There's a paper published by Danilow in Chemische Berichte in 1930 on the subject of synthesis of phenylacetylcarbinol from benzylglycolaldehyde (2-hydroxy-3-phenylpropanal) by treatment sulfuric acid in an unspecified alcohol and heating to 130-135C for an unspecified period of time, providing an 80% yield of crude product. By distillation a 47% yield of phenylacetylcarbinol was recovered, among with a smattering of other products (page 2767-2770).
Curiously, there is a chapter from a book that mentions effecting the same transformation in alcohol using only a few drops of sulfuric acid, with no mention of heating. Cited is a Russian paper, also by Danilow (assuming Danilow is Germanized "Danilov"), published in the same year (S. N. Danilov and E. D. Venus-Danilova, J. Russ. Chem. Soc., 62, 1697 (1930)) but Assyl can't even begin to read this - his German is bad enough - so someone else would need to provide a translation. Perhaps the conditions there are milder or higher yielding, though it's possible the Chemische Berichte paper is simply a German translation of the original. But "a few drops" of sulfuric acid doesn't sound much like a 14% solution, so hopefully the book's citation isn't erroneous and procedure is different, and superior.
Aldehydes can generally be rearranged to ketones by treatment with sulfuric acid, sometimes requiring varying degrees of forcing conditions (heat, concentration of acid, duration of reaction, etc.). Now, this begets the question: what aldehydes other than benzylglycolaldehyde and hydratropic aldehyde can be rearranged to something useful?
The obvious and most accessible candidates for such a transformation are hydrocinnamaldehyde (3-phenylpropanal) and 3-hydroxy-3-phenylpropanal, the former via reduction of cinnamadehyde, and the latter by hydrolysis of cinnamaldehyde.
3-hydroxy-3-phenylpropanal appears the most likely to follow a mechanism similar to that of the rearrangement of benzylglycolaldehyde suggested by Danilow. Alternatively, perhaps some sort of alpha-keto rearrangement occurs with benzylglycolaldehyde (followed by steps that are not immediately obvious), or it simply suffers carbocationic rearrangement after protonation of the hydroxyl, effectively leading to the same intermediate that would result from protonation of 3-hydroxy-3-phenylpropanal, followed by a classical acid catalyzed aldehyde rearrangement.
Danilow also mentions that methylbenzoylcarbinol is rearranged by acid to phenylacetylcarbinol, offering another potential precursor. This is just an alpha-keto rearrangement.
Either way, these are certainly interesting possible routes to phenylacetylcarbinol (and P2P) to explore, but the question remains: how to prepare the starting materials?
Improvements on the Danilow method of preparing benzylglycolaldehyde must be possible somehow - the method requires preparation of 2-bromo-3-phenylpropanal, and hydrolyzing this with barium carbonate in water, affording a 65% yield (page 2773-2774). Perhaps some sort of alpha-hydroxylation of hydrocinnamaldehyde may be done, or chlorination with TCCA and hydrolysis using something more widely available and less toxic than barium carbonate. The benzyl Grignard with glyoxal is another possibility.
Methylbenzoylcarbinol could be prepared by similar means, except by using propiophenone as opposed to hydrocinnamaldehyde as a starting material. The synthesis of propiophenone is already a widely covered topic, so it will not be discussed here.
3-hydroxy-3-phenylpropanal can be prepared by the hydrolysis of cinnamaldehyde. Base catalyzed methods exist, though these are particularly prone to the retro-aldol reaction, affording benzaldehyde and acetaldehyde from cinnamaldehyde. A potential alternative is found in an obscure food chemistry paper, where cinnamaldehyde is hydrolyzed in high yield by various amino acids buffered to pH 10 in water at near room temperature, in some cases with complete conversion after just one hour. The retro-aldol reaction is still present here, though different amino acids provide varying selectivity (ranging from approximately 2-20% benzaldehyde produced on a molar basis). Higher temperatures favor the retro-aldol reaction. The main issue will likely be complex aldol condensations occurring when higher concentrations of cinnamaldehyde are used, resulting in a lot of tars - perhaps avoiding proline, and running it under a weak vacuum to pull off acetaldehyde, would limit this somewhat, should it prove to be a problem. Lysine is probably the ideal choice of catalyst, since it is extremely soluble in water, and 3-hydroxy-3-phenylpropanal is as well, while phenylacetylcarbinol is not (and has a m.p. of 172C) - it should simplify the workup.
Hydrocinnamaldehyde can be prepared in so many ways that the topic (and the possibility of rearrangement to P2P) is worthy of a thread in itself. As a teaser, one method involves passing cinnamaldehyde and methanol, ethanol, or propanol over alumina at 330C and could perhaps be ran at a sub-industrial scale. Conversion was reported at 35%, and 60% of converted material was found to be hydrocinnamaldehyde - a 20% yield of hydrocinnamaldehyde from cinnamaldehyde, in total. There's a lot of room to improve the percent converted and the yield itself with respect to converted material. More to come.
Curiously, there is a chapter from a book that mentions effecting the same transformation in alcohol using only a few drops of sulfuric acid, with no mention of heating. Cited is a Russian paper, also by Danilow (assuming Danilow is Germanized "Danilov"), published in the same year (S. N. Danilov and E. D. Venus-Danilova, J. Russ. Chem. Soc., 62, 1697 (1930)) but Assyl can't even begin to read this - his German is bad enough - so someone else would need to provide a translation. Perhaps the conditions there are milder or higher yielding, though it's possible the Chemische Berichte paper is simply a German translation of the original. But "a few drops" of sulfuric acid doesn't sound much like a 14% solution, so hopefully the book's citation isn't erroneous and procedure is different, and superior.
Aldehydes can generally be rearranged to ketones by treatment with sulfuric acid, sometimes requiring varying degrees of forcing conditions (heat, concentration of acid, duration of reaction, etc.). Now, this begets the question: what aldehydes other than benzylglycolaldehyde and hydratropic aldehyde can be rearranged to something useful?
The obvious and most accessible candidates for such a transformation are hydrocinnamaldehyde (3-phenylpropanal) and 3-hydroxy-3-phenylpropanal, the former via reduction of cinnamadehyde, and the latter by hydrolysis of cinnamaldehyde.
3-hydroxy-3-phenylpropanal appears the most likely to follow a mechanism similar to that of the rearrangement of benzylglycolaldehyde suggested by Danilow. Alternatively, perhaps some sort of alpha-keto rearrangement occurs with benzylglycolaldehyde (followed by steps that are not immediately obvious), or it simply suffers carbocationic rearrangement after protonation of the hydroxyl, effectively leading to the same intermediate that would result from protonation of 3-hydroxy-3-phenylpropanal, followed by a classical acid catalyzed aldehyde rearrangement.
Danilow also mentions that methylbenzoylcarbinol is rearranged by acid to phenylacetylcarbinol, offering another potential precursor. This is just an alpha-keto rearrangement.
Either way, these are certainly interesting possible routes to phenylacetylcarbinol (and P2P) to explore, but the question remains: how to prepare the starting materials?
Improvements on the Danilow method of preparing benzylglycolaldehyde must be possible somehow - the method requires preparation of 2-bromo-3-phenylpropanal, and hydrolyzing this with barium carbonate in water, affording a 65% yield (page 2773-2774). Perhaps some sort of alpha-hydroxylation of hydrocinnamaldehyde may be done, or chlorination with TCCA and hydrolysis using something more widely available and less toxic than barium carbonate. The benzyl Grignard with glyoxal is another possibility.
Methylbenzoylcarbinol could be prepared by similar means, except by using propiophenone as opposed to hydrocinnamaldehyde as a starting material. The synthesis of propiophenone is already a widely covered topic, so it will not be discussed here.
3-hydroxy-3-phenylpropanal can be prepared by the hydrolysis of cinnamaldehyde. Base catalyzed methods exist, though these are particularly prone to the retro-aldol reaction, affording benzaldehyde and acetaldehyde from cinnamaldehyde. A potential alternative is found in an obscure food chemistry paper, where cinnamaldehyde is hydrolyzed in high yield by various amino acids buffered to pH 10 in water at near room temperature, in some cases with complete conversion after just one hour. The retro-aldol reaction is still present here, though different amino acids provide varying selectivity (ranging from approximately 2-20% benzaldehyde produced on a molar basis). Higher temperatures favor the retro-aldol reaction. The main issue will likely be complex aldol condensations occurring when higher concentrations of cinnamaldehyde are used, resulting in a lot of tars - perhaps avoiding proline, and running it under a weak vacuum to pull off acetaldehyde, would limit this somewhat, should it prove to be a problem. Lysine is probably the ideal choice of catalyst, since it is extremely soluble in water, and 3-hydroxy-3-phenylpropanal is as well, while phenylacetylcarbinol is not (and has a m.p. of 172C) - it should simplify the workup.
Hydrocinnamaldehyde can be prepared in so many ways that the topic (and the possibility of rearrangement to P2P) is worthy of a thread in itself. As a teaser, one method involves passing cinnamaldehyde and methanol, ethanol, or propanol over alumina at 330C and could perhaps be ran at a sub-industrial scale. Conversion was reported at 35%, and 60% of converted material was found to be hydrocinnamaldehyde - a 20% yield of hydrocinnamaldehyde from cinnamaldehyde, in total. There's a lot of room to improve the percent converted and the yield itself with respect to converted material. More to come.
Amino acid-catalysed retroaldol condensation.pdf