Ab initio analysis on metal ion catalysis in the enolization reactions of some acetylheterocycles: Kinetics of the enolization reactions of 3-acetyl-5-methylisoxazole, 5-acetyl-3-methylisoxazole and 3(5)-acetylpyrazole

Article abstract of DOI:10.1002/poc.474

Kinetic data on the enolization reaction of 3-acetyl-5-methylisoxazole, 5-acetyl-3-methylisoxazole, 3(5)-acetylpyrazole and some previously studied acetylheterocycles have been the object of a comprehensive ab initio analysis. Enolization rate constants were measured spectrophotometrically by the halogen trapping technique at 25 °C and ionic strength of 0.3 mol dm^-3 in water, in acetate buffers, in dilute hydrochloric acid, in dilute sodium hydroxide and in the presence of some metal ion salts. In the spontaneous (water) and base (acetate) catalysed reactions the ketones investigated are generally more reactive than acetophenone, according to the electron-withdrawing effect of the heterocyclic ring compared with the benzene ring. In particular, acetylisoxazoles, 3(5)-acetylpyrazole and acetylthiazoles are more reactive than acetylfurans, 2-acetylpyrrole and acetylthiophenes respectively, and this can be attributed to the additional effect of the second heteroatom in the heterocyclic moiety. On the other hand, the compounds investigated are generally less reactive than acetophenone in the H₃O⁺ -catalysed reaction. Ab initio calculations on the relative stability of the protonated and unprotonated forms of the substrates investigated have been compared with the kinetic results obtained in acid solutions. As far as the metal ion catalysis is concerned, an approach in terms of ΔΔE can give an estimate of the combined contributions of charge densities and frontier orbitals to the interaction of the substrate with the metal ion catalyst. A comparison of experimental metal activating factor values and ab initio calculations outlines the importance of charge density on the carbonyl oxygen atom of acetylheterocycles with one heteroatom. An analogous comparison for acetylheterocycles with two heteroatoms suggests the formation of a chelate complex or transition state, in which the metal ion coordinates both the carbonyl oxygen and the aza nitrogen, whenever these two atoms are suitably placed within the molecular structure of the acetylheterocycle. Copyright

Full text of DOI:10.1002/poc.474

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2002; 15: 247–257
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.474
Ab initio analysis on metal ion catalysis in the enolization
reactions of some acetylheterocycles: kinetics of the
enolization reactions of 3-acetyl-5-methylisoxazole, 5-
acetyl-3-methylisoxazole and 3ꢀ5)-acetylpyrazole
Antonella Fontana,1* Paolo De Maria,1 Marco Pierini,1 Gabriella Siani,1 Simona Cerritelli2 and
Gabriella Macaluso3
1Dipartimento di Scienze del Farmaco, UniversitaÁ `G. d'Annunzio', Via dei Vestini 31, I Nucleo Didattico, 66013 Chieti, Italy
2Dipartimento di Chimica, Ingegneria Chimica e Materiali, UniversitaÁ degli Studi dell'Aquila, via Vetoio±Coppito due, 67010 L'Aquila, Italy
3Dipartimento di Chimica Organica, UniversitaÁ di Palermo, Viale delle Scienze, Parco D'Orleans, 90128 Palermo, Italy
Received 14 August 2001; revised 3 December 2001; accepted 13 December 2001
ABSTRACT: Kinetic data on the enolization reaction of 3-acetyl-5-methylisoxazole, 5-acetyl-3-methylisoxazole,
epoc
3(5)-acetylpyrazole and some previously studied acetylheterocycles have been the object of a comprehensive ab initio
analysis. Enolization rate constants were measured spectrophotometrically by the halogen trapping technique at 25°C
and ionic strength of 0.3 mol dmÀ3 in water, in acetate buffers, in dilute hydrochloric acid, in dilute sodium hydroxide
and in the presence of some metal ion salts. In the spontaneous (water) and base (acetate) catalysed reactions the
ketones investigated are generally more reactive than acetophenone, according to the electron-withdrawing effect of
the heterocyclic ring compared with the benzene ring. In particular, acetylisoxazoles, 3(5)-acetylpyrazole and
acetylthiazoles are more reactive than acetylfurans, 2-acetylpyrrole and acetylthiophenes respectively, and this can be
attributed to the additional effect of the second heteroatom in the heterocyclic moiety. On the other hand, the
‡
compounds investigated are generally less reactive than acetophenone in the H3O -catalysed reaction. Ab initio
calculations on the relative stability of the protonated and unprotonated forms of the substrates investigated have been
compared with the kinetic results obtained in acid solutions. As far as the metal ion catalysis is concerned, an
approach in terms of DDE can give an estimate of the combined contributions of charge densities and frontier orbitals
to the interaction of the substrate with the metal ion catalyst. A comparison of experimental metal activating factor
values and ab initio calculations outlines the importance of charge density on the carbonyl oxygen atom of
acetylheterocycles with one heteroatom. An analogous comparison for acetylheterocycles with two heteroatoms
suggests the formation of a chelate complex or transition state, in which the metal ion coordinates both the carbonyl
oxygen and the aza nitrogen, whenever these two atoms are suitably placed within the molecular structure of the
acetylheterocycle. Copyright  2002 John Wiley & Sons, Ltd.
Additional material for this paper is available from the epoc website at http://www.wiley.com/epoc
KEYWORDS: metal ion catalysis; ab initio calculations; Enolization reactions; acid–base catalysis
INTRODUCTION
upon ‘hardness’ and ‘softness’ of the carbonyl group and
of the metal ion catalyst. With the use of a simple relative
number, the metal activating factor (MAF) value, i.e. the
ratio between the catalytic constants for metal ion
catalysis kM and that for proton catalysis kH, we have
defined as ‘soft’ the carbonyl groups of 3-acetyl-N-
methylpyrrole2b (3AP) and 2-acetylpyrrole2c (2APH) as
they enolize equally well (or faster) in the presence of the
above-mentioned metal ions and in the presence of the
hydronium ion. On the contrary, the carbonyl groups of
acetophenone, 2- and 3-acetylthiophenes2a (2AT and
3AT), 2-acetylfuran2b (2AF) and 5-acetylthiazole2c
(5ATZ) can be defined ‘hard’, as the hydronium ion is
a much more effective catalyst than any of the metal ions
investigated in the enolization reaction. The formation of
a chelate complex between the metal ion and the
Lewis acid catalysis by metal ions provides alternative
pathways for specific acid-catalysed enolization reactions
of ketones and comparisons of catalysis by metal ions and
hydronium ion is a subject of considerable interest.1 We
have recently investigated2 the effect of the addition of
metal ions (e.g. Cu2‡, Zn2‡, Ni2‡) in the enolization
reaction of acetophenone and a number of aromatic five-
membered acetylheterocycles in aqueous solutions. The
effectiveness of metal ion catalysis depends primarily
*Correspondence to: A. Fontana, Dipartimento di Scienze del
Farmaco, Universita` ‘G. d’Annunzio’, Via dei Vestini 31, I Nucleo
Didattico, 66013 Chieti, Italy.
E-mail: fontana@unich.it
Contract/grant sponsor: MURST (Rome).
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

248
A. FONTANA ET AL.
salt solutions) were followed under pseudo-zero-order
conditions2 and the rate law had the form shown in Eqn.
(1), where S refers to the substrate and [I2]tot refers to the
total concentration of iodine ([I2] ‡ [I3À]):
Àd‰I2Štot=dt ˆ ke‰SŠ
ꢀ1†
Scheme 1
Under the experimental conditions adopted, the rate-
determining step is the formation of the enol or enolate
ion and there is no evidence of reversibility of the
iodination reaction.
The bromination reactions (in dilute sodium hydroxide
solution) were followed under pseudo-first-order condi-
tions in solutions containing a slight excess of bromine as
previously described2a–c,e and the rate law was given by:
acetylheterocycle, enolizing much faster than the un-
complexed substrate, was detected only for 2-acetylpy-
ridine3 and 2-acetylimidazole (2AI),2d probably due to
the relatively high basicities of these substrates. On the
other hand, the relatively high MAF values measured for
the less basic 2- and 4-acetylthiazole (2ATZ and 4ATZ)
were attributed2c to a transition state in which the metal
cation coordinates both the carbonyl oxygen and the aza
nitrogen atoms. 3-Acetyl-5-methylisoxazole (3AIO), 5-
acetyl-3-methylisoxazole (5AIO) and 3(5)-acetylpyra-
zole [3(5)APz] possess (Scheme 1), as 2-acetylpyridine,
2AI, 2ATZ and 4ATZ do, ‘pyridine-like’ nitrogen
atoms. In two of them, i.e. 3AIO and the tautomer
3APz, this aza nitrogen atom is located in a position that
could assist proton transfer by forming with the metal
cation a chelate complex or transition state. Within this
framework we have studied the enolization reaction of
the title ketones in water, in dilute hydrochloric acid, in
dilute sodium hydroxide, in acetate buffer solutions and
in the presence of Zn2‡, Ni2‡ and Cu2‡ solutions, at
25.0 Æ 0.1°C and at a constant ionic strength of
0.3 mol dmÀ3 (KCl or NaClO4).
Ab initio calculations on the same three and some
previously investigated acetylheterocycles have been
carried out: (i) to evaluate the influence of the thermo-
dynamic contribution to the acid-catalysed enolizations;
(ii) to correlate the calculated charge density on the
carbonyl oxygen atoms and the energy of the substrate–
metal ion interaction with the experimental rate constants
for metal ion catalysis; (iii) to test further the validity of
the previously suggested mechanisms2c for metal ion
catalysis via the formation of a chelate complex (or
transition state) in which the metal ion coordinates both
the carbonyl oxygen and the aza nitrogen whenever these
two atoms are suitably placed within the molecular
structure of the acetylheterocycle; (iv) to estimate the
relative stability of the two tautomeric forms of 3(5)APz
of Scheme 1.
Àd‰OBrÀŠ=dt ˆ ke‰SŠ
ꢀ2†
Enolization of 3AIO, 5AIO and 3ꢀ5)APz in dilute
hydrochloric acid. Reaction rates were measured with a
concentration of 3AIO of about 1 Â 10À2 mol dmÀ3
,
5AIO of about 2 Â 10À3 mol dmÀ3 and 3(5)APz of about
5 Â 10À3 mol dmÀ3; [HCl] was in the range 0.0025–
0.25 mol dmÀ3 and [IÀ] = 2, 5 or 10 Â 10À3 mol dmÀ3
.
The observed rate law was given by Eqn. (3) with ke
being of the form shown in Eqn. (1):
‡
ke ˆ kH O ‡ kH‰H Š
ꢀ3†
2
The intercepts, kH O, of Eqn. (3) measure the ‘water’ rate
2
‡
constants, whereas kH refers to the H3O -catalysed rate
constants. Values of ke calculated from Eqn. (3) using
evaluated kH O and kH values agree with experimental
2
‡
values to within 5% in an interval of H3O concentration
at least of a factor of 100.
The results obtained are collected in Table 1, together
with analogous results previously obtained for acetophe-
none and some other acetylheterocycles.
Enolization of 3AIO, 5AIO and 3ꢀ5)APz in acetate
buffers. The observed rate constants ke [Eqn. (1)] were
of the form shown in Eqn. (4), where kAcO represents the
second-order rate constant for the acetate-catalysed
reaction and the intercept ko represents the rate constant
for the water and the (very small) specific catalysis
contributions.
ke ˆ ko ‡ kAcO‰AcOÀŠ
ꢀ4†
RESULTS
Kinetics
Several sets of ke values were measured for the three
ketones at buffer ratios r = [AcOÀ]/[AcOH] = 0.5, 1, 3
and 5 for 5AIO, 1 and 5 for 3AIO and 3 and 5 for
3(5)APz, each set at four or more different concentrations
of acetate, [AcOÀ], in a range of approximately a factor
of ten. The values of ke showed no evidence of general
acid catalysis by the acidic component of the buffer, as
The enolization reactions were followed spectrophoto-
metrically by the well-established halogen-trapping
procedure as described previously.2 The iodination
reactions (in hydrochloric acid, in acetate buffers and in
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

METAL ION CATALYSIS IN THE ENOLIZATION REACTIONS OF ACETYLHETEROCYCLES
249
Table 1. Acid dissociation constants pKBH
‡
and catalytic constants k5dm3 molÀ1 sÀ1) for the enolization of acetophenone and
some acetylheterocycles at 25°C and ionic strength 0.3 mol dmÀ3 5NaCl or KClO4)
Substrate
kH O (10À9)/sÀ1
kH (10À6
)
kAcO (10À7
)
kOH (10À1
)
pKBH
‡
Ref.
2
3AIO
86.7
2.16
435
1.93
À3.79
this work
5AIO
193
1.29
5.97
1510
30.9
37.4
–
À3.58
this work
this work
3(5)APz
16.5
0.09
2c
2c
2ATZ
4ATZ
70.5
28.3
3.19
6.89
328
3.16
0.68
À0.69
À0.45
45.0
2c
5ATZ
2AF
279
2.69
11.8
14.8
83.6
10.7
4.38
1.89
4.08
0.22
0.45
2b
2c
6.72
33.3
3AF
2c
2a
2a
2AS
2AT
3AT
36.0
2.00
2.20
2.99
3.95
11.8
6.41
5.80
6.60
2.00
2.75
2.46
À4.14a
2c
2b
2APH
3AP
27.7
154
3.43
10.6
3.91
250
–
1.59
2a
A
4.1
10.0b
8.40b
2.37c
À3.46d
a
Ref. 4.
Ref. 5.
Ref. 6.
Ref. 7.
b
c
d
Enolization of 3AIO, 5AIO in dilute sodium hydro-
they were found to be independent of r. The correlation
coefficient of the straight line of Eqn. (4) was, for each
substrate, higher than 0.999 and the standard error of kAcO
was within 5%.
A slight linear dependence of ke on the concentration
of [I2]tot was observed in the case of 3AIO and kAcO
values for this ketone were extrapolated to zero [I2]tot
concentration. The reason of this dependence is not
understood.
xide. Ionization rates in dilute aqueous hydroxide [NaOH
concentration in the range (0.1–2.5) Â 10À2 mol dmÀ3
]
were measured as rates of bromination using the rate law
given by Eqn. (2).
The second-order rate constants kOH obtained from
Eqn. (5) are reported in Table 1.
ke ˆ ko0 ‡ kOH‰OHÀŠ
ꢀ5†
The results obtained for acetate catalysis are reported
in Table 1.
Values of ke calculated from Eqn. (5) using experimental
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

250
A. FONTANA ET AL.
k'o and kOH values agree with experimental values to
within 5%. No OHÀ catalysis was detected for 3(5)APz,
as this compound is probably acidic enough to be
appreciably dissociated under the basic conditions
adopted [pyrazole has pKa 14.28a and 3(5)APz should
be even more acidic].
A similar mechanism could be proposed for the metal
ion catalysis; however, no experimental evidence for the
formation of a complex between the substrate and the
metal ion has been found for a series of acetylhetero-
cycles with one ring heteroatom2 (with the notable
exception of 2-acetylpyridine3). The observed metal ion
catalysis can be interpreted according to the perturbation
teory,11 which provides semi-quantitative interpretation
of relative reactivities in terms of the interactions of
frontier orbitals [highest occupied molecular orbital
(HOMO)–lowest unoccupied molecular orbital (LUMO)]
and/or the electrostatic interactions between the species
involved. Ab initio calculations on the acetylheterocycles
investigated have been run and charge densities on their
carbonyl oxygen atoms (and aza nitrogen atoms if
appropriate) as well as the energies of their HOMOs
have been derived. The results obtained are collected in
Table 2.
Metal-ion-catalysed enolization of 3AIO, 5AIO and
3ꢀ5)APz. The effect of the presence of Zn2‡, Ni2‡ and
Cu2‡ on the rates of enolization of the three substrates
was studied in unbuffered solutions (pH 4–6). Substrate
concentrations were ca (2–5) Â 10À3 mol dmÀ3 and the
concentrations of I2 and IÀ were the same as those for the
reactions in the absence of metal ions. Rates were
measured at several metal ion concentrations in the range
(2.5–75)  10À3 mol dmÀ3. In the case of Cu2‡ the
concentration range (0.25–7.5) Â 10À3 mol dmÀ3 was
chosen instead, in order to avoid precipitation of CuI2
(and/or Cu2I2).
Values of ke for 3AIO and 3(5)APz showed a linear
increase with increasing metal ion concentration accord-
ing to:
DISCUSSION
Inspection of the rate constants collected in Table 1
shows that 3(5)APz, 3AIO and 5AIO are more reactive
than acetophenone in the ‘water’ and acetate-catalysed
reactions. This fact can be easily accounted for by the
electron-withdrawing effects of the heterocyclic rings,
compared with the benzene ring, which should make the
acetyl hydrogen atoms more easily removable by a base.
This behaviour (with a few exceptions for the acetate
reaction)2b,c appears to be quite general for acetylhetero-
cycles compared with acetophenone. By the same token,
the presence of the second ring heteroatom explains why
3(5)APz is eight times more reactive than 2APH in the
acetate-catalysed reaction. Similarly 3AIO and 5AIO are
respectively 40 and 140 times more reactive than 2AF,
and 2ATZ and 4ATZ are eight and 56 times more
reactive than 2AT in the same reaction. The prevalence
of the inductive over the resonance effect in the
protonation and the electrophilic substitution reactions
of azoles compared with pyrrole, furan and thiophene has
already been pointed out.8
ke ˆ ko ‡ kM‰M2‡
Š
ꢀ6†
where [M2‡] represents the molar concentration of Zn2‡
,
Ni2‡ or Cu2‡ and kM represents the second-order rate
constant for the metal-ion-catalysed reaction. The
correlation coefficients of the straight lines of Eqn. (6)
were, in all cases, higher than 0.990 and the standard
error of kM was within 5%.
There was no evidence of saturation even at the highest
metal ion concentration used, which is at variance with
the behaviour previously observed with more basic
substrates such as 2-acetylpyridine3 and 2-acetylimida-
zole.2d On the other hand, no metal ion catalysis was
observed for 5AIO.
The results obtained are collected in Table 2.
Ab initio calculations. As the accepted mechanism9 for
the acid-catalysed enolization reaction involves a pre-
equilibrium proton transfer to the carbonyl group, with
subsequent C—H ionization being assisted by the solvent
or the conjugate base of the acid catalyst, the overall rate
constant can be dissected into a kinetic term, kv, and a
The rate constant for the OHÀ-catalysed reaction of
5AIO is about 20 times higher than that of 3AIO, and this
is in agreement with the strong inductive electron-
withdrawing effect of the ring oxygen atom and the fact
that kOH of 2AF is 19 times higher than kOH of 3AF.2e
However, the kOH values of the different acetylhetero-
cycles are generally difficult to rationalize as they are
rather insensitive2 to the structure of the ketone. Indeed,
for most of the substrates in Table 1 the kOH values vary
within a factor of about six (notable exceptions are the
values of 5AIO and 3AF discussed above).
thermodynamic term, 1/KBH
‡
(where KBH is the acid
‡
dissociation constant of the protonated acetylhetero-
cycle), according to:
‡
kH ˆ kv=KBH
ꢀ7†
The relative stability of the unprotonated (B) and
‡
protonated (BH ) forms of some 2-and 3-acetylhetero-
cycles (pyrrole, thiophene and furan derivatives) can be
evaluated by ab initio calculations according to the
aqueous solvation model.10 The results are collected in
Table 3.
As expected, in the light of the accepted mechanism9
for the acid-catalysed enolization reaction, 3(5)APz,
3AIO, 5AIO and most acetylheterocycles are vice versa
less reactive than acetophenone in the hydronium-
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

METAL ION CATALYSIS IN THE ENOLIZATION REACTIONS OF ACETYLHETEROCYCLES
251
Table 2. Metal-ion-catalysed rate constants, MAF values, charge densities on the carbonyl oxygen and the aza nitrogen atoms
and calculated HOMO energies for some acetylheterocycles with one ring heteroatom 5a) and with two ring heteroatoms 5b)
kZn/10À8 dm3
molÀ1 sÀ1
Substrate (MAF values)
kCu/10À8 dm3
molÀ1 sÀ1
(MAF values)
kNi/10À8 dm3
molÀ1 sÀ1
Charge on oxygen
HOMO/hartrees
(kcal molÀ1
(MAF values) (Charge on nitrogen)
)
kCu/kZn
One ring heteroatom
2AF
2AT
3AT
3.56 (0.011)
20.9 (0.059)
19.0 (0.016)
0.00
26.6 (0.079)
0.00
À0.348
À0.359
À0.354
À0.2193 (À137.6)
À0.2226 (À139.7)
À0.2183 (À136.9)
–
23.5 (0.059)
14.1 (0.012)
1.10
0.700
0.00
2AS
3AF
147 (0.49)
281 (0.19)
94.6 (0.32)
792 (0.54)
0.00
À0.362
À0.370
À0.2238 (À140.4)
À0.2101 (À131.8)
0.600
2.80
50.6 (0.03)
2APH
3AP
441 (1.30)
6790 (20:0)
1060 (3.09)
31800 (30)
À0.391
À0.372
À0.2128 (À133.5)
15.4
3200 (3.00)
788000 (740)
À0.2004 (À125.8) 246
Two ring heteroatoms
5AIO
0.00
0.00
0.00
À0.330
À0.2345 (À147.1)
À0.2266 (À142.2)
–
3AIO
3730 (17.3)
1090 (5.05)
261 (1.21) À0.334 (À0.085)
0.300
2ATZ
4ATZ
5510 (17.0)
1530 (2.20)
131000 (410)
299000 (430)
19600 (61.4) À0.348 (À0.224)
6470 (9.39) À0.349 (À0.211)
À0.2318 (À145.5)
23.8
À0.2226 (À139.7) 195
3(5)APz
1360 (2.28)
48800 (81.7)
1590 (2.66) À0.350a (À0.184)a,b À0.2145a (À134.6)a 35.9
À0.364c (À0.168)c,b À0.2349c (À147.4)c
a
Values calculated for the 3APz tautomer.
Charge on N–2.
Values calculated for the 5APz tautomer.
b
c
Table 3. Calculated energies [Density Functional Theory 5DFT) ‡ hydration; hartrees] for the most stable conformation of the
‡
protonated 5BH ) and unprotonated 5B) forms of some acetylheterocycles
‡
BH
À383.26169
À382.83673
0.42496
À461.93115
À.461.50161
0.42954
À706.27506
À705.85518
0.41988
À706.27558
À705.84917
0.42641
À363.41189
À362.98411
0.42778
À402.73025
À402.29405
0.43620
B
DEBÀBH
‡
(kcal molÀ1
)
(266.67)
(269.54)
(263.48)
(267.58)
(268.44)
(273.72)
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

252
A. FONTANA ET AL.
catalysed reaction, again due to the electron-withdrawing
effect of the heterocyclic rings. In particular, in substrates
with only one ring heteroatom, the more basic 3-acetyl
derivatives react faster than their 2-acetyl isomers. This
interpretation of the results for acid catalysis is supported
by ab initio calculations, which show that the energy
difference between the unprotonated and protonated
related to the hardness and softness of their carbonyl
groups.2c,e
Ab initio calculations on 2AF, 3-acetylfuran (3AF),
2AT, 3AT, 2-acetylselenphene (2AS) and 3AP have
highlighted the existence of a linear correlation (n = 6,
r = 0.954) between the rate constant for the Zn2‡
-
catalysed reaction and the relative charge density on
the carbonyl oxygen atom (Fig. 1). The rate constants for
the Zn2‡-catalysed reaction have been conveniently
forms of 3-acetyl heterocycles {DDE(BÀBH )3
than that of the corresponding 2-acetyl isomers
2.87 kcal molÀ1
‡
} is higher
{DDE(BÀBH )2
‡
} [i.e. DDE(BÀBH )(3 À 2)
‡
,
expressed as the logarithm of the rate constant [ln(kZn i
kZn ref)] with reference to that of 3AP (kZn ref).
/
4.10 kcal molÀ1 and 5.28 kcal molÀ1 for the acetylfurans,
the acetylthiophenes and the acetylpyrroles respectively
(see Table 3)]. However, for the acetylheterocycles with
two ring heteroatoms, generally there is not a correspon-
dence between thermodynamic and kinetic basicities, as
for most of the substrates investigated the thermody-
namic protonation site is the aza-nitrogen rather than the
On the other hand, a comparison of the above catalytic
effects of Zn2‡ with those observed for Cu2‡ points to an
additional contribution arising from the interaction
between frontier orbitals. According to Klopman13
(based on the ionization potentials and the electronic
affinities reported by Moore)14 Cu2‡ is a softer cation
than Zn2‡ or Ni2‡ (En = À2.16 eV, À1.57 eV and
À0.24 eV respectively). In fact, catalysis by Cu2‡ is
more efficient than that by Zn2‡ for 3AP and 3AF, in
agreement with the higher energy values of the HOMOs
of these substrates compared with the HOMOs of the
other substrates containing one ring heteroatom in Table
2. A correlation between charge density and rate
constants for the Cu2‡-catalysed reaction, similar to that
discussed above for Zn2‡ catalysis, shows a strong
deviation from linearity for 3AP (see Fig. 2). This is the
substrate with the HOMO of highest energy and the
contribution due to the interaction of the frontier orbitals
clearly cannot be neglected for this acetylheterocycle.
In conclusion, the linearity of the correlations of Figs 1
and 2 outlines the paramount importance of charge
density on the carbonyl oxygen atom as a molecular
carbonyl oxygen atom. Therefore, similar DDE(BÀBH )
‡
calculations were not performed for the latter com-
pounds.
Interestingly for the substrates with an anomalously
high (5AIO) or low (3AF) kOH value, an anomalously
low (5AIO) or high (3AF) kH value has also been found.
The overall variation of kH for the investigated
compounds so far (Table 1) is only a factor of 11, as
‡
expected from the well-known fact that the H3O -
catalysed enolization reaction of monocarbonyl com-
pounds is not very sensitive12 to the molecular structure
of the substrate.
As for metal-ion catalysis, let us first discuss the
acetylheterocycles with one ring heteroatom (Table 2).
Rate enhancements for these substrates were previous-
ly2a,b,e attributed to coordination of the carbonyl oxygen
atom by the metal cation in the transition state of the
enolization reaction and the effectiveness of catalysis was
Figure 2. Linear correlation between rate constants for the
Cu2‡ -catalysed enolization reaction and differences in
charge densities on the carbonyl oxygen atom, with
reference to 3AP, for some acetylheterocycles with one ring
heteroatom. Open circles 5*) refer to compounds used for
drawing the line. 3AP 5*) has been excluded from the
correlation
Figure 1. Linear correlation between rate constants for the
Zn2‡ -catalysed enolization reaction and differences in
charge densities on the carbonyl oxygen atom, with
reference to 3AP, for some acetylheterocycles with one ring
heteroatom
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

METAL ION CATALYSIS IN THE ENOLIZATION REACTIONS OF ACETYLHETEROCYCLES
253
ring heteroatoms (Table 2). Calculations have been
performed on 3AIO, 5AIO, 3(5)APz, 2ATZ and
4ATZ. Interestingly, the charge density on the carbonyl
oxygen atoms of these substrates is lower (less negative
values) than that calculated for the one ring heteroatom
substrates in Table 2. Thus the present ab initio
calculations support the previous2c interpretation that
the exceptionally high metal ion catalysis for acetyl-
heterocycles with two heteroatoms comes from the
formation (where possible) of a chelate complex or
transition state (the two pathways are kinetically indis-
tinguishable) in which the Zn2‡ ion coordinates both the
carbonyl oxygen atom and the aza group. In particular,
for 3(5)APz, the tendency to form a chelate complex is so
strong that the tautomeric equilibrium of Scheme 1 is
completely shifted towards the 3-acetyl derivative,
which, according to our calculations (see Table S1
available as supplementary material at www.wiley.
com/epoc), per se is less stable than the 5-acetyl tautomer
in the gas phase, and is only slightly more abundant at
equilibrium in aqueous solution.
Figure 3. Linear correlation between rate constants for the
Zn2‡-catalysed enolization reaction and thermodynamic
stability of the substrate±metal ion complexes with reference
to 3AP
descriptor of the Lewis (but not Brønsted) basicity of the
acetylheterocycles containing one ring heteroatom of
Table 2 (with the exception of 3AP toward Cu2‡).
A closer insight into the Zn2‡ catalysis in the
enolization reaction has been achieved from the linear
correlation (n = 6, r = 0.949) obtained by plotting the
relative reactivity, [ln(kZn i/kZn ref)], of the acetylhetero-
cycles (containing one ring heteroatom) in Table 2
against the energy difference for the interaction of these
substrates with the Zn2‡ cation, calculated in the vacuum
with reference to 3AP (Fig. 3).
This approach in terms of DDE, unlike that in terms of
charge density, can give an estimate of the combined
contributions of charges and frontier orbitals to the
interaction of the substrate with the metallic catalyst. The
DDEiÀref. values were evaluated by neglecting the
contribution due to solvation (responsible for the
exoergonicity of the process), which should be similar
for the different substrates and probably cancel out* in
the calculation of DDEiÀref.. It was also assumed that the
DDE values are proportional to the real DDG# as the
process under investigation occurs in a condensed phase
(PV = 0, DU ꢁ DH) and it also seems reasonable to
neglect both the perturbation induced by the coordinating
metal ion on the vibrational motion (DDEvibrational) and
the entropy variation (DDS about zero at room tempera-
ture).
Our ab initio calculations also account for the absence
of metal ion catalysis in the enolization reaction of 5AIO.
Indeed, the formation of the complex 5AIO–Zn2‡ would
be characterized by the highest absolute value of
DDEiÀref (calculated from data of Table S1 available as
supplementary material at www.wiley.com/epoc) and the
highest residual positive charge density on the zinc cation
[Mulliken charges on Zn2‡ in the substrate–Zn2‡
complex are: 5AIO, 1.224; 3AIO, 1.131; 3(5)APz,
1.112; 4ATZ, 1.108; 2ATZ, 1.111]. This is in agreement
with a poor charge delocalization within the complex
(due to the impossibility for the aza nitrogen atom to
chelate the metal cation) as well as with the low energy
value of the HOMO of 5AIO.
Only a weak metal ion catalysis by Ni2‡ was
evidenced for the title substrates, as well as for some
other acetylheterocycles (i.e. 2AT,2a 3AT,2a 2AS,2e and
3AF.2e) The lack of an exhaustive series of kinetic data
did not allow a proper ab initio analysis of Ni2‡ catalysis.
In general, the stabilities of the complexes between the
acetylheterocycles with two ring heteroatoms in Table 2
and metal ions depend on the Lewis basicities of both the
pyridine-like nitrogen and the oxygen carbonyl atoms.
Consequently, the complex can be qualitatively analysed
in terms of electrostatic interactions and of overlapping
of frontier orbitals. The presence of a significant negative
charge density on both the carbonyl oxygen and the aza
nitrogen atoms is an essential factor in determining the
existence of a chelate metal catalysis, whereas a high
value of the HOMO energy suggests an effective
overlapping of the frontier orbitals and a considerable
reduction of the ion–substrate distance in the chelate
complex. Thus 2ATZ, which is the most sensitive
substrate to Zn2‡ catalysis among the acetylheterocycles
with two ring heteroatoms, is characterized by the highest
values of charge density on the aza nitrogen atom (Table
2APH has been excluded from the correlations of Figs
1–3 because of the unique hydrogen-bonding properties
of its N—H group with the solvent.
Let us now discuss the acetylheterocycles with two
*Additional calculations [suggested by one reviewer and performed
using a more recent program (GAMESS)] that take into account
differential solvation effects give a very similar linear plot, albeit with
a worse correlation coefficient (r = 0.88).
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

254
A. FONTANA ET AL.
Scheme 2. Classi®cation of substrates with reference to
metal ion caralysis
particular, for 3AP the high metal ion catalysis appears
to be due to a dominant contribution from the overlapping
of frontier orbitals.
The substrates of group II, characterized by relatively
high MAF values (2 ꢃ MAF ꢃ 430), display high charge
densities on both the aza nitrogen and the oxygen
carbonyl atom. The proposed classification is summar-
ized in Scheme 2.
Figure 4. Linear correlation between rate constant for the
Cu2‡-catalysed enolization reaction and thermodynamic
stability of the substrate±metal ion complexes of 3AP,
2ATZ, 4ATZ and 3ꢀ5)APz, with reference to 3AP
2). On the other hand, 4ATZ, which is the most sensitive
substrate to Cu2‡ catalysis among the acetylheterocycles
with two ring heteroatoms, has a high value for its
HOMO energy (Table 2).
Finally, the good correlation coefficient (r = 0.991) of
the linear plot of ln(kCu i/kCu ref) against DDEiÀref. (Fig. 4)
for 3AP, 2ATZ, 4ATZ and 3(5)APz (i.e. for the most
sensitive substrates to Cu2‡ catalysis) highlights the
importance of the proposed approach in terms of DDE to
estimate the combined contributions of charges and
frontier orbitals to the interaction of the acetylhetero-
cycles with the Cu2‡ cation.
Finally, the ratio of the rate constants for Cu2‡- and
Zn2‡-catalyses can be taken as an indication of the
importance of the overlapping of frontier orbitals in the
interaction of the substrate with the metal cation in the
enolization reaction. High energies of the frontier orbital
(e.g. those of 3AP, 3(5)APz and 4ATZ) are generally
associated with a large (i.e. >30) ratio of the two
experimental rate constants [kCu/kZn are 246, 35.9 and
195, for 3AP, 3(5)APz and 4ATZ respectively].
EXPERIMENTAL
Instruments
CONCLUSIONS
The kinetic experiments were carried out with a Jasco V-
550 UV–Vis spectrophotometer equipped with a Hi-tech
rapid kinetic accessory for the faster reactions or with a
Varian Cary 1E spectrophotometer. A 93 Radiometer pH-
The efficiency of metal ion catalysis in the enolization
reaction of acetophenone and a number of acetylhetero-
cycles was previously2c,e rationalized in terms of ‘hard’
and ‘soft’15 carbonyl groups. In this paper we have shown
that ab initio calculations can improve the understanding
of the influence of the factors affecting relative reactivity.
The substrates investigated can consequently be divided
in two groups: group I acetylheterocycles, which form
transition states or intermediate complexes by interaction
of the metal cation with the oxygen carbonyl atom, and
group II acetylheterocycles, which form a chelate
complex or transition state with the metal ion.
In turn, compounds of group I can be subdivided in
three sub-groups: group Ia contains less reactive
substrates (2AT, 3AT, 2AF, 5AIO, 5ATZ), character-
ized by ‘hard’ carbonyl groups, with MAF ꢂ1; group Ib
substrates (2AS, 3AF) characterized by ‘borderline’
carbonyl groups, for which 0.1 < MAF < 1; group Ic
contains more reactive substrates bearing a ‘soft’
carbonyl group (2APH,2c 3AP), with MAF > 1. In
1
meter was used for the pH measurements. H and 13C
NMR spectra were recorded on a Varian VXR 300-MHz.
Materials
All inorganic salts [KCl, NaClO4, KI, NaBr, ZnCl2,
NiCl2, Cu(NO3)2] and halogens (I2 and Br2) were
samples of AnalaR grade (Aldrich, Merck or Carlo Erba)
and were used without further purification.
3-Acetyl-5-methylisoxazole ꢀ3AIO)
3AIO was prepared by essentially the same procedure
described in Ref. 16. Acetonylacetone (Aldrich; 2 ml),
HNO3 (d = 1.40 g dmÀ3; 4 g) and water (6 ml) were
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

METAL ION CATALYSIS IN THE ENOLIZATION REACTIONS OF ACETYLHETEROCYCLES
255
heated to boiling. The reaction mixture was then allowed
DB is the absorbance of the unprotonated substrate, DBH
‡
that of the conjugate acid and C are molar concentrations
of the specified species.
to cool at room temperature. After the addition of 6 ml of
water the compound was extracted with diethyl ether,
dried and distilled under reduced pressure (b.p. 68–69°C
at 20 mmHg; lit.16 65–70°C at 20 mmHg). The identity
of the product was confirmed from its 1H NMR spectrum.
‡
‡
I ˆ CBH =CB ˆ ꢀDB À D†=ꢀD À DBH
†
ꢀ8†
The data were analysed by both the Hammett20a and
the Cox and Yates ‘excess acidity’20b methods. With the
5-Acetyl-3-methylisoxazole ꢀ5AIO)
former method, pKBH was obtained by plotting log I
against the acidity function HA according to:
‡
20a
5AIO was prepared following the procedure described in
Ref. 17 A solution of nitroethane (Aldrich; 7.2 g,
96 mmol) and triethylamine (Aldrich; 20 drops) in dry
benzene was added dropwise to a solution of phenyl
isocyanate (Aldrich; 20.7 g, 174 mmol) and 3-buten-2-
one (Aldrich; 6.7 g, 95 mmol) in dry benzene (35 ml).
After stirring for 1 h, the reaction mixture was refluxed
for one additional hour and cooled. The solid was filtered
off and the filtrate was concentrated to give a yellow oil
which was distilled to afford 5-acetyl-3-methyl-4,5-
dihydroisoxazole (i). Compound i (2.0 g, 15.7 mmol) in
dry benzene (50 ml) and active g-manganese dioxide (5.0
g) were refluxed for 3 h, while the water produced was
removed by means of a Dean–Stark trap. The solid was
filtered through Celite and washed with dry benzene.
Evaporation of the filtrate gave 5-acetyl-3-methylisoxa-
zole, which was then purified by sublimation at 40°C/
20 mmHg: m.p. 72–74°C (lit. m.p. 73–74°C).17 The
‡
‡
log I ˆ logꢀCBH =CB† ˆ ÀmHA ‡ pKBH
ꢀ9†
With the latter method, pKBH was obtained as the
‡
intercept by plotting the left-hand side of Eqn. (10)
against excess acidity values X20b of the HClO4 solutions:
log I À log CH ˆ mÃX ‡ pKBH
ꢀ10†
‡
‡
The average values obtained by the two methods were
as follows:
pKBH
pKBH
‡
‡
(3AIO) À 3.79(Æ0.11); (m 1.12, m* 0.55)
(5AIO) À 3.58(Æ0.34); (m 0.93, m* 0.50)
A pKBH of 0.09 for 3(5)APz was interpolated from a
Hammett correlation reported for a series of monosub-
‡
stituted pyrazoles.21
1
identity of the product was confirmed by its H and 13C
Though the site of protonation of 3(5)APz is the aza
nitrogen,21 it is not quite clear which is the site of
protonation of 3AIO and 5AIO. However, the following
pieces of evidence point to a carbonyl oxygen rather than
nitrogen protonation: (1) the absorbance of the substrate
increases upon protonation, and this is typical of a,b-
unsaturated carbonyl compounds,22a whereas the absor-
bance of thiazoles, which are protonated on the aza
nitrogen atom, decreases upon protonation;22b (2) the
present acetylisoxazoles, like most a,b-unsaturated
carbonyl compounds,22a follow the HA acidity function
(whereas thiazoles22b and pyridines22c follow the H0
NMR spectra.
3ꢀ5)-Acetylpyrazole [3ꢀ5)APz]
3(5)APz was prepared according to a literature proce-
dure.18 An ethereal solution of diazomethane was added
dropwise to an ethereal solution of methylethynylketone
at 0°C. After 12 h, evaporation of the solution gave a pale
yellow solid, which on crystallization from petroleum
ether yielded needles of 3(5)-acetylpyrazole: m.p. 100°C
(lit. m.p. 100–101°C).19 The identity of the product was
acidity function); (3) by comparing the values of pKBH
of different heterocyclic compounds for which the aza
nitrogen atom is the site of protonation, it transpires that
‡
1
confirmed by its H NMR spectrum.
the pKBH
‡
of the acetyl-substituted derivative is much
pKBH‡ measurements
lower (from 2.5 to 4 logarithmic units) than that of the
parent compound [e.g. pKBH
versus pKBH
(pyridine) 5.2;8a pKBH
versus pKBH
(thiazole) 2.5;23 pKBH
(imidazole) 7.1;23 pKBH
(pyrazole) 2.5224]. The pKBH
‡
(2-acetylpyridine) 2.643
The thermodynamic acid dissociation constants of 3AIO
and 5AIO were determined spectrophotometrically. Acid
solutions of appropriate concentrations were made up by
diluting concentrated perchloric acid (ca 11 mol dmÀ3).
The absorbances D of equivalent amounts of substrate in
aqueous HClO4 of different molarities were measured at
appropriate wavelengths, near lmax of the free base B.
The spectra were recorded against blank HClO4 solutions
of the same molarity and the absorbance values D were
taken immediately after the addition of the substrate.
Ionization ratios I were calculated from Eqn. (8), where
‡
‡
(4ATZ) À 0.452c
‡
‡
(2AI) 3.122d versus
pKBH
pKBH
‡
‡
‡
(3(5)APz) 0.09 versus
of 5-methyl-
‡
(À2.01)25 and 3-methyl-isoxazole (À1.62)25 are only
2.17 and 1.57 units higher than those of 3AIO and 5AIO
respectively; nevertheless, both heterocycles behave as
Hammet bases (i.e. follow the H0 acidity function); (4)
the pKBH
‡
values obtained of 3AIO and 5AIO fall within
values of ketones (i.e.
the ‘sensible pattern’7 of pKBH
‡
from À3.1 for acetone to À4.7 for benzophenone) and the
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

256
A. FONTANA ET AL.
‡
Figure 5. Structures and conformations 5syn/anti) of some acetylheterocyles and their adducts with metal 5M2‡) or proton 5H )
ions considered for DFT calculations
observed m* values are only slightly higher than that
found for acetophenone (i.e. 0.46).
Fig. 5) for all the acetylheterocycles were considered and
the most stable one was selected.
As expected the pKBH
‡
values of the two isoxazole
Some molecular and atomic properties of the structures
of minimal energy were evaluated (Mulliken atomic
charge and HOMO OM energies). The lowest energy
structure among the four possible structures of acetyl-
heterocycle–Zn2‡ adducts, as depicted in Fig. 5, was
selected and used in the DDEiÀref calculations (see
Discussion section). The structures of the acetylhetero-
cycle–Cu2‡ adducts have been obtained by replacing
Cu2‡ for Zn2‡ ion in each lowest-energy complex
previously selected and the geometries and energies of
the adducts were re-optimized. All lowest-energy
structures of acetylheterocycles and their proton or metal
ion adducts were identified as minima points by
calculating harmonic vibrational frequencies at the DFT
pBP/DN** level. For the basicity comparison of the 3-
and 2-acetylheterocycles (see Results and Discussion
isomers are quite similar, as the aza nitrogen is either in
position ortho-like or para-like with respect to the
carbonyl group.
Ab initio calculations
The DFT ab initio method (non-local density calculations
using the BP86 functional), as implemented in the PC
Spartan Pro v.1.0.3 package, was used throughout this
work. Structures and energies of all acetylheterocycles
and their adducts with the proton and metal ions were
optimized at the SCF level with the pBP/DN** numerical
polarization basis set (equivalent in size to the 6-31G**
Gaussian basis set). Both syn and anti conformations (see
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257

METAL ION CATALYSIS IN THE ENOLIZATION REACTIONS OF ACETYLHETEROCYCLES
257
9. Bell RP. The Proton in Chemistry, 2nd edn. Chapman and Hall:
London, 1973.
10. Chambers CC, Hawkins GD, Cramer CJ, Truhlar DG. J. Phys.
Chem. 1996; 100: 16 385.
sections) the hydration energy has also been estimated for
the protonated and the unprotonated forms, using the
SM5.4 procedure of Cramer and co-workers.10
11. Klopman G. Chemical Reactivity and Reaction Paths. Wiley
Interscience: New York, 1974.
12. (a) Keeffe JR, Kresge AJ. In The Chemistry of Enols, Patai S,
Rappoport Z (eds). Wiley: Chichester, 1990; 399–480. (b) Carey
ARE, More O’Ferrall RA, Murphy MG, Murray BA. J. Chem.
Soc., Perkin Trans. 2 1994; 2471.
REFERENCES
13. Klopman G. J. Am. Chem. Soc. 1968; 223.
1. Eustace SJ, McCann GM, More O’Ferrall RA, Murphy MG,
Murray BA, Walsh SM. J. Phys. Org. Chem. 1998; 11: 519 and
references cited therein.
14. Moore CE. Analyses of Optical Spectra, NSRDS-NBS 34. Office of
Standard Reference Data, National Bureau of Standards: Wash-
ington, DC.
15. Pearson RG. Hard and Soft Acids and Bases Principle in Organic
Chemistry. Academic Press: New York, 1977.
2. (a) De Maria P, Fontana A, Spinelli D. J. Chem. Soc., Perkin
Trans. 2 1991; 1067; (b) De Maria P, Fontana A, Frascari S,
Ferroni F, Spinelli D. J. Chem. Soc., Perkin Trans. 2 1992; 825; (c)
De Maria P, Fontana A, Arlotta M, Chimichi S, Spinelli D. J.
Chem. Soc., Perkin Trans. 2 1994; 415; (d) De Maria P, Fontana A,
Spinelli D, Macaluso G. Gazz. Chim. Ital. 1996; 126: 45; (e) De
Maria P, Fontana A, Siani G, Spinelli D. Eur. J. Org. Chem. 1998;
1867.
3. Cox BG. J. Am. Chem. Soc. 1974; 96: 6823.
4. Noto R, Gruttadauria M, Arnone C, Lamartina L, Spinelli D. J.
Chem. Res. (S) 1993; 274.
5. Cox BG, De Maria P, Fini A, Hassan AF. J. Chem. Soc., Perkin
Trans. 2 1981; 1351.
6. Jones JR, Marks RE, Subba Rao SC. Trans. Faraday Soc. 1965;
61: 111.
7. Bagno A, Scorrano G, More O’Ferrall RA. Rev. Chem. Intermed.
1987; 7: 313.
8. (a) Joule JA, Mills K, Smith GF. Heterocyclic Chemistry, 3rd edn.
Chapman & Hall: London, 1995; (b) Katritzky AR, Lagowski JM.
The Principles of Heterocyclic Chemistry. Methuen & Co. Ltd:
London, 1967.
16. Schmidt J, Widmann KT. Berichte 1909; 42: 1869.
17. Chimichi S, Cosimelli B. Synth. Commun. 1992; 22(20): 2909.
18. Bowden K, Jones ERH. J. Chem. Soc. 1946; 953.
19. Panizzi L, Benati O. Gazz. Chim. Ital. 1946; 76: 66.
20. (a) Rochester CH. Acidity Functions. Academic Press: London,
1970. (b) Cox RA, Yates K. J. Am. Chem. Soc. 1978; 100: 3861.
21. (a) Elguero J, Gonzalez E, Jacquier R. Bull. Chim. Soc. Fr. 1968;
12: 5009; (b) Broughton HB, Green SM, Rzepa HS. J. Chem. Soc.,
Perkin Trans. 2 1995; 431.
22. (a) Zalewski RI, Dunn GE. Can. J. Chem. 1968; 46: 2469; (b)
Forlani L, De Maria P. J. Chem. Soc., Perkin Trans. 2 1979; 163;
(c) Cook MJ, Dassanyake NL, Johnson CD, Katritzky AR, Toone
TW. J. Chem. Soc., Perkin Trans. 2 1974; 1069.
23. Gilchrist TL. Heterocyclic Chemistry, 3rd edn. Longman: 1997.
24. Katritzky AR. Handbook of Heterocyclic Chemistry. Pergamon
Press: New York, 1985.
25. Clementi S, Forsythe PP, Johnson CD, Katritzky AR, Terem B. J.
Chem. Soc., Perkin Trans. 2 1974; 399.
Copyright  2002 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2002; 15: 247–257