Vanillin chemistry: 3,4,5-trimethoxyphenethylamine

Welcome to my second vanillin chemistry thread,
not to be confused with my previous and, at the time of writing this, still ongoing piperonal thread.

In this one we will concentrate on another synthesis that is frequently encountered in the clandestine context - the legendary psychedelic colloquially known as mescaline. As far as my undertakings go, this is the lengthiest and most difficult one so far and there is no doubt in my mind that I will be discovering new and innovative ways to fail on every opportunity. Or, at the very least, I may be the first to document those failures online. With pictures. But that's not the point. The point is to be that much better in the end.

"Failure is success in progress."
– Albert Einstein, apparently


The roadmap

This is all subject to change, as I cannot anticipate every obstacle with my current knowledge and experience.

Step 1: Vanillin is brominated to yield 5-bromovanillin
Step 2: Dehalogenation and subsequent methoxylation to yield syringaldehyde
Step 3: The 4-hydroxy is methylated using iodomethane to yield 3,4,5-trimethoxybenzaldehyde
Step 4: The aldehyde is reduced using sodium borohydride to yield 3,4,5-trimethoxybenzylalcohol
Step 5: The alcohol is brominated using phosphorus tribromide to yield 3,4,5-trimethoxybenzylbromide
Step 6: The bromide is reacted with sodium cyanide to yield 3,4,5-trimethoxyphenylacetonitrile
Step 7: The acetonitrile is reduced using lithium aluminium hydride to yield 3,4,5-trimethoxyphenethylamine

Failing this, another attractive (sounding) route is via the Henry condensation of the benzaldehyde with nitromethane.

And without further filibustering, let us embark on this magical journey of learning and fuckery.



Step one: 5-bromovanillin

In a 500 mL RBF, vanillin (43.2 g, 284 mmol) was dissolved in 225 mL glacial acetic acid and, with strong stirring, bromine (50 g, 626 mmol) was added dropwise while maintaining the temperature well below 20°C. After the addition, the reaction mixture was stirred for a few more hours.



The mixture was then poured in approximately 600 mL of ice cold water. The resulting precipitate was filtered and washed with cold water followed by 200 mL of -22°C ethanol and dried to yield 65.1 grams of cream-colored, slightly acetic-smelling crude product. The crude product was recrystallized from ethanol, resulting in 57.6 grams of off-white 5-bromovanillin with a sharp melting point of around 165°C and a wonderful, rich vanilla aroma.






Well, that went great! If this reaction is anything to go by, the rest is going to go just f–
Step two: Syringaldehyde

This is where it starts to get difficult. Apparently. Anhydrous methanol and sodium metal are generally employed to form methoxide in situ, which then gets exchanged with bromine forming sodium bromide and syringaldehyde or, 4-hydroxy-3,5-dimethoxybenzaldehyde. The reaction is catalyzed by a Cu(I) halide which is reportedly stable when dimethylformamide is used as the solvent. At the moment I do not have DMF so I've decided to explore other options in the meantime.

My first attempt was to try DMSO in place of DMF, solely on the grounds that copper(I) bromide seemed soluble in it. Sodium methoxide was prepared by adding freshly purified sodium metal (1.64g) to 32 mL methanol and refluxing for 30 minutes. 10 grams of methanol was distilled off and a solution of CuBr (0.97g) and 5-BrV (5.73g) in 24 grams of DMSO was added in one go. More methanol was distilled off until the temperature of the mixture reached ~100°C and the heating was discontinued.

Solution of 5-BrV and CuBr in DMSO


DMSO solution added to the methoxide solution


Final appearance of the reaction mixture



The reaction mixture was poured into a mixture of conc. HCl (22 mL) and snow (50 g). This was extracted with ethyl acetate (3x25 mL) and the solvent was distilled off in a hot water bath. The residue was recrystallized from ethanol to yield 4.02 grams of sauerkraut-smelling crystals with the melting point of 5-bromovanillin. Evaporation of the mother liquor yielded more of the same. DMSO seems like a no-go.


Crystals of unreacted 5-BrV


Above crystals exhibiting a strange color when dissolved in hot ethanol. Went back to yellow on cooling.



For the second attempt, because I only have a few grams of sodium at the moment, I attempted to use a sodium methoxide solution prepared by sequestering water from a solution of sodium hydroxide (20g) in methanol (100g) over 48 hours using 100 (50+50) grams of 3Å molecular sieves. This time, instead of DMF or DMSO, the reaction would take place in an (hopefully) approximately 5M solution of sodium methoxide in methanol.

According to this reference, a stable complex would form when Cu(I) halide and ethyl acetate are employed in these conditions.

EtOAc (2.51 g, 28 mmol), 5-BrV (10 g, 43 mmol), CuCl (0.87 g, 8.8 mmol) were added to 5M MeONa in MeOH (87 mL) and the mixture was refluxed for 16 hours.

Freshly prepared copper(I) chloride


Beginning of reaction


>10 minutes into the reaction


8 hours into the reaction


Work up was the same as before. Once more, only unreacted 5-BrV was isolated. It seemed clear from the color change of the reaction mixture that the complex had decomposed. This was likely due to bad sodium methoxide and there was still hope for the ester-catalyzed approach.

Unreacted 5-BrV



It's getting late so I'll wrap this up for now. The third attempt yielded some encouraging results and I'm currently in the middle of my fourth one. More on those soon.

[Edited on 13-4-2021 by Benignium]

Thank you everyone for the wonderful responses!
zed - Immensely interesting as always! Your friend sounds like one hell of a chemist. I'll settle for any identifiable amount.
monolithic - Useful indeed! Thank you for bringing this to my attention!

Now, for the third attempt at producing syringaldehyde.

While drying more methanol by refluxing it with magnesium, a thought occurred to me. Perhaps I could use magnesium to more reliably prepare the sodium methoxide for this reaction, without using up my remaining few grams sodium. It's not exactly smart as magnesium is not that much cheaper than sodium, and could be used to convert sodium hydroxide to metallic sodium with relative ease. It could save me some time, though, as I have plenty of it on hand. And besides, chemistry is an experimental science.

Let's go over a few equations. Feel free to correct me if I'm wrong.
NaOCH3 + H2O <-> NaOH + CH3OH
Mg(OCH3)2 + H2O -> MgO + 2CH3OH
Mg(OCH3)2 + 2NaOH -> Mg(OH)2 + 2NaOCH3

Adding magnesium metal to a solution of sodium hydroxide in methanol almost certainly wouldn't work, and didn't (I tried this later on), but magnesium dimethoxide should be an entirely different story. Now let's find out if there is in fact pudding in this pudding.

Magnesium (12.27 g, 0.5 mol) was added in a 250 mL Erlenmeyer and the vessel was filled to approximately 150 mL with methanol. The flask was fitted with a 300 mL coil condenser and placed in a lukewarm water bath with strong stirring for a couple of hours.

Magnesium metal reacting with methanol


Meanwhile, sodium hydroxide (20 g, 0.5 mol) was dissolved in 100 mL of methanol.

The solution of sodium hydroxide was added to the stirred magnesium methoxide still in its water bath. Only mild exotherm was observed, and the rate of addition was controlled sufficiently to avoid the formation of a temporary gelatinous precipitate which hindered agitation.

A droplet of methanolic NaOH entering the methoxide mixture


The mixture was vacuum filtered from a covered Büchner funnel to retain 100 mL of a clear-ish methanolic solution of what was hopefully sodium methoxide in sufficient concentration (3-5M). So far so good. The whole preparation took about 12 hours from start to finish. And now for the actual reaction.

Ethyl acetate (4.91 g), 5-bromovanillin (11.40 g) and an unrecorded yet appropriate amount of copper(I) chloride were added to the 100 mL of sodium methoxide solution prepared earlier. The mixture was kept at reflux for 28.5 hours.

Beginning of reaction


4.5 hours into the reaction


13.5 hours into the reaction


Cessation of heating and stirring


Cooled reaction mixture


During this work up I tried some new things, and a lot of it was completely useless. The gist of it, though, is that after dilution, acidification, and filtration of the reaction mixture, most of the unreacted 5-BrV was first extracted from the solid by dissolving everything in ethanol, filtering and adding water to the filtrate until precipitation of the 5-BrV occurred. After filtering, the filtrate as well as the remainder of the reaction mixture were stripped of most alcohol by distillation in a boiling water bath, filtered again and extracted with ethyl acetate. The ethyl acetate extracts were dried and the residues pooled and extracted with copious amounts of boiling heptanes. After chilling the alkane solution to -20°C, the solvent was decanted off, and the tan solids were dissolved in minimal ethanol and left to evaporate slowly in a loosely covered beaker. Some initial crystals formed with a boiling point of 5-BrV. Then, a different kind of crystal growth was observed. Thin, transparent plate-like crystals, with small amounts of coprecipitated impurity.

Acidification of diluted reaction mixture in progress


Acidified reaction mixture


Evaporation of ethyl acetate extract


Strange crystals



A melting point test was carried out. Some initial melting occurred at 74-75°C, 6-7 degrees below the melting point of vanillin (81°C), which is apparently a characteristic side product of this reaction. The majority, however, only started melting at above 100°C, with the whole sample turning to liquid by 104.9°C, 5-8 degrees below the reported melting point of syringaldehyde (110-113°C). Coincidence? I think possibly. The amount was also very little. Still, enough provocation for me to try again.












Right, so. Attempt number four. This one is a lot like the third attempt, with the exception that instead of 28.5 hours I'm going to let it go for a whopping week. I'll close with some pictures of the procedure so far and come back with another update once the week is up and I've gotten around to working up the reaction.

Magnesium in methanol


Filtered methoxide solution


Fresh copper(I) chloride


Beginning of reaction


58 hours into the reaction


82 hours into the reaction


103 hours into the reaction



Cheers!

zed - I'm primarily interested in syringaldehyde specifically because of its potential in trying out different 4-substitutuents. You're completely right in that the fewer steps the better.

njl - I don't think it does form an azeotrope, no. However, the idea is to react any water from the equilibrium with magnesium methoxide which, unlike sodium methoxide, reacts irreversibly. I neglected to actually explain what I meant by the equations. Apologies.

myr - You may be right. Methanol is an attractive choice of solvent for me, though, which is part of the reason why I moved on so quickly from DMSO. I'd describe the smell of that sample as that of vanilla-scented cardboard.

Newton2.0 - You're very welcome!

154 hours into the reaction


End of reaction. Strangely, the color had turned back toward a lighter green in the final hours after pausing the reaction.



After 164 hours of stirring and heating, a two hour pause and six more hours of stirring and heating after that, the reaction mixture was cooled, diluted, acidified and vacuum filtered to remove a brown copper-based precipitate. Everything else dissolved, which was a clear sign that no 5-bromovanillin remained in the mixture. The filtrate was extracted with DCM (3x30 mL) and the solvent was distilled off leaving 7.8 grams a striking reddish-orange residue with a melting point of 107.9-109.4°C.

Diluted and acidified reaction mixture


Diluted and acidified reaction mixture


Solids and filtrate


Flame test on the above solid material


First DCM extraction


Second DCM extraction


Third DCM extraction


Extracted reaction mixture and combined extracts


Saturated NaCl wash of the DCM solution


Residue after stripping DCM


Melting point test on the above residue











The residue was placed in a soxhlet apparatus and extracted with heptanes. After 24 hours, the hot extract was decanted off co-extracted precipitated impurities, heated once more to redissolve everything and slowly cooled to -10°C. The solvent was decanted off and the solids dried for a partial yield of 3.9 grams of a voluminous neon orange material. 3.4 grams of this material was recrystallized from aqueous methanol (~7.3% per weight) for a nearly quantitative yield of vibrant orange crystals.

Crude product loaded into a soxhlet apparatus


Crystals from boiling heptanes


Crystals from boiling heptanes. Considerably more orange in person.


Recrystallization from aqueous methanol


Recrystallized product. Quite close to the color of ammonium dichromate.



The reaction seems to have been successful, with a complete conversion of the starting material and an excellent yield. Granted, there is a lot of the product still stuck with the red impurity and the purification needs more work. I'm going to try out some things and repeat the melting point test for the cleanest sample I can obtain. While I doubt that it was necessary to let the reaction go for a week, it evidently still takes rather long to go to completion. I recall reading that the rate accelerates at a point. I am also glad to see that the magnesium dimethoxide trick seems to have worked.

I was contacted by The Vespiary user Sawdust and Honey who advised me to look into the route involving the Henry condensation of 3,4,5-trimethoxybenzaldehyde with nitromethane to yield beta-nitro-3,4,5-trimethoxystyrene, which could then be reduced to 3,4,5-trimethoxyphenethylamine in one step by the action of NaBH4 and CuCl2 in aqueous isopropyl alcohol, in very good yield. Reference for the reduction can be found on page 20 of this document.

This was an excellent suggestion, and I've decided to change the planned route accordingly.

That's all for now!

Quote: Originally posted by zed

Making Anhydrous Sodium Methoxide solutions, without Sodium Metal is possible, but as I recall, the guys have reported problems pulling it off. [Edited on 14-4-2021 by zed]

Quote: Originally posted by Benignium

I was contacted by The Vespiary user Sawdust and Honey who advised me to look into the route involving the Henry condensation of 3,4,5-trimethoxybenzaldehyde with nitromethane to yield beta-nitro-3,4,5-trimethoxystyrene, which could then be reduced to 3,4,5-trimethoxyphenethylamine in one step by the action of NaBH4 and CuCl2 in aqueous isopropyl alcohol, in very good yield. Reference for the reduction can be found on page 20 of this document.

Quote: Originally posted by dicyanin

The nitroalkane can than be efficiently reduced using a dissolving metal reduction, such as acidic (AcOH) Al/Hg. Or using zinc.

It's finally time for an update!

dioxine - I have stored the 5-bromovanillin in an airtight amber glass bottle at room temperature for several months and have not observed decomposition.


Step three: 3,4,5-trimethoxybenzaldehyde

In a 100 mL Erlenmeyer, 3.00 grams (16.5 mmol) of syringaldehyde was dissolved in 30 mL of dimethylformamide. To this were added, in order, 3.78 grams (27.4 mmol) of K2CO3 beads and 6.10 grams (43 mmol) of methyl iodide. The mixture was refluxed for 48 hours after which it was diluted with water to a total volume of approximately 80 mL and heated at near-boiling temperatures without a condenser for a few hours. Finally, the mixture was diluted to 100 mL and cooled to room temperature. After some time crystals formed and the mixture was vacuum filtered to yield 1.95 grams of dry crystalline solid. On further dilution, more crystal formation was observed and the mixture was extracted with DCM (no amounts recorded). The solvent was evaporated and the residue was dissolved in minimal boiling water. From the cooled mixture, an additional 0.96 grams of crystals were obtained by vacuum filtration. Melting point 70.6-73.0°C.

Yield: 2.91 grams (89.9%)









Step four: 3,4,5-trimethoxynitrostyrene

While the methylation step was delightfully straightforward, I found myself having trouble with this next one. The Henry condensation, as it's commonly referred to, of benzaldehydes with nitroalkanes is notoriously finicky when it comes to the base catalyst it requires. Too little, and the yield would suffer; too much, and the product would get polymerized into a red tar. Properties of the catalyst also play a big part in determining the extent of side reactions taking place, and therefore yield.
By now I had under my belt a good few experiments using ethanolamine in modest relative molar quantities with great success. But it became apparent that something was strikingly different with the 3,4,5-trimethoxy substitution pattern.

Experiment 1
One gram (5.1 mmol) of the benzaldehyde was dissolved in 10 grams of isopropanol, followed by 500 milligrams (8.2 mmol) of nitromethane, 300 milligrams of acetic acid, and 140 milligrams (2.3 mmol) of ethanolamine. The mixture was heated on a 100°C hotplate with stirring for 30 minutes, cooled for 30 minutes and diluted with ~15 mL of water. The resulting crystalline mass was filtered, washed with copious water and some cold methanol and air dried to yield 731 milligrams of crude product.

Experiment 2
0.94 grams (4.8 mmol) of the benzaldehyde was dissolved in 7.44 grams of isopropanol, followed by 390 milligrams (6.4 mmol) of nitromethane, 300 milligrams of acetic acid, and 131 milligrams (2.1 mmol) of ethanolamine. The mixture was heated on a 100°C hotplate with stirring for 40 minutes, cooled for 10 minutes and diluted with ~30 mL of water. The crystals were filtered, washed like before and air dried to yield 570 milligrams of crude product.

Cooled reaction mixture


Diluted reaction mixture


Dried crude product


At this point things seemed to have gone according to plan. The crude product was moved to storage pending purification. There are two things I've since learned about these nitroalkene compounds. One, they're unstable, and it's generally advised to store them in a refrigerator. Two, residual impurities from the condensation reaction promote degradation. I ended up needing to store the material for several months, and for much of this time it was kept at room temperature. Finally, when recrystallization from methanol was attempted, some of the material would not dissolve, and most of what did precipitated as a pale seemingly amorphous solid on cooling.

Precipitate from attempted recrystallization of impure 3,4,5-TMBA from methanol


With only a gram of the benzaldehyde that I intended to keep as such remaining, the remainder of my 5-bromovanillin (excluding a similar small sample) was used in an attempt to prepare more syringaldehyde, this time utilizing sodium metal in methanol, DMF, freshly prepared elemental copper powder and copper(I)iodide in place of the chloride. Somewhat surprisingly this experiment converted all of the 5-BrV to a waxy substance that had a broad melting point lower than that of vanillin and could not be resolved to isolate any desired product. The material was insoluble in water but completely extracted from an organic solvent by dilute bisulfite solution. Its components were never identified. I did, however, as a last-ditch effort try to methylate it hoping any syringaldehyde it may contain would convert and crystallize from boiling water. No such luck, and the material was eventually disposed of.

Product from a failed attempt to produce more syringaldehyde



The project was put on hold until I would get around to synthesizing more 3,4,5-trimethoxybenzaldehyde from scratch. Eventually I managed to directly purchase more and after months of nothing happening the experiments could continue. To preface this next one: going into it I was under the impression that I had used too much catalyst in the previous experiments.


Experiment 3
0.98 grams (5 mmol) of 3,4,5-trimethoxybenzaldehyde was dissolved in 5 grams of ethanol. To this were added 330 milligrams (5.5 mmol) of nitromethane, 100 milligrams of acetic acid and a solution of 41 milligrams (0.7 mmol) of ethanolamine in 1409 milligrams of ethanol. The mixture was stirred on a 85°C hotplate for 2 hours, cooled and pipetted into a 5% bisulfite solution. Copious precipitation occurred in the bright yellow mixture, which quickly cleared up considerably, simultaneously becoming colorless. Definite crystals of the nitrostyrene were observed which strangely also eventually disappeared into the colorless solution.

Reaction mixture shortly after adding the catalyst


Last of the precipitate in before disappearing into the bisulfite solution


I had theorized that by adding the reaction mixture into a dilute bisulfite solution, any unreacted benzaldehyde would dissolve as the adduct and only the nitroalkene would remain to be filtered out. Seeing the product, which should be practically insoluble as well as unreactive toward bisulfite, disappear without even its color remaining was baffling. It seems likely that the presence of other reagents from the condensation mixture is the cause. It was time to take a closer look at what had worked for other people in this specific case. I wanted to make ethanolamine work as the catalyst, but I also had cyclohexylamine which I recalled Alexander Shulgin using in PiHKAL. That would be my fallback.

I found two particularly reputable accounts of the 3,4,5-TMNS being formed in excellent yields.
The first one was from hyperlab.info user tetraedr.
The second one comes from carl on thevespiary.org.
Aside from the catalyst, the most notable feature shared by theirs, but not mine, was the use of acetic acid as the solvent. My current understanding is that any free acetic acid in the mixture will likely alter the reaction dynamics once water is produced as a side product and enables its behavior as an acid, but I cannot speak as to how exactly that might be in the grand scheme of things (I would love to know!). All I knew was that it would likely be different from using an alcohol solvent and that it shouldn't hurt.


Experiment 4
2.04 grams (10 mmol) of benzaldehyde was dissolved in a mixture of 5 grams of acetic acid and 1.47 grams (24 mmol) of nitromethane. With stirring there was added 0.90 grams (15 mmol) of ethanolamine. The mixture was heated on a 100°C hotplate for 40 minutes with stirring, after which the heating was switched off. Stirring was continued until about 7 hours later when the mixture was diluted with ~20 mL of water and filtered. The solids were thoroughly washed with ~100 mL of water, evenly wetted with exactly one gram of isopropanol from a pasteur pipette and sucked dry, followed by a final washing with water. After air drying, there was obtained 2.11 grams (84.8%) of bright yellow crystalline powder with a melting point of 117-123°C (lit. 120-121°C).

Reaction mixture of the fourth experiment after cessation of heating


Crude product from the fourth experiment



I had initially planned to use 90 milligrams of the catalyst, as it would have been roughly equivalent to the amounts of ethylenediamine and ethylamine used in the references, but I managed to mistakenly use a tenfold excess. The mixture seemed to darken relatively quickly but did not become dark red right off the bat as I expected. It was therefore allowed to react for the intended 40 minutes with surprisingly encouraging results; the melting point also indicated a very passable purity. At this point I took another look at PiHKAL entry #96 and saw that there, as well, cyclohexylamine was used in a similar excess, with acetic acid as the solvent. Based on what's presented here it seems that acetic acid as the solvent is in many cases essential for a fast and high-yielding reaction, but the required amount of a given catalyst might be independent of this fact.

Two more experiments were carried out. They were very similar to each other with the exception that one used ethanolamine as the catalyst while the other was done with the molar equivalent of cyclohexylamine instead.

Experiment 5 proportions
2.06 g 3,4,5-TMBA
5.15 g acetic acid
0.91 g ethanolamine
1.48 g nitromethane

Experiment 6 proportions
2.06 g 3,4,5-TMBA
5.15 g acetic acid
1.46 g cyclohexylamine
1.51 g nitromethane

Both experiments were heated on a hotplate set at 85°C for 20 minutes, followed by 20 minutes on the 100°C setting. The mixtures were then cooled and stirred for 11 hours before being worked up exactly as in experiment 4.

Results
Experiment 5
Yield: 75.6%
Melting point: 120-125°C

Experiment 6
Yield: 60.9%
Melting point: 118-124°C

This simply suggests that ethanolamine is likely the more efficient of the two.

Crude products from experiments 6 (left) and 5 (right)




Purification
The fifth and final step calls for a nitrostyrene that is very pure and in particular free of the benzaldehyde it was prepared from. Therefore the product was recrystallized twice.

The total of 5.55 grams of crude product from experiments 4-6 was placed in a 50 mL flat bottomed boiling flask along with a stir bar and 5 mL of methanol for each gram of material to be recrystallized. A 200mm Liebig was attached and through it additional methanol was added while refluxing the mixture until at 8.2 mL/g a clear mixture was obtained. The mixture was slowly cooled first to room temperature, kept in the refrigerator for a moment and finally brought to -10°C in the freezer. The cold mixture was vacuum filtered and the crystals were washed with three generous portions of water followed by a solution of 0.55 grams of potassium metabisulfite in 25.33 grams of water. Three more portions of water were used to rinse the crystals before air drying to obtain 5.11 grams of crystals.

The second recrystallization was carried out by similarly dissolving these crystals in boiling methanol and cooling down the mixture. The crystals were filtered and washed with water as before, followed by a solution of 0.25 grams of K2S2O7 in 9.75 grams of water, and rinsed with three portions of water. These crystals were then air dried, transferred to a small beaker in which they were coarsely ground in preparation for the upcoming reduction using a stir bar retriever. 4.68 grams (62.4% overall yield).

Combined crude product


Crystals from first recrystallization


Ground consistency of purified material to be used in the reduction step




Step five: 3,4,5-trimethoxyphenethylamine

This one is exceedingly interesting. A relatively recent development, this partially copper(II) chloride-catalyzed procedure represents an extremely attractive alternative to many previously tried and tested methods for reducing both the double bond and the nitro group of a nitroalkene in one go. While it strikes a great balance between accessibility, convenience, efficiency and hazardousness, it does present a formidable challenge for those that wish to master it.

To me, the considerations that seem most pertinent for success are the use of clean reagents and ensuring that the reaction does not run out of sodium borohydride, the reducing agent that powers both phases. To that end, the main steps that have to be taken seem to be rigorous purification of the nitroalkene and sufficient cooling of the reaction, particularly in the beginning. This is why the 3,4,5-trimethoxynitrostyrene in the last step was recrystallized twice and washed with bisulfite solution both times.

Apparatus
The reaction took place in a 250 mL RBF fitted with a Claisen adapter. In the joint above the flask there was a powder funnel through which the borohydride and nitrostyrene were added. For the addition of the copper catalyst, this powder funnel was replaced with a pressure equalizing addition funnel pre-loaded with the catalyst. In the sidearm of the Claisen there was fitted a 300mm coil condenser. The RBF was submerged up to its neck in 600 mL of cold water in a 1000 mL beaker which sat placed on a hotplate stirrer. A digital thermometer was placed in the water bath to monitor its temperature. Both the water bath and reaction mixture were stirred at 1150 rpm for the duration of the whole experiment.

Reaction
In a 250 mL RBF there was added 40 grams of isopropanol and 20 grams of water. Both the solvent and water bath were cooled so that when the reaction was begun by adding 4.47 grams (118 mmol) of sodium borohydride, the temperature of the water bath was at 4.1°C. After three minutes the hotplate was set to heat the water bath on the 550°C setting and the nitroalkene addition was started with the first of seven roughly equal portions. When the color from the first addition had disappeared, a second portion was added and the heating was switched off. All 4.68 grams (19.6 mmol) of the nitroalkene were added in 12 minutes, during which the water bath temperature had reached 32.5°C. The funnel and Claisen were rinsed with 20 mL of isopropanol and a small amount of water. Heating was resumed, and when a total of 19 minutes had elapsed, a further 1.61 grams (43 mmol) of NaBH4 was added, followed by 0.70 grams (4.1 mmol) of copper(II) chloride dihydrate as a 17% solution in 50% aqueous isopropanol. With sufficient heating, refluxing was continued for 110 minutes after which the heating was discontinued and the reaction mixture stirred for a moment longer until the evolution of hydrogen had died down. The water in the bath was replaced with cold water from the tap and the reaction mixture was allowed to cool to 19.7°C.

Reaction mixture before addition of catalyst


Reaction mixture following addition of catalyst


Reaction mixture refluxing


Reaction mixture refluxing


Black oily droplets appearing during reflux


Reaction mixture refluxing


Green color appearing in the lower phase at the very end of reflux (the blue parts seemed black to the human eye)


Workup
Vacuum filtration of the reaction mixture was attempted, but a dense black sludge blocked the filter paper instantly. To try and avoid having to clean out the Büchner, 50 grams of acetic acid was added straight in the funnel. With occasional light scratching of the filter paper, the filtration speed picked up. The filtered mixture was distilled until distillate was collected at 100°C. 28 grams of NaOH was then added as 16.9% solution in water, which brought the pH up to above 10 and caused the precipitation of a dense solid. After allowing it to cool back down to room temperature, the mixture was extracted with four 50mL portions of butyl acetate, the first of which caused a suspension of filmy solids that necessitated vacuum filtration of the whole biphasic mixture. The organic extracts were combined and it was noted that more of the filmy substance was produced when the extract came into contact with deionized water. Shaking the extract with ~100 mL of water and once more vacuum filtering the mixture removed nearly all of the bothersome solid, as well as the green color, leaving a clear dirty orange/brown organic phase above the colorless water layer. 5.51 grams of approximately 4.5M sulfuric acid was added and the mixture was shaken to move the product into the aqueous phase. The organic phase was washed with two 25 mL portions of water and the combined aqueous phases were washed with three 10 mL portions of DCM. The aqueous mixture was made strongly basic by adding 10 grams of NaOH as 16.9% solution and extracted with three 33 mL portions of DCM. The pooled extracts were directly distilled in a warm water bath to remove most of the solvent and finally a vacuum of 50-70 mmHg was pulled to remove as much of the remainder as possible, leaving in the flask a viscous amber oil weighing 3.81 grams.

Acidified reaction mixture being distilled


Sodium hydroxide being poured into distilled acidic reaction mixture


Precipitate from basic reaction mixture


Precipitate from basic reaction mixture


Bothersome filmy mess during solvent-solvent extraction


Filtered biphasic mixture


More filmy mess from adding water to organic extracts


Aqueous extract before washing with DCM


Aqueous extract after washing with DCM


Combined DCM extracts before removal of solvent


DCM extracts after atmospheric distillation


DCM extracts after pulling a vacuum for some minutes


The impure mescaline base was divided into two portions by pipetting 1.00 grams into a 25 mL beaker and transferring the remaining 2.81 grams into a 50 mL beaker using 29.98 grams of isopropanol. 5.53 grams of approximately 2.25M sulfuric acid was prepared, 2.99 grams of which was used to bring the pH of the larger alcoholic portion of oil to just below 7. 65 milligrams of oil from the smaller portion was required to back the larger portion up to above 7 once more. Finally, 0.95 grams of the acid was added to the remainder of the small portion to leave it ever so slightly basic.

Precipitate from dropwise addition of H2SO4 into soln. of amine in IPA





The neutralized mixtures were combined and the resulting slurry was poured on a ceramic plate and dried in the oven at 50-75°C. The dried light brown material was separated from a dark red oil that had mostly dried around it on the plate, and transferred to a 100 mL beaker where it was triturated under acetone using the stir bar retriever. This was then filtered and the powder washed with several portions of acetone. Once dry, the now-white crystalline powder weighed 3.67 grams.

Nearly dry impure product after oven drying


Acetone-washed product


To further purify the product, it was dissolved in a little over 15 mL of hot water and placed in a 56°C water bath. To this was then added 30 mL of hot acetone which caused the immediate formation of crystals. The contents were cooled and filtered to yield 3.48 grams of perfectly colorless, annoyingly glitter-like crystals melting at 163-174°C. Assuming this is mescaline (hemi)sulfate dihydrate, this would correspond to a yield of 63.8%. But is it?

The precursor was verified. The reduction is established and an amine was produced and isolated; an amine that generally seemed to be a solid at room temperature, even as the base. A spot analysis was performed with the Marquis reagent, producing a positive (orange) result. The amount of acid used to neutralize the amine base corresponds fairly well to there being two moles of product for every mole of acid. A colorless end product was achieved and, given more time and space to form, the crystals seem fairly uniform and needle-like. Condensation inside the capillary tube where a sample was melted seems to suggest that there is some water of crystallization. The logical thing to do would be to form the hydrochloride whose melting point can reliably be determined. But for now I will choose to be convinced.

Acetone-washed product dissolving in water


Recrystallized product




[Edited on 7-3-2022 by Benignium]

Hard to take eyes off your posts

Cheers guys! Good to see familiar names still following along!

Metacelsus - You wouldn't happen to have any specifics to share about your experiences with ammonium acetate? I've read that KOH also works OK for this one; a weird selection of popular choices.


I recrystallized the product from plain hot water to get something less airborne and less affected by static electricity. These crystals are still very tiny, but less messy to handle and more representative of their proper geometry. Here are a couple of macro shots:

Quote: Originally posted by Benignium

Cheers guys! Good to see familiar names still following along! Metacelsus - You wouldn't happen to have any specifics to share about your experiences with ammonium acetate? I've read that KOH also works OK for this one; a weird selection of popular choices.