OK, I did two different runs.
Run A
The same procedure as Bandil described. The Zn-salt filtrate was stored in the fridge (4°C) for two-three hours (there is no need to do this; I just had other things to do as well, but this "storing" might have had his effect on the reaction mixture content).
The acidified Zn-filtrate was reduced in volume and extracted with DCM; the organic layer was dried over Na2SO4 and a sample analyzed with GC/MS. The following substances have been retrieved (percentage stands for AREA% of TIC (40->500 amu)):
- phenyl-2-nitropropane (62.81%)
- phenyl-2-nitropropene (7.62%)
- phenyl-2-propanone (5.70%)
- N-formyl-amphetamine (4.69%)
The aqueous layer was made alkaline with KOH and allowed to cool down to RT. The turbid mixture was extracted with DCM (3 x 75 mL) and the combined organic layers dried over Na2SO4 to yield 2.7 g of a yellow oil. GC/MS analysis of this oil:
- amphetamine (40.66%)
- phenyl-2-propanoxime (28.87%)
- amphetamine-P2P imine (18.66%)
- amphetamine-benzaldehyde imine (1.04%)
- phenyl-2-propanol (1.02%)
Detected as well: N-[2-(1-phenylpropyl)]-ethylidenimine (cf Comments & Additional Information at the end of this post).
Run B
A second run was performed. The working concept is still the same, but conclusions I made based on the results from A enforced me to modify the procedure.
150 mmol phenyl-2-nitropropene (recrystallized twice after synthesis, high GC purity) was suspended in a mixture of 150 mL IPA and 50 mL dH2O. The suspension was stirred vehemently while 2 mol eq sodium borohydride was added over 20 minutes. The reaction was allowed to continue for another 45 minutes, after which GAA was added to neutralize residual NaBH4; the liquid phase was filtered off and the borate cake rinsed with a small amount of IPA/dH2O. The filtrate was treated with table salt and the IPA layer isolated.
The IPA layer was introduced in a three-necked 1-L RB equiped with stirbar and reflux condenser, which was charged with 10 mol eq untreated Zn powder (relative to the amount of phenyl-2-nitropropene). A total of 3 mol eq HCOOH (98%, relative to the amount of Zn) was added over 20 minutes.
The Zn-salts were filtered off and the filtrate reduced in volume. Its colour was a deep red by now. It was acidified with 10% HCl and extracted 2 x 150 mL DCM. The combined DCM extracts were dried over Na2SO4 and analyzed via GC/MS:
- N-formylamphetamine (48.20%)
- phenyl-2-propanoxime (17.88%)
- phenyl-2-propanone (5.94%)
- amphetamine-P2P imine (5.96%)
Also detected (in trace amounts): phenyl-2-nitropropane, amphetamine.
The aqueous layer was made alkaline by addition of KOH and extracted with 3 x 100 mL DCM. After evaporation of the solvent, a yellowish oil remained. Weight: 11.1 g amphetamine freebase. GC/MS analysis indicated the presence of phenyl-2-propanoxime and amphetamine-P2P imine, but their joint AREA% counted for less than 2%. The "missing" AREA% is taken by amphetamine.
Neutralizing amphetamine by 0.5 mol eq concentrated H2SO4 in two times its volume IPA. Some IPA and ether added after complete addition of the sulfuric acid. Crystals give a fairly pure look...
Comments & Additional Information
N-[2-(1-phenylpropyl)]-ethylidenimine
Where the fuck does this come from? This substance is reported to be found when amphetamine is dissolved in EtOH (it has m/z 70 - N-ethyl, like MDEA, has m/z 72). However, I did not use EtOH as solvent, but IPA. Contamination from the IPA? Possible, but unlikely (certainly since GC/FID did not indicate its presence when analyzed). However, there is something else that bothers me: amphetamine-benzaldehyde imine. This imine is formed by condensing amphetamine with benzaldehyde, an impurity that often is found in nitro-related reductions (concerning PEAs). But benzaldehyde? Yes, could be an impurity from the initial synthesis step, but I still have difficulties believing that this substance has survived two recrystallizations and NaBH4. Yes, there probably are some molecules of benzaldehyde still wandering around from the initial step, but it seems to much to be true. However, coming back to the "EtOH-artifact"... where does EtOH come from? Nowhere... The artifact is formed by condensation of P2P (also present as impurity) and ethylamine. Ethylamine? Yes, ethylamine, which originates from nitroethane. You see, I suspect that phenyl-2-nitropropene is partially hydrolyzed in benzaldehyde and nitroethane, after which nitroethane is reduced to ethylamine and benzaldehyde condensed with amphetamine. I have no other explanation for the presence of this artifact.
High mol eq of NaBH4 and Zn
You might notice that I use more NaBH4 and Zn than has been reported. However, using only a slight excess of sodium borohydride resulted in much unreacted phenyl-2-nitropropene. The latter disturbs your reaction in the Zn/HCOOH mediated reduction of phenyl-2-nitropropane. The nitropropene is reduced to phenyl-2-aminopropene. Sounds good, no? No! This compound will tautomerize and hydrolyze in an acidic environment (clearly present) and result in phenyl-2-propanone. This is a side reaction which is present in LAH reductions of phenyl-2-nitropropene as well. However, I believe that this side reaction has to be minimized (since P2P can condense with amphetamine), hence the excess of sodium borohydride.
I also opted for a large excess of Zn. It seemed to me that their was still alot of phenyl-2-propanoxime left, which usually indicates an incomplete reduction reaction. Since Zn is rather cheap, I just threw in 10 mol eq. As you can see in my results, there is a marked difference in yield (A: 2.7 g oil from 50 mmol phenyl-2-nitropropene, impure amphetamine - B: 11.1 g oil from 150 mmol phenyl-2-nitropropene, rather pure amphetamine)
Compared to LAH
I haven't performed the reduction of phenyl-2-nitropropene with LAH myself, but the literature describes that large excesses of LAH have to be used or phenyl-2-propanoxime will be the major substance (and NOT amphetamine!). Considering the price and availability of LAH, I guess it is time to wave LAH goodbye. Bye bye...