molecules of aniline complex with one molecule of Hg-salt. I don't know whether more basic amines like ammonia will complex with even more amines per Hg-salt. This means the excess amine is necessary. Otherwise the reaction will be slowed down.

> On a related note, how about using an aq. Ammonia sol'n for hypothetically
> creating MDA in this reaction.

Great idea! That way you won't have to go through all that trouble making aqueous methylamine. I guess it'll work. Otherwise, bubble dry NH₃ into a mixture of dry THF and a mercury salt, drip in the alkene. If an insoluble complex forms, add a little perchloric acid and watch it dissolve. The water in the mix is used to make the amine-mercury complex less stable.
I've done a little literature searching on aminomercuration and found some more refs:
Synth. Comm. 26, 4279-4288 + Synth. Comm. 1507-1516 (1996)
Here they're talking about creating an mercurinium ion between mercury and the olefin, which is trapped with a nucleophile, like amines. So no direct addition of the amine-mercury complex to the double bond as mentioned in JOC 44, 3580 (1979)
J. Organomet. Chem. 78, 177-84 (1974)
→ not so interesting, dealing mainly with the stereochemistry of aminomercuration
Tet. Lett. 51, 5165-8 (1967) + Compt. Rend. 262, 1591 (1966)
→ Using NaBH₄ instead of LAH produces less anti-Markownikoff compound.
Synthesis 375 (1981) + Synthesis 919 (1982)

• Reduction and deaminomercuration are always competitive processes
• Use of dichloroethane as a solvent for the aminomercuration reaction

Reflux:

I understand exactly what you mean about the excess of amine needed, since more than one amine complexes per mercury, even though only one gets added on, and since all the complex formation is at the begininng, excess of amine is required.
Now, as to the anion of the mercuric salt, I'm a little confused (picture ReFlux scratching his head!) If, for example, Mercuric Perchlorate is used; when it is dissolved in H₂O:THF sol'n. you have mercury ions and perchlorate ions, correct? Now when the amine is added and it complexes with the mercury, what happens to the perchlorate ions?
I too have faith in the Aq. NH₃ route and wish to use it as a test bed for initial dreaming about this reaction. (No sense in wasting hypothetical MeNH₂ on working out the bugs in our system, cause even in dreams, MeNH₂ is hard to get!!)
I'll check out the synthesis and synth. comm. refs that you posted. They both seem very interesting! You know I've been looking for an alternative to THF for this reaction from the begining!
Also, do you have any info on solubility of MeNH₂ and ammonia in THF/DCE? Ideas on where to look?

Labrat:

ReFlux - I've read the article you mentioned: Tetrahedron 40, 1199-1204 (1984)
And I guess you're right what the reversibility of the aminomercuration concerns: if you use the mercury salt of a strong acid, the aminomercuration is reversible.
The authors in this article mention the contradictory results of Perie and Lattes [Bull.Soc.Chem.Fr. 583 (1970)] and Griffith et al [JOC 44, 3580 (1979)] concerning the rate of aminomercuration with use of the various mercury-II-salts. I'd follow the instructions of Griffith et al, since they've tried it on allylbenzenes.
The authors state that: "..aminomercuration is reversible only when the mercury(II)salt derives from a strong acid. By the contrary the aminomercuration of olefins with mercury(II)acetate has been found an irreversible process which only leads to the kinetically controlled products." It's very worthwhile to consider using mercury-II-acetate. You'll have to expect a reaction time of several days/weeks to get a decent yield of product. Refluxing will speed up the process.
In the article I read that: "The rate of aminomercuration increases with the ionic character of the mercury salt and the polarity of the solvent, but the extent to which the deaminomercuration takes place largely depends on temperature and reaction time"
Hmmm, I wonder, what if we used mercury(II)acetate in combination with a little perchloric acid as done in Tetrahedron 34, 1943-50 (1978)? Probably the aminomercuration will proceed with a faster rate and higher yield. Well, let's do it then!
I'm very curious what will happen if somebody used the above mentioned protocol to brew some MDA. It's probably best to use dry ammonia for starters, you can always add a little perchloric acid (or water) if a insoluble complex precipitates. I'd certainly consider using mercury(II)acetate, since this gives irreversible aminomercuration.

Osmium:

2-acetamidohexane [N-(2-hexyl)acetamid]

To a ice-cooled solution of 64.8g (0.20mol) of Hg(NO₃)₂ in 100ml MeCN 16.8g (0.20mol) of 1-hexene are slowly added keeping the temp below 30°C. The clear yellow solution is stirred for 1 h at RT. With ice-cooling 200ml 3N NaOH are added, followed by a solution of 3.80g (0.10mol) of NaBH₄ in 200ml 3N NaOH. Again stir for another hour. Decant the aq/organic layer leaving the elemental Hg behind, saturate the aq. layer with NaCl and extract with Et₂O (2×100ml). The extracts are dried with Na₂SO₄, the solvent removed and the residue distilled. 22.8g (80%) amide, bp 93-96°C/1.6mmHg.
Ref: J. Am. Chem. Soc. 91, 5647 (1969)
Hydrolysis (use the crude amide):
20.0g (0.14 mol) of the above amide, 15.7g (0.28mol) KOH in 25ml H₂O and 62.5 ml ethyleneglycol are refluxed for 24 h with good stirring. After cooling to RT extract with ether (3×100ml), back-extract the amine with 3×100ml 2N HCl. From the ether 3.08g of unconverted amide are obtained. The acidic aq. extracts are basified with 30g NaOH and the amine is extracted with ether (2×150ml). The extract is dried with K₂CO₃, the ether evaporated and the residue distilled at bp 114-116°C, giving 7.50g (53%) amine (62% on reacted amide).

Ritter:

Amidomercuration - Nitriles work too!
Many moons ago an associate successfully performed an amidomercuration with acetonitrile as source of amide. Technically I guess this is completely different than a true amidomercuration reaction, however the reaction did indeed produce a good yield of N-acetyl MDA which is indeed an amide and was hydrolyzed to MDA in a good yield. This reaction is detailed in M. V. Smiths' book. Safrole is reacted with acetonitrile and Hg(NO₃)₂ then demercurated with NaBH₄. The resulting acetylamide was hydrolyzed with 4 N HCl. This reaction is good for small batches - scaling above 100 mmol can result in a run-away exotherm during the mercuration. Not to mention no one should be messing with this toxic shit when there are better methods available anyway.