Observable enols of anhydrides: Claimed literature systems, calculations, and predictions

Article abstract of DOI:10.1002/1522-2675(20010613)84:6<1405::AID-HLCA1405>3.0.CO;2-G

The literature describing the observation of enols of carboxylic anhydrides and mixed carboxylic-sulfuric anhydrides was examined. In the phenylbutyric anhydride system, the alleged enol was shown to be ethylphenylketene, and the monoenol EtC(Ph)=C(OH)OC(=O)CH(Ph)Et (5) and the dienol (6) should not be observed according to calculations. Calculations also show that the claimed enols H₂C=C(OH)OSO₂Y, Y= SO₃^-, Ac (15) and the enol of 2H-pyran-2,6(3H)-dione (7) are too unstable to be observed. The bulky enols of β,β-ditipylacetic formic (35a) or trifluoroacetic (35b) anhydride were calculated to be unstable with pK_Enol=7.7 (6.2). The suggestion that compounds with the 3-acyl or 3-aroyl-2H-pyran-2,6(3H)-dione skeleton are enolic was examined. In the solid state, all the known structures show that enolization takes place on C(5)=O. However, B3LYP/6-31G** calculations show that, for 3-acetyl-4-methyl-2H-pyran-2,6(3H)-dione (10, R¹=Me, R²=H), which is completely enolic, the enol on the acetyl group (cf. 12) is only 0.9 kcal/mol more stable than the enol on the anhydride (cf. 11). Calculations also revealed that 3-(trifluoroacetyl)-2H-pyran-2,6(3H)-dione (28) should exist in nearly equal amounts of the enol of anhydride (cf. 30) and the enol of the acyl group (cf. 29), whereas the enol of anhydride (cf. 32) is the only stable species for 3-(methoxycarbonyl)-2H-pyran-2,6(3H)-dione (31). Furan-2,5-diol (27) and 5-hydroxyfuran-2-one (26) are calculated not to give observable isomers of succinic anhydride (25) (pK_Enol=30 and 18, resp.) in spite of the expected aromatic stabilization of 27. Surprisingly, the calculations reveal that the enol (NC)₂C=C(OH)OCHO (38) is less stable than its tautomeric anhydride (37) (pK_Enol=1.6). Comparison of calculated pK_Enol values for (NC)₂CHC(=O)X (41) and MeC(=O)X indicates that the assumption that substitution by two β-CN groups affects similarly all the systems regardless of X is incorrect. A pK_Enol((NC)₂CHC(=O)X) vs. pK_Enol(MeC(=O)X) plot is linear for most substituents with severe and mild negative deviations, respectively, for X=NH₂ and MeO. Appropriate isodesmic reactions have shown that the β,β-(CN)₂ substitution increases the stabilization of the enol of amide (X=NH₂) by 14.6 kcal/mol over that for the anhydride (X=OCHO), whereas the amide form is 7.1 kcal/mol less destabilized than for the anhydride. The pK_Enol value for (MeOCO)₂CHCOOCHO (43) is 3.6, i.e., stabilization by these βelectron-withdrawing groups is insufficient to make the enols observable.

Full text of DOI:10.1002/1522-2675(20010613)84:6<1405::AID-HLCA1405>3.0.CO;2-G

Helvetica Chimica Acta ± Vol. 84 (2001)
1405
Observable Enols of Anhydrides: Claimed Literature Systems, Calculations,
and Predictions
a
a
b
by Zvi Rappoport )*, Yi Xiong Lei ), and Hiroshi Yamataka )*
a) Department of Organic Chemistry and the Minerva Center for Computational Quantum Chemistry,
The Hebrew University of Jerusalem, Jerusalem 91904, Israel
b) Institute of Scientific and Industrial Research, Osaka University, Ibaraki, Osaka 567-0047, Japan
Dedicated to Prof. Edgar Heilbronner on the occasion of his 80th birthday
The literature describing the observation of enols of carboxylic anhydrides and mixed carboxylic-sulfuric
anhydrides was examined. In the phenylbutyric anhydride system, the alleged enol was shown to be
ethylphenylketene, and the monoenol EtC(Ph)ˆC(OH)OC(ˆO)CH(Ph)Et (5) and the dienol (6) should not
be observed according to calculations. Calculations also show that the claimed enols H
2
CˆC(OH)OSO Y, Y ˆ
2
�
SO
3
, Ac (15) and the enol of 2H-pyran-2,6(3H)-dione (7) are too unstable to be observed. The bulky enols of
b,b-ditipylacetic formic (35a) or trifluoroacetic (35b) anhydride were calculated to be unstable with pKEnol ˆ 7.7
(6.2). The suggestion that compounds with the 3-acyl or 3-aroyl-2H-pyran-2,6(3H)-dione skeleton are enolic
was examined. In the solid state, all the known structures show that enolization takes place on C(5)ˆO.
1
However, B3LYP/6-31G** calculations show that, for 3-acetyl-4-methyl-2H-pyran-2,6(3H)-dione (10, R ˆ Me,
2
R ˆ H), which is completely enolic, the enol on the acetyl group (cf. 12) is only 0.9 kcal/mol more stable than
the enol on the anhydride (cf. 11). Calculations also revealed that 3-(trifluoroacetyl)-2H-pyran-2,6(3H)-dione
(
28) should exist in nearly equal amounts of the enol of anhydride (cf. 30) and the enol of the acyl group (cf. 29),
whereas the enol of anhydride (cf. 32) is the only stable species for 3-(methoxycarbonyl)-2H-pyran-2,6(3H)-
dione (31). Furan-2,5-diol (27) and 5-hydroxyfuran-2-one (26) are calculated not to give observable isomers of
succinic anhydride (25) (pKEnol ˆ 30 and 18, resp.) in spite of the expected aromatic stabilization of 27.
Surprisingly, the calculations reveal that the enol (NC)
2
CˆC(OH)OCHO (38) is less stable than its tautomeric
anhydride (37) (pKEnol ˆ 1.6). Comparison of calculated pKEnol values for (NC)
MeC(ˆO)X indicates that the assumption that substitution by two b-CN groups affects similarly all the systems
regardless of X is incorrect. A pKEnol((NC)
CHC(ˆO)X) vs. pKEnol(MeC(ˆO)X) plot is linear for most
substituents with severe and mild negative deviations, respectively, for X ˆ NH and MeO. Appropriate
isodesmic reactions have shown that the b,b-(CN) substitution increases the stabilization of the enol of amide
X ˆ NH ) by 14.6 kcal/mol over that for the anhydride (X ˆ OCHO), whereas the amide form is 7.1 kcal/mol
less destabilized than for the anhydride. The pKEnol value for (MeOCO) CHCOOCHO (43) is 3.6, i.e.,
stabilization by these b-electron-withdrawing groups is insufficient to make the enols observable.
2
CHC(ˆO)X (41) and
2
2
2
(
2
2
1. Introduction. ± Enols of carboxylic acid derivatives are rare species [1]. The
calculated energy differences between the favored tautomer, i.e., the parent (b,b-di-H)
carboxylic acid or its derivatives (2) and its enols (1) are 24 ± 33 kcal/mol, depending on
the derivative and the method of calculation [2]. Nevertheless, enols of carboxylic
acids, esters, and amides were observed in recent years, based on three approaches.
Their generation by a fast reaction technique such as flash photolysis as short lived
species was developed by Kresge and Wirz (for a review, see [1b]), their formation from
bulky b,b-diaryl-substituted ketenes as observable species, although of short or moderate
life-time was developed by Hegerty and co-workers [3] and by us [4], and their generation
as stable species was developed by us, using systems carrying strongly b-electron-

1406
Helvetica Chimica Acta ± Vol. 84 (2001)
withdrawing groups (EWGs) [5]. Such enols, e.g., of esters were also suggested as unob-
servable intermediates on the basis of kinetic and stereochemical evidence (see, e.g., [6]).
The calculations of the parent system give a rough order of the relative stability,
measured as pKEnol (ˆ � log KEnol) of Scheme 1, as a function of the a-substituent X [2].
The order of pKEnol values for X in CH ˆC(OH)X is: H (9.9) < Me(11.5) < OCHO
2
(
(
17.6) < Br (18.7) ꢀ Cl (18.6) < F (19.8) < NMe (21.2) ꢀ NH (21.3) < OH (22.0), OMe
2
2
22.3) at B3LYP/6-31G**. In practice, when a competition between an amide and an
ester group is possible, formation of enols of amides is more facile than formation of
enols of esters [5a,b] as expected from the calculations.
Scheme 1
According to these calculations, enols of anhydrides should be relatively more
stable than the other enols of carboxylic acid derivatives 1 [2]. Hence, their preparation
is potentially more favored than that of enols of systems with, e.g., X ˆ NHR or MeO,
which were already observed.
Several suggestions for observation, detection, or intermediacy of enols of
anhydrides appear in the literature. As early as 1911 and 1912, Thorpe and co-workers
[
7] had written 3-substituted glutaconic anhydrides in the enol form, although with no
‡
evidence. Deno et al. suggested in 1970 [8] that alkylation of Ac O with Ph C cation
2
3
proceeds by electrophilic reaction on traces of the enol of the anhydride 3 (Scheme 2).
Except for the product, there is no evidence for this suggestion, and the fact that
‡
acetone is not alkylated by Ph C under more drastic conditions does not support this
3
assumption. Note that the percentage of 3 in Ac O is extremely low according to our
2
calculations [2].
Scheme 2
In two works, the isolation and observation of an enol of carboxylic acid anhydride
was reported. Gordon and co-workers [9] distilled optically active 2-phenylbutyric acid
anhydride and found that it undergoes racemization, that the distillate is bright yellow,
and that the color disappears after a few hours. They interpreted the NMR and IR
spectra of the freshly distilled anhydride as indicative of the presence of three species,
the diketone (i.e., anhydride, 4), keto-enol (i.e., monoenol, 5) and the yellow dienol (6;
Scheme 3), and that an equilibrium between the species is established in CCl4.
Kagan et al. [10], in a re-analysis of tautomerism and isomerism in the a- and g-
methylglutaconic acid and derivatives, prepared g-methylglutaconic anhydride (5-
1
methyl-2H-pyran-2,6(3H)-dione) 7 and identified it by its H-NMR spectrum (d 2.30
(
Me), 3.58 (CH ), and 6.47 (ˆCH)). The compound did not isomerize on distillation,
2
in spite of an earlier claim [11], neither when HCl was bubbled into its CDCl solution
3

Helvetica Chimica Acta ± Vol. 84 (2001)
1407
Scheme 3
nor when heated in TFA at 1008 for 60 h. However, when a drop of conc. H SO was
2
4
1
added to the TFA solution, the H-NMR changed (d 2.28 (Me), 6.72, 8.38 (2d, J ˆ
8
Hz)). It was deduced from the spectrum that enolization to rapidly exchanging
isomeric enols of the anhydride 8 and 9 took place (Scheme 4). Analogs such as the
parent compound (i.e., with 5-H) or the 4-Me derivative were also written as existing in
an equilibrium analogous to that of 8 and 9, although with little evidence [12].
Scheme 4
From our calculations and experience, it is unlikely that species 9 was indeed
observed. However, a family of 3-acetyl-2H-pyran-2,6(3H)-diones 10 was suggested to
exist in the H-bonded enol forms 11/12 [13], on the basis of rather convincing NMR, IR,
and UV-spectral evidence. Enols 11 belong to the general case of systems carrying b-
electron-withdrawing groups, of which we previously observed enols of amides and
esters [5], and hence enols 11 are predicted to be relatively stable. The enolic structure
was suggested also in some related structures [14].
A complexed enol of an anhydride 13 was suggested by Hunt and Satchell [15] on
the basis of IR spectra.
The observation of mixed acetic sulfuric anhydrides was also reported. Truce and
Olsen [16] suggested an intermediate in the side-chain sulfonation of phenylalkanoic
acids with SO3 in dioxane, which has one resonance hybrid of the structure

1408
Helvetica Chimica Acta ± Vol. 84 (2001)
PhCH CHˆC(OH)OSO H. Montoneri and co-workers [17] have suggested that, in
2
3
equimolar SO /Ac O or SO /AcOH mixtures the keto form 14 exists in equilibrium
3
2
3
with the enol 15 of the mixed anhydrides at ꢁ 253 K (Scheme 5), and that 15 is
observable by NMR and IR spectroscopy. The enol was suggested to be involved in
elimination to ketene [14] and in chlorination of AcOH [17b]. From NMR band areas,
the 15b/14b ratio is 2 in Ac O and the 15a/14a ratio is 0.67 in AcOH. The d values for
2
the vinylic protons of 15 are assigned as 2.57 and 2.75, and a rapid exchange between
1
4a and 15a was suggested. The assignment of structure 15b in SO /Ac O is based also
3 2
�
1
on assignment of IR absorptions at 1600 and 1555 cm , but with no unequivocal evidence.
A low field OH signal is observed at d 15.13 and, since its position is concentration-
independent, an intramolecular H-bond (cf. 16b) is suggested. The enol is not observed
at room temperature. Additional species are formed in some of these reactions.
Scheme 5
In AcOH/SO 1:1 in SO , the changes in the IR spectrum after several hours at
3
2
ꢂ 253 K suggest the establishment of an equilibrium 14a > 15a, although this was not
established unambiguously.
Ogata and Adachi [18] have suggested on the basis of some kinetic evidence that a-
halogenation of aliphatic acids proceeds via the analog of 16 with Yˆ H and H Cˆ
2
replaced by RR'Cˆ. According to them, enols of carboxylic anhydrides are not the
intermediates in halogenation of carboxylic acids.
Consequently, we conclude that the present experimental evidence for the
observation of enols of anhydrides is meager and mostly questionable. We, therefore,
analyze below the literature data with the aid of former and new calculations as well as
with a few experiments to exclude unequivocally the formation of these species in some

Helvetica Chimica Acta ± Vol. 84 (2001)
1409
cases and to search for evidence for them in other cases, and as a guide for attempts to
identify such species unequivocally in the future.
2
. Results and Discussion. ± The 2-Phenylbutanoic Anhydride System. Our
experience and the results of the calculations mentioned above indicate that it is
extremely unlikely that 4 will enolize to an observable mono-enol 5 and certainly not to
a bis-enol 6. Mandelic acid and 2-cyano-2-phenylacetic acid, whose skeletons contain
the EWGs OH or CN, respectively, instead of the Et group of 18 (cf. Scheme 6), have
pKEnol values in H O of 15.6 [19a] and 7.2 [19b], and even if the anhydride pKEnol is 4.4
2
units lower than that for the acid (disregarding the less-polar solvent) no enol is
supposed to be detected by IR or NMR.
The spectral evidence brought in support of the enol structure in the freshly distilled
�
1
sample includes disappearance of a strong sharp IR peak at 2100 cm that disappears
�
1
at the same rate as one at 1700 cm . No specific assignment could be made for this
�
1
peak [9]. A broad disappearing peak at 3000 cm was ascribed to an intermolecular H-
bond between 4 and 6. The enol OH signal was assumed to be at d 11.5.
We suggest that the process occurring on distillation is an elimination of the
carboxylic acid (18) with formation of the ketene 17 (Scheme 6). This is the reversal of
the general method of addition of a carboxylic acid to a ketene to form the anhydride.
�
1
The ꢀmost interestingꢁ but unassigned sharp peak at 2100 cm is most likely due to the
characteristic ketene absorption, which appears in this region. The absorptions at
�
1
� 1
1
700 cm and the broad peak at 3000 cm will then be those of the carboxylic acid.
The two elimination products react relatively rapidly with one another with a
temperature-dependent rate to regenerate the anhydride, and hence the disappearance
�
1
of the peaks at 1700 and 2100 cm occurs at the same rate.
Scheme 6
Likewise, the lowest field signal at d 11.5 is not due to enol signals. The H enol
signals in the ditipylacetic acid enol [4] or in various enols of amides or esters carrying
EWGs [5] appear< 15 ppm. In contrast, the OH signal of carboxylic acids appears in
the observed region of d 11.5. Moreover, each of the Me, CH , and OH signals of
2
alleged 5 and 6 are given at the same d values, and it is clear that they belong to one
compound rather than two.
Although we believe that this interpretation accounts for all the observations, we
conducted several experiments and calculations in order to completely exclude the
presence of the anhydride enols 5 and 6. When the reaction was repeated and the
13
C-NMR spectrum determined at 298 K in CDCl , the main signals were those of the
3
anhydride, but also ten new signals were observed. The most revealing ones were at
2.10 and 205.5 ppm, which almost coincide with those reported for ethylphenylketene
4

1410
Helvetica Chimica Acta ± Vol. 84 (2001)
at 42.1 (C(a)) and 205.6 (C(b)) [20], which is characteristic for Ph-substituted ketenes
21].
The 13C-NMR spectra of the mixture were also compared with that of 2-
[
phenylbutanoic acid. Authentic 2-phenylbutanoic acid at 240 K in CDCl displays Et
3
signals at d 12.15 (Me), 26.00 (CH ), 53.05 (CH), 127.45 ± 128.58 (o-, m-, p-C), 137.96
2
1
(
Cipso), 181.12 (CO) (with the expected coupling in the H-coupled spectrum). The
signals in the distilled mixture that are not identified with those of the ketene or the
anhydride appear at the same positions as those of the acid, except for the CˆO signal,
which appears at 180.60 ppm, and the 0.5 ppm difference can be ascribed to different
type of H-bonding. Regardless of the explanation, neither the anhydride nor the ketene
has a CˆO signal at ca. 180 ppm.
The ratio ketene 17/acid 18 was 1:1 as expected. The 4/17/18 ratios changed from
1.0 :0.3 :0.3 at 240 K to 1.0 :0.2 :0.2 at 298 K. The higher percentage of the anhydride 4,
and the temperature effect are ascribed to a rapid reaction of 17 and 18 to form 4 before
the NMR experiment was conducted. In the crude distillate, a very broad signal at d 12
is observed, and from its position we ascribe it to the OH group of the acid 18.
Three attempts to capture the pyrolysis products of the reaction in Scheme 6 were
made. Addition of cyclopentadiene to the distillate at 253 K did not give an expected
cycloadduct of cyclopentadiene and ketene 17. Apparently, the reaction was too slow
compared with other reactions in the system. Following the work of Tidwell and co-
workers [22] on reactions of ketenes with radical reagents, we attempted the reaction
with TEMPO (2,2,6,6-tetramethylpiperidine N-oxide). The NMR of the crude product
shows an aromatic AB quadruplet, suggesting that a reaction occurred at the para-
position, with elimination of the Et group rather than reaction on the CˆC bond. This
reaction, which cannot be an indicator for the nature of the ketene, was, therefore,
abundant without identifying the product. A reaction with CH N gave methyl 2-
2
2
phenylbutanoate, thus capturing the acid 18, which was formed in the thermal reaction
of 4. No reaction product with the ketene was observed. However, the capture of the
acid, the observation of an acidic OH, and the coincidence of the signals of the distillate
with those of the ketene 17 indicate that reaction according to Scheme 6 takes place.
The experimental results are supported by calculations on all three species 4, 5 and
6
(Table 1). Several conformations are at local minima for each species both at HF/3-
21G and B3LYP/6-31G** level of theories. The anhydride has two stereogenic centers,
and the (S,S)-conformer (Fig. 1,a) was found to be the most stable one, but only by
.1 kcal/mol over the (RS) conformer (Fig. 2,a) at B3LYP. Another conformer
0
1
(
Fig. 2,b) is 1.4 kcal/mol less stable ).
The mono-enol displays four local minima, all with energies higher than for the
neutral anhydride by 13.6, 15.1, 15.6, and 18.1 kcal/mol. The most stable one has (Z)-
value is 10.0, which
conformation at the enol CˆC bond (Fig. 1,b). The derived pK
Enol
is in the range roughly estimated from the values for the parent anhydride super-
imposed on the stabilization by a b-Ph group.
The bis-enol 6 has four different stable conformations, and, as could have been
guessed, it has a much higher energy than the monoenol. The conformers are 39.3, 39.6,
1)
Fig. 1 contains all the most stable conformers of the species calculated. Structures of higher-energy isomers
are shown in Fig. 2.

Helvetica Chimica Acta ± Vol. 84 (2001)
1411
Table 1. Calculated Electronic Energy, Gibbs Energy (in Hartree), and pKEnol Values in Keto-Enol Isomer-
a
ization of Acid Anhydrides at B3LYP/6-31G** )
Compound
Electronic energy
DE [kcal/mol]
Gibbs energy
DG [kcal/mol]
pKEnol
4
5
6
7
8
9
0 )
1 )
2 )
4a
5a
4b
5b
2
3
4
5
6
7
8
9
0
1
2
5a
6a
5b
6b
7
8
3
4
� 1001.113189
� 1001.090718
� 1001.04839
� 457.935248
� 457.927299
� 457.924989
� 610.576297
� 610.589077
� 610.591355
� 1476.164359
� 1476.128416
� 1005.516951
� 1005.483403
� 646.484100
� 646.507575
� 646.509512
� 380.528779
� 380.490623
� 380.462622
� 908.272549
� 908.287612
� 908.287732
� 685.805644
� 685.820859
� 1512.195161
� 1512.180267
� 1849.217237
� 1849.205205
� 526.851463
� 526.850750
� 798.147897
� 798.141859
0.0
14.1
40.0
0.0
5.0
6.4
� 1000.792497
� 1000.770754
� 1000.729875
� 457.856922
� 457.847271
� 457.845836
� 610.465336
� 610.476195
� 610.477673
� 1476.127610
� 1476.092275
� 1005.444761
� 1005.408971
� 646.396741
� 646.417574
� 646.419902
� 380.480526
� 380.442149
� 380.415101
� 908.188470
� 908.200922
� 908.201137
� 685.690833
� 685.703055
0.0
13.6
39.3
0.0
6.1
7.0
10.0
28.8
4.4
5.1
b
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
4
4
0.0
0.0
b
� 8.0
� 9.5
0.0
22.6
0.0
21.1
0.0
� 14.7
� 16.0
0.0
23.9
41.5
0.0
� 9.5
� 9.5
0.0
� 6.8
� 7.7
0.0
22.2
0.0
22.5
0.0
� 13.1
� 14.5
0.0
24.1
41.1
0.0
� 7.8
� 8.0
0.0
� 5.0
� 5.7
b
16.3
0.0
16.5
� 9.6
� 10.7
17.7
30.1
� 5.7
� 5.8
� 9.6
� 7.7
� 5.6
7.7
c
0.0
0.00 )
c
9.4
0.0
7.6
0.0
0.5
0.0
3.8
10.5 )
0.0 )
8.5 )
0.0
2.2
0.0
5.0
c
c
6.2
� 526.820157
� 526.816714
� 798.035909
� 798.028025
1.6
3.6
a) DH and DG are relative energies (in kcal/mol) with respect to the most stable anhydride form. ) R ˆ Me,
b
1
2
c
R ˆ H. ) Relative Gibbs energy with respect to the most stable anhydride form at B3LYP/6-31G** calculated
from the relative electronic energy at B3LYP/6-31G** and the relative thermochemical contribution obtained at
HF/6-31G** frequency analysis.
4
2.5, and 43.2 kcal/mol above the stable anhydride conformer, and their CˆC� O� H
conformations are (E(syn),Z) (E(syn),E), (Z,Z), and (E(anti),Z), respectively. Here,
(
E/Z) refers to the configuration at the enolic CˆC bonds, and syn/anti represents the
direction of OH with respect to the Ph group. The most stable (E(syn),Z)-conformer is
shown in Fig. 1,c, and others in Fig. 2. Solvent effects will not change this value much,
and, in H O, when the anhydride can take a more stable conformation with a single
2
H O molecule bridging the two CˆO groups, the pK
value will be higher. We
2
Enol
conclude, on both experimental and theoretical grounds, that neither 5 nor 6 is formed
in observable concentrations in solution, and that the report of their observation [9] is
erroneous.

1412
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 1. Structures of the most stable conformers for each species optimized at B3LYP/6-31G**
The 5-Methyl-2H-pyran-2,6(3H)-dione ((Z)-g-Methylglutaconic Anhydride) Sys-
tem. There are some difficulties with the interpretation of the data for 7 ± 9. The role of
the acid according to Scheme 4 is of a keto ! enol catalyst, and it is unclear why the
presence of HCl or TFA at 1008 for a long time does not lead to any enolization,
especially since it is implied that the enols 8/9 are thermodynamically more stable than
7
. A possible protonation on the CˆC bond is not mentioned at all, in spite of the
known protonation of enones [23].
On the other hand, compounds 8 and 9 are conjugated and, judging from data on
simple enols, conjugation increases the stability of the enols by a few kcal/mol [24].

Helvetica Chimica Acta ± Vol. 84 (2001)
1413
Fig. 1 (cont.)
Moreover, the electron-withdrawing carbonyl group at the end of the conjugated
system should make it more enolic [5], with a consequent decrease of pKEnol. Since the
probability that 8/9 will be observable is larger than for other systems, calculations to
check this possibility were conducted. Several possible conformations for 7 ± 9 were
examined. In the most stable conformers (Fig. 1,d ± f), the Me H-atom that lies on the
molecular plane points away from the anhydride moiety; the other conformation in
which the H-atom is directed to the CˆO is the transition state for Me rotation. For 8
and 9, the enolic OH can be placed either toward or away from the other CˆO group,
and the former conformation is more stable by 2.9 (for 8) and 3.5 (for 9) kcal/mol. The

1414
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 1 (cont.)
most stable conformers of 8 and 9 are less stable than 7 by 6.1 and 7.0 kcal/mol,
respectively (Table 1), indicating that the anhydride form is significantly more stable
than its enols.
The 3-Acyl-2H-pyran-2,4(3H)-dione System. In contrast with the 5- (or 3-)
unsubstituted pyran-2,4-dione system, the substituted 3-acyl system is capable of
forming the anhydride enol for two reasons: a) The acyl group is b to the potentially
enolizable dione of the anhydride while the other b-substituent is an enone moiety
attached via its vinylic C-atom. Consequently, the system is a b,b-EWG-substituted
anhydride system, that can enolize according to our generalization [2]. b) The enolic

Helvetica Chimica Acta ± Vol. 84 (2001)
1415
Fig. 1 (cont.)
OH group derived from the anhydride can form a stable six-membered ring via H-bond
to the acyl CˆO group. However, there are features that operate against or compete
with such enolization. i) The enolizable H-atom at C(5) can enolize not only on the
anhydride C(6)ˆO but also on the acyl CˆO substituting C(5), and, a priori,
enolization on the latter is favored. ii) Stabilization of the latter enol is sometimes
possible by H-bonding to another site, i.e., to a C(4)ˆO, and, in this case, the anhydride
CˆO will have a relatively lower probability of being involved in the enolization.

1416
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 1 (cont.)
Inspection of all the anhydride structures in the Cambridge Structural Database
(CSDB) shows no solid-state structure with an enol of anhydride moiety, including
compounds with a H-atom at C(3) or C(5). There are six structures with a formal acyl
or aroyl group on the C-atom neighboring the anhydride CˆO, and all have an enol
structure, but significantly in all of them the acyl or aroyl group is enolized (cf.

Helvetica Chimica Acta ± Vol. 84 (2001)
1417
Fig. 1 (cont.)
structures 19 [25], 20 [26] and 21 [27]). Moreover, in structure 19, which was previously
regarded as a mixture of isomers in the solid (based on IR spectra) [13c] and in solution
(
based on NMR spectra) [13c] the two (former Ac) enol groups are H-bonded to the
C(4)ˆO according to recent X-ray study [25]. In the other five structures there are
O� H ´´´ OˆC (anhydride) H-bonds, but definitely the H-atom is bonded to the former
Ac group: the O ´´´ O, acetyl O� H, and anhydride CˆO ´´´ H distances in Š are 2.496,
0
2
.965, and 1.607 in 20a, 2.496, 0.955, and 1.588 in 20b, 2.514, 0.963, and 1.623 in 20c, and
.537, 0.850, and 1.780 in 21. Consequently, there is no evidence for an enol of anhydride
structure in the solid in these systems, either.
The situation is more complex in solution. The various authors who dealt with the
structure, write it frequently [14] as the enol of anhydride, with or without a H-bond, or
sometimes with an enolized acyl, usually in a H-bonded form. The NMR data seem to
favor the anhydride enol data, or at least an important contribution from it as discussed
by Tan and co-workers [13], and Hansen et al. [14d]. Note, however, that the latter
[
14d] had calculated the relative contributions of the enol of anhydride vs. enol of CˆO
structure and found that the former is 3.3 kcal/mol less stable, in contrast to
experiment, but in line with our assumption.
The energy difference between 11 and 12, R ˆ Me, R ˆ H was, therefore,
calculated by B3LYP/6-31G**. The most stable conformers for 10, 11, and 12 are shown
in Fig. 1,g ± i. The enol 12 on the acyl group was found to be only 0.9 kcal/mol more
stable than the enol in the anhydride 11, i.e., in the gas phase at 298 K, 11 will consist
1
2

1418
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 2. Structures and relative energies of unstable isomers of all species calculated at B3LYP/6-31G**. Numbers
in parentheses are relative Gibbs energies with respect to the most stable conformers.
22% of the 11/12 equilibrium mixture. It is, therefore, clear that a minor structural
variation can make the enol of anhydride more stable.
The 3-acetyl tetrahydropyran-2,4,6-trione system (22, Fig. 1,n) was also examined
by B3LYP calculations, which revealed that two enol forms of former Ac, one H-
bonded to anhydride (23, Fig. 1,o) and the other H-bonded to the CˆO group (24,
Fig. 1,p) were at a local minimum. Both are more stable than the keto form 22 by 14.5
and 13.1 kcal/mol, respectively. In contrast, the attempted optimization of the enol of
anhydride led to 23. Thus, introduction of additional CˆO increases the stability of the

Helvetica Chimica Acta ± Vol. 84 (2001)
1419
Fig. 2 (cont.)
enols on the acyl group, resulting in an overall reduced possibility for the formation of
the enol of anhydride.
The Mixed Acetic Sulfuric Anhydride System. In view of the very high pKEnol values
for the b,b-unsubstituted systems [2], the assignment of structures 15 seems
questionable. It is true that pKEnol seems to decrease with the electron withdrawal of
�
the substituent at C(a), that the Ac and OSO substituents are electron-withdrawing,
3

1420
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 2 (cont.)
and that H-bonding stabilizes the enols. Hence, these features will increase the enol
stability, but they do not seem to be sufficient to stabilize the enol to such an extent that
it will be observable.
We again used calculations of the pair 14/15 to probe this problem. The results are
listed in Table 1. B3LYP/6-31G** Calculations reveal five stable conformations for
diacetyl sulfate 14b with relative energies of 0.0, 2.2, 3.2, 5.3, and 5.8 kcal/mol. Four
stable conformations were calculated for the enol 15b with energies of 22.5, 24.9, 25.5,

Helvetica Chimica Acta ± Vol. 84 (2001)
1421
Fig. 2 (cont.)
and 27.8 kcal/mol above that of the most stable enol. The most stable conformations of
4b and 15b are shown in Fig. 1,k and m, and other conformations are given in Fig. 2.
1
When the DG value is converted, it gives pKEnol ˆ 16.5, which is the lowest value
calculated so far for a b,b-unsubstituted system, but is still very far from making the
enol observable. From past experience on solvent effect on equilibria of enols [2a][28],
we do not believe that this value, as well as other values calculated below, will be

1422
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 2 (cont.)
changed by more than 1 ± 2 units in a polar solution compared with the gas-phase
calculated value. The same conclusions apply for 14a/15a (Fig. 1,j and l) for which a
pKEnol value of 16.3 was calculated. Consequently, we are in doubt that these enols are
the species observed.

Helvetica Chimica Acta ± Vol. 84 (2001)
1423
Fig. 2 (cont.)
Predictions for Observable Enols of Anhydrides. The results presented above
indicate a high probability that none (but one) of the few systems reported to display
observable enols, indeed form them in observable concentrations. Hence, the question
arises of what type of systems will be able to generate observable enols. Consequently,
pKEnol values for three systems were calculated: i) for a system analogous to the cyclic
7
± 9 system ii) for a b,b-diaryl-substituted anhydride, and iii) for anhydrides carrying
two b,b-strongly EWGs.
i) Succinic Anhydride (Tetrahydrofuran-2,5-dione)/5-Hydroxyfuran-2(3H)-one/
Furan-2,5-diol. This system is analogous to 7 ± 9, except that the bis-enol 27 should

1424
Helvetica Chimica Acta ± Vol. 84 (2001)
gain aromatic stabilization of the furan ring. B3LYP Calculations showed that the
energy differences between succinic anhydride (25; Fig. 1,q) and the most stable
conformations of 5-hydroxyfuran-2(3H)-one (26; Fig. 1,r), and furan-2,5-diol (27;
Fig. 1,s) are 24.1 and 41.1 kcal/mol, which amounts to pKEnol values of ca. 18 and 30,
respectively. Hence, the situation resembles that for systems 4 ± 6 and certainly the
dienol and highly probably the monoenol will not be observable.
3
-Acetyl-2H-pyran-2,6(3H)-dione. B3LYP/6-31G** Calculations show that 3-ace-
1
2
tyl-4-methyl-2H-pyran-2,6(3H)-dione (10; R ˆ Me, R ˆ H) is much less stable than its
1
2
enolic counterparts, the enol on the anhydride (11, R ˆ Me, R ˆ H) and the enol on
1
2
the Ac group (12, R ˆ Me, R ˆ H), by 6.8 and 7.7 kcal/mol, respectively, giving pK
Enol
values of � 5.0 for 11 and � 5.7 for 12. However, the most stable species in this system
is the enol on the Ac group.
3
-(Trifluoroacetyl)-2H-pyran-2,6(3H)-dione. Replacement of the CH CO group of
3
1
0 by a CF CO group makes the enol on the acyl group less stable, and then the enol of
3
anhydride may exist as the most stable tautomer, even though the H-bond to a CF CO
3
group will be weaker than that to a CH CO group. B3LYP Calculations revealed that
3
both the enol of the acyl group (29; Fig. 1,u) and the enol of anhydride (30; Fig. 1,v)
are more stable than the anhydride form (28; Fig. 1,t) by 7.8 and 8.0 kcal/mol,
respectively, as expected. The difference between the two enols is not large, the enol of
anhydride being more stable than the enol of the acyl group by only 0.2 kcal/mol. The
pKEnol values for 29 and 30 are � 5.7 and � 5.8, respectively. Hence, 30 should be one of
the observable species in the isomers mixture.
3
-(Methoxycarbonyl)-2H-pyran-2,6(3H)-dione. Since an ester has the least pro-
pensity to become an enol among carboxylic acid derivatives, 3-(methoxycarbonyl)-
H-pyran-2,4(3H)-dione (31; Fig. 1,w) should have a higher probability than 28 to exist
2
as the enol of anhydride (32; Fig. 1,x). B3LYP Calculations, indeed, indicated that,
among the systems calculated, 32 is the only enol optimized and is 7.7 kcal/mol more
stable than 31, corresponding to pKEnol of � 5.6.
ii) b,b-Ditipylacetic Anhydride and its Enol. b,b-Ditipylacetic acid 33 is 12.7 kcal/
mol more stable than its enol 34 according to B3LYP/6-31G** calculations [29].

Helvetica Chimica Acta ± Vol. 84 (2001)
1425
However, 34 was observed by NMR and can be retained for a long time at 255 K, when
generated from ditipyl ketene and H O under kinetically controlled conditions [4c].
2
Since the pKEnol value of the parent enol of anhydride is 4.4 units lower than that of the
parent acid [2a], it is expected that the enol 36 of the mixed anhydride of b,b-
ditipylacetic acid with another acid, e.g., AcOH, i.e., 35 (R ˆ Me) will have pK ꢁ 9.9
Enol
and may hopefully be observed at least as a kinetically stable species.
Calculation on the model system 35a and 36a, i.e., the formate derivatives, by
B3LYP/6-31G** showed two conformations each for 35a and 36a. For 35a, a
conformation in which two carbonyl O-atoms are oriented in the same direction
(
Fig. 2) is 3.5 kcal/mol less stable than the most stable conformer with the CˆO group
directed in opposite ways (Fig. 1,y). For 36a, a conformation with the CˆO/C� O
bonds in the same direction is the species of lower energy (by 2.1 kcal/mol) since CˆO
´
´´ H� O H-bonding is possible for this conformer (Fig. 1,z). The difference in B3LYP
electronic energy of the two more-stable conformations is 9.4 kcal/mol in favor of 35a.
Addition of thermochemical contributions obtained from the frequency calculations at
HF/3-21G to the B3LYP electronic energy gives a Gibbs energy difference of 10.5 kcal/
mol, corresponding to a pKEnol value of 7.7. This value resembles that for a-cyanoacetic
acid, whose enol had been observed by Kresge and co-workers by flash photolysis
[19b], and the value is lower than any value calculated so far for a b,b-di(bulky)aryl
substituted enol [29]. In this respect, the effect of the b,b-ditipyl substitution is
qualitatively in the expected direction. In view of the observation of 34, it is likely that,
if 36a could be prepared under kinetically controlled conditions, it will be observable.
The situation was found to be even more favorable for 35b/36b (Fig. 1,aa and bb).
The calculated energy difference is 8.5 kcal/mol, giving a pKEnol value of 6.2. This
prediction could not yet be confirmed experimentally since attempts to prepare 35b
have failed so far.
iii) b,b-Dicyano- and b,b-(Dimethoxycarbonyl)acetic Anhydrides and Their Enols.
The B3LYP calculations indicate that b,b-dicyano substituents strongly stabilize the
enol of anhydride. Thus, enol 38 (Fig. 1,dd) is only 2.2 kcal/mol less stable than the
anhydride isomer 37 (Fig. 1,cc), giving pKEnol of 1.6. The pKEnol for 37 is 16.0 units

1426
Helvetica Chimica Acta ± Vol. 84 (2001)
smaller than for the parent CH COOCHO system. Nevertheless, since the amide 39
3
was found by calculations to be nearly of the same energy as 40 [5b], it is surprising that
the b,b-(CN) substitution does not make 38 the exclusive species in view of the
2
significantly lower pKEnol value of the parent acetic anhydride compared with the
acetamide.
The pKEnol values for the series 41/42 were calculated by B3LYP/6-31G**
(
(
Scheme 7). The derived pKEnol values for X are: H (� 5.2) < NH (� 0.3) < OCHO
2
1.6) < Br (2.7) < Cl (3.0) < F (3.4) < OMe (4.4). The difference between the most-
and the least-stabilized enols is 9.6 units. Consequently, the enol of the aldehyde (X ˆ
H) is practically the only species that will be observed in the mixture of 41 and 42,
whereas the corresponding methyl ester will be nearly exclusively in the ester form. An
interesting result is the 4.6 pKEnol units difference between XˆNH and XˆOMe,
2
which indicates that enolization on an amide CˆO is significantly preferred over that of
an ester CˆO. This is consistent with the observation that, in amido esters
(
[
EWG)CH(CO R')CONHR, the enolization is exclusively on the amide group
5a,b]. The much smaller DpKEnol of 1.0 for the parent compounds 1/2 let us to expect
2
that, in a competitive intramolecular enolization, the ester may give a small percentage
of observable enol.
Scheme 7
Fig. 3 is a plot of pKEnol(41) vs. pKEnol(2) with a slope of 0.81. The plot is linear for
six substituents including MeO, which shows a small deviation, but NH shows a
2
significant deviation (of ca. 4 pKEnol units) from the line. Clearly the effect of the b,b-
(
CN) substitution does not reduce pK
to the same extent for different X groups, but
2
Enol
there is a rather significant effect of X on pKEnol
.
To understand this behavior, appropriate isodesmic reactions 1 ± 4, based on either
CH or H CˆCH as the reference system, were used to calculate the separate effect of
4
2
2
the b-substituents on the enol and the anhydride (or amide) species, and the results are
listed in Table 2.

Helvetica Chimica Acta ± Vol. 84 (2001)
1427
Fig. 3. A plot of pKEnol for 41 vs. pKEnol for 2
Table 2. Gibbs Energies (in kcal/mol) for the Isodesmic Reactions 1 ± 4
Reactions:
1
2
3
4
(EWG)
(EWG)
(EWG)
(EWG)
2
2
2
2
CˆC(OH)X ‡ CH
CHC(ˆO)X ‡ CH
4
4
> H
2
CˆC(OH)X ‡ CH
(EWG)
CˆC(OH)X ‡ (EWG)
> MeC(ˆO)X ‡ (EWG) CˆCH
2 2
(EWG)
> MeC(ˆO)X ‡ CH
2
2
CˆC(OH)X ‡ H
CHC(ˆO)X ‡ H
2
CˆCH
2
> H
2
2
CˆCH
2
2
CˆCH
2
2
2
Reaction
EWG
X
DG
1
2
3
4
a
b
c
a
b
c
a
b
c
a
b
c
CN
CN
AcO
CN
CN
AcO
CN
CN
AcO
CN
CN
NH
OCHO
OCHO
NH
OCHO
OCHO
NH
2
OCHO
OCHO
NH
OCHO
OCHO
2
26.3
11.7
13.9
� 2.6
� 9.9
� 5.0
16.3
1.7
7.3
� 12.9
� 20.0
� 11.6
2
2
AcO
Since the quantitative differences between the CH and H CˆCH references are
4
2
2
the same for each EWG/X pair, we will use the values for the latter (Table 2,
Reactions 3 and 4) in our discussion.
The effect of the b,b-(EWG) substitution on the enols and the parent acid
2
derivatives, is, as shown previously [2], composed of significant but opposite effects on
both species. For EWG ˆ CN, the enol of the amide (X ˆ NH ) is stabilized by
2
16.3 kcal/mol (Table 2, Reaction 3a), but the anhydride (X ˆ OCHO) enol is stabilized

1428
Helvetica Chimica Acta ± Vol. 84 (2001)
by only 1.7 kcal/mol (Reaction 3b). For Reaction 3c (EWG ˆ MeOOC, X ˆ OCHO),
the stabilization is higher, 7.3 kcal/mol. However, the acid derivative tautomer is
destabilized by 12.9, 20.0, and 11.6 kcal/mol for Reactions 4a (EWG ˆ CN, X ˆ NH ),
2
4
b (EWG ˆ CN, X ˆ OCHO), and 4c (EWG ˆ MeOCO, X ˆ OCHO), respectively.
The large enol stabilization, coupled with acid derivative destabilization for the
dicyanoamide makes the pKEnol value lower for 39/40 (pKEnol ˆ � 0.3) compared with
the 1/2 system (pKEnol ˆ 21.3).
For EWG ˆ CN, X ˆ OCHO, the dissection of the effects shows that they behave in
an opposite manner. The enol stabilization is minor (1.7 kcal/mol), which is insufficient
to make the enol the dominating species at equilibrium.
Bis(methoxycarbonyl) substitution (43/44; Fig. 1,ee and ff) on the anhydride is
more balanced: 7.3 kcal/mol enol stabilization, coupled with 11.6 kcal/mol anhydride
isomer destabilization. The overall of 18.9 kcal/mol is the lowest of the three values and
the b,b-(MeO C) substituted anhydride is the thermodynamically less stabilized enol
2
2
of an anhydride (pKEnol ˆ 3.6). Consequently, the expectation of enol of anhydride
stabilization by two b-EWGs based on our general considerations seems (computa-
tionally) fulfilled for the anhydrides, but in contrast with the amides, so far such
stabilization is insufficient to make the enols observable.
It is interesting that similar doubts concerning the ability to observe enols of
carboxylic acid halides by introducing b,b-(EWG) substituents could be raised on the
2
basis of the calculations of Scheme 7 for the 41/42 pairs with X ˆ Cl, Br, I.
3
. Summary. ± Analysis of the literature claims for the observation of enols of
carboxylic acid anhydrides, mostly by calculations but also by experiment, suggests that
such enols are mostly unlikely to be observed, except perhaps in the 3-acyl-2H-pyran-
2
2
,6(3H)-dione system. B3LYP/6-31G** Calculations suggest that 3-(trifluoroacetyl)-
H-pyran-2,6(3H)-dione will exist partially and the 3-(methoxycarbonyl) analog will
exist completely in the enol of anhydride form. b,b-Dicyano or b,b-dimethoxycarbonyl
substitution of the parent system showed that the species stable by calculation is still the
anhydride rather than its enol.
We are indebted to Prof. M. Kaftory for his assistance with the CSDB. and to the Israel Science Foundation
for support. Numerical calculations were carried out at the Research Center for Computational Science,
Okazaki, Japan.
Experimental Part
General. M.p., and IR and NMR spectra were recorded as described in [4a].
The Reaction of 2-Phenylbutanoic Anhydride (4). 2-Phenylbutanoic anhydride was prepared according to
the literature [30] by converting commercial 2-phenylbutanoic acid (18) to both its sodium salt and the
corresponding acyl chloride and allowing the two reagents to react. The yellow anhydride was distilled at
reduced pressure according to the literature; it formed a yellow mixture whose color faded rapidly. The IR
spectra of the distillate in the absence of a solvent (3082w, 3058w ± m, 3027m, 2966s, 2934s, 2873s, 2092s, 1815s,
1741s, 1700m), which were taken at intervals of a few min after the distillation, showed a decrease in intensity of

Helvetica Chimica Acta ± Vol. 84 (2001)
1429
�
1
the signals at 2093 cm (ketene) and 2500 ± 2900 and 1700 (acid), e.g., from transmittance of 0.3 to 0.65 after
1
13
1
2 min. The H- and C-NMR spectra of pure 4 and acid 18 were recorded and compared with those of the
distillate.
Acid 18: H-NMR (CDCl
1
3
, 240 K): 0.91 (2 overlapping t, Me); 1.78, 2.10 (2 oct., CH
2
); 3.46 (2 overlapping t,
1
CH); 7.25 ± 7.36 (m, 5 arom. H); 12.68 (br., 0.75 H, OH). H-NMR (CCl
4
, 298 K): 0.11 (t, Me); 0.98, 1.29 (sept.,
, 240 K): 12.16 (Me); 26.02
CH
2
); 2.57 (t, CH); 6.36 ± 6.45 (s, 5 arom. H); 11.54 (s, OH). 13C-NMR (CDCl
3
(
CH
2
); 53.05 (CH); 127.45, 127.94, 128.58, 137.96 (arom. C); 181.13 (CˆO).
1
Anhydride 4: H-NMR (CDCl
3
, 240 K): 0.82 (t, J ˆ 5.7, 2 Me), 1.73, 2.03 (m, 2 CH
2
); 3.42 (2 overlapping t,
13
2
5
CH); 7.12 (s, 4 arom. H); 7.30 (s, 6 arom. H). C-NMR (CDCl
3.77 (CH); 127.45, 127.51, 127.88, 127.92, 128.62, 128.68, 136.66, 136.72 (arom. C); 168.96, 169.10 (CˆO). We
3
): 11.74, 11.79 (Me); 25.35, 25.50 (CH ); 53.66,
2
note that most signals are doubled, and this can be due to different conformations of the two anhydride
moieties.
The identification of the ketene signals is based on the NMR spectra of the reaction mixture after
1
subtracting the signals for the anhydride and the acid. Ketene 17: H-NMR (CDCl
3
, 240 K): 1.24 (t, J ˆ 6.9, Me);
2
1
.44 (q, J ˆ 7.0, CH
23.52, 123.88, 128.82, 132.58 (arom. C); 205.50 (CˆO).
2 3 2
, 240 K): 12.14 (Me); 16.54 (CH
); 7.05 (d, 2 arom. H). 13C-NMR (CDCl
); 42.10 (Cˆ);
Attempts to Capture the Ketene 17. a) Reaction of the Distillate with Cyclopentadiene. Anhydride 4 (5 g,
6 mmol) was distilled in vacuo, and to the fresh distillate in Et O (20 ml) at 253 K was added freshly distilled
1
2
cyclopentadiene (1.0 g, 15.2 mmol) at 253 K. The mixture was stirred for 2 h and left overnight at r.t. After
1
evaporation of the solvent at reduced pressure, the H-NMR spectrum of the residue showed that no reaction
had taken place.
b) Reaction with TEMPO. A soln. containing 4 (0.5 g, 16 mmol) and TEMPO (0.6 g, 3.9 mmol) in p-xylene
(
10 ml) was refluxed for 48 h. TLC showed no signals of 4 and only one spot corresponding to product. This was
1
separated by chromatography (silica gel; AcOEt/petroleum ether 1:9). The solvent was removed, and H-NMR
spectrum of the residue showed an aromatic AB quadruplet and no Et group. Since the ketene structure had not
remained intact, the reaction was not investigated further.
2 2
c) Reaction with CH N . Anhydride 4 (3 g, 9.7 mmol) was distilled in vacuo into a dry-ice cooled beaker.
CH
2
N
2
Soln. in Et
2
O (ca. 12 mg/ml, 22 ml) was added to this soln. at 273 K, and the mixture was kept at 273 K for
1
1
day. The solvent was evaporated, and the H-NMR spectrum of the crude product showed the presence of at
least three products: 4, methyl 2-phenylbutanoate and a very broad signal at ca. 12 ppm, indicating the presence
1
of 18. Chromatography (silica gel; Et
2
O/petroleum ether 1:9) gave the ester (0.55 g, 32%). H-NMR (CDCl
3
,
2
98 K): 0.89 (t, J ˆ 7.4, Me); 1.79, 2.10 (2m, J ˆ 7.5, CH
2
); 3.45 (t, J ˆ 7.7, CH); 3.65 (s, MeO); 7.22 ± 7.35
, 298 K): 12.11 (qq, J ˆ 126, 3.6, Me); 26.70 (t quint., J ˆ 129, J ˆ 4.2, CH );
1.82 (q, J ˆ 147, MeO); 53.33 (br. d, J ˆ 126, CH); 127.14 (dt, J ˆ 160, J ˆ 4.5, arom. C); 127.90 (dt, J ˆ 158,
ˆ 5.2, arom. C); 128.52 (dd, J ˆ 160, 5.9, C of Ph); 139.07 (s, Cipso); 174.48 (CˆO).
Calculations. The compounds studied in the present paper were optimized by the Gaussian 98 program [31]
1
3
(
5 arom. H). C-NMR (CDCl
3
t
q
2
5
d
t
d
J
t
p
both at the ab initio HF/3-21G and the hybrid-density-functional B3LYP/6-31G** method [32]. The latter
method was extensively used in recent calculations of enols of carboxylic acid derivatives and was shown to give
pKEnol values reliable to within 2 units [2a][28][33]. All optimized structures, including higher-energy
conformers, were verified by means of their Hessian matrices to be local minima on the potential-energy
surfaces at both levels of theories, except for the ditipyl derivatives, for which frequency calculations at the
higher level are too demanding. Structures determined at the B3LYP level showed the same geometric
characteristics as in the HF level. The calculated energies of the compounds are given in Table 1. Structures of
the most stable conformer for each species are illustrated in Fig. 1, and structures of higher energy isomers are in
Fig. 2.
REFERENCES
[
[
[
1] a) A. F. Hegarty, P. OꢁNeill, in ꢀThe Chemistry of Enolsꢁ, Ed. Z. Rappoport, Wiley, Chichester, 1990, Chapt.
1
0, p. 639; b) A. J. Kresge, Chem. Soc. Rev. 1996, 25, 275.
2] a) S. Sklenak, Y. Apeloig, Z. Rappoport, J. Am. Chem. Soc. 1998, 120, 10359; b) P. Perez, A. Toro-Labb e , J.
Phys. Chem. A 2000, 104, 1557.
3] P. J. OꢁNeill, A. F. Hegarty, J. Chem. Soc., Chem. Commun. 1987, 744; B. M. Allen, A. F. Hegarty, P.
OꢁNeill, M. T. Nguyen, J. Chem. Soc., Perkin Trans. 2 1992, 927; B. M. Allen, A. F. Hegarty, P. OꢁNeill, J.
Chem. Soc., Perkin Trans. 2 1997, 2733.

1430
Helvetica Chimica Acta ± Vol. 84 (2001)
[
[
[
4] a) J. Frey, Z. Rappoport, J. Am. Chem. Soc. 1995, 117, 116; b) J. Frey, Z. Rappoport, J. Am. Chem. Soc. 1996,
1
18, 3994; c) J. Frey, Z. Rappoport, J. Am. Chem. Soc. 1996, 118, 5169, 5182; d) Z. Rappoport, J. Frey, M.
Sigalov, E. Rochlin, Pure Appl. Chem. 1997, 69, 1933.
5] a) J. K. Mukhopadhyaya, S. Sklenak, Z. Rappoport, J. Am. Chem. Soc. 2000, 122, 1325; b) J. K.
Mukhopadhyaya, S. Sklenak, Z. Rappoport, J. Org. Chem. 2000, 65, 6856; c) Y. X. Lei, G. Cerioni, Z.
Rappoport, J. Org. Chem. 2000, 65, 4028.
6] H. E. Zimmerman, T. W. Cutshall, J. Am. Chem. Soc. 1958, 80, 2893; H. E. Zimmerman, T. W. Cutshall, J.
Am. Chem. Soc. 1959, 81, 4305; L. Graham, D. L. H. Williams, J. Chem. Soc., Chem. Commun. 1991, 407;
D. L. H. Williams, L. Xia, J. Chem. Soc., Perkin Trans. 2 1992, 747; D. L. H. Williams, L. Xia, J. Chem. Soc.,
Perkin Trans. 2 1993, 1429.
[7] F. B. Thole, J. F. Thorpe, J. Chem. Soc. 1911, 99, 2208; N. Bland, J. F. Thorpe, J. Chem. Soc. 1912, 101,
8
56.
[
[
8] W. C. Deno, W. E. Billups, R. E. DiStefano, K. M. McDonald, S. Schneider, J. Org. Chem. 1970, 35, 278.
9] J. E. Hendon, A. W. Gordon, M. Gordon, J. Org. Chem. 1972, 37, 3184.
[
[
[
10] J. Kagan, L. Tolentino, M. G. Ettlinger, J. Org. Chem. 1970, 45, 3085.
11] J. S. Fitzgerald, G. A. R. Kon, J. Chem. Soc. 1937, 725.
12] M. E. Jung, R. W. Brown, Tetrahedron Lett. 1981, 22, 3355; S. F. Tan, Aust. J. Chem. 1972, 25, 1367; P. F.
Vlad, M. Z. Krimer, Org. Prep. Proced. Inter. 1998, 30, 657.
[
13] a) S. F. Tan, K.-P. Ang, H. Jayachandran, J. Chem. Soc., Perkin Trans. 2 1986, 973; b) S. F. Tan, K. P. Ang, H.
Jayachandran, J. Chem. Soc., Perkin Trans. 2 1983, 473; c) S. F. Tan, K. P. Ang, H. L. Jayachandran, A. J.
Jones, W. R. Begg, J. Chem. Soc., Perkin Trans. 2 1982, 513; d) A. K. Kiang, S. F. Tan, W. S. Wong, J. Chem.
Soc. C 1971, 2721.
[
14] a) L. Crombie, D. E. Games, M. H. Knight, J. Chem. Soc. C 1967, 763; b) L. Crombie, R. V. Dove, J. Chem.
Soc., Chem. Commun. 1987, 438; c) L. Crombie, R. V. Dove, J. Chem. Soc., Perkin Trans. 1 1996, 1695;
d) P. E. Hansen, S. Bolvig, T. Kappe, J. Chem. Soc., Perkin Trans. 2 1995, 1901.
15] P. Hunt, D. N. P. Satchell, J. Chem. Soc. 1964, 5437.
[
[
[
16] W. E. Truce, C. E. Olsen, J. Am. Chem. Soc. 1953, 75, 1651.
17] a) E. Montoneri, E. Tempesti, L. Giuffre, A. Castoldi, J. Chem. Soc., Perkin Trans. 2 1980, 662; b) L.
Guiffre, E. Montoneri, E. Tempesti, V. Pozzi, F. Rossi, Chim. Indus., Milano 1981, 63, 85.
18] Y. Ogata, K. Adachi, J. Org. Chem. 1982, 47, 1182.
19] a) Y. Chiang, A. J. Kresge, P. Pruszynski, N. P. Schepp, J. Wirz, Angew. Chem., Int. Ed. 1990, 29, 792; Y.
Chiang, A. J. Kresge, V. V. Popik, N. P. Schepp, J. Am. Chem. Soc. 1997, 119, 10203; b) J. Andraos, Y.
Chiang, A. J. Kresge, I. G. Pojarlieff, N. P. Schepp, J. Wirz, J. Am. Chem. Soc. 1994, 116, 73.
20] J. Firl, W. Runge, Z. Naturforsch. 1974, 29B, 393; J. Firl, W. Runge, Angew. Chem., Int. Ed. 1973, 12,
[
[
[
6
18.
[21] G. Cerioni, A. Plumitallo, J. Frey, Z. Rappoport, Magn. Reson. Chem. 1995, 33, 669.
[22] A. D. Allen, B. Cheng, M. H. Fenwick, W.-W. Huang, S. Misshia, D. Tahmassebi, T. T. Tidwell, Org. Lett.
1
999, 1, 693.
[23] K. Müllen, P. Wolf, in ꢀThe Chemistry of Enonesꢁ, Eds. S. Patai, Z. Rappoport, Wiley, Chichester, 1989,
Chap. 12, pp. 513 ± 558.
[24] B. Capon, in ꢀThe Chemistry of Enonesꢁ, Eds. S. Patai, Z. Rappoport, Wiley, Chichester, 1989,
p. 1063.
[
[
25] V. Bertolasi, P. Gilli, V. Ferretti, G. Gilli, J. Chem. Soc., Perkin Trans. 2 1997, 945.
26] a) R. Kakou-Yao, A. Saba, N. Ebby, M. Pierrot, J. P. Aycard, Acta Crystallogr., Sect. C 1999, 55, 1591; b) R.
Kakou-Yao, E. Degny, N. Ebby, J. P. Aycard, Z. Kristallogr. ± New Cryst. Struct. 1999, 214, 483; c) R.
Kakou-Yao, A. Saba, N. Ebby, J. P. Aycard, Z. Kristallogr. ± New Cryst. Struct. 1999, 214, 481.
27] K. A. Parker, G. A. Breault, Tetrahedron Lett. 1986, 27, 3835.
[
[
28] J. Gao, J. Mol. Struct. 1996, 370, 203; Y. Apeloig, D. Arad, Z. Rappoport, J. Am. Chem. Soc. 1990, 12, 9131;
S. Sklenak, Y. Apeloig, Z. Rappoport, J. Chem. Soc., Perkin Trans. 2 2000, 2269.
[
[
[
29] H. Yamataka, Z. Rappoport, J. Am. Chem. Soc. 2000, 122, 9818.
30] H. Falk, K. Schlögl, Monatsh. Chem. 1965, 96, 276; A. Horeau, Tetrahedron Lett. 1962, 965.
31] Gaussian 98, Revision A.6, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M.
Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B.
Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K.
Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B.

Helvetica Chimica Acta ± Vol. 84 (2001)
1431
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T. Keith,
M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W.
Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle, J. A. Pople, Gaussian, Inc., Pittsburgh
PA, 1998.
[
[
32] a) A. D. Becke, Phys. Rev. 1988, A38, 3098; b) C. Lee, W. Yang, R. G. Parr, Phys. Rev. 1988, B37, 785.
33] P. G. Conrad, R. S. Givens, B. Hellrung, C. S. Rajesh, M. Ramseier, J. Wirz, J. Am. Chem. Soc. 2000, 122,
9
346.
Received March 26, 2001