Keto/enol tautomerism in phenylpyruvic acids: Structure of the o- nitrophenylpyruvic acid

Article abstract of DOI:10.1016/S0022-2860(99)00323-3

The synthesis of a tautomeric keto/enol mixture of o-nitrophenylpyruvic acid followed the acid hydrolysis of the azlactone of o-nitrobenzaldehyde was carried out. The structures of the two tautomeric forms were assigned by NMR spectroscopy. X-ray diffraction of a single crystal revealed that the crystalline form corresponds to the keto tautomer. Quantum mechanics calculations in the gas phase confirmed the experimental findings in solution. (C) 2000 Elsevier Science B.V.

Full text of DOI:10.1016/S0022-2860(99)00323-3

Journal of Molecular Structure 520 (2000) 191–198
www.elsevier.nl/locate/molstruc
Keto/enol tautomerism in phenylpyruvic acids: structure
of the o-nitrophenylpyruvic acid
a,
b
a,c
d
*
A.J.M. Carpy , P.P. Haasbroek , J. Ouhabi , D.W. Oliver
a
Laboratoire de Physico- & Toxico-Chimie des Syst e` mes Naturels (LPTC), UPRES-A 5472, CNRS, Universit e´ de Bordeaux I, 351, Cours de la
Lib e´ ration, 33405 Talence Cedex, France
b
Department of Pharmacy, Medical School, University of the Witwatersrand, 7 York Road, Parktown 2193, South Africa
c
D e´ partement de Chimie, FST-Gu e´ liz, Universit e´ Cadi Ayyad, Avenue Abdelkrim Khattabi, B.P. 618, Marrakech, Morocco
d
Department of Pharmacology, Potchefstroom University for C.H.E., Potchefstroom 2520, South Africa
Received 7 May 1999; received in revised form 2 July 1999; accepted 12 July 1999
Abstract
The synthesis of a tautomeric keto/enol mixture of o-nitrophenylpyruvic acid followed the acid hydrolysis of the azlactone of
o-nitrobenzaldehyde was carried out. The structures of the two tautomeric forms were assigned by NMR spectroscopy. X-ray
diffraction of a single crystal revealed that the crystalline form corresponds to the keto tautomer. Quantum mechanics
calculations in the gas phase confirmed the experimental findings in solution. ᭧ 2000 Elsevier Science B.V. All rights reserved.
Keywords: Phenylpyruvic acids; Azlactones; Tautomerism; NMR; X-ray; Quantum mechanics
1
. Introduction
tautomers could only be observed in hydrolysis
products of azlactones with moderate (chlorine) or
strong electron-withdrawing groups (nitro) in the
ortho-position of the phenyl ring. Their para-substi-
tuted isomers formed only minute quantities of the
keto tautomers, indicating that the inductive effect
plays a major role in the equilibrium formation of
the enol/keto tautomeric mixture when the electron-
withdrawing group is in the ortho-position. This
investigation also found that the formation of these
enol tautomers and their derivatives proceed with
retention of configuration of the Z-geometry of the
parent azlactones [2–5].
Understanding tautomeric equilibria is of central
importance in synthetic chemistry and especially in
medicinal chemistry. During an investigation of the
hydrolysis products (phenylpyruvic acids) of different
ring-substituted aromatic azlactones, tautomerism
was found to be an important aspect of phenylpyruvic
acid formation [1] (Scheme 1). This study has shown
that phenylpyruvic acids obtained from basic and
acidic hydrolysis of these azlactones exist over-
whelming in the enol form. The type and position of
the substituent group in the phenyl ring of the azlac-
tone have a definite influence on the ratio of enol/keto
tautomer formation. Significant amounts of the keto
Tautomerism in phenylpyruvic acids obtained from
azlactones was first reported in 1943 [6]. The spectral
data reported on the keto and enol forms of such acids,
in some cases, seem rather confusing and inconsistent
*
Corresponding author. Tel.: ϩ 33-556-848944; fax: ϩ 33-
1
regarding tautomerism. A H NMR spectrum of p-
5
0
56-848948.
E-mail address: a.carpy@lptc.-u-bordeaux.fr (A.J.M. Carpy).
hydroxyphenylpyruvic acid in DMSO ϩ CDCl3
022-2860/00/$ - see front matter ᭧ 2000 Elsevier Science B.V. All rights reserved.
PII: S0022-2860(99)00323-3

1
92
A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
and the X-ray study performed on the keto tautomer.
In addition, computational methods were used to gain
information about this phenomenon and the predict-
ability of keto/enol tautomerism as was demonstrated
experimentally by the current study.
CH
C
COOH
CH
C
O
COOH
2
R
R
OH
enol isomer
keto isomer
Scheme 1.
shows only the enol form but the incorrect keto form
was assigned to it [7]. The structure of the enol form
of p-hydroxyphenylpyruvic acid was confirmed by X-
ray crystallography [8]. Cyanation-hydrolysis of
2
. Experimental
2
.1. Synthesis
PhCH COCl was reported to give 23% of the keto
The synthesis of the tautomeric mixture 3a (enol)
2
form of phenylpyruvic acid, whereas the synthesis
from the azlactone gave the enol form [9]. More
recently the formation and H NMR (in DMSO)
and 3b (keto) involved the formation of the azlactone
of o-nitrobenzaldehyde 2 (4-(o-nitrobenzylidene)-2-
methyl-5-oxazolone) from o-nitrobenzaldehyde 1,
followed by acid (HCl/AcOH) hydrolysis. Their struc-
tures were investigated by NMR spectroscopy and X-
ray crystallography. However, X-ray data could only
be obtained on the keto tautomer. Difficulties in the
preparation and hydrolysis of 2-nitrophenyl azlactone
derivatives were reported in the literature [11,12].
1
data of the enol tautomers of some ring substituted
(mono-, di- and tri-substituted, mainly hydroxy and
methoxy) phenylpyruvic acids synthesized from the
azlactones were also reported [10]. Displacement of
the keto/enol equilibrium towards the keto form can
be accomplished by treating the potassium salt of the
enol with H SO (98%) in methanol [9], or in acetate
2
4
buffer (pH 6) where the rate of tautomerism to the
keto form is slow [8]. A H NMR spectrum of o-
1
2.1.1. Synthesis of azlactone of o-nitrobenzaldehyde 2
A mixture of o-nitrobenzaldehyde 1 (7.55 g;
nitrophenylpyruvic acid in DMSO ϩ CDCl shows
3
0
.05 mol), N-acetylglycine (5.85 g; 0.05 mol), anhy-
the presence of only the keto form [7], whereas in
1
13
drous sodium acetate (8.2 g; 0.1 mol) and acetic anhy-
dride (25 ml) was heated at 80–85ЊC with stirring.
After 2 h, the dark brown mixture was poured into a
mixture (60 ml) of equal volumes of water and
ethanol, stirred for 30 min and cooled in ice. The
crystals that formed were filtered off, washed with
the water/ethanol mixture and dried at 45ЊC (4.24 g;
our study the H NMR and C NMR spectra (in
DMSO) of the same product obtained from the acid
hydrolysis of the azlactone, clearly show the presence
of the tautomeric mixture of the keto and enol forms.
This paper reports on the synthesis and NMR data
of the tautomeric phenylpyruvic acid mixture
obtained from the azlactone of o-nitrobenzaldehyde
0
.018 mol). Treatment with activated charcoal and
crystallization from a mixture of benzene (4 parts)
and n-hexane (1 part) gave yellow crystals of 2 with
m.p. 115–116ЊC.
Table 1
H NMR data of 3a and 3b
1
Hydrogen atom
H-3 (Ar)
3a (ppm)
3b (ppm)
8.27 (dd,
8.15 (dd,
2.1.2. Synthesis of the tautomeric mixture of
o-nitrophenylpyruvic acid 3a and 3b
J ˆ 7:9 Hz,
J ˆ 7:9 Hz,
0
0
a
J
ˆ 1:1 Hz)
J
ˆ 1:1 Hz)
m
m
The azlactone 2 (1.5 g; 6.46 mmol) was boiled
under reflux with stirring in glacial acetic acid
(4.5 ml) and concentrated hydrochloric acid
(10.5 ml). After 6 h, the mixture was poured into
water (15 ml), stirred well and cooled in a refrigerator
overnight. The crystals that formed were filtered off,
washed with water and dried at 45ЊC (0.76 g;
3.63 mmol). Recrystallization from benzene gave
H-4 (Ar)
H-5 (Ar)
H-6 (Ar)
7.39–7.80 (m)
7.39–7.80 (m)
7.94 (dd,
7.39–7.80 (m)
7.39–7.80 (m)
7.39–7.80 (m)
J ˆ 8:1 Hz,
0
J
ˆ 1:1 Hz)
m
Ar-CHyC
6.70 (s)
–
–
Ar–CH
2
4.65 (s)
a
dd: double doublet; m: multiplet; s: singlet.

A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
193
Table 2
C NMR data of 3a and 3b
Table 4
1
3
4
Atomic coordinates ( × 10 ) and isotropic equivalent displacement
2
2
2
2
˚
parameter Beq (A ) defined as: 4=3‰a b ϩ b b ϩ c b
ϩ
1
;1
2;2
3;3
Carbon atom
3a (ppm)
3b (ppm)
ab cos gb1;2 ϩ ac cos bb ϩ bc cos ab2;3Š for 3b
1;3
a
CO–COOH
COOH
–
192.2 (s)
–
x
y
z
Beq
165.8 (s)
–
CO–COOH
C-1 (Ar)
C-2 (Ar)
C-3 (Ar)
C-4 (Ar)
C-5 (Ar)
C-6 (Ar)
Ar–CHyC
Ar–CHyC
162.0 (s)
128.5 (s)
148.4 (s)
124.2 (d)
131.0 (d)
134.1 (d)
127.6 (d)
–
C(1)
C(2)
C(3)
C(4)
C(5)
C(6)
N(7)
O(8)
O(9)
C(10)
C(11)
O(12)
C(13)
O(14)
O(15)
6594(3)
7317(3)
7398(4)
6807(4)
6105(4)
5987(3)
6433(3)
7025(3)
5703(3)
8008(3)
8981(3)
9399(3)
9480(3)
9009(3)
5544(8)
3579(8)
3347(9)
4960(10)
6891(9)
7161(9)
5978(7)
4817(8)
7529(7)
1789(8)
3123(8)
5148(6)
1762(8)
1425(3)
1987(3)
2940(3)
3311(3)
2734(3)
1784(3)
404(2)
2.21(8)
2.27(8)
3.0(1)
3.3(1)
3.2(1)
2.80(9)
2.92(8)
4.90(8)
4.79(9)
2.83(9)
2.57(9)
3.58(7)
2.70(9)
3.83(7)
4.02(7)
129.7 (s)
148.2 (s)
124.9 (d)
132.6 (d)
134.1 (d)
128.7 (d)
144.8 (s)
102.2 (d)
–
68(2)
–
Ϫ 81(2)
1647(3)
1526(3)
1914(2)
871(3)
Ar–CH
2
43.7 (t)
a
s: singlet; d: doublet; t: triplet.
Ϫ 166(6)
413(2)
854(2)
10390(3)
2833(6)
cream-colored crystals of 3a and 3b with m.p. 120–
1
22ЊC.
Table 3
Crystal data and structure refinement for 3b
2
.2. Spectroscopic analysis
Chemical formula
Formula weight
Crystal size (mm)
Crystal system
Space group
Unit-cell dimensions
C
9
H
7
NO
5
NMR analysis of the tautomeric mixture, 3a and
209.16
1
0:32 × 0:10 × 0:07
3b, was performed at 200 MHz ( H) and 50 MHz
1
3
Monoclinic
(
C) on a Bruker AC-200 spectrometer with
1
P2 /c
DMSO-d as solvent and TMS as internal standard
6
ꢀ
a ˆ 12:765ꢀ5† A,
(
Tables 1 and 2). The melting points (uncorrected)
were determined on an electrothermal melting point
apparatus.
ꢀ
b ˆ 5:149ꢀ12† A,
ꢀ
c ˆ 14:951ꢀ3† A, b ˆ
1
13:63ꢀ1†Њ
3
˚
Volume (A )
900.4(1)
4
Z
2.3. X-ray analysis
3
Density calculated (Mg/m )
1.54
10.6
Absorption coefficient m
A pale yellow crystal (dimensions 0:32 × 0:10 ×
Ϫ1
(
cm )
0
:07 mm) was mounted on an Enraf-Nonius CAD-4
F(000)
432
diffractometer using graphite-monochromated Cu Ka
radiation. The crystal data are listed in Table 3.
Lattice parameters were determined by least-squares
refinement of 25 reflections with 20Њ Ͻ u Ͻ 35Њ.
Intensities were collected with v Ϫ 2u scan mode,
up to u ˆ 65Њ. The intensities of two standard reflec-
tions, measured every 90 min, showed no significant
deviation. The data were corrected for Lorentz and
Temperature (K)
Wavelength (A)
u range for data collection
Index ranges
296(2)
1.54178
3–65Њ
0 Ͻ h Ͻ 15; 0 Ͻ k Ͻ 5;
Ϫ17 Ͻ l Ͻ 16
1786
˚
Unique reflections collected
Observed reflections
Refinement method
Data/restraints/parameters
Final R indices [I Ն 3sꢀI†]
R indices (all data)
1270
Full-matrix least-squares
1270/0/164
2
w
R ˆ 0.061, R ˆ 0:068
polarization effects but not for absorption
R ˆ 0:077
Ϫ1
(
m ˆ 10:6 cm ). A total of 1786 unique reflections
Largest diff. peak and hole
0.38(8) and 0.00(8)
Ϫ3
were collected, of which 1270 with I Ն 3sꢀI† were
˚
(
e A )
considered as observed. The structure was solved by

1
94
A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
Table 5
Bond lengths (A) for 3b
˚
ءء
ءء
RX
AMPAC-AM1
RHF/6-31G
RHF/6-31G
B3LYP/6-31G
C(1)–C(2)
C(1)–C(6)
C(1)–N(7)
C(2)–C(3)
C(2)–C(10)
C(3)–C(4)
C(4)–C(5)
C(5)–C(6)
N(7)–O(8)
N(7)–O(9)
C(10)–C(11)
C(11)–O(12)
C(11)–C(13)
C(13)–O(14)
C(13)–O(15)
1.399(5)
1.385(7)
1.473(5)
1.391(6)
1.498(7)
1.381(7)
1.382(6)
1.375(6)
1.219(6)
1.221(5)
1.493(7)
1.208(5)
1.536(7)
1.219(5)
1.295(6)
1.409
1.407
1.490
1.402
1.488
1.392
1.394
1.391
1.202
1.201
1.501
1.225
1.509
1.233
1.354
1.398
1.388
1.450
1.391
1.512
1.385
1.386
1.381
1.231
1.223
1.510
1.207
1.511
1.210
1.332
1.396
1.385
1.460
1.389
1.512
1.384
1.384
1.380
1.198
1.191
1.517
1.183
1.532
1.187
1.312
1.408
1.397
1.473
1.400
1.508
1.393
1.395
1.389
1.236
1.229
1.526
1.208
1.543
1.212
1.339
direct methods using MolEN [13] and refined aniso-
tropically. It was determined that the crystal corre-
sponded to the 3b keto form. The final residuals
2.4. Computational analysis
Semi-empirical calculations, geometry optimizations
and frequency calculations, were performed on 3b and
3a with the AMPAC 6.0 package [14] and the AM1
Hamiltonian. The corresponding ab initio calculations
were carried out by using the gaussian94 program
package [15] at Hartree–Fock (RHF) (6-31G and
were R ˆ 0:061 (R ˆ 0:068). The final fractional
w
coordinates for 3b and equivalent displacement para-
meters are listed in Table 4. The bond lengths and
angles are given in Tables 5 and 6. The selected
torsion angles are shown in Table 7.
Table 6
Bond angles (Њ) for 3b
ءء
ءء
RX
AMPAC-AM1
RHF/6-31G
RHF/6-31G
B3LYP/6-31G
C(2)–C(1)–C(6)
C(2)–C(1)–N(7)
C(6)–C(1)–N(7)
C(1)–C(2)–C(3)
C(1)–C(2)–C(10)
C(3)–C(2)–C(10)
C(2)–C(3)–C(4)
C(3)–C(4)–C(5)
C(4)–C(5)–C(6)
C(1)–C(6)–C(5)
C(1)–N(7)–O(8)
C(1)–N(7)–O(9)
O(8)–N(7)–O(9)
C(2)–C(10)–C(11)
C(10)–C(11)–O(12)
C(10)–C(11)–C(13)
O(12)–C(11)–C(13)
C(11)–C(13)–O(14)
C(11)–C(13)–O(15)
O(14)–C(13)–O(15)
122.6(4)
121.1(4)
116.3(3)
116.0(4)
125.2(4)
118.8(3)
122.1(4)
120.2(4)
119.5(5)
119.6(4)
120.0(3)
118.0(4)
121.9(4)
112.9(4)
124.6(4)
116.5(4)
118.9(4)
119.5(4)
115.0(4)
125.4(5)
120.7
122.0
117.3
118.1
123.7
118.2
121.2
120.1
120.0
119.8
119.9
118.3
121.7
113.7
126.3
112.4
121.3
127.2
114.6
118.2
122.3
121.4
116.3
116.6
124.7
118.8
121.8
120.3
119.3
119.7
119.0
118.4
122.6
112.5
123.6
115.6
120.7
122.1
113.9
123.9
122.5
121.3
116.1
116.4
124.8
118.8
121.9
120.3
119.3
119.6
118.4
117.7
123.9
112.3
124.3
115.1
120.5
122.1
113.1
124.9
122.3
121.6
116.0
116.3
124.6
119.1
122.1
120.1
119.4
119.7
118.7
117.8
123.5
112.7
124.4
114.3
121.2
123.0
112.2
124.7

A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
195
Table 7
Selected torsion angles (Њ) for 3b
ءء
ءء
RX
Ϫ 7.1(6)
Ϫ 108.2(4)
Ϫ 160.0(3)
Ϫ 173.7(3)
AMPAC-AM1
RHF/6-31G
RHF/6-31G
B3LYP/6-31G
O(9)–N(7)–C(1)–C(6)
18.5
Ϫ 108.7
161.2
Ϫ 16.7
Ϫ 106.2
Ϫ 151.2
167.8
Ϫ 19.6
Ϫ 103.9
Ϫ 154.2
165.7
Ϫ 14.3
Ϫ 107.2
Ϫ 154.7
173.5
C(3)–C(2)–C(10)–C(11)
C(2)–C(10)–C(11)–C(13)
C(10)–C(11)–C(13)–O(15)
122.9
ءء
1
6
(
-31G basis sets) and density functional (DF)
1. Investigation of the H NMR spectrum of the
hydrolysis product of the azlactone of o-nitrobenzal-
dehyde revealed the presence of a mixture of the enol
3a and keto 3b tautomers of o-nitrophenylpyruvic
acid. The signal at 6.70 ppm accounts for the single
benzylic proton of the enol tautomer, and the signal at
4.65 ppm for the two benzylic protons of the keto
tautomer. Peak intensity calculations show that this
tautomeric mixture consists of approximately 59%
of the enol form and 41% of the keto form.
ءء
B3LYP/6-31G [16,17]) levels. The geometries were
optimized using Berny’s algorithm [18]. Following the
geometry optimizations, analytical frequency calcula-
tions were carried out at the RHF/6-31G level only,
using standard procedures [19], to obtain thermoche-
mical properties. The tautomeric keto , enol equili-
brium in the phenylpyruvic acid was presented in
Scheme 1.
The ratio of the two tautomers is described by the
equilibrium constant K , which is classically calcu-
lated from their free energy difference DG
1
3
The C NMR data of 3a and 3b are given in Table
2. Two sets of chemical shift values could clearly be
distinguished for the two tautomeric forms. The signal
at 144.8 ppm (a-carbon, Ar-CHyC), and the signal at
102.2 ppm (benzylic carbon, Ar-CHyC) are evidence
for the presence of the enol tautomer 3a. These values
are in accordance with values that have already been
observed for other enol tautomer derivatives
[1,2,8,10]. The low field signal at 192.2 ppm was
assigned to the carbon atom of the a-carbonyl group
–CO–COOH) and the high field signal at 43.7 ppm
triplet) to the benzylic carbon (Ar–CH ) of the keto
tautomer 3b. These chemical shift values were found
to be characteristic of the keto form and represent
therefore a unique and effective way to identify the
presence of the keto form, even in trace amounts [1].
T
K ˆ expꢀϪDG=RT†
DG is expressed as a function of enthalpy difference
DH and entropy difference DS
ꢀ1†
T
DG ˆ DH Ϫ TDS
The enthalpy difference DH at 298 K is given by
ꢀ2†
298
0
v
298
v
298
298
DH ˆ DE ϩ DE ϩ DꢀDE † ϩ DEr ϩ DEt
T
(
(
2
ϩ DꢀPV†
ꢀ3†
0
v
where E is the total energy, E the zero-point energy
ZPE), DEv the change in vibrational energy when T
T
298
(
298
298
increases from 0 to 298.15 K, Er and Et are the
rotational and translational energies, and PV is the
work term. In the tautomeric equilibria, the last
three terms of Eq. (3) are all zero. For comparison,
the same calculations were performed on the phenyl-
pyruvic acid itself and on the o-chloro substituted
compound.
3
.2. Structure assignment of 3b by X-ray diffraction
The keto structure of 3b was established by X-ray
diffraction on a single crystal. An ORTEP view of two
molecules paired by H-bonds involving the carboxylic
acid groups, and showing the atomic numbering is
depicted in Fig. 1. Bond lengths between C(11) and
˚
C(10) (1.493(7) A) and between C(11) and O(12)
1.208(5) A) clearly indicate that the selected crystal
3
. Results and discussion
˚
(
3
.1. Spectroscopic analysis
existed with the keto form. The orientation of the
trans-extended side chain (including the acid moiety)
towards the phenyl ring is defined by the value of the
1
The H NMR data of 3a and 3b are given in Table

1
96
A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
Fig. 1. ORTEP view of a pair of 3b molecules linked by H-bonds.
ءء
31G , the other three methods tend to overestimate
bond lengths. It is particularly true for B3LYP/6-
torsion angle Cꢀ3†–Cꢀ2†–Cꢀ10†–Cꢀ11† ˆ Ϫ108:2ꢀ4†Њ
Table 7). The dihedral angle between the phenyl
ring and the plane containing the keto group [C–
C(yO)–C] is equal to 80.3(1)Њ. The o-nitro group is
nearly coplanar with the aromatic ring with a torsion
angle Oꢀ9†–Nꢀ7†–Cꢀ1†–Cꢀ6† ˆ Ϫ7:1ꢀ6†Њ.
(
ءء
31G which overestimate 12 of the 15 bond lengths.
However, the average difference between bond
lengths calculated with this method and experimental
˚
ones is only 0.013 A, intermediate between RHF/6-
ءء
˚
1G results (0.011 A) and RHF/6-31G results
0.015 A). The same data, calculated with AMPAC,
3
(
˚
3.3. Computational analysis
˚
is 0.018 A. For all methods, the highest individual
differences between experimental and calculated
bond lengths are found in the terminal carboxylic
group (H-bonds are not considered in the calculations)
(Table 5). The calculated bond angles are statistically
over- and underestimated. The average difference
between experimental and calculated bond angles
are 1.7Њ (AMPAC-AM1), 0.6Њ (RHF/6-31G and
The geometries of the two tautomeric forms were
optimized using AMPAC-AM1 and gaussian94 at
ءء
Hartree–Fock (basis sets 6-31G and 6-31G ) and
ءء
density functional (B3LYP/-631G ) levels. Only the
geometry of 3b, calculated with these methods, is
reported in Tables 5–7 for comparison with experi-
mental results. Although it is trivial to notice that ab
initio calculations reproduced the geometry better
than the semi-empirical method, the advantages of
AMPAC (diffusion of the program, computational
time, etc.) have to be considered. Except RHF/6-
ءء
ءء
6-31G ) and 0.9Њ (B3LYP/6-31G ). At the
Hartree–Fock level, this difference is very close to
the average s.d. of experimental bond angles (0.4Њ)
(Table 6). Finally, the trans extended conformation
Table 8
Total energies and total energy differences (entropies and zero-point energy (ZPE) plus thermal corrections)
Ϫ1
Ϫ1
Ϫ1
E (a.u.)
DE (kcal mol
)
Entropy (cal mol
K
)
ZPE ϩ thermal correct
Ϫ1)
(kcal mol
Keto form
Enol form
Keto Ϫ enol
Keto form
Enol form
Keto form
Enol form
Phenylpyruvic acid
o-Nitro derivative
o-Chloro derivative
569.82230
Ϫ 773.17193
Ϫ 1028.69901
Ϫ 569.83287
Ϫ 773.17034
Ϫ 1028.70682
6.63
Ϫ 1.00
4.89
105.795
113.075
108.809
98.075
100.962
105.547
110.038
113.470
104.545
110.558
112.630
105.039

A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
197
Table 9
Tautomer equilibrium constant K
T
of the reaction keto form $ enol form
Ϫ1
Ϫ1
DH (kcal mol
)
DG (kcal mol
)
K
T
Phenylpyruvic acid
o-Nitro derivative
o-Chloro derivative
Ϫ 6.11
0.16
Ϫ 4.40
Ϫ 3.81
3.77
Ϫ 3.43
624.8
1.73 × 10
326.8
Ϫ3
of the side chain is only reproduced by the ab initio
methods. AMPAC tilted the nitro group from the
phenyl ring by an angle almost equal in value, but
opposite in sign, to the angle found by the ab initio
methods; the torsion angle defining the position of the
OH group of the carboxy moiety is closest to 90Њ
(3b/3a) of o-nitrophenylpyruvic acid resulted from
the acid hydrolysis of the azlactone of o-nitrobenzal-
dehyde. The structures of the two forms were identi-
fied by NMR spectroscopy. Their proportions in the
1
mixture were based on the peak intensities of the H
NMR spectra of 3a and 3b. Approximately 59% of the
enol form and 41% of the keto form are present in the
mixture in DMSO at room temperature. Increasing
the temperature of this mixture increased the percen-
tage of the keto form. Subsequent NMR studies have
shown that the keto/enol ratios are solvent and
temperature dependent [1,20].
Following the crystallization process, only the keto
form 3b crystallized. Its structure was unambiguously
determined by X-ray crystallography.
(122.9Њ) than to 180Њ (trans extended as found in the
crystals and by ab initio calculations).
According to these results, the analytical frequency
and thermodynamic calculations were only performed
with the RHF/6-31G method. For comparison, the
same calculations were also performed on the phenyl-
pyruvic acid itself, on the o-nitrophenylpyruvic acid
and on the o-chlorophenylpyruvic acid. The total
energies and the total energy differences between
the two tautomeric forms ‰DEꢀketo Ϫ enol†Š are listed
in Table 8. The enol forms of phenylpyruvic acid and
of the o-chloro derivative are more stable than the
keto forms as previously observed [1]. The energy
differences for phenylpyruvic acid and the o-chloro
The phenylpyruvic acids synthesized from the
corresponding azlactones, favored their existence in
the enol form, in the unsubstituted as well as in
different substituted phenylpyruvic acids and only
electron-withdrawing groups in the ortho position of
the phenyl ring causes displacement of the equili-
brium towards keto formation [1–5]. The tautomeric
equilibrium was investigated in the gas phase by
quantum mechanics calculations. Three potential
keto/enol mixtures were investigated: the phenyl-
pyruvic acid itself, the o-nitrophenylpyruvic acid
and the o-chlorophenylpyruvic acid. Ab initio calcu-
lations showed that the unsubstituted acid exists
predominantly in the enol form confirming our experi-
mental findings. The strong electron-withdrawing
nitro group in the ortho position has a significant
influence on the formation of the keto/enol tautomeric
mixture, by shifting the equilibrium towards keto
formation. As expected, the chlorine atom in the
ortho position being a moderate electron-withdrawing
group had a lesser influence than the nitro group.
Consequently the equilibrium is still in favor of the
enol form as found experimentally (ϳ88% of the enol
tautomer and 12% of the keto tautomer are present in
DMSO at room temperature) [1]. In conclusion, this
derivative are, respectively, around 6.63 and
Ϫ1
4
.89 kcal mol . However, the energy difference
between the 3b (keto form) and 3a (enol form)
Ϫ1
favored the more stable keto form by 1 kcal mol .
Values of entropies and ZPE plus thermal correc-
tions at T ˆ 298 K are listed in Table 8. From Eqs.
(1)–(3) and the values extracted from Tables 8 and 9,
the tautomeric equilibrium constants could be calcu-
lated. The calculations show a large preference for the
enol tautomers relative to the keto tautomers in the
case of the non-substituted- and the o-chloro-substi-
tuted molecules (K Ͼ 1). Conversely, the tautomeric
T
constant K of the equilibrium 3b $ 3a is equal to
T
Ϫ3
1
:73 × 10
showing a large preference for 3b
Ϫ1
ꢀ
K
ˆ 579:3†.
T
4. Conclusion
The synthesis of a tautomeric keto/enol mixture

1
98
A.J.M. Carpy et al. / Journal of Molecular Structure 520 (2000) 191–198
study has shown that substitution type and substitution
position clearly play an important role in the keto/enol
tautomerism of phenylpyruvic acid derivatives.
[11] J.M. Gulland, K.I. Ross, N.B. Smellie, J. Chem. Soc. (1931)
885.
2
[12] H. Burton, J. Chem. Soc. (1935) 1265.
[13] MolEN, An Interactive Structure Solution Procedure, Enraf-
Nonius, Delft, The Netherlands, 1990.
[14] AMPAC 6.0, Semichem, 7128 Summit, Shawnee, KS 66216,
USA.
References
[
15] M.J. Frisch, G.W. Trucks, H.B. Schlegel, P.M.W. Gill, B.G.
Johnson, M.A. Robb, J.R. Cheeseman, T. Keith, G.A.
Peterson, J.A. Montgomery, K. Raghavachari, M.A. Al-
Laham, V.G. Zakrewski, J.V. Ortiz, J.B. Foresman, C.Y.
Peng, P.Y. Ayala, W. Chen, M.W. Wong, J.L. Andres, E.S.
Replogle, R. Gomperts, R.L. Martin, D.J. Fox, J.S. Binkley,
D.J. Defrees, J. Baker, J.P. Stewart, M. Head-Gordon, C.
Gonzalez, J.A. Pople, gaussian94, revision B.3, Gaussian,
Inc., Pittsburgh, PA, 1995.
[
[
[
[
[
1] P.P. Haasbroek, Ph.D. Thesis, Potchefstroom University for
C.H.E., Potchefstroom, South Africa, 1997.
2] D.W. Oliver, P.P. Haasbroek, J.M. Leger, A.J.M. Carpy, J.
Chem. Crystallogr. 24 (1994) 665.
3] P.P. Haasbroek, D.W. Oliver, A.J.M. Carpy, J. Chem. Crystal-
logr. 28 (1998) 139.
4] P.P. Haasbroek, D.W. Oliver, A.J.M. Carpy, J. Chem. Crystal-
logr. 28 (1998) 193.
5] P.P. Haasbroek, D.W. Oliver, A.J.M. Carpy, J. Chem. Crystal-
logr. 28 (1998) 811.
[
16] R.G. Parr, W. Yang, Density-Functional Theory of Atoms and
Molecules, Oxford University Press, Oxford, 1989.
17] A.D. Becke, J. Chem. Phys. 98 (1993) 5648.
18] H.B. Schlegel, J. Comput. Chem. 3 (1982) 214.
19] W.J. Hehre, L. Radom, P.v.R. Schleyer, J.A. Pople, Ab Initio
Molecular Orbital Theory, Wiley, New York, 1986, chap. 6.
20] Details to be published elsewhere.
[
[
6] L.C. Raiford, C.H. Buurman, J. Org. Chem. 8 (1943) 466.
7] C.J. Pouchert, J.R. Campbell, The Aldrich Library of NMR
Spectra, VI, Aldrich Chemical Company, 1974, p. 127.
8] M.C. Pirrung, J. Chen, E.G. Rowley, A.T. McPhail, J. Am.
Chem. Soc. 115 (1993) 7103.
[
[
[
[
[
[
9] I.J. P e´ rez, S.M. Rosell, Afinidad 36 (1979) 332.
[10] H.N.C. Wong, Z.L. Xu, H.M. Chang, C.M. Lee, Synthesis 8
(1992) 793.