phencyclidine sans cyclohexanone

[ning]

K bees, since I've been doing a lot of studying on the interesting and amazing powers of nitro-compounds, this one just hit me, and I want to know what you think about it:

Short synthesis of PCP without organometallics, or cyclohexanone:

Since I learned how to use the modeller recently, it will bee graphical:

I. nitromethylbenzene (aka alpha-nitrotoluene) + 1,5-dichloropentane --> err...something

Molecule:

first step ("c1ccccc1CN(=O)=O.ClCCCCCCl>>c1ccccc1C2(N(=O)=O)CCCCC2")


This works just like an enolate synth, because nitro compounds act very much like carbonyls, except they are much more acidic. So you should only need NaOH to effect this condensation (if that)

II. nitro cyclohexyl thingy gets reduced to phenylcyclohexylamine thingy

Molecule:

reduce with iron or zinc in HCl ("c1ccccc1C2(N(=O)=O)CCCCC2>>c1ccccc1C2(N)CCCCC2")



Reduction of nitros to amines is not very hard, how lucky for us...

III. Final condensation

Molecule:

PCP! ("c1ccccc1C2(N)CCCCC2.ClCCCCCCl>>c1ccccc1C2(N3CCCCC3)CCCCC2")



That's it. The cool thing here is that there are only two compounds involved, to form both rings you use the same dichloropentane. So, there are two questions: where do you get alpha-nitrotoluene, and where do you get 1,5-dichloropentane? To answer the first question, here are several possibilities:

1. Benzyl chloride + NaNO2
2. Bromo or iodobenzene + nitromethane (enolate synth!)
3. benzaldehyde + ammonia + some oxidizer (more on this later)

That's assuming you can't buy it.
As for making 1,5 dichloropentane, I bet some sort of 5-carbon diol or cylic ether could bee converted into it by HCl. That's the hardest one, but I also bet its pre-precursors aren't watched all that much.
That all for now, folks. Ning out.

[Rhodium]

Where are your references for this reaction sequence? As I have told you before, ideas like these are a dime a dozen, while tried and true synthetic details for the actual transformations are the only thing that counts and are of any use...

For example, I can think of a lot of pitfalls in the synthesis you give above - what's stopping polymerization from happening in both reactions using 1,5-dichloropentane?

Please do not post anything which you don't have references for - with solvents, reagents, reaction times and workup. If it is a very novel synthesis where the precursors aren't in the literature, it is okay if you reference analogous preparations - for example, if a reaction works for a 2,4,5-trimethoxy compound, it is likely to work with a 2,5-dimethoxy compound too - so even if your proposal concerns a 2,5-MeO-something, give the literature details for the corresponding 2,4,5-MeO-compound, as a guide for other users.

[Lilienthal]

But it's a cool idea. I can't see why it shouldn't work. It might be beneficial to use KI as a catalyst (Finkelstein conditions).

[ning]

It is a cool idea, and one of the reasons I posted it is because I CAN'T find all of the reaction times, solvents, etc. But maybe some other bees can. And if what forum is better for this "wild guessing" than novel discourse? The couch?

Why doesn't it polymerize? I wondered that same thing in the fentanyl thread with regards to the mannich reaction part. Why don't they all just wad up into one big crosslinked ball? And the answer appears to be this:
Once one end of the chain has attached, statistically, there are many more molecules of inert solvent than there are reactable molecules. This means that the molecule is undergoing strong thermal agitation, and it is relatively unlikely to meet other unattached chains. That long chain attached to it is quite floppy. It is also already tied to the molecule at one end. Also, the cyclohexane ring has angles that are very comfortable for the carbons involved. I think it's fair to say that there is a huge thermodynamic and kinetic advantage for the other end of that ring to hook up with the amine group, rather than for another new molecule to attach.

If this isn't formal enough for you, I am sorry. When time presents itself, I will dig up some relevant refs, which I'm sure exist. But if I knew it all already, why would I bother posting anything to the hive? We all work TOGETHER here, that's why this place is so powerful.

Edit:

Rhodium, just for your interest and edification, I refer you to the PCP document on

your own website

(https://www.thevespiary.org/rhodium/Rhodium/chemistry/pcp/pca_synth.html), where, under "Reaction Scheme VI. PCP via phenylacetonitrile", you will see phenylacetonitrile condensed with 1,5-dibromopentane, which should sound familiar. Furthermore, this route produces phenylcyclohexylamine, which, according to scheme V, ritter reaction, is turned into PCP by reacting it with, yes, you guessed it: 1,5-dibromopentane.

So yes, it was a wild guess. But seeing as how, right now, the major "novel substituttion" is a nitro group substituting for a nitrile, it looks like my "wild guess" was probably not so far off. Could you please bee a little more considerate and think some more before you attack ideas you see posted? As "Chief Bee", you are in a position of great power and responsibility. What you do has a wide and far reaching effect in the hive.

That said, I am flattered you bother to read my threads and reply to them at all

[ning]

For those bees who do not wish to go sort through the PCP page on Rhodium's site, here is the relevant bits:

I. 1,5 dibromopentane condensing with the benzylcyanide

Chemistry: In the first step, phenylacetonitrile is reacted with 1,5-dibromopentane and strong base to yield 1-phenyl-1-cyclohexanecarbonitrile. The base can be either NaNH2 (sodamide; 65% yield) or sodium hydride in anhydrous DMSO (better). If NaNH2 is used it may be purchased commercially or prepared from sodium metal and liquid ammonia (-30 C).
...
Step 1: 1-phenylcyclohexanecarbonitrile: A mixture of 222 gm (1.9 moles) of phenylacetonitrile and 400 gm (1.71 moles) of 1,5-dibromopentane was added over 6 hrs. to a refluxing suspension of 593 gm (7.54 moles) of sodamide in 4.5 L of ether. Reflux was continued overnight, followed by cooling in an ice bath and addition of 1.6 L of water over a period of 2.5 hrs, keeping the temperature at 21-24 C. The mixture was then refluxed for 1 hr. and cooled. Following filtration and seperation of the aqueous layer, the ether layer was washed with 1 L of water and 1 L of 3 N HCl. The solvent was then distilled to give 234 (72% yield) of crude product. Purification by fractional distillation (102 C / 0.1 mm Hg) gave 62% of pure nitrile


Here is the first major advantage of using a nitro group, which is much stronger than a nitrile: Where a nitrile requires a REALLY strong base like NaNH2 to deprotonate, a nitro will deprotonate quite well with, say, lye. That simplifies the procedure considerably, and may raise yields. Often only room temperature stirring is required.

The second major advantage is that to convert a nitro to an amine is one step and high yielding, while the nitrile needs two steps, with trifluoroacetic acid, and bromine. Yuck!

 1,5 dibromopentane condensing with phenylcyclohexylamine:
(Yield 70%)

Preparation of PCP from PCA: A mixture of 8.69 gm of PCA, 11.5 gm of 1,5-dibromopentane, and 8.0 gm of anhydrous K2CO3 in 50 mL of dry DMF was stirred and heated. At 50-55 C an exothermic reaction took place and the temperature rose to 95-100 C. The flask was heated for 1 hr on a steam bath, poured into ice cold water and extracted with ether, followed by distillation and recrystallization to give the final compound.


And that's it. Not so wild, really. Just easier and better. Actually, though I didn't know before, it seems that the two syntheses are in essence exactly the same, both using electron withdrawing groups that are convertible to amines. Yet for some reason, the scientists that used the nitrile route chose a harder group to use than they had to. I have a strange feeling that maybe many chemists don't know nitro chemistry very well. It's true that even ning's better college textbooks don't have much information on the subject. That's a shame, since nitros are so versitile, and require such gentle conditions to do what they do.
Here is a very interesting pdf file on nitroalkane chemistry. Rhodium, I think it should bee on your website:

http://www.iptonline.com/articles/public/IPTFOUR79.pdf

. Get it quick, I don't know how long it will bee available.

[Nicodem]

Ning, the reaction scheme you posted looks so nice and easy that it breaks my heart to inform you that it wont work. The reason is that the alkyl halogenides O-alkylate the nitroalkanes and not C-alkylate as you proposed. There are only few exceptions to this rule as I already explained in

Post 473764

(Nicodem: "Something more about the nitro group", Chemistry Discourse).
Ironicaly it is an another nitro group that changes the mechanism of the reaction to allow C-alkylation.
If this was so easy then we could wonder why Shulgin made almost all of his PEAs from benzaldehides and not benzyl halogenides as it would be much easier. The botom line is: carbonyl compound C-alkylate, alkyl halogenides O-alkylate. So what would hapen in the first stage is

Molecule:

  ("c1ccccc1CN(=O)=O.ClCCCCCCl>>C(=[N+](/[O-])OCCCCCCl)\c1ccccc1")


in the next stage an internal redox reactions form an aldehide and so on.

[Nicodem]

I might have been to harsh criticizing Ning’s idea. I’m always optimistic and convinced that every problem can be solved, so here comes my solution of the “two precursors PCP synthesis” challenge. This modification does use only two organic compounds but off course more inorganic compounds and solvents are needed. Here it is, criticize at will:

First, let me tell you that the preparation of phenylnitromethane is described in
J. Am. Chem. Soc. 78, 1497-1501 (1956)
(

https://www.thevespiary.org/rhodium/Rhodium/pdf/nitroethane.kornblum.dmf.pdf

) posted in

Post 474663

(Rhodium: "original nitroethane literature", Methods Discourse).

The Michael addition of nitroaliphatics to vinylketones is well known. In this modification the phenylnitromethane reacts with divinylketone to form a cyclohexanone ring. For similar reactions see:
Britten-Kelly M., Willis B.J, Barton D.H.R.; Michael additions to alkyl substituted divinyl ketones. Synthesis (1980) 1, 27-29.
Acetone and formaldehide under certain conditions might be used instead of divinylketone, but probably giving much lower yield (besides requiring 3 precursors instead 2).

Molecule:

double Michael addition ("c1(ccccc1)C[N+]([O-])=O.C=CC(C=C)=O>>C2CC(CCC2(c1ccccc1)[N+]([O-])=O)=O")



In the second reaction the nitro group gets selectively reduced. I think SnCl₂/HCl would do the trick.

Molecule:

selective reduction of the nitro group ("C2CC(CCC2(c1ccccc1)[N+]([O-])=O)=O>>C2CC(CCC2(c1ccccc1)N)=O")



The third reaction is quite analog to the first with the amino group undergoing the double Michael reaction giving the piperidinone ring.

Molecule:

double Michael addition, again, but on the amino group ("C2CC(CCC2(c1ccccc1)N)=O.C=CC(C=C)=O>>C3CC(CCN3C2(c1ccccc1)CCC(CC2)=O)=O")



It would sure be interesting to research the pharmacology of the last compound, but since the goal was set to be PCP, the two carbonyls must be eliminated. The Clemensen reduction is probably a wrong bet here since aminoketones often give some rearrangement products and besides this, the benzylic amine could get debenzylated. The best choice would probably the Wolf-Kishner reduction with hydrazine and KOH at elevated temperatures.

Molecule:

Wolf-Kishner reduction ("C3CC(CCN3C2(c1ccccc1)CCC(CC2)=O)=O>>C3CCCCN3C2(c1ccccc1)CCCCC2")



But even if it would work I don't see it as an elegant synthesis. The diketo-PCP compound might be intersting, who knows?

[ning]

It's so sweet of you ...
And you did have me worried for a little while
since it seems the literature is scarce with examples of nitroalkane alpha-substitution (yay! ning learned more chemist jargon!)...

But then, ning found this:

Molecule:

What does this mean?? ("c1ccccc1CN(=O)=O.BrCCBr>>c1ccccc1C2(CC2)N(=O)=O")



Yield is only 18%, but still, it's something. And I can always handwave and say that poor yield is due to ring strain...
Unfortunately, can't get the paper. Maybee someone else can?

Russian Journal of Organic Chemistry 1981, 1277
aka Zh. Org. Kh., p. 1435

And ning's hope was buoyed further by this paper:
JACS 1929, 2151, where they add a chloro compound to nitromethylbenzene with a yield of 82%.

Molecule:

Page 2151, NaOMe+NaOAc ("ClCCC(=O)c1ccccc1.c1ccccc1CN(=O)=O>>c1ccccc1C(N(=O)=O)CCC(=O)c1ccccc1")



A very wierd paper, actually. They ALSO perform a michael addition of the vinyl ketone version of that haloketone (like your idea), to give the exact same product.

Molecule:

Page 2151 ("C=CC(=O)c1ccccc1.c1ccccc1CN(=O)=O>>c1ccccc1C(N(=O)=O)CCC(=O)c1ccccc1")



Here is what they say:

"In the presence of alkaline reagents, phenylnitromethane combined with beta-chloropropiophenone to give the same saturated gamma-nitro-ketone as was obtained with phenyl vinyl ketone. Its structure was proved by formation of dibenzoylethane when a solution of the sodium derivative was decomposed with cold dilute hydrochloric acid.
...
The addition product with phenylnitromethane
gamma-nitro-gamma-phenylbutyrophenone:
Fifteen grams of beta-chloropropiophenone and 12 g. of fused potassium acetate were dissolved in 60 cc. of hot methyl alcohol. Then 12 g. of phenylnitromethane in 20 cc. of alcohol was added and the whole made faintly alkaline to litmus by admitting a 5% solution of sodium methylate(!), drop by drop. The reddish mixture was refluxed for 40 minutes, then acidified with acetic acid and poured into 120 cc. of cold water. After several hours, the precipitated addition product was filtered; the average yield of several precipitations was 82%--this was true only when the phenylnitromethane was redistilled.
...now the reaction with vinyl ketone...
To 3.8 g. of the ketone was add 10 cc. of absolute methyl alcohol; most of it immediately polymerized to a white, insoluble mass. The whole was warmed and the clear solution decanted from the polymer into a clean flask. Two grams of phenylnitromethane was added, then enough dilute sodium methylate to give an alkaline reaction to litmus, and the mixture warmed on the steam bath for fifteen minutes. It was next acidified with acetic acid; an oil separated and crystallized on cooling. This was purified as described above and identified as the same substance by a comparison of melting points, mixed melting points and solubilities."

Two exciting points here: One, that it sounds like the substitution had a better yield than the Michael addition (that part about "decanting the liquid from polymers" didn't sound too good), and Two, that it is very likely that NaOMe is not necesary at all (enough was added to give alkaline reaction to litmus??)--probably it was used because it was soluble in methanol, and they had lots of sodium lying around. Probably NaOH would do nicely.

Actually, upon reading in the org chem books, it seems that nitro and carbonyl groups are not so different. With carbonyls there is also a competition between attack on the carbonyl oxygen, and alpha substitution, right? How are they so different?

Anyway, if some bee could dig up that paper and post it, it would be real great.

Edit:

And of course, we should not ignore this fine post, made not many days ago, directly related to our topic, though going in reverse...

Post 475109

(Lego: "Amphetamines/PEAs w/o benzaldehyde or nitroethane", Novel Discourse) Lego, cool idea!

[Nicodem]

Yield is only 18%, but still, it's something. And I can always handwave and say that poor yield is due to ring strain...

Yes, this is what I was talking about when I said that the nitroaliphatics get O-alkylated with alkylhalogenides. The yield of the C-alkylation products rarely exceed 10%. In this case the yield must be a little better because of the phenyl ring making the carboanion relatively stable (but still not as stable as the anion in the nitronium tautomere). Don’t get to hopeful by thinking that in the JACS 1929, 2151 you have an example of halogen substitution. In the basic conditions they used, the first thing that happens is the elimination of the HCl giving exactly the phenylvinyl ketone they used later. Both reactions are basically the same since the alkylation of the phenylnitromethane goes trough the Michael addition in both cases.
And, no, the ring strain is not responsible for the low yield. I doubt that in our case the yield would be any different. Once the first bromine gets substituted the next will follow, regardless of the ring strain (as long the ring is not to big). This is because an internal conversion of a molecule has much more probability to happen than a transformation involving two or more molecules. Actually, the only way to prevent the ring closure is to use two differently reactive halogens, like with 1-bromo-5-chloro-pentane, but we don’t want that.
What a chemist can do to improve the C to O alkylation ratio is therefore to make the carboanion more stable, like for example in the ethyl nitroacetate, which can be effectively C-alkylated.
This can be done, in our case, by making the phenyl ring more electron-poor:
-   By attaching a strong electron-withdrawing group (-NO2, -COOR, -COR...) para or ortho to the nitromethyl group.
-   By using an electron-poor heterocycle instead of phenyl (like pyridine, pirazine, pyrimidine, quinoline...).
This would probably allow using 1,5-dibromopentane for constructing the desired cyclohexane ring, but then you would not end up with PCP but with an analogue.
For example if you would use (p-nitro-pheny)-nitromethane you would end up with two amino groups after the reduction of the nitros. And then? Even if you could selectively substitute the aromatic amino group (trough the Sendmayer reaction) with a halogen, hydrogen, hydroxy or alkyltio group and at the end get a PCP analogue it would still make a really irrational route especially if you consider that the required phenylnitromethanes are not commercially viable.

About your comparison of carbonyl to the nitro, read this:
Nitroalkane chemistry (

http://www.iptonline.com/articles/public/IPTFOUR79NP.pdf

)
They are quite different.
Btw, do you live in a library or what? I just admire your tenacity   .

[Rhodium]

I think it is very likely that the  beta-chloropropiophenone is dehydrohalogenated in situ to the vinyl ketone (look at all the added base!), and that the vinyl ketone then adds michael style to the phenylnitromethane. Thus it is probably not a true alkyl halide alkylation.

[Nicodem]

Exactly what I already told Ning. That's why they added the sodium methoxyde until they got the indicator to the basic reaction. If there would be no elimination the lakmus would get purple after one drop or even without the methoxyde since sodium acetate is already slightly basic.
In conclusion: This is not a halogen substitution!

[ning]

I'm still not convinced that it's impossible, though I see where your point is. I'll keep my trap shut until I find something more conclusive for you doubters. But anyway, I found some more hot shit for ya to chew on.

Nicodem: yes, I do more or less live in the library. They hate me, I'm always the last to leave when they want to close at night

So, I found that russian paper, the one with the cyclopropanation, and yes, it had shitty, low yields. Fine. So I see that they were referencing another paper on that procedure. It's in the same journal, by the same guy, a couple years earlier. So I look it up too. And what I see is this:

Molecule:

oh, my gawd... ("c1ccccc1CC#N.BrCCBr>>c1ccccc1C2(C#N)CC2")



Except this time it's in 85% yield. And you want to know why this is good? Because the base they used...guys, it wasn't NaNH3...it wasn't NaOEt. No, it was NaO-fucking-H.
And triethylbenzylammonium chloride. Holy crap. Now, that's a HUGE improvement over the nitrile procedure. And I say that ring strain is still a factor, and a 6-membered ring will have even higher yields than that funky little cyclopropane did.

Here's the ref:

The author's name was Synchkova, in
Russian Journal of Organic Chemistry 1980, 1775
In the russian version (ZOK), it was pg. 2086

Check it out, it looks like, though the nitro route may not pan out, there is lying in the ancient russian journals a significant improvement to the current procedure.

BTW, nicodem, I posted just that link on nitroalkanes before. If you remove the "NP" from the end, you can download a pdf that lets you print it...sometimes humans are so predictable

[Nicodem]

Nice article for the pethidine fans, but I thing the use of a PTC catalist for its synthesis has already discused in another thread and a better article found (for the N-methyl-piperidine ring).
But, again, why does this makes you think that it would work with the nitro compound? I did my best to explain why it does not work with regular nitro compounds (though I still admire your optimism).
Sorry about the link, I did know you already found it, I should have known   .
Here is another hint for you: ever heard about the Schmidt Reaction or Hoffman degradation? You can make an amino group out of that nitrile, either by partial hydrolysis to the amide and then using either NaOCl or NaOH/Br2, or complete hydrolysis to the carboxy and then NaN3/H2SO4.
Just something for you to chew on.

[ning]

Well, I agree, I am a little overoptimistic perhaps. The reason I wanted to use nitro was because it is 1. OTC, and 2. Very Powerful, i.e. weak base needed compared to nitrile.
Perhaps the first reason is slightly moot, as it seems actually easire to make your own nitriles than nitro compounds, but as for the second...well, I guess I need to read up more on that PTC stuff.

Was reading about nitriles and hofmann rearrangements just recently, sounds good. Would like to see some refs on the use of hypochlorite for that, actually.

The reason I'm still optimistic is because I have seen some papers where they do attach halo-things to the nitro like we want, iirc, using weaker bases, so it probably isn't michael addition. Not sure yet, though, need to check it out.

Don't know if someone has covered this yet, I'm checking on the feasibility of a cool conversion of allybenzenes direct to methamphetamine derivatives using the ritter reaction followed by the hoffmann rearrangement, like this:

Molecule:

allylbenzene+acetonitrile ("c1ccccc1CC=C.CC#N>>c1ccccc1CC(C)NC(=O)C")

Molecule:

hoffmann rearrangement ("c1ccccc1CC(C)NC(=O)C>>c1ccccc1CC(C)NC")



It's not so useful for meth as it would be for safrole, if yields aren't too bad. I never heard whether the strong acid required for the ritter reaction made it not work for safrole. The md-p2p-ol doc by zwitterion/eleusis seems to suggest that carefully done, strong sulfuric acid won't trash safrole, but of course, I'm not the hive's expert on such matters either.

Too bad there's not more fentanyl or sernyl fans out there...seems only the really serious chemists come here...

[Rhodium]

Post 280053

(Rhodium: "Why the Ritter Reaction Fails for Safrole", Chemistry Discourse)

Also, this is the

Hofmann Rearrangement

(http://orgchem.chem.uconn.edu/namereact/hofmannrear.html)

It will not magically turn an N-acetyl to an N-methyl.

[Nicodem]

Yes, the Hoffmann rearrangement works only on unsubstituted amides. Presumably because the intermediate is a nitrene CO-N:, which can only be formed if there are no alkyls on the -CO-NH₂ group. Ning, Here you have an example of its use for an amphetamine (you probably found this already):

https://www.thevespiary.org/rhodium/Rhodium/chemistry/pma.hwe-rxn.html