Dakin-West synthesis of β-aryl ketones

Article abstract of DOI:10.1021/jo0607966

Triethylamine and 1-methylimidazole were found to be selective catalysts for the Dakin-West synthesis of diaryl ketones and aryl methyl ketones, respectively. In the 1-methylimidazole-catalyzed reaction, catalysis is due to the simultaneous formation of both an effective acylating agent, 1-acyl-3-methylimidazolium, and a base, carboxylate anion. Hydrocinnamic acid, a compound previously reported to be unreactive under Dakin-West conditions, forms 4-phenyl-2-butanone when the reaction is catalyzed by 1-methylimidazole.

Full text of DOI:10.1021/jo0607966

Dakin-West Synthesis of â-Aryl Ketones
Khanh-Van Tran and David Bickar*
Department of Chemistry, Smith College,
Northampton, Massachusetts 01063
FIGURE 1. The reaction of amino acids with acetic anhydride reported
by Dakin and West .
1
dbickar@email.smith.edu
ReceiVed April 14, 2006
1-methylimidazole (MIM), the scope of the Dakin-West
reaction can be extended to compounds previously believed to
be unreactive. We also suggest an alternative to the previously
1
2,13
proposed mechanisms for this reaction.
The most efficient catalyst identified to date for this reaction
is probably 4-(dimethylamino)pyridine (DMAP),1
4,15
but as
shown below, several compounds can be used to catalyze the
acid-ketone transformation (Table 1). Of these, we found that
MIM has several advantages. It is less expensive and less toxic
than DMAP and, as a liquid, often easier to use. Like DMAP,
it permits the Dakin-West reaction to proceed at room
temperature with excellent overall yields. With MIM, hydro-
cinnamic acid (entry 6), a compound that is unresponsive to
Triethylamine and 1-methylimidazole were found to be
selective catalysts for the Dakin-West synthesis of diaryl
ketones and aryl methyl ketones, respectively. In the 1-meth-
ylimidazole-catalyzed reaction, catalysis is due to the simul-
taneous formation of both an effective acylating agent,
1
1
pyridine catalysis, was converted to 4-phenyl-2-butanone in
50% yield, and phenylacetic acid (entry 2), 4-nitrophenylacetic
∼
1-acyl-3-methylimidazolium, and a base, carboxylate anion.
acid (entry 9), and 4-methoxyphenylacetic acid (entry 10) gave
almost quantitative yields. In contrast, 2-phenylbutanoic acid
Hydrocinnamic acid, a compound previously reported to be
unreactive under Dakin-West conditions, forms 4-phenyl-
(entry 11) showed only 6% conversion to 3-phenyl-2-pentanone,
2-butanone when the reaction is catalyzed by 1-methylim-
and ethyl 2-phenylacetate (entry 12) showed no reaction even
with prolonged heating.
idazole.
Although we found 1-benzylimidazole (entry 3) to be an
effective, if somewhat slower, catalyst than MIM, 1,2-dimeth-
ylimidazole (entry 4) was much less efficient, giving only 12%
product yield and poorer selectivity than MIM and 1-benzylim-
idazole under the same reaction conditions.
For phenylacetic acid (1), the ratios of products 3 and 5
differed with the catalyst employed. Using MIM or 1-benzyl-
imidazole, the ratios of products 3 and 5 were 94:6 and 90:10,
respectively (entry 2 and entry 3). Using 1,2-dimethylimidazole
or triethylamine (TEA), the ratios of products 3 and 5 were
The Dakin-West reaction, originally carried out in acetic
anhydride and pyridine, is the best known route for the synthesis
1
,2
of â-acetoamido ketones from R-amino acids (Figure 1).
Among other applications,3 Woodward used the Dakin-
,4
5
West reaction in his classic synthesis of strychnine, and more
recently, Fu used a modified Dakin-West reaction to asym-
6
metrically C-acylate acyclic silyl ketene acetals. The Dakin-
West reaction also has been used successfully to make ketones
from aryl acetic acids (Figure 2),7 and this reaction is the
subject of our study.
-9
7
2:28 and 6:94, respectively (entry 4 and entry 5). No reaction
was observed for hydrocinnamic acid with TEA alone (entry
). However, reactions carried out with both TEA and MIM
Application of the Dakin-West reaction to aryl acetic acids
has had two important limitations: (1) it required that the
8
present produced more 1,5-diphenyl-3-pentanone than did
reactions with only MIM present (entry 6 and entry 7).
In some trials, reactions with 1 gave product mixtures
containing over 30% of 3 and 5 as their enol esters (Figure 3).
Surprisingly, although the enol esters were sometimes a large
fraction of the product mixture, little or no enol ester was
7
carboxylic acid form an anhydride with an acidic R-CH group,
and (2) the reaction yielded two ketone products; for example,
phenylacetic acid (1) and acetic anhydride (2) with pyridine
(
entry 1) give phenyl-2-propanone (3) and 1,3-diphenyl-2-
10,11
propanone (5).
In this work we show that the major product can be
determined by the choice of catalyst, and that by using
1
13
detected in the H NMR or C NMR spectra of the reaction
medium during the course of the reaction. Instead, it appears
that the enol ester formed during the workup from residual acetic
acid and the ketone products. The enol ester content could be
reduced to <2% of the total product by thoroughly washing
the initial organic phase with aqueous potassium bicarbonate
before removing the solvent.
(
1) Dakin, H. D.; West, R. J. Biol. Chem. 1928, 78, 91.
(2) Khodaei, M.; Khosropour, A.; Fattahpour, P. Tetrahedron Lett. 2005,
4
1
995, 36, 4797.
(
4) Kawase, M.; Hirabayashi, M.; Koiwai, H.; Yamamoto, K.; Miyamae,
H. Chem. Commun. 1998, 641.
5) Woodward, R. B.; Cava, M. P.; Ollis, W. D.; Hunger, A.; Daeneker,
H. U.; Schenker, K. Tetrahedron 1963, 19, 247.
1
We used H NMR to monitor the disappearance of phenyl-
(
acetic acetic anhydride (20) and 1 and the formation of 3 and
5
. The time course of reaction showed apparent first-order
(
(
(
(
(
(
6) Mermerian, A.; Fu, G. J. Am. Chem. Soc. 2005, 127, 5604.
7) Buchanan, G. L. Chem. Soc. ReV. 1988, 17, 91.
8) Nicholson, J. J. Chem. Educ. 2004, 81, 1362.
9) Smith, G. G. J. Am. Chem. Soc. 1952, 75, 1134.
10) King, J.; McMillian, F. J. Am. Chem. Soc. 1954, 77, 2814.
11) King, J.; McMillian, F. J. Am. Chem. Soc. 1951, 73, 4911.
(12) Smith, G. G.; Fahey, D. J. Am. Chem. Soc. 1958, 81, 3391.
(13) Allinger, N.; Wang, G.; Dewhurst, B. J. Org. Chem. 1973, 39, 1730.
(14) Hofle, G.; Steglich, W. Angew. Chem., Int. Ed. Engl. 1969, 8, 981.
(15) Scriven, E. F. Chem. Soc. ReV. 1983, 12, 129.
10.1021/jo0607966 CCC: $33.50 © 2006 American Chemical Society
6
640
J. Org. Chem. 2006, 71, 6640-6643
Published on Web 07/15/2006

FIGURE 2. The Dakin-West reaction of phenylacetic acid.
TABLE 1. Reaction Conditions, Yields, and Product Ratios for Dakin-West Syntheses of â-Aryl Ketones
a
The reaction medium contained 5 equiv of acetic anhydride. See experimental procedure for other details. b MIM ) 1-methylimidazole; TEA )
c
1
d
triethylamine. Yields were calculated based on the isolated product mixture, except entry 9 which was determined by H NMR. The numbers in parentheses
indicate the relative mole percentage of that compound in the isolated product mixture. The reaction mixture was refluxed for the duration of the reaction.
e
kinetics with respect to 20 with rate constants of (4.2 ( 0.4) ×
the carboxylic acid concentration.11,16 This suggested to previous
investigators that pyridine was acting as a general base,
catalyzing the reaction by removing a proton from the R-carbon
of the â-aryl acetic anhydride. However, we found this assump-
tion to be less tenable after observing effective catalysis with
-
3
-1
-3
-1
1
0
min and (1.7 ( 0.2) × 10 min for the 1-methylim-
idazole- and 1-benzylimidazole-catalyzed reactions, respectively
Figure 4). This reaction order is consistent with that reported
(
by other investigators using pyridine to catalyze the Dakin-
West reaction.1
2,13
Pyridine catalysis of the Dakin-West reaction requires the
catalyst to be present in large excess, typically 6 or more times
(
16) Godfrey, A.; Brooks, D.; Hay, L.; Peters, M.; McCarthy, J.; Mitchell,
D. J. Org. Chem. 2003, 68, 2623.
J. Org. Chem, Vol. 71, No. 17, 2006 6641

such as DMSO and acetonitrile,17,18 suggesting that 4a would
be capable of abstracting the R-hydrogen of 20 to form 20a in
acetic anhydride. The subsequent reaction of 20a with 19a forms
the tetrahedral intermediate 21 (Scheme 1).
It is possible to predict the rate-limiting step in Scheme 1.
The first step of this mechanism is unlikely to be rate limiting
because the reaction is first order with respect to 20. The second
step is also unlikely because the anhydride transfer is relatively
rapid, reaching equilibrium within a few minutes in our reaction
conditions. Later steps f, g, and h are probably not rate limiting
because the transient formation of 22, 23, or 24 was not observed
in the NMR spectra of the reaction medium during the reaction.
By elimination, either the formation or breakdown of 21 is likely
to be the rate-limiting step (steps d or e of Scheme 1). Previous
studies have shown that for acyl transfer reactions in aprotic
solvents, the formation or breakdown of a tetrahedral intermedi-
FIGURE 3. Enol esters 17 and 18, formed by reaction of acetic acid
with 3 and 5, respectively.
1
7,19,20
ate is frequently rate limiting.
In either case, the reaction
rate would depend on the concentrations of the enolate and the
6
acyl-imidazolium precursors. This is consistent with our
observation that both nitro and methoxy groups promote the
reaction (entries 9 and 10), presumably by stabilizing the enolate
by resonance or induction, and with our observation that the
reaction is first order with respect to 20. With reactants that are
less able to form an enolate, such as ethyl phenylacetate, the
reaction proceeds poorly or not at all. The poor yield of
3-phenyl-2-pentanone from phenylbutyric acid probably reflects
the steric hindrance at the R-carbon as well as the relative
instability of the tertiary carbanion formed as an intermediate.
In contrast to catalysis with MIM, TEA catalysis gives a diaryl
ketone as the major product. Results similar to those with TEA
have been observed using sodium acetate as the catalyst,1
suggesting that these catalysts promote a different mechanism
of acid-ketone conversion. We propose that TEA and acetate
may act by general base catalysis rather than nucleophilic
catalysis. The difference is notable; for the reaction of 1,
nucleophilic (MIM) catalysis gives a product ratio of 3 to 5 of
1,21
FIGURE 4. Concentrations of reactants and products over the course
of the reaction: (A) MIM catalyzed; (B) 1-benzylimidazole catalyzed.
Data shown are for ()) phenylacetic acid (1), (O) phenylacetic acetic
anhydride (20), (1) phenyl-2-propanone (3), and (2) 1,3-diphenyl-2-
propanone (5). After equilibration of the phenylacetic acid and acetic
anhydride, the reaction was initiated by the addition of 0.5 equiv of
catalyst. Samples of the reaction mixture were removed at regular
intervals after initiation, and the reactants’ and products’ relative
9
4:6 while base (TEA) catalysis gives a 3 to 5 ratio of 6:94.
Previously, Smith and Fahey proposed a general base mecha-
1
2
nism for pyridine catalysis. In this broadly accepted mecha-
nism, the Dakin-West reaction is a base-catalyzed condensation
of anhydrides that forms a methyl aryl ketone as the major
product. Our results suggest that the predominant reaction Smith
and Fahey observed was actually nucleophilic catalysis by
pyridine rather than base catalysis and that base catalysis alone
leads to the diaryl ketone as the major product.
1
concentrations were determined by H NMR, and normalized to the
initial concentration of phenylacetic acetic anhydride and phenylacetic
acid.
MIM even when its concentration was half that of the carboxylic
acid present in the reaction mixture. Moreover, although a
general base catalyst must be able to abstract a proton from the
R-carbon of 20 to make 20a (Scheme 1), we found no
correlation between the effectiveness of the catalyst and the pKa
of its conjugate acid. For example, MIM is a better catalyst
than TEA, but its conjugate acid has a lower pKa.
MIM, 1-benzylimidazole, DMAP, and to a lesser extent,
pyridine, give a methyl ketone as their major product with acetic
anhydride (Table 1). These catalysts can all react with an
anhydride to generate a powerful acylating agent and a car-
boxylate anion. In keeping with this, we propose that for aryl
acetic acids, the Dakin-West reaction occurs as shown in
Scheme 1.
Base catalysis may follow a similar reaction sequence to that
shown in Scheme 1, but with the critical tetrahedral intermediate
formed from 20a and 20 (Figure 5). Alternatively, as has been
previously suggested, this reaction may occur by a mechanism
involving either an intramolecular OsC migration (Figure 5)
5,11,22
or the bimolecular equivalent.
Whatever the mechanism,
for reasons that remain unclear, base catalysis generates the
diaryl ketone (5) as the major product.
Compared to MIM and 1-benzylimidazole, 1,2-dimethylim-
idazole catalyzes the reaction of 1 much less effectively. This
implies that the methyl group on the 2-position of 1,2-
(17) Hajdu, J.; Smith, G. J. Am. Chem. Soc. 1981, 103, 6192.
(18) Kolthoff, I. M.; Chantooni, M. K.; Bhowmik, S. J. Am. Chem. Soc.
1967, 90, 23.
In this mechanism, MIM is actually a pre-catalyst, that upon
reaction with 2 produces the true catalysts of the reaction, acyl-
imidazolium cation (19a) and carboxylate anion (4a). The pKa
of acetic acid is between 12 and 22 in aprotic organic solvents
(19) Fersht, A.; Jencks, W. J. Am. Chem. Soc. 1970, 92, 2622.
(20) Fersht, A.; Jencks, W. P. J. Am. Chem. Soc. 1970, 92, 5432.
(21) Hurd, C.; Thomas, C. J. Am. Chem. Soc. 1936, 58, 1240.
(22) Buchanan, G. L.; McArdle, J. J. Chem. Soc. 1952, 2944.
6
642 J. Org. Chem., Vol. 71, No. 17, 2006

a
SCHEME 1. Mechanism for Nucleophilic Catalysis of the Dakin-West Reaction by 1-Methylimidazole
a
Reaction steps are (a) formation of the acyl-imidazolium, 19a, and the carboxylate anion, 4a; (b) acid-anhydride acetyl transfer; (c) abstraction of an
R-hydrogen from the â-aryl anhydride 20 to form the enolate 20a; (d) nucleophilic attack on 19a by 20a to form a tetrahedral intermediate 21; (e) breakdown
of 21 by the loss of 19 to form 22; (f) attack on the acetyl carbonyl of 22 by 19 to form 23; (g) formation of 24 by the transfer of an acetyl group to 19;
(h) the subsequent decarboxylation of 24 to release CO2; and (i) protonation of the enolate 3a to form phenyl-2-propanone, 3.
Experimental Section
General Procedure. In a typical reaction, an aryl carboxylic
acid (0.05 mol, 1 equiv) was dissolved in 2 (0.25 mol, 25 mL, 5
equiv), at room temperature, and the solution was stirred and purged
with N
addition of catalyst (0.025 mol, 0.5 equiv), and the reaction was
continuously purged with a slow flow of N over the course of the
2
for several minutes. The reaction was initiated by the
2
FIGURE 5. (a) Tetrahedral intermediate (25) proposed for a stepwise
mechanism for base catalysis similar to the mechanism for nucleophilic
catalysis shown in Scheme 1; (b) mechanism for base catalysis via a
reaction. At regular intervals during the reaction, 100 µL samples
of the reaction mixture were transferred to an NMR tube containing
, and their 1H and C NMR spectra were immediately
13
CDCl
3
5
,11,22
concerted O-C transfer.
recorded. Reactions were typically allowed to run for 12 or 48 h,
depending on the catalyst and reactants. For the reactions using
MIM, NMR results showed that in most cases the reaction reached
completion in 10 h or less.
Reactions with hydrocinnamic acid, 3-phenylbutyric acid and
ethyl phenylacetate (entries 6-8, 11, and 12) followed the procedure
described above except the reaction mixture contained more catalyst
dimethylimidazole hinders the formation or reaction of the acyl-
imidazolium ion, thus limiting nucleophilic catalysis without
affecting base catalysis. These results are consistent with reports
that 2-methyl substituted pyridines are much less effective as
nucleophilic reagents.15,23
(0.1 mol, 2 equiv) and was heated to reflux over the course of the
Simple general base catalysis is less efficient than nucleophilic
catalysis. For 1, the reactions were slower and had lower yields
when catalyzed by TEA than by MIM. Hydrocinnamic acid with
acetic anhydride showed no reaction when treated with TEA
but underwent an acid-ketone transformation with MIM. King
and McMillan reported that hydrocinnamic acid remained
reaction.
1
H NMR Kinetics. The relative concentrations of 1, 3, 5, and
1
2
0 in the reaction mixture were determined using their H NMR
peak areas. This information was used to generate a best-fit curve
for the reaction time course and to determine the reaction order
for 20.
10,11
unchanged using pyridine catalysis.
Our results suggest that
Workup. After completion, water (10 mL) was added to the
reaction flask to hydrolyze 2. The reaction mixture was extracted
with ethyl acetate (3 × 50 mL), and the extracts combined and
washed with saturated potassium bicarbonate (2 × 50 mL) followed
by water (2 × 50 mL), then dried over magnesium sulfate and
filtered. Removing the solvent by rotary evaporation gave the
with an appropriate catalyst, this reaction can be applied to a
broader range of compounds than previously believed.
In conclusion, we show that by choosing the right catalyst it
is possible to obtain either a methyl aryl ketone or a diaryl
ketone as the major product of the Dakin-West reaction. A
catalyst that can generate an effective acylating agent and a
strong base in aprotic solvents (such as MIM and DMAP) will
convert anhydrides with only mildly acidic R-carbons to aryl
methyl ketones. In contrast, catalysts that cannot facilitate
nucleophilic catalysis (such as TEA or sodium acetate) require
more acidic C-H groups for reaction and give diaryl ketones
as their major products.
1
13
product mixture, which was characterized by H NMR, C NMR,
and GC/MS.
Acknowledgment. We thank the Howard Hughes Medical
Institute and Smith College for funding and Charles Amass for
technical assistance in this project.
Supporting Information Available: 1H NMR and 13C NMR
of products. This material is available free of charge via the Internet
at http://pubs.acs.org.
(23) Butler, A. R.; Gold, V. J. Chem. Soc. 1961, 4362.
JO0607966
J. Org. Chem, Vol. 71, No. 17, 2006 6643