Revised mechanism and improved methodology for the Perkin condensation. Resuscitation of the mechanism involving benzal acetate and the improbability of the enolate of acetic anhydride

Article abstract of DOI:10.1016/j.tetlet.2006.01.114

The Perkin condensation most likely occurs via the initial formation of a gem-diacetate from the aromatic aldehyde and acetic anhydride reactants. The key reactive nucleophile appears to be the enolate of the gem-diacetate rather than of acetic anhydride. The diacetate from PhCHO may be converted to cinnamic acid under a variety of (relatively) mild basic conditions.

Full text of DOI:10.1016/j.tetlet.2006.01.114

Tetrahedron Letters 47 (2006) 2249–2251
Revised mechanism and improved methodology
for the Perkin condensation. Resuscitation of the mechanism
involving benzal acetate and the improbability of
the enolate of acetic anhydride
Sosale Chandrasekhar* and Phaneendrasai Karri
Department of Organic Chemistry, Indian Institute of Science, Bangalore 560 012, India
Received 18 August 2005; revised 9 January 2006; accepted 18 January 2006
Abstract—The Perkin condensation most likely occurs via the initial formation of a gem-diacetate from the aromatic aldehyde and
acetic anhydride reactants. The key reactive nucleophile appears to be the enolate of the gem-diacetate rather than of acetic anhy-
dride. The diacetate from PhCHO may be converted to cinnamic acid under a variety of (relatively) mild basic conditions.
Ó 2006 Elsevier Ltd. All rights reserved.
Although the venerable Perkin condensation reaction
has retained its practical importance to this day,1–3 an
analysis of its accepted mechanism raises serious doubts.
The reaction is generally believed to occur via the initial
addition of the enolate (I) of acetic anhydride (1) to
benzaldehyde (2) (Scheme 1), and yet this key step seems
unlikely considering that the weakly basic acetate ion is
employed to generate I. [The pKa of acetic anhydride is
apparently unknown, but may be estimated at ꢀ20,
rather lower than that of the analogous carboxylic ester
(ꢀ25.0);4a and that of acetic acid4b is ꢀ4.5.] Impor-
tantly, the enolate I—even were it to be generated—is
expected to be highly unstable, fragmenting rapidly to
ketene II and an acetate ion under the relatively high
temperatures employed in the Perkin reaction
(ꢀ180 °C).5 (Although the enolates of cyclic anhydrides,
e.g., homophthalic anhydride,3 are reportedly stable, in
these the fragmentation to ketene is presumably rapidly
reversed intramolecularly.)
O
O
O
AcO
O
OAc
I
In fact, a preliminary study indicated that acetic anhy-
dride (1) was stable per se to the conditions of the Perkin
reaction, i.e. in the absence of benzaldehyde, thus
ruling out the possibility that I could be formed under
these conditions. (It is noteworthy that acetic anhydride
is purified by distillation from sodium acetate!6) Also,
treating 1 with potassium tert-butoxide (at room
temperature) led to its rapid decomposition, strongly
indicating the instability of the enolate I—when it is
indeed generated.
1
O
O
O
O
O
+
Ph
H
Ph
OH
OAc
Ph
OAc
O
2
3
I
O
OAc
II
I
Scheme 1.
The essential mechanistic problem with the Perkin reac-
tion, therefore, lies in proposing a viable two-carbon
nucleophilic species—formally, an acetic acid dianion
equivalent—under the conditions employed. Although
an alternative to I is not obvious, it is noteworthy that
benzaldehyde is known to react readily with excess
Keywords: Acetic anhydride; gem-Diacetate; Cinnamic acid; Enolate;
Perkin condensation.
*
Corresponding author. Tel.: +91 80 2293 2689/2402; fax: +91 80
3600 529; e-mail: sosale@orgchem.iisc.ernet.in
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.01.114

2250
S. Chandrasekhar, P. Karri / Tetrahedron Letters 47 (2006) 2249–2251
Table 1. Yields of cinnamic acid (3) produced from the following
O
OAc
OAc
reactants under the conditions indicated:a,b,c,d Ac2O (1), PhCHO (2),
PhCH(OAc)2 (4)b
+ Ac2O
Ph
Ph
H
O
1
2
4
Entry Reactants
Conditions
% Yield of 3
1a
2
1+2+KOAc
1+4+KOAc
2+4+KOAc
4+KOAc+DMF
Reflux
Reflux
Reflux
Reflux
60
30
75
70
O
O
O
O
O
Ph
O
Ph
3
O
4c
5e
H
OAc
IV
4+t-BuOK+THF Room temperature 70
III
a Entry 1 refers to the classical Perkin reaction; the base and the solvent
employed in the other cases are indicated with the reactants. All the
above reactions were carried out for the same period of time (5 h).
b Diacetate 4 was prepared in 80% yield as reported,8 from 1, 2 and
anhydrous FeCl3 at 0 °C; 4 was identified from its mp 44–46 °C (lit.8
44–46 °C), and by IR and NMR (1H and 13C) spectroscopy.
c Typical procedure (with DMF)—A mixture of diacetate 4 (1.0 mmol),
fused potassium acetate (1.0 mmol) and dry dimethyl formamide
(5 ml) was heated at 180 °C in a 10 ml round bottom flask carrying a
water condenser and guard tube (CaCl2). After being heated for 5 h,
the reaction mixture was cooled to room temperature, treated with
saturated NaHCO3 (2 ml) to ensure alkalinity and concentrated by
evaporation in vacuo. The mixture was then taken into a mixture of
diethyl ether and 6 N HCl, and the organic layer separated and
washed with water. The ether extract was dried (Na2SO4) and dis-
tilled in vacuo to remove the volatiles. The residue of crude cinnamic
acid (3) was purified by column chromatography over silica gel to
give pure 3 (0.66 mmol, 66%), identified by comparison with an
authentic sample employing the mp (132 °C, lit.d 133 °C), IR and
300 MHz 1H NMR spectra.
O
OAc O
O
O
O
O
+
Ph
O
Ph
H
Ph
Ph
O
V
2
IV
(-2)
OAc
O
O
OAc
3
CO2
Ph
Ph
O
Ph
(+ AcO )
VII
VI
Scheme 2.
acetic anhydride to furnish the gem-diacetate 4 in high
yields (Scheme 2):7,8 this then could well be the de facto
substrate in the Perkin reaction. Interestingly, this possi-
bility had indeed been considered very early but only to
be abandoned, apparently because of the poor yields of
cinnamic acid (3) obtained.3 In the present study, we
have shown that the yields of 3 can indeed be vastly
improved with 4 as the substrate if appropriate conditions
are employed. Thus, employing excess benzaldehyde, or
DMF as the solvent, or potassium tert-butoxide (KO-
But) as the base, led to excellent yields of cinnamic acid
(3) (Table 1, entries 3–5). Intriguingly, an excess of 1
leads to poor yields (Table 1, entry 2); this is explained
further below. (Higher yields indicate that a relatively
favourable pathway exists for the transformation.)
These considerations lead to the following reasonable
mechanistic conclusions which, although perhaps tenta-
tive, are at least compatible with the available evidence,
and can also form the basis of further work.
d Vogel’s Textbook of Practical Organic Chemistry, 5th ed.; Addison
Wesley Longman: Harlow, 1989; p 1038.
e t-BuOH may be substituted for THF as the solvent.
of acetic acid (ꢀ4.5). (The estimated pKa of IV is based
on the reported pKa of ꢀ12 for the analogous hemiace-
tal grouping in the sugars, e.g., 12.34 for D-glucose;9
alcohols possess pKa’s ꢀ17,4b so the inclusion of addi-
tional b oxygen atoms substantially lowers the pKa.
Note also that the electron withdrawing enol ether
moiety in IV would lower the pKa further.) Thus, the
process shown in III would be relatively favoured
thermodynamically, and possibly lead to ‘viable’
amounts of the incipient IV. (In fact, additional electro-
philic assistance by 1 itself could provide a further
kinetic advantage to the above deprotonation of 4 (cf.
Scheme 3); this would lead to the O-acetyl derivative
VIII, which could be O-deacetylated with acetate to
yield IV for further reaction with 2.)
The fact that employing excess benzaldehyde (2) leads to
vastly improved yields, suggests that 2 is the electro-
philic species involved. Furthermore, the high yields of
3 obtained from the reaction of diacetate 4 with KOBut
prove that it is possible to generate the enolate anion of
4, and effect the Perkin reaction with it. Of course, this is
not necessarily valid under the conditions of the original
Perkin reaction: clearly, the weakly basic acetate ion
normally employed therein could not normally deproto-
nate 4 (estimated pKa ꢀ 24).4 However, an interesting
assisted mechanism is possible, in which the ‘spectator’
acetate moiety in 4 participates electrophilically in the
deprotonation as shown in III, leading to the orthofor-
mate anion IV (Scheme 2).
The final formation of 3 follows from the initial addition
of IV to 2 yielding V, and the succeeding reactions via
VI and VII in Scheme 2. It is interesting and noteworthy
that the formation of VII from VI is accompanied by the
O
OAc
O
AcO
O
-AcO-
-AcOH
OAc
O
O
Interestingly, in IV the enolate moiety would gain
thermodynamic stability by covalent bonding of its
oxygen atom to the orthoformate carbon atom. In fact,
IV may be estimated to possess a pKa of ꢀ11, a value
closer—than that of 1 (pKa possibly >20)—to the pKa
Ph
O
Ph
O
O
H
VIII
III
Scheme 3.

S. Chandrasekhar, P. Karri / Tetrahedron Letters 47 (2006) 2249–2251
2251
release of a molecule of benzaldehyde (2), which thus
apparently acts as both substrate and catalyst. This indi-
cates that the reaction would slow down markedly as it
progresses, and also explains the improved yields
obtained with excess benzaldehyde.
acetic anhydride itself—adds to the aldehyde in the key
step. The deprotonation of the diacetate to the enolate
appears to be electrophilically assisted by the neighbour-
ing acetate group. An important practical consequence
of these studies is the development of relatively mild con-
ditions for effecting the overall transformation of the
Perkin condensation (employing the diacetate as the
substrate).
Clearly, the above mechanism requires the presence of
both benzaldehyde (2) and diacetate 4, which presum-
ably exists in equilibrium under the reaction conditions
(cf. Scheme 2). The poor yields with excess acetic anhy-
dride (1), mentioned above, most likely result from a
shift of this equilibrium towards 4, with a corresponding
depletion of 2. In fact, since 4 is quite stable and easily
isolable, this equilibrium would largely favour 4 under
normal conditions, only to be reversed to form viable
amounts of 2 (and 1) at the high temperatures employed
in the Perkin reaction. The higher yields obtained with
excess benzaldehyde (2) are compatible with the possi-
bility that the addition of IV to 2 is the rate determining
step in the overall reaction. However, the dependence of
the overall rate on [2] is likely to be complex, in view of
the catalytic role for 2 (proposed above).
Acknowledgements
We are grateful to the editor and several referees of this
journal for constructive criticism and insightful com-
ments, which greatly aided in the preparation of the final
manuscript.
References and notes
1. Rosen, T. In Comprehensive Organic Synthesis; Trost, B.
M., Fleming, I., Heathcock, C. H., Eds.; Pergamon:
Oxford, 1991; Vol. 2, pp 395–408.
The interesting formation of 3 from 4 and KOBut (Table
1, entry 5) possibly occurs via an alternative mechanism,
and is apparently only obliquely relevant to the present
discussion. In addition, the formation of 3 from 4 under
a variety of relatively mild conditions (Table 1, entries
3–5), is noteworthy. As 4 is readily prepared and iso-
lated from 1 and 2 under relatively mild conditions, this
(effectively) defines a considerable improvement in
methodology for the Perkin condensation: this essen-
tially avoids high temperatures in excess acetic anhy-
dride, and should suit relatively sensitive substrates.
2. House, H. O. Modern Synthetic Reactions, 2nd ed.; W. A.
Benjamin, Inc.: Menlo Park, 1972, pp 660–663.
3. Johnson, J. R. Org. React. (N.Y.) 1942, 1, 210–265, and
references cited therein.
4. (a) Amyes, T. L.; Richard, J. P. J. Am. Chem. Soc. 1996,
118, 3129–3141, and references cited therein; (b) Ref. 2, p
494.
5. Brady, W. T.; Gu, Y.-Q. J. Org. Chem. 1988, 53, 1353–
1356.
6. Tietze, L. E.; Eicher, T. Reactions and Syntheses in the
Organic Chemistry Laboratory; University Science Books:
Mill Valley, 1989, p 567.
7. Man, E. H.; Sanderson, J. J.; Hauser, C. R. J. Am. Chem.
Soc. 1950, 72, 847–848.
8. Kochhar, K. S.; Bal, B. S.; Deshpande, R. P.; Rajadhyak-
sha, S. N.; Pinnick, H. W. J. Org. Chem. 1983, 48, 1765–
1767.
In conclusion, the available evidence indicates that the
accepted mechanism of the Perkin reaction needs to be
substantially revised. Essentially, it appears that the eno-
late of the gem-diacetate derived from the aromatic alde-
hyde and acetic anhydride—rather than the enolate of
9. Dictionary of Natural Products; Buckingham, J., Ed.;
Chapman and Hall: London, 1994; Vol. 3, p 2578.