Solvation and nucleophilic reactivity of conjugate-base anions of 5-methyl Meldrum's acid and of 3,3-dimethylbarbituric acid in acetonitrile-methanol

Article abstract of DOI:10.1039/a900283i

Conjugate-base anions of 5-methyl Meldrum's acid and of 1,3-dimethylbarbituric acid give comparable nucleophilic reactivity toward ethyl iodide in acetonitrile and also have specific interaction enthalpies, Δ_tH_SI^AN→MeOH comparable with those of 3,5-dinitrobenzoate ion and of phthalimidide ion, while pK_a values in the aqueous phase vary significantly among the series. The results lend support to the view that nucleophilic reactivity in acetonitrile is controlled mainly by the partial desolvation of nucleophilic anions accompanying activation. Variation of the specific interaction enthalpy for the present reactions on going from the initial to the transition-state is much smaller by comparison to those for imidide ion and carboxylate ion reactions. Theoretical analysis of the enthalpy with use of MNDO/PM3 procedures indicates that the steric inhibition of the approaching solvent molecule to the carbonyl oxygen by the group coordinated to the central atom is the main factor bringing about the smaller variation.

Full text of DOI:10.1039/a900283i

View Article Online / Journal Homepage / Table of Contents for this issue
Solvation and nucleophilic reactivity of conjugate-base anions of
-methyl Meldrum’s acid and of 3,3-dimethylbarbituric acid in
5
acetonitrile–methanol
a
b
a
a
a
Yasuhiko Kondo,* Kouji Yano, Miyuki Urade, Aki Yamada, Ruiko Kouyama and
Tatsuya Takagi
c
a
Department of Natural Sciences, Osaka Women’s University, Daisen-cho, Sakai, Osaka 590,
Japan
b
Department of Applied Chemistry, Faculty of Engineering, Osaka University, Suita,
Osaka 565, Japan
c
Graduate School of Pharmaceutical Sciences, Osaka University, Suita, Osaka 565, Japan
Received (in Cambridge) 11th January 1999, Accepted 22nd March 1999
Conjugate-base anions of 5-methyl Meldrum’s acid and of 1,3-dimethylbarbituric acid give comparable
nucleophilic reactivity toward ethyl iodide in acetonitrile and also have specific interaction enthalpies,
AN→MeOH
∆tHSI
comparable with those of 3,5-dinitrobenzoate ion and of phthalimidide ion, while pK values in the
a
aqueous phase vary significantly among the series. The results lend support to the view that nucleophilic reactivity
in acetonitrile is controlled mainly by the partial desolvation of nucleophilic anions accompanying activation.
Variation of the specific interaction enthalpy for the present reactions on going from the initial to the transition-
state is much smaller by comparison to those for imidide ion and carboxylate ion reactions. Theoretical analysis
of the enthalpy with use of MNDO/PM3 procedures indicates that the steric inhibition of the approaching solvent
molecule to the carbonyl oxygen by the group coordinated to the central atom is the main factor bringing about
the smaller variation.
In solution chemistry it has long been accepted that oxygen
acids are more acidic than nitrogen acids and nitrogen acids are
more acidic than carbon acids. Recently quantitative support
has been given to this notion through equilibrium measure-
Table 1 Enthalpies of solution, ∆ H in acetonitrile–methanol mix-
s
Ϫ1
tures at 25 ЊC (in kJ mol
)
TMA
TEA
MA
TMA
PhMA
TMA
DMBarb
1
χ
MeOH
MMA
ments in DMSO. In this respect the high acidity of 2,2-
dimethyl-1,3-dioxane-4,6-dione (Meldrum’s acid) and of
0
14.5
15.1
Ϫ1.92
Ϫ4.10
Ϫ2.27
0.10
19.0
6.49
3.13
4.16
7.81
14.3
15.0
pyrimidine-2,4,6-(1H,3H,5H)-trione (barbituric acid) i.e., pK ’s
a
0.10
0.25
0.50
0
1.0
Ϫ3.96
Ϫ7.35
Ϫ5.55
Ϫ2.01
4.45
Ϫ5.75
Ϫ8.98
Ϫ6.12
Ϫ1.16
6.37
2
3
in H O are 4.8 and 4.04, is very intriguing since both are
2
carbon acids although activated by geminal carbonyl groups,
and attention has been paid to interpreting the high acidity
.75
2–7
6.03
from a physico-chemical viewpoint.
Quantitative evaluation of nucleophilic reactivity has been
TMA, tetramethylammonium; TEA, tetraethylammonium; MMA,
conjugate-base anion of 5-methyl Meldrum’s acid; MA, conjugate-base
anion of Meldrum’s acid; PhMA, conjugate-base anion of 5-phenyl
Meldrum’s acid; DMBarb, conjugate-base anion of 1,3-dimethyl-
barbituric acid; χMeOH, mole fraction of methanol.
8–12
a long standing subject in physical organic chemistry.
Recently, empirical correlations between the logarithmic rate
and the specific interaction enthalpy for a nucleophile,
AN→MeOH
∆tHSI
(a measure of the hydrogen-bond accepting bas-
icity for a nucleophile) have been documented for nucleophilic
13,14
substitution reactions in acetonitrile.
From the correlations
it has been concluded that the nucleophilic reactivity of anions
is more influenced by a partial desolvation of the nucleophile
on going from the reactant to the transition state by com-
13,14
parison to intrinsic properties of the nucleophile.
Comparative studies on kinetic reactivities as well as on
equilibrium reactivities for the conjugate-base anion of
Meldrum’s acid as a nucleophile with those for carboxylate ions
and for imidide ions would be quite worthwhile for deduc-
ing essential features of nucleophilic substitution reactions in
solution.
will be discussed on the basis of semi-empirical molecular
orbital calculations.
Ϫ
Ϫ
Nu ϩ Et–I → Nu–Et ϩ I
In this work, following the enthalpy of solution measure-
ments for the tetraalkylammonium salts containing conjugate-
base anions of Meldrum’s acid and its derivatives, A (X = H,
Me, Ph) and of 1,3-dimethylbarbituric acid, B, rates of reaction
of the conjugate-base anion of 5-methyl Meldrum’s acid and of
Results
Enthalpies of solution, ∆ H for the tetraalkylammonium salts
s
containing the conjugate-base anion of Meldrum’s acids and of
1,3-dimethylbarbituric acid have been measured in acetonitrile–
methanol mixtures and are summarized in Table 1. Single-ion
1
,3-dimethylbarbituric acid with ethyl iodide will be determined
in acetonitrile–methanol mixtures and the solvation properties
J. Chem. Soc., Perkin Trans. 2, 1999, 1181–1186
1181

View Article Online
AN→mix
Table 2 Single ion enthalpies of transfer from acetonitrile to acetonitrile–methanol mixtures, ∆ H
and interaction parameters
t
Ϫ
Ϫ
χMeOH
MMA
0
MA
PhMA
0
DMBarb
0
TS (MMA)
TS (DMBarb)
0
0
0
0
0
0
0
0
1
.10
.25
.50
.75
.0
Ϫ17.6
Ϫ16.6
Ϫ18.25
Ϫ17.0
Ϫ14.75
Ϫ10.1
11.4
Ϫ21.5
42.0
—
Ϫ11.6
Ϫ19.85
Ϫ3.7
Ϫ2.85
0.8
Ϫ5.65
Ϫ7.20
Ϫ4.10
2.80
15.7
31.7
Ϫ20.55
Ϫ19.7
Ϫ16.8
Ϫ12.75
11.75
Ϫ24.5
30.0
Ϫ14.6
Ϫ14.5
Ϫ11.5
Ϫ7.4
12.1
Ϫ19.5
18.0
Ϫ22.7
Ϫ20.8
Ϫ16.5
Ϫ11.3
15.2
Ϫ26.5
40.0
2.5
16.85
25.85
Ϫ9.0
18.0
Ϫ0.6
AN→MeOH
∆t
∆tHSI
Kse
H
PHYS
AN→MeOH
Ϫ16.0
12.0
Z
2.2
—
2.3
1.5
Ϫ
Ϫ
βPHYS’s are 395 and 445 K for MMA and DMBarb reaction. TS (MMA) and TS (DMBarb), transition-state anions for the reaction of the
conjugate-base anion of 5-methyl Meldrum’s acid and of 1,3-dimethylbarbituric acid.
Table 3 Rate constants and activation parameters in acetonitrile–methanol mixtures (30 ЊC)
5
-Methyl Meldrum’s acid anion ϩ EtI
1,3-Dimethylbarbiturate ϩ EtI
Ϫ6
3
Ϫ1 Ϫ1
‡
Ϫ1
‡
Ϫ1
Ϫ1
Ϫ6
3
Ϫ1 Ϫ1
‡
Ϫ1
‡
Ϫ1
Ϫ1
χMeOH
k/10 dm mol
s
∆H /kJ mol
∆S /J K mol
k/10 dm mol
s
∆H /kJ mol
∆S /J K mol
3
3
0
1.43 × 10
1.11 × 10
29.0
8.64
3.83
1.24
67.7
81.6
85.5
88.6
87.9
98.9
Ϫ76.2
1.80 × 10
1.42 × 10
40.9
12.9
5.24
1.73
68.2
82.4
83.8
85.3
88.4
96.8
Ϫ72.5
2
2
0
0
0
0
1
.10
.25
.50
.75
.0
Ϫ51.6
Ϫ49.9
Ϫ49.7
Ϫ58.8
Ϫ31.9
Ϫ46.6
Ϫ52.5
Ϫ57.0
Ϫ54.5
Ϫ35.9
AN→MeOH
enthalpies of transfer from acetonitrile to mixed solvents,
negative value of ∆ H
, that is to say, is the most
t
SI
AN→mix
∆t
H
have been calculated on the basis of the tetrabutyl-
hydrogen-bond accepting of all (see Table 2).
15,16
ammonium/tetrabutyl borate (TBA/TBB) assumption
and
Rate constants and activation parameters for the reaction
of ethyl iodide with the conjugate-base anion of 5-methyl
Meldrum’s acid (5-methyl Meldrum’s acid anion) and with that
of 1,3-dimethylbarbituric acid (1,3-dimethylbarbiturate ion)
have been determined in acetonitrile–methanol mixtures and
are summarized in Table 3. The rate constant shows a very sig-
nificant decrease, while the activation enthalpy undergoes a
sharp increase, at low methanol mole fractions, in a way that is
are summarized in Table 2. All these enthalpies indicate a
sharp decrease at a low methanol mole fraction, followed by
a curvi-linear increase with increasing methanol mole fraction,
a typically observed pattern for anions which are stabilized
15–17
through hydrogen-bonding interactions with methanol.
The results are well simulated by eqns. (1) and (2), where χAN
AN→mix
reminiscent of the pattern observed for the single ion enthalpy
∆t
H
=
AN→mix
of transfer for anions, ∆ H
. In contrast to these, the acti-
t
AN→MeOH
∆tHPHYS
× χMeOH × [1 Ϫ 1.23 × χMeOH × (1 Ϫ χMeOH)]
vation entropy goes through a maximum. These are all typically
observed trends for the reactions in which the hydrogen-
bonding interactions of a nucleophile with methanol play a
On the basis of a thermodynamic cycle
using the enthalpies of transfer for ethyl iodide in the solvent
mixtures, single ion enthalpies of transfer for the transition-
state anion were calculated and are summarized in Table 2. The
interaction parameters for the transition-state anion were also
calculated as described above and are given in Table 2. The
results indicate that although specific interactions play a signifi-
AN→MeOH
ϩ ∆ H
× K × χMeOH/(χAN ϩ K × χMeOH) (1)
se se
t
SI
15,17
significant role.
AN→MeOH
AN→MeOH
AN→MeOH
∆t
H
= ∆ H
ϩ ∆ H
SI
(2)
t
PHYS
t
16
and χMeOH stand for the mole fraction of acetonitrile and of
methanol in the solvent mixtures, and Kse stands for the
equilibrium constant for a solvent exchange process on the
15–17
solvation site around an ion.
The first term on the right
hand side of eqn. (1) simulates the curvi-linear increase of the
cant role even at the transition state, the amount of variation of
AN→mix
AN→MeOH
AN→MeOH
enthalpy, ∆ H
and the term, ∆ H
indicates
the specific interaction enthalpy, δ∆ H
, on going from
t
t
PHYS
t
SI
the enthalpy of transfer from acetonitrile to methanol arising
from more “physical” interactions such as electrostatic, proto-
phobic and cavity forming interactions. The second term
the reactant to the transition state, i.e., 15.5 for the 5-methyl-
Meldrum’s acid anion and 10.5 for the 1,3-dimethylbarbiturate
ion reaction, is much smaller by comparison to those for
imidide ion and carboxylate ion reactions, i.e., 22.9 for the
simulates the sharp decrease of the enthalpy at low methanol
AN→MeOH
18
19
Ϫ1
mole fractions and the term ∆ H
indicates more
former and 27.2 for the latter (all in kJ mol ).
t
SI
“
specific” or more “chemical” interactions such as hydrogen-
bonding and charge-transfer interactions which are observed
15–17
Discussion
for a specific pair of solute and solvent.
The search for a
AN→MeOH
consistent set of interaction parameters, i.e., ∆ H
,
Logarithmic rates for the nucleophilic substitutions in aceto-
nitrile have been documented to be linearly correlated with the
specific interaction enthalpy for the relevant nucleophile (a
scale of hydrogen-bond accepting basicity), ∆ H
From this observation, together with the concurrent analysis of
t
PHYS
AN→MeOH
∆tHSI
and K has been performed through curve-fitting
se
AN→MeOH
procedures, systematically varying ∆ H
and K , until
t
SI
se
AN→MeOH 13
the optimum fit of the calculated values to the experimental one
was attained. One of the most plausible sets of parameters is
given in Table 2. Usually experimental results were simulated
with these eqns. within the maximum deviation of ±1.6 kJ
.
t
SI
the reaction enthalpy, ∆ H, nucleophilic reactivity in
R
acetonitrile (as log kAN) is concluded to be controlled mainly by
the partial desolvation of anions accompanying activation by
comparison to the intrinsic properties of the central atom in
Ϫ1
mol . Amongst all the ions studied in this work, the conjugate-
base anion of 1,3-dimethylbarbituric acid indicates the largest
1
182
J. Chem. Soc., Perkin Trans. 2, 1999, 1181–1186

View Article Online
Table 4 Comparison of various reactivity parameters
AN→MeOH
Ϫ1
Ϫ1
Nucleophiles
3 ϩ log kAN
∆ H
t
/kJ mol
pK (in H O)
pK (in DMSO)
∆ H/kJ mol
R
SI
a
2
a
a
a
b
c
a
3
1
5
,5-Dinitrobenzoate
,3-Dimethylbarbiturate
-Methyl Meldrum’s acid anion
0.471
Ϫ24.0
Ϫ26.5
Ϫ24.5a
Ϫ26.0
2.81d
4.68
4.8
9.15
Ϫ60.0
—
Ϫ110.7
Ϫ134.7
c
0.252
0.153a
1.808
8.4f
e
7.4
13.7
g
h
a
Phthalimidide
8.30
a
b
c
d
e
f
g
h
Ref. 14. Ref. 24. Ref. 2. Ref. 6. The value for Meldrum’s acid, ref. 2. Ref. 4. Ref. 25. Ref. 26.
13,14
‡
anions.
This suggests that if comparisons of reactivity are
δ∆Y =
‡
AN→mix
Ϫ
AN→mix
Ϫ
made for the reaction of nucleophiles of comparable specific
interaction enthalpy, the intrinsic reactivity of the central atom
could be evaluated without being modified by a solvational
term. In order to critically examine the credibility of this sup-
position, as well as of various scales of reactivity, various scales
of nucleophilic reactivity have been summarized in Table 4 for
δ∆Y
ϩ ∆ Y
(TS ) Ϫ ∆ Y
(Nu ) (3)
PHYS
t
SI
t
SI
exchange model with the use of the interaction parameters
derived above, can be carried out through the following
procedures.
1. Activation enthalpies arising from “more physical” inter-
actions can be derived through eqn. (4), substituting the rele-
2
1,22
the reactions of nucleophiles which have comparable values for
AN→MeOH
the enthalpy, ∆ H
, even though they have different
t
SI
‡
reaction centers.
These nucleophiles, although phthalimidide ion has a rate
that is faster by one logarithmic unit, indicate comparable
δ∆H PHYS
=
AN→MeOH
PHYS
Ϫ
AN→MeOH
PHYS
Ϫ
[∆
H
(TS ) Ϫ ∆
H
(Nu )] × χMeOH
×
t
t
AN→mix
[1 Ϫ 1.23 × χMeOH × (1 Ϫ χMeOH)] Ϫ ∆
H
(EtI) (4)
t
reactivity as anticipated from the similar hydrogen-bond
AN→MeOH
vant values given in Table 2 and enthalpies of transfer for ethyl
accepting basicity, ∆ H
. pK values in the aqueous
t
SI
a
15
iodide into eqn. (4).
. In the reactions for which an isokinetic relationship holds
phase and in DMSO have been suggested as scales expressing
nucleophilic reactivity, but they indicate a large variation along
the series, from 3,5-dinitrobenzoate to phthalimidide, i.e., 5.5
pK units for the former and 6.3 pK units for the latter and the
2
between the activation parameters arising from “more physi-
cal” interactions, eqn. (5) holds. Thus, activation entropies
a
a
reaction enthalpy in acetonitrile, ∆ H as well, in contrast to the
R
‡
‡
δ∆S
= δ∆H PHYS/βPHYS
(5)
PHYS
much smaller variation in logarithmic rate, 3 ϩ log k and in
AN
AN→MeOH
the enthalpy, ∆ H
. It is to be noted that the trend
t
SI
arising from “more physical” interactions can be derived
through the division of the enthalpies calculated above by an
assumed value for the isokinetic temperature due to “more
observed for the reaction enthalpy in acetonitrile, ∆ H could be
R
understood according to the principle that the weaker the acid,
the more significant is the nucleophilic reactivity of the conju-
gate base but this applies only to the relation between the
equilibrium properties, pK (in H O) and ∆ H. 1,3-Dimethyl-
physical” interactions, βPHYS
.
3
. Single ion enthalpies of transfer arising from specific
interactions can be calculated through the substitution of the
relevant quantities given in Table 2 into eqn. (6).
a
2
R
barbituric acid has acidity comparable to diethylacetic acid in
6
the aqueous phase, i.e., pK ’s are 4.68 for the former and 4.73
a
20
for the latter, whereas the hydrogen-bond accepting basicity
∆tHSIAN→mix
=
AN→MeOH
for the barbiturate ion is much weaker, i.e., ∆ H
t
SI
AN→MeOH
Ϫ1 14
∆
t
H
SI
× Kse × χMeOH/(χAN ϩ Kse × χMeOH
)
(6)
values are Ϫ26.0 and Ϫ48.0 kJ mol . The logarithmic rate in
acetonitrile is larger by 2 units for the diethylacetate ion in
comparison to the 1,3-dimethylbarbiturate ion, i.e., 3 ϩ log
kAN’s are 0.252 and 2.55 for the 1,3-dimethylbarbiturate and
diethylacetate ion reactions. A similar situation holds for 5-
methyl Meldrum’s acid and pivalic acid: they have comparable
4
. According to the solvent exchange model the quantity
AN→MeOH
∆tHSI
can be separated into components, the number
of methanol molecules participating in specific interactions
with the relevant solute, Z, and the enthalpy of the solvent
exchange process on the solvation site around the anion, ∆Hse,
according to eqn. (7). Single ion entropies of transfer arising
14
aqueous pK values, while the hydrogen-bond accepting basicity
a
as well as the nucleophilic reactivity for pivalate ion is much
larger; relevant values for the pivalic acid and pivalate ion reac-
AN→MeOH
∆tHSI
= Z∆Hse
(7)
2
0
AN→MeOH 14
tions are 5.03, Ϫ43.5 and 2.456 for pK , ∆ H
,
and
a
t
SI
14
3
ϩ log kAN
, respectively. It has long been accepted that O–H
from specific interactions can be derived through eqn. (8) by
acids are more acidic in comparison to N–H acids, and N–H
acids are more acidic in comparison to C–H acids. The results
presented above are likely to give further support for the view
that nucleophilic reactivity in acetonitrile is controlled mainly
by the partial desolvation of anions accompanying activation
AN→mix
∆ S
=
t
SI
(Z∆H /T) × (K χMeOH)/(χAN ϩ KseχMeOH) ϩ
se
se
ZRln (χAN ϩ KseχMeOH
)
(8)
and that the nucleophiles having comparable values for the
substituting the relevant quantities given in Table 2, together
AN→MeOH
enthalpy, ∆ H
leads to comparable nucleophilic
with the assumed values of Z and ∆H into eqn. (8).
t
SI
se
reactivity irrespective of the intrinsic properties of the nucleo-
philic central atom.
5. Substitution of these derived quantities into eqn. (3) gives
the desired thermodynamic quantities. The procedures 1–5 were
repeated until optimum fits between the calculated and experi-
mental values were attained. One of the most plausible sets of
parameters, solvation numbers Z for nucleophilic anions and
for the transition state anions are also given in Table 2; the final
values of βPHYS are 395 and 445 K for the reaction of 5-methyl
Meldrum’s acid anion and for the reaction of 1,3-dimethyl-
barbiturate ion.
Quantitative separation of the activation parameters into
constituents is an essential part of an analysis for the under-
standing of reaction behaviors in mixed solvents. Variation of
activation parameters in solvent mixtures is brought about by
differential responses of two constituent interactions to solvent
composition, i.e., more “physical” and specific interactions, as
well as differential contributions for the transition-state anion
from those at nucleophiles and is expressed by eqn (3), where Y
stands for relevant thermodynamic quantities. Reconstruction
of activation parameters, which is guided by the solvent
General trends observed for the activation parameters
‡
‡
(S-shaped character of δ∆H vs. δ∆S correlation) could be
reproduced by the procedures as described above (see Fig. 1).
J. Chem. Soc., Perkin Trans. 2, 1999, 1181–1186
1183

View Article Online
Fig. 1 Activation enthalpy vs. activation entropy correlations for the
reaction of 1,3-dimethylbarbiturate ion plus ethyl iodide in aceto-
nitrile–methanol mixtures: ᭺, overall quantity; ᭹, “more physical”
interaction quantity; curve, calculated values (see text). 1, Acetonitrile;
2
, methanol.
Fig. 2 Empirical correlation between the specific interaction enthalpy,
AN→MeOH
∆tHSI
and the atomic charge on carbonyl oxygen: ᭺, nucleo-
, transition-state anions; , effective atomic charge
These are other examples which indicate that the linear
enthalpy vs. entropy correlation does not necessarily hold for
observed overall quantities, even when it holds between particu-
lar constituent quantities, i.e., “more physical” interaction
quantities.
Statistical analysis on specific interaction enthalpies with
respect to the number of solvent molecules which participate in
specific interaction, Z, leads to eqn. (9). This implies that the
philic anions;
on carbonyl oxygen in transition-state anions. 1, 1,3-dimethyl-
barbiturate ion; 2, conjugate-base anion of 5-methyl Meldrum’s acid; 3,
conjugate-base anion of Meldrum’s acid; 4, conjugate-base anion of
5-phenyl Meldrum’s acid; 1TS, transition-state anion of 1,3-dimethyl-
barbiturate ion reaction; 2TS, transition-state anion of 5-methyl
Meldrum’s acid anion reaction.
the one between the specific interaction enthalpy for the
nucleophilic anion, ∆ H
the carbonyl oxygen as shown in Fig. 2.
In this correlation the enthalpies for the transition state
AN→MeOH
and the atomic charge on
t
SI
AN→MeOH
∆tHSI
= Ϫ2.2 Ϫ 10.2 × Z
(9)
n = 4, r = 0.99
anion (
) are likely to indicate an upward deviation from the
enthalpy change for the solvent exchange process on the sol-
regression line as had been observed for the imidide ion and the
22,23
vation site, ∆H for relevant anions studied here, Ϫ10.2 kJ
carboxylate ion reactions.
In order to allow for the effects of
se
Ϫ1
mol is fairly close to those for imidide ion and for carboxylate
23
the methyl group in the incoming methyl iodide, the geometry
was optimized for the transition-state anion complexed with
methanol, with C (reaction center of nucleophilic anion)–C (in
methyl iodide) and C (in methyl iodide)–I distances being fixed
at the values which have been optimized in advance for uncom-
plexed transition-state anions. One example of optimized struc-
tures, i.e., the transition-state anion for 5-methyl Meldrum’s
acid anion reaction is shown in Scheme 1.
22
ion reactions, Ϫ11.6 for the former and Ϫ12.9 for the latter
reactions. The result is insufficient to provide a satisfactory
interpretation of the smaller differential specific interaction
enthalpy on going from the initial to the transition states for the
AN→MeOH
present reactions, δ∆ H
by comparison to those for
t
SI
imidide ion and carboxylate ion reactions.
In order to explore solvation effects on reactions in solution,
information on molecular properties for relevant species in the
gas phase is indispensable. Semi-empirical molecular orbital
calculations were performed for nucleophilic anions as well as
Electrostatic energy, ESE for the four atom systems shown by
arrows in Scheme 1 can be expressed by eqn. (10), where Q
and
i
27
ESE = Q Q /r ϩ Σ Q Q /r
for transition-state anions with MNDO/PM3 procedures. For
the reactions of the present study, methyl iodide, the electro-
phile approaches the central carbon from the direction expected
O
HЈ O–HЈ
HЈ
H
HЈ–H
= (Q /r
) [Q ϩ Σ Q /(rHЈ–H/rO–HЈ)] (10)
O H
HЈ O–HЈ
3
for sp -hybridization at the central carbon atom (see Scheme 1).
ri–j stand for the atomic charge on atom i, and the distance
between atom i and atom j, and the suffixes O, H, and HЈ stand
for the carbonyl oxygen, the hydrogen atoms in the incoming
methyl group and the hydroxy hydrogen in methanol. The sec-
ond term in the square bracket goes to zero, when the distance
rHЈ–H becomes infinity, that is, at the initial state, and the term in
the square bracket would be taken as an effective atomic charge
on carbonyl oxygen being sensed by the hydrogen atom in
methanol. When the effects of the two hydrogen atoms nearest
to the carbonyl oxygen are taken into account, effective atomic
charges on oxygen shift to the position denoted by ( ) in
Fig. 2. The regression equation is given by eqn. (11).
Using the calculated results, various correlations were tried
empirically. One of the seemingly successful correlations is
AN→MeOH
∆tHSI
=
3
1.0 ϩ 110.7 × [Q ϩ ΣQ /(rHЈ–H/rO–HЈ)] (11)
O
H
n = 6, r = 0.94
AN→MeOH
Scheme 1
The eqn. (11) predicts that the enthalpy, ∆ H
t SI
1
184
J. Chem. Soc., Perkin Trans. 2, 1999, 1181–1186

View Article Online
Table 5 Solvents for recrystallization and elementary analysis
Obs.(%)
C
Calc.(%)
Solvents
H
N
Formula
C
H
N
Tetramethylammonium 5-methyl Meldrum’s acid anion
Tetraethylammonium Meldrum’s acid anion
Tetramethylammonium 5-phenyl Meldrum’s acid anion
Tetramethylammonium 1,3-dimethylbarbiturate
Acetone
57.2
60.4
65.3
51.8
9.01
9.88
7.94
8.32
6.14
5.03
4.83
C H NO
57.1
61.5
65.5
52.4
9.15
9.96
7.90
8.35
6.06
5.12
4.77
1
1
21
4
4
4
Acetonitrile
Acetonitrile
Acetonitrile–THF
C H NO
1
4
27
C H NO
16 23
18.1
C H N O
3
18.3
1
0
19
3
becomes zero at effective atomic charge = Ϫ0.280. The value is
very close to the charge on the carbonyl oxygen of uncharged
compounds for which the effect of hydrogen-bonding inter-
actions cannot usually be detected through enthalpy of transfer
of elementary analysis are also shown in Table 5. Other
15,17
materials were treated as described elsewhere.
Enthalpy of solution measurements
28
analysis. The slope value, 110.7 is ca. one-third of those
for imidide and carboxylate ion reactions, i.e., 369.2 for the
Enthalpies of solution, ∆ H for tetraalkylammonium salts were
s
measured at 25.0 ± 0.1 ЊC with a Tokyo Riko twin isoperibol
22
23
former and 333.8 for the latter reactions. This means that
for the reactions in the present study molecular properties such
as atomic charges on the carbonyl oxygen are not effectively
reflected in solute–solvent interactions.
1
5,16
calorimeter.
Final concentration ranges were (0.4–1.5) ×
Ϫ2
Ϫ3
1
0
mol dm and the experimental errors were ca. 0.7 kJ
mol . The enthalpies of reaction, ∆ H have also been deter-
mined with the calorimeter as described elsewhere.
∆RH’s are Ϫ101.9 kJ mol and Ϫ110.7 kJ mol for
Ϫ1
R
14
In carboxylate and imidide ions, three and two lone pair
orbitals at the reaction center are open for reaction as well as
for molecular interactions and these do not prohibit the
approach of solvent molecules to an adjacent carbonyl oxygen.
In contrast, in charge-delocalized carbanions as studied in this
work, three σ orbitals at the reaction center are used for bonding
with three atoms, only the lone pair p orbital is free for molecu-
lar interaction, and the groups coordinated to the central
carbon, i.e., hydrogen atom, methyl and phenyl groups, are
likely to hinder the approach of solvent to an adjacent carbonyl
oxygen, that is, only part of the surface of the carbonyl oxygen
remains free for solute–solvent interactions. The transition-
state anion for the charge-delocalized carbanion reaction is
coordinated with four atoms, leading to further limited surface
being open. Thus, even when the atomic charge on carbonyl
oxygen varies significantly on going from reactant to transition-
state anion as indicated by quantum-mechanical calculations
Ϫ1
Ϫ1
Meldrum’s acid and 5-methyl Meldrum’s acid anion reactions
Ϫ1
and experimental errors are ca. 5 kJ mol .
Product analysis and kinetic procedures
Stock solutions of ethyl iodide and of the relevant tetraalkyl-
ammonium salt were mixed in a round bottomed flask and kept
overnight. After near completion of the reaction, the reaction
mixtures were carefully evaporated to dryness and solid precipi-
tates were washed several times with several portions of ether.
The solvent, ether, was evaporated to near dryness and the
1
uncharged reaction product was dissolved in chloroform. H
NMR spectra of the uncharged reaction product for the
5-methyl Meldrum’s acid anion reaction agreed with that of
1
2,2-dimethyl-5-ethyl-1,3-dioxane-4,6-dione. H NMR spectra
of the uncharged reaction product of the 1,3-dimethyl-
barbituric acid anion reaction agreed with that of tautomeric
mixtures of 1,3-dimethyl-5-ethylpyrimidine-2,4,6-(1H,3H,5H)-
trione (trioxo form) and 1,3-dimethyl-5-ethylpyrimidine-2,
4(1H,3H)-dion-6-ol (dioxo-monoenol form). Thus ethylation
reactions seem to proceed at the carbon atom for both the
Meldrum’s acid anion reaction and for the 1,3-dimethyl-
barbituric acid anion reaction, as far as can be detected.
Reaction rates were calculated through the determination of
iodide ion formed by potentiometric titration using silver
(
see the abscissa of Fig. 2), the effects would not effectively be
transmitted to surrounding solvents, resulting in the rather
AN→MeOH
minor changes in ∆ H
. Thus, the rather smaller vari-
t
SI
ation in the specific interaction enthalpy on going from the
reactant to the transition-state as well as the smaller coefficient
in the correlation, eqn. (11), for the present reactions in com-
parison to those for carboxylate and imidide ion reactions
would be described as resulting from steric inhibition of sol-
vation by the coordinated atom and group at the central
carbon.
21,23
nitrate solution
and rates were measured at four of the fol-
lowing temperatures, 0.0, 20.0, 30.0, 40.0, 50.0 and 60.0 ЊC.
Experimental errors were estimated to be ca. 2%, 0.5–1.0 kJ
mol and 1.7–3.0 J K mol for rate constants, activation
enthalpies and activation entropies, respectively.
Conclusion
Ϫ1
Ϫ1
Ϫ1
In the nucleophilic substitution of ethyl iodide by the
conjugate-base anion of 5-methyl Meldrum’s acid and of 1,3-
dimethylbarbituric acid, the effects of solvent composition on
activation enthalpies are less marked by comparison to those of
imidide ion and carboxylate ion reactions. The origin of the
effects has been ascribed to the steric inhibition of solvation by
the groups coordinated to the reaction center. Even for these
more complicated systems, the specific interaction enthalpy for
nucleophilic anion, ∆ H
scale for molecular mechanistic calculations.
Calculations
Semi-empirical molecular orbital calculations were carried out
using the MNDO/PM3 method with most of the parameters
27
kept at the default values. In order to shorten the computation
time, the following restrictions were imposed; the frameworks
AN→MeOH
of the anion were assumed to have C symmetry, the methyl
s
, seems to be a very useful
t
SI
group to have C3v symmetry and in the calculation methyl
iodide was used as an electrophile instead of ethyl iodide which
was used in experiments. The structures of the transition state
anion hydrogen bonded with methanol were optimized with
C–C and C–I bond distances being fixed at the values which
had in advance been determined for uncomplexed transition
structure optimizations.
Experimental
Materials
Tetraalkylammonium salts containing the conjugate-base
anions of Meldrum’s acid and of 1,3-dimethylbarbituric acid
were prepared from tetraalkylammonium hydroxide and a rele-
vant acid in methanol as described previously and recrystal-
lized three times from the solvents shown in Table 5. The results
29
Acknowledgements
The authors thank Dr M. H. Abraham, University College
J. Chem. Soc., Perkin Trans. 2, 1999, 1181–1186
1185

View Article Online
London for his helpful comments on the original version of this
manuscript.
15 Y. Kondo, M. Ittoh and S. Kusabayashi, J. Chem. Soc., Faraday
Trans., 1982, 78, 2793.
6 Y. Kondo, T. Fujiwara, A. Hayashi and S. Kusabayashi, J. Chem.
Soc., Faraday Trans., 1989, 85, 2931.
17 Y. Kondo, T. Fujiwara, A. Hayashi, S. Kusabayashi and T. Takagi,
1
References
J. Chem. Soc., Perkin Trans. 2., 1990, 741.
1
F. G. Bordwell, H. E. Fried, D. L. Hughes, T.-Y. Lynch, A. V. Satish
and Y. E. Whang, J. Org. Chem., 1990, 55, 3330.
1
1
2
8 The value was calculated from the data given in ref. 20 and 21.
9 The value was calculated from the data given in ref. 17 and 22.
0 J. W. Larson and L. G. Hepler, in Solute–Solvent Interactions,
ed. J. F. Coetzee and C. D. Ritchie, Marcell Dekker, New York,
2
3
4
F. G. Bordwell, Acc. Chem. Res., 1988, 21, 456.
X.-M. Zhang and F. G. Bordwell, J. Org. Chem., 1994, 59, 6456.
E. M. Arnett, S. G. Maroldo, S. L. Schilling and J. A. Harrelson,
J. Am. Chem. Soc., 1984, 106, 6759.
1
969, vol. 1, ch. 2.
2
2
2
1 Y. Kondo, K. Kondo and S. Kusabayashi, J. Chem. Soc., Perkin
Trans. 2, 1993, 1141.
2 Y. Kondo, O. Nonaka, K. Iwasaki, T. Kuwamoto and T. Takagi,
J. Chem. Soc., Perkin Trans. 2, 1994, 473.
3 Y. Kondo, W. Sugitani, M. Tokui and T. Takagi, J. Chem. Soc.,
Perkin Trans. 2, 1995, 1049.
5
6
7
M. J. Bausch, C. Guadalupe-Fasano, R. Gostowski, D. Selmarten
and A. Vaughn, J. Org. Chem., 1991, 56, 5640.
F. J. Millero, J. C. Ahluwalia and L. G. Hepler, J. Chem. Eng. Data,
1
965, 10, 199.
M. J. Bausch, B. David, P. Dobrowolski, C. Guadalupe-Fasano,
R. Gostowski, D. Selmarten, V. Prasad, A. Vaughn and L.-H. Wang,
J. Org. Chem., 1991, 56, 5643.
2
2
4 J. F. J. Dippy and S. R. C. Hughes, Tetrahedron, 1963, 19, 1527.
5 I. Koppel, J. Koppel, I. Leito, V. Pihl, L. Grehn and U. Ragnarsson,
J. Chem. Res. (S), 1994, 212.
8
9
J. N. Bronsted and K. J. Pedersen, Z. Phys. Chem., 1924, 108, 185.
C. G. Swain and C. B. Scott, J. Am. Chem. Soc., 1953, 75, 141.
2
2
2
2
6 E. M. Arnett, S. Venimadhavan and K. Amarnath, J. Am. Chem.
Soc., 1992, 114, 5598.
1
1
1
0 J. O. Edwards and R. G. Pearson, J. Am. Chem. Soc., 1962, 84, 16.
1 C. D. Ritchie, Can. J. Chem., 1986, 64, 2239.
2 E. S. Lewis, T. A. Douglas and M. L. McLaughlin, in
Nucleophilicity, eds. J. M. Harris and S. P. McManus, Advances in
Chemistry Series, No. 215, American Chemical Society, Washington
DC, 1987, ch. 3 and references cited therein.
7 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 209. MOPAC93
program package, Dr J. J. P. Stewart and Fujitsu Ltd., Tokyo, Japan.
8 Y. Kondo, O. Nonaka, K. Iwasaki, T. Kuwamoto and T. Takagi,
J. Chem. Soc., Faraday Trans., 1994, 90, 121.
9 Y. Kondo, S. Kusabayashi and T. Mitsuhashi, J. Chem. Soc.,
Perkin Trans. 2, 1988, 1799.
1
3 Y. Kondo, T. Tsukamoto and C. Moriguchi, J. Chem. Soc., Perkin
Trans. 2, 1996, 1699.
1
4 Y. Kondo, T. Tsukamoto and N. Kimura, J. Chem. Soc., Perkin
Trans. 2, 1997, 1765.
Paper 9/00283I
1
186
J. Chem. Soc., Perkin Trans. 2, 1999, 1181–1186