L. N. D’yakonova-Shul’ts

On the Mechanism of the Perkin Reaction

With regard to the mechanism of the synthesis of cinnamic acid by Perkin’s method, the chemical literature contains a great variety of explanations. On the basis of experimental investigations, authors give different explanations, but to this day the mechanism of the reaction remains not fully elucidated.

Perkin¹, in his works, reports in detail on the synthesis of unsaturated aromatic acids. Thus, he obtained cinnamic acid from two parts of benzaldehyde, one part of sodium acetate, and two parts of acetic anhydride. Using, instead of sodium acetate, other salts of fatty acids, the author comes to the conclusion that condensation occurs exclusively between the aldehyde and the anhydride; the salt plays the role of a dehydrating agent. The nature of the salt is not reflected in the composition of the product. From benzaldehyde and acetic anhydride only cinnamic acid is obtained, irrespective of the salt of the fatty acid taken.

Conrad² likewise asserts that the reaction occurs between the aldehyde and the anhydride. But Fittig³ and his co-workers, on the basis of experimental data, find that the course of the reaction is different from that described by Perkin. They assert that condensation occurs not between the aldehyde and the anhydride, but between the aldehyde and the salt; the anhydride acts as a dehydrating agent. The composition of the product depends on the composition of the salt, and not on the anhydride.

Experiments were carried out by Ch. Stuart and F. Slocum. Ch. Stuart⁴, heating benzaldehyde, sodium propionate, and acetic acid at 180–200°, obtained a large amount of phenylcrotonic acid and a little cinnamic acid. F. Slocum⁵ heated benzaldehyde, acetic anhydride, and sodium butyrate at various temperatures; at 100° for 60 hours he obtained a large amount of phenylangelic acid, and little cinnamic acid. At 150° the amount of cinnamic acid increases; at 180° cinnamic acid was obtained in 10 parts, and phenylangelic acid in 1 part.

On the basis of experimental investigations the authors give the following explanation: at 100° a reaction occurs between the aldehyde and the salt. At high temperature, interaction first takes place between acetic anhydride and the salt—as a result, the anhydride of butyric acid and sodium acetate are obtained; the latter then reacts with benzaldehyde.

F. Tiemann and R. Kraatz¹ suppose, as does Fittig, that at high temperature there first occurs an interaction between the anhydride and the salt, and then condensation sets in, but not between the aldehyde and the salt, as Fittig thinks, but between the aldehyde and the newly formed anhydride, according to the equation:

\[ \mathrm{C_6H_5-CH{=}O} + \begin{matrix} \mathrm{R-CH_2-C} & \displaystyle {^{\mathrm{O}}_{\backslash \mathrm{O}}} \\ \mathrm{R-CH_2-C} & \displaystyle {_{\backslash \mathrm{O}}} \end{matrix} \;\longrightarrow \]

\[ \longrightarrow\; \mathrm{C_6H_5CH=C-C-O-C-CH_2-R+H_2O} \]
\[ \begin{matrix} & \mathrm{|} & \displaystyle \overset{\mathrm{O}}{\parallel} & & \displaystyle \overset{}{\parallel} \\ & \mathrm{R} & & & \mathrm{O} \end{matrix} \]

On decomposition with water, cinnamic acid or cinnamic acid with a side chain is obtained.

A. Michael², on the basis of experimental data, like the preceding authors, comes to the conclusion that the reaction takes place between the aldehyde and the anhydride. He first carries out a series of experiments with two components: anhydride and salt, and then with three components: anhydride, salt, and aldehyde. The experiments showed that the interaction between acetic anhydride and sodium butyrate begins at \(100^\circ\). Butyric anhydride reacts with sodium acetate at \(150^\circ\), with very little acetic anhydride being formed; at \(180^\circ\) the yield of acetic anhydride increases considerably. Acetic anhydride reacts with the sodium salt of caproic acid immediately at \(100^\circ\). Further, by heating benzaldehyde, butyric anhydride, and sodium acetate at \(100^\circ\) for 225 hours, the author obtained 1 part of cinnamic acid and 14 parts of phenylangelic acid. On heating benzaldehyde, acetic anhydride, and sodium butyrate at \(100^\circ\) for 225 hours, he obtained cinnamic and phenylangelic acids in the ratio \(1:2\). Subjecting benzaldehyde, caproic anhydride, and sodium acetate to heating for 300 hours at \(100^\circ\), he obtained cinnamic and butylcinnamic acids in the ratio \(1:5\). At the same time, from acetic anhydride and sodium caproate he obtained 5 parts of the former and 9 parts of the latter.

On the basis of the results presented, the author comes to the conclusion that the reaction takes place between the aldehyde and the anhydride.

J. Nef³, from his experiments, concludes that in the Perkin reaction benzylidene diacetate is first formed as an intermediate product, which subsequently reacts with the salt of the fatty acid.

J. Nef⁴, studying the reaction between benzaldehyde and acetic anhydride under various conditions, comes to the conclusion that benzylidene diacetate is formed in large quantity upon the addition of acetic acid; in the absence of acetic acid, benzylidene diacetate is not obtained either at ordinary temperature or upon heating.

A. Geuther¹) succeeded in obtaining benzylidene diacetate in a negligible amount from benzaldehyde and anhydride by heating in a sealed vessel at \(230^\circ\).

Quite different results are obtained upon the addition of acetic acid; the reaction takes place on standing in a sealed tube even at room temperature. But if the acetic acid is replaced by a salt, cinnamic acid is obtained. Cinnamic acid is obtained from benzylidene diacetate and sodium acetate, but only in small amount; upon addition of acetic acid the amount of cinnamic acid increases.

When a mixture of benzaldehyde, acetic anhydride, and acetic acid is heated to \(180^\circ\), cinnamic acid is not formed. When a mixture of benzaldehyde, acetic acid, and sodium acetate is heated to \(180^\circ\), cinnamic acid likewise is not formed.

From the results obtained the author concludes that, in the Perkin reaction, the anhydride and the diacetate take part.

H. M. Meyer and R. Beer²) state in their work that, when benzaldehyde, acetic acid, and sodium acetate are heated to \(180^\circ\), cinnamic acid is not formed; but if, instead of benzaldehyde, another more active aldehyde is used, for example \(o\)-chlorobenzaldehyde, or if the cation of the sodium salt is replaced by a more energetic cation, cinnamic acid can be obtained without participation of the anhydride. Thus benzaldehyde with 2 mol. of potassium acetate and 3 mol. of acetic acid gives, on heating for 26 hours, 30–40% cinnamic acid. \(o\)-Chlorobenzaldehyde with potassium acetate in the presence of acetic acid gives 78.6% chlorocinnamic acid. On the basis of experimental data the authors find that the Perkin reaction can proceed without participation of the anhydride.

J. Reich and P. Chaskelis³) find that the Perkin reaction can proceed in two ways, depending on the salt. If one takes the sodium salt of malonic acid, which contains a very mobile hydrogen atom, it is capable of reacting without the aid of acetic anhydride; but if one takes a salt with a slightly mobile hydrogen atom \((\mathrm{CH_3COONa})\), then participation of acetic anhydride is necessary. In this case a diacetate is formed, which then reacts with sodium acetate.

P. Kalnin⁴) asserts that condensation proceeds not between the salt and the aldehyde, but between the aldehyde and the anhydride. The salt plays the role of a catalyst; it can be replaced by other catalysts that are not derivatives of carboxylic acids. Tertiary bases are especially suitable for this purpose: pyridine, quinoline, diethylbenzylamine, benzylpiperidine, and triethylamine. Further, the author believes that acetic anhydride reacts in its enolic form, while the aldehyde adds to the enolic form of acetic anhydr-

¹) A. Geuther, Ann 106, 251.
²) H. Meyer und R. Beer C. 1913, 1, 2121. Monatshefte f. Ch. 34, 649.
³) J. Reich und P. Chaskelis C. 1916, 11, 905; Bull. soc. Chem. de France 19, 287.
⁴) P. Kalnin C. 1929, 11, 562; C. 1932, I, 816; B. 64, 2935 Helv. chim. Acta. 11, 977.

series, in which the aldehydic hydrogen has decisive significance. According to this explanation, he gives the following scheme for the Perkin reaction:

\[ \begin{gathered} \ce{ (CH3CO)2O -> CH3-C(=O)-CH2-C(=O)OH + C6H5CHO -> CH3-C(=O)-CH2-CH(OH)-C(=O)-C6H5 }\\[4pt] \ce{ -> CH3-C(=O)-O-CH(OH)-CH=C(OH)-C6H5 -> C6H5-C(=O)-CH=CH + CH3-COH }\\[4pt] \ce{ -> C6H5-CH=CH-C(=O)OH } \end{gathered} \]

R. Kuhn and S. Ishikawa\(^{1}\), on the basis of their experiments, adhere to the view expressed by Kalinin.

E. Müller\(^{2}\), carrying out experiments with homophthalic anhydride in the presence of triphenylmethylsodium, found that at first the enolate is obtained

\[ \begin{gathered} \ce{C6H4<\!\!\begin{matrix} CH=CO-Na\\[-2pt] \big|\\[-2pt] C(=O)-O \end{matrix}} \end{gathered} \]

which, on interaction with benzoic aldehyde, according to Kalinin should give benzylidenehomophthalic acid; however, the compound obtained is:

\[ \begin{gathered} \ce{ CH(ONa)-C6H5\\[-2pt] \big|\\[-2pt] C6H4<CH-C(=O)-O\\[-2pt] \phantom{C6H4<}C(=O) } \end{gathered} \]

1) R. Kuhn und J. Ishikawa. C. 1932, I, 60; B. 64, 2347.
2) E. Müller, C. 1932, I, 815; Ann. 491, 251.

The latter, upon saponification, gives the oxy acid:

\[ \begin{array}{c} \mathrm{COOH}\\ |\\ \mathrm{C_6H_4}\!\!<\!\begin{array}{c} \mathrm{CH{-}CHOH{-}C_6H_5}\\ \mathrm{C{-}OH}\\ \|\\ \mathrm{O} \end{array} \end{array} \]

but not in the free form, rather in the form of a lactone

\[ \begin{array}{c} \mathrm{COOH}\\ |\\ \mathrm{C_6H_4}\!\!<\!\begin{array}{c} \mathrm{CH{-}CH{-}C_6H_5}\\ \mathrm{C{-}O}\\ \|\\ \mathrm{O} \end{array} \end{array} \]

On the basis of the data obtained, Müller finds that his experiments confirm Kalnin’s and Kuhn’s view regarding enol as the reacting form of acetic anhydride, but do not confirm Kalnin’s further explanations concerning the addition of benzaldehyde to the enol form of acetic anhydride. In addition to the experiments described, Müller points to benzonitrile, which reacts with acetic anhydride despite the absence of an aldehydic hydrogen atom. The author finds that for the Perkin synthesis the presence of a carbonyl group and of an anhydride capable of tautomerically passing into the enol form is necessary. First an oxy acid is obtained, which at high temperature splits off water. In the absence of the necessary hydrogen atom, the reaction stops at the formation of the oxy acid.

I agree with the opinion expressed by Müller regarding the addition of the anhydride to benzaldehyde, and not the reverse. Kalnin points to the mobility of the hydrogen atom of the methyl group of the anhydride, but if the indicated hydrogen atom is mobile and capable of shifting to oxygen, then with equal success it can add to the carbonyl group of the aldehyde. On the basis of Müller’s experiments, one may rather come to the conclusion that, for the Perkin synthesis, a tautomeric rearrangement of acetic anhydride into the enol form is not at all necessary.

Only a great mobility of the hydrogen atom in the neighboring unit is necessary¹. The parallelism with the capacity for formation of the enol form

is quite natural, since the latter requires the mobility of the same hydrogen. Thus Möller obtained the compound

\[ \chemfig{*6(---(-[:0]CH(-[:90]CH(ONa)-[:0]C_6H_5)-[:0]C(=O)-[:270]O)-C(=O)---)} \quad \text{from} \quad \chemfig{*6(---(-[:0]CH=CONa-[:270]O-C(=O)---)-C(=O)---)} \quad \text{and} \quad \mathrm{C_6H_5-CH{=}O} \]

apparently according to the scheme:

\[ \chemfig{*6(---(-[:0]CH(-[:90]\uparrow)-[:0]CONa(-[:90]\uparrow)-[:270]O-C(=O)---)-C(=O)---)} \;+\; \chemfig{C_6H_5-CH(-[:90]\uparrow)(-[:270]O(-[:270]\uparrow))^{+}} \;\longrightarrow \]

\[ \longrightarrow\; \chemfig{*6(---(-[:0]CH(-[:90]CH(-[:90]C_6H_5)-[:0]O)-[:0]CONa(-[:270]O-C(=O)---))-C(=O)---)} \;\longrightarrow\; \chemfig{*6(---(-[:0]CH(-[:90]CHONa-[:90]C_6H_5)-[:0]C(=O)-[:270]O)-C(=O)---)} \]

but the formation

\[ \chemfig{*6(---(-[:0]CH(-[:90]CH(-[:90]C_6H_5)-[:0]O)-[:0]CONa(-[:270]O-C(=O)---))-C(=O)---)} \]

with a four-membered ring

is possible for it with sodium at carbon, since sodium also must readily jump from place to place, just as hydrogen does.

\[ \chemfig{CH_3-[:270]CONa=[:270]CH-[:270]COOC_2H_5} \;+\; \mathrm{CH_3J} \;\longrightarrow\; \chemfig{CH_3-[:270]C(-[:30]ONa)(-[:330]J)-[:270]CH(-[:0]CH_3)-[:270]COOC_2H_5} \;\longrightarrow\; \chemfig{CH_3-[:270]C(=O)-[:270]CH(-[:0]CH_3)-[:270]COOC_2H_5} \;+\; \mathrm{NaJ} \]

seems improbable; most likely, the reaction proceeds as follows:

\[ \begin{aligned} &\ce{ \begin{matrix} & CH=CONa\\ C6H4< & |\\ & C-O\\ & ||\\ & O \end{matrix} } \rightleftarrows \ce{ \begin{matrix} & CHNa-C=O\\ C6H4< & |\\ & C-O\\ & ||\\ & O \end{matrix} } + \\[2em] &\ce{ +C6H5-CH } \longrightarrow \ce{ \begin{matrix} & C6H5\\ & |\\ & CHONa\\ & |\\ & CH-C=O\\ C6H4< & |\\ & C-O\\ & ||\\ & O \end{matrix} } \end{aligned} \]

i.e.

\[ \ce{ \begin{matrix} & CH=CONa\\ C6H4< & |\\ & C-O\\ & ||\\ & O \end{matrix} } \]

reacts not in the enol form.

From the data of J. Nef, A. Geuther, J. Reich, and others, one may suppose that the Perkin reaction proceeds simultaneously in two directions, depending on temperature and other conditions:

\[ \begin{aligned} 1.\quad &\ce{ C6H5-CH + HCH2-C-O-C-CH3 ->} \\[-0.2em] &\ce{ -> C6H5-CHOH-CH2-C-O-C-CH3 ->} \\[-0.2em] &\ce{ -> C6H5-CH=CH-C-O-C-CH3 + H2O ->} \\[-0.2em] &\ce{ -> C6H5-CH=CH-C-OH + CH3COOH } \end{aligned} \]

\[ \begin{aligned} 2.\quad &\ce{ C6H5-CH + CH3-COO-CH3 ->} \end{aligned} \]

\[ \begin{gathered} -\mathrm{C_6H_5}-\mathrm{CH} \left[ \begin{array}{c} \mathrm{O}-\mathrm{C}(=\mathrm{O})-\mathrm{CH_3}\\ \mathrm{O}-\mathrm{C}(=\mathrm{O})-\mathrm{CH_3} \end{array} \right] +\mathrm{H_2CH}-\mathrm{COONa} \longrightarrow \\[1em] \longrightarrow \mathrm{C_6H_5}-\mathrm{CH}=\mathrm{CH}-\mathrm{COONa} +2\,\mathrm{CH_3}-\mathrm{C}(=\mathrm{O})\mathrm{OH}. \end{gathered} \]

1. These considerations may also be applied to all reactions of a similar type. The question is, for example, whether syntheses with acetoacetic and malonic esters really proceed through the enol form. After all, addition of alkyl halides to double bonds, it seems, has never been observed. The possibility is not excluded that here too the sodium compound of acetoacetic ester simply reacts according to an entirely pos- ↩↩↩

2. A. Michael, B. 34, 918 (1901). ↩↩

3. J. Nef, Ann. 298, 309. ↩↩

4. J. Nef, Ann. 298, 277. ↩↩

5. F. Slocum, Ann, 227, 53. ↩