Proton inventory study of the base-catalyzed hydrolysis of formamide. Consideration of the nucleophilic and general base mechanisms

Article abstract of DOI:10.1021/ja021055z

An NMR study of the rates of hydroxide-promoted hydrolysis of formamide in aqueous media of varying mole fraction D₂O (n) was performed at [LO^-] = 1.42 M, T = 25 °C, to shed light on whether the mechanism involves a nucleophilic attack of HO^- on the C = O or HO^- acting as a general base to remove a proton from an attacking water. The solvent deuterium kinetic isotope effect under these conditions is inverse, k_OH/k_OD = 0.77 ± 0.02 or k_OD/k_OH = 1.30 ± 0.03. Proton inventory analysis of the k_n versus n data was undertaken through NLLSQ fits to equations representing four possible mechanisms encompassing nucleophilic and general base ones with waters of solvation on the attacking hydroxide, and with or without waters of solvation on the developing amide hydrate oxyanion. Both nucleophilic and general base mechanisms can be accommodated, but there are restraints on each that are discussed. The preferred mechanism is a nucleophilic one proceeding through a transition state having two solvating waters remaining on the attacking hydroxide and three additional waters attached to the developing amide hydrate oxyanion.

Full text of DOI:10.1021/ja021055z

Published on Web 01/23/2003
Proton Inventory Study of the Base-Catalyzed Hydrolysis of
Formamide. Consideration of the Nucleophilic and General
Base Mechanisms
H. SÄ lebocka-Tilk, Alexei A. Neverov, and R. S. Brown*
Contribution from the Department of Chemistry, Queen’s UniVersity,
Kingston, Ontario, Canada, K7L 3C1
Received August 5, 2002 ; E-mail: rsbrown@chem.queensu.ca
Abstract: An NMR study of the rates of hydroxide-promoted hydrolysis of formamide in aqueous media of
varying mole fraction D2O (n) was performed at [LO-] ) 1.42 M, T ) 25 °C, to shed light on whether the
mechanism involves a nucleophilic attack of HO- on the CdO or HO- acting as a general base to remove
a proton from an attacking water. The solvent deuterium kinetic isotope effect under these conditions is
inverse, kOH/kOD ) 0.77 ( 0.02 or kOD/kOH ) 1.30 ( 0.03. Proton inventory analysis of the kn versus n data
was undertaken through NLLSQ fits to equations representing four possible mechanisms encompassing
nucleophilic and general base ones with waters of solvation on the attacking hydroxide, and with or without
waters of solvation on the developing amide hydrate oxyanion. Both nucleophilic and general base
mechanisms can be accommodated, but there are restraints on each that are discussed. The preferred
mechanism is a nucleophilic one proceeding through a transition state having two solvating waters remaining
on the attacking hydroxide and three additional waters attached to the developing amide hydrate oxyanion.
Scheme 1
Introduction
The mechanisms for acid- and base-catalyzed hydrolysis of
simple amides and esters have received much attention due to
their relevance to various biological processes, and in general
it can be said that these mechanisms are among the best
understood of any chemical process.1 The generally accepted
mechanism for the base-catalyzed hydrolysis involves the
process depicted in Scheme 1, wherein a hydroxide nucleophile
reversibly adds to the CdO to produce an anionic tetrahedral
formamide. They concluded that the 18O-isotope effects observed
for basic hydrolysis in a medium containing 18O-labeled water
are nicely accommodated by a mechanism that involves the HO-
acting as a general base (GB) to remove a proton from one of
-
intermediate (TI ). This can decompose to products via at least
four pathways involving a water-promoted spontaneous reaction,
and those promoted by acidic and basic components of buffers,
or a second hydroxide. This mechanism is supported by 18Od
C exchange experiments, which show the reversible formation
-
its waters of solvation during formation of TI . A GB mechanism
was also proposed in Marlier’s earlier report of the heavy atom
isotope effects for the alkaline hydrolysis of methyl formate4
and was supported by a more recent analysis of a proton
inventory study of the hydrolysis of ethyl acetate reported by
Mata-Segreda.5 Interestingly, a very recent heavy atom isotope
effect study of the base-catalyzed hydrolysis of the phospho-
diester, thymidine-5′-p-nitrophenyl phosphate, indicates a nu-
cleophilic role for the hydroxide and that the dkie is slightly
inverse at kOD/kOH ) 1.11, suggesting a different behavior
toward HO- from formamide and ethyl acetate.6 Nevertheless,
the GB mechanism seems attractive in that it does not require
desolvation of HO-(H2O)3 to form HO-(H2O)2 with an available
pair of electrons on the hydroxide which would be required for
direct nucleophilic attack. Such a mechanism could be a
potentially important process and certainly deserves additional
-
of TI and solvent deuterium kinetic isotope experiments (dkie)
of various amides, which indicate that kOH/kOD is generally unity
or slightly inverse, consistent with a direct nucleophilic attack
of hydroxide.1 A number of studies of ester saponification also
demonstrated inverse dkie values1d,2 in the range of kOD/kOH
)
1.3-1.4, consistent with the widely held view that DO- is a
better nucleophile in D2O than is HO- in H2O.
Marlier and co-workers3 recently reported an interesting study
of the heavy atom kinetic isotope effects for the hydrolysis of
(1) (a) Brown, R. S.; Bennet, A. J.; SÄlebocka-Tilk, H. Acc. Chem. Res. 1992,
25, 481. (b) Bennet, A. J.; Brown, R. S. Physical Organic Chemistry of
Acyl Transfer Reactions. In ComprehensiVe Biological Catalysis:
A
Mechanistic Reference; Sinnott, M., Ed.; Academic Press: New York, 1997;
Vol. 1, pp 293-326. (c) Brown, R. S. Studies of the Hydrolysis of Amides.
In The Amide Linkage: Selected Structural Aspects in Chemistry, Bio-
chemistry, and Materials Science; Greenberg, A., Brenneman, C. A.,
Liebman, J. F., Eds.; Wiley-Interscience: New York, 1999; pp 85-114.
(d) Kirby, A. J. In ComprehensiVe Chemical Kinetics; Bamford, C. H.,
Tipper, C. F. H., Eds.; Elsevier: Amsterdam, 1972; Vol. 10, pp 57-207.
(2) (a) Wynne-Jones, W. F. K. Chem. ReV. 1935, 17, 115. (b) Reitz, O. Z.
Phys. Chem. 1936, 177, 85. (c) Jencks, W. P.; Carriuolo, J. J. Am. Chem.
Soc. 1960, 82, 675.
(3) Marlier, J. F.; Dopke, N. C.; Johnstone, K. R.; Wirdzig, T. J. J. Am. Chem.
Soc. 1999, 121, 4356.
(4) Marlier, J. F. J. Am. Chem. Soc. 1993, 115, 5953.
(5) Mata-Segreda, J. F. J. Am. Chem. Soc. 2002, 124, 2259.
(6) Cassano, A. G.; Anderson, V. E.; Harris, M. E. J. Am. Chem. Soc. 2002,
124, 10964.
9
10.1021/ja021055z CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 1851-1858
1851

A R T I C L E S
SÄlebocka-Tilk et al.
Table 1. Second-Order Rate Constants (kn) for Base-Catalyzed
Hydrolysis of Formamide in Aqueous Media of Different Mole
Fraction (n) D, T ) 25 °C, [LO-] ) 1.42 M
investigation because it is phenomenologically quite different
from the normal direct nucleophilic attack mechanism.
Recently we reported some studies on the hydrolysis of
formamide wherein we determined the activation parameters
for the acid and base processes7 as well as the experimental
values for the activation parameters and rate constant for the
water reaction (kw ) 1.1 × 10-10 s-1 (t1/2 ) 199 years) at 25
°C), which was previously reported as being important at 80
°C by Hine and co-workers.8 As part of that study,7 we also
determined an inverse dkie for the hydroxide reaction, kOH/kOD
) 0.77 ( 0.06 at [OL-] ) 1.47 M, the inverse value being
consistent with what is observed for other simple amides and
esters.1 Thus, our preferred direct nucleophilic mechanism for
the base-catalyzed hydrolysis of formamide seems at variance
with the Marlier3,4 and Mata-Segreda5 analyses and suggested
to us that further studies with formamide were required to
resolve this important dilemma.
n
103 × kn (M-1 s-1 a
)
n
103 × kn (M-1 s-1 a
)
0.03
0.103
0.21
0.31
0.41
0.465
0.52
0.52
3.17 ( 0.05
3.18 ( 0.04
3.22 ( 0.04
3.48 ( 0.04
3.54 ( 0.03
3.57 ( 0.03
3.55 ( 0.04
3.67 ( 0.05
0.62
0.72
0.83
0.93
1.0
3.68 ( 0.05
3.85 ( 0.03
4.00 ( 0.16
3.99 ( 0.04
4.13 ( 0.03
(3.07 ( 0.1)
(4.03 ( 0.1)
0.03b
1.0b
a Errors determined from the standard deviations of the linear regressions
of the plots of the integrated intensities of the H-C(dO)-X NMR versus
time data (see Experimental Section). b Data at 27 °C as determined from
ref 7.
(b) Kinetic Experiments. The rates of hydrolysis of formamide were
determined by 1H NMR analyses at 25 ( 0.2 °C using a Bruker DMX-
Avance 500 spectrometer equipped with a broadband inverse probe.
The 1H NMR spectra in water were accumulated using a standard
presaturation water suppression technique. Solutions of base (1.47 M)
in either H2O or D2O were prepared under CO2-free conditions and
stored under Ar. Base concentrations were determined by titration with
standardized 1.0 and 0.5 N HCl, phenolphthalein indicator. In pure
water, the formamide H-CdO signal at δ 7.55 appears as a doublet
(J ) 14.9 Hz) coupled to one of the NH protons, while the two NH
protons appear as two broad triplets coupled to 14N. The one coupled
to the formamide proton appears at δ 7.17 (JN-H ≈ 60 Hz), while the
other NH appears as a sharper triplet at δ 7.55 (JN-H ≈ 62 Hz).
The main difference between the nucleophilic and general
base-catalyzed processes stems from consideration of the
respective simplified transition states (TS 1 and 2) shown in
eqs 1 and 2 (L ) H, D). In either mechanism, the reactant state
Solvent deuterium kinetic isotope experiments were undertaken in
D2O/H2O mixtures, the mole fraction D varying from ∼0 to 1. The
solutions for individual runs were prepared by adding precise volumes
of base ([NaOL]i ) 1.47 M, total volume 0.60 mL) and were
thermostated in the spectrometer probe at 298 K for ∼10 min after
which 20 µL of a 10-2 M formamide/D2O solution was added,
[NaOL]final ) 1.42 M. The pseudo-first-order rate constants (kobs) for
formamide hydrolysis were obtained by observing the rate of increase
of the intensity of the signal at δ 8.46 attributable to the OdC-H proton
of the formic acid formed (A) and the decrease of the intensity of the
signal at δ 7.55 for that of formamide (B). NMR data were acquired
continuously, and every 16 scans were summed separately (the time
being recorded as the midpoint of the number of scans utilized). This
process was repeated up to at least two hydrolysis half times. Pseudo-
first-order rate constants (kobs) were evaluated from the slopes of the
ln(A/(A + B)) versus time plots with the errors in kobs being determined
as the standard deviation of the linear regression lines. Between 15
and 20 experimental points were used for each plot.
is the same, a hydroxide with three solvating waters plus
formamide. In eq 1, the TS incorporates a hydroxide having
one of its solvating waters removed to expose the nucleophile
lone pair, while the TS shown in eq 2 incorporates a proton in
flight between hydroxide and the attacking water. The bold L
protons are those undergoing changes in their bonding. Each
process explicitly shows the solvating waters of the hydroxide,
but for the moment ignores the solvation of the developing (-)
on the carbonyl oxygen which, in either mechanism, should be
the same. The proton in flight in TS 2 should be subject to a
primary dkie, but there is no equivalent in TS 1, so it seems
likely that these two processes should be distinguishable on the
basis of detailed dkie experiments, a distinction that was noted
by Mata-Segreda for his proton inventory studies with ethyl
acetate.5 In an attempt to determine which of these two
mechanisms applies to formamide, we have conducted proton
inventory analyses of a series of kinetic experiments on its base-
catalyzed hydrolysis in media of varying mole fraction of D2O.
The following describes our findings.
Results and Discussion
Shown in Table 1 are the second-order rate constants for
lyoxide-catalyzed hydrolysis of formamide at 25 ( 0.2 °C (kn
) kobs/[LO-], [LO-] is 1.42 M, corrected for the additional 20
µL of D2O in which the formamide is added) at different mole
(n) fractions D. Also included in the table are two previous
values for the hydrolysis in pure H2O and D2O at 27 °C reported
in our preliminary work.7 Error limits presented in the table
are derived from the standard deviations of the linear regressions
of the ln(A/(A + B)) versus time plots to determine kobs, where
A and B, respectively, are the integrated NMR intensities of
the formate and formamide H-C(dO)- protons. The 25 °C
data are graphically presented in Figure 1, which, upon cursory
inspection, form a rather featureless straight line relationship
without a definite curvature upward or downward. The best-fit
linear regression has a slope of (1.02 ( 0.05) × 10-3, r2 )
Experimental Section
(a) Materials. H2O was made free from dissolved CO2 and stored
under Ar. D2O (CDN Isotopes, 99.9 atom % D) was used as supplied,
as was formamide (99.5+%, A.C.S. reagent grade, Aldrich).
(7) SÄlebocka-Tilk, H.; Sauriol, F.; Monette, M.; Brown, R. S. Can. J. Chem.
2002, 80, 1343.
(8) Hine, J.; King, R. S.-M.; Midden, R.; Sinha, A. J. Org. Chem. 1981, 46,
3186.
9
1852 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

Study of the Base-Catalyzed Hydrolysis of Formamide
A R T I C L E S
be taken as the solvent dkie largely on k1. We note that these
conditions are substantially the same as those used previously
for determination of the solvent oxygen nucleophile isotope
effect (1.7 M) where “the attack of nucleophile is largely rate-
determining”.3
Proton Inventory Analysis. The proton inventory technique12
provides information about the number of protons undergoing
a significant change in bonding in the transition state (TS)
relative to the ground state. Equation 5, sometimes referred to
as the Gross-Butler equation, expresses the relationship
between the rate constant observed in mixtures of L2O (L ) H,
D) with known isotopic composition and the fractionation factors
(φ) for the exchangeable protons; n is the atom fraction of D in
the medium, while i and j represent the contributing hydrogens
in the transition and reactant states.
Figure 1. Plot of the second-order rate constants for base-catalyzed
hydrolysis of formamide as a function of mole fraction of D2O (n). Dashed
line, NLLSQ fit of the data to eq 15; solid line, fit of the data to eq 16.
Best fit parameters are given in Table 3.
TS
RS
0.9724. Also included in the figure are two lines relating to
NLLSQ9 fits to eqs 15 and 16 described below.
k ) k
(1 - n + nΦ )/
(1 - n + nΦ )
(5)
∏
∏
n
o
i
j
i
J
Rate Constants at n ≈ 0 and 1.0. Marlier3 and, earlier,
Kirsch10 have shown that the rate law for alkaline hydrolysis
of formamide contains both first- and second-order terms in
hydroxide, which is analyzed in terms of the mechanism
presented in eq 3 for which steady-state analysis gives the
expression in eq 4. Kirsch’s analysis10 of the secondary DKIE
The fractionation factors for hydrogens refer to the tightness of
their bonding and are significantly less than unity for H’s being
transferred or “in flight” between O and N, or O and O, as part
of the rate-limiting step. In these cases, normal primary dkie’s
of kH/kD > 1 are expected unless other compensating factors,
such as changes in solvation, are at play. In hydrogen-bonding
situations where the overall bonding is loose, the fractionation
factors are less than unity, also giving rise to normal dkie’s of
kH/kD > 1. The latter contribute secondary effects of solvation
and can significantly alter the overall dkie.13 Generally speaking,
plots of kn versus n have one of three shapes: bowed upward,
linear, and bowed downward. The latter two scenarios are
analyzed in terms of a single proton, or two or more protons
undergoing significant changes in bonding in passing from
reactant to transition state.12
for hydroxide attack on HC(O)NH2 and DC(O)NH2 further
indicates that the transition state for the addition of hydroxide
must be very late.
k1[HO-](k3 + k4[HO-])
In analyzing the reactant state, we employ procedures we used
before,1 which follow the suggestions of Gold and Grist14 for
hydroxide solvated by three waters (3). All conceivable mech-
khyd
)
(4)
k2 + k3 + k4[HO-]
The mechanism shown in eq 3 is a simplified version of the
general one for amide hydrolysis presented in Scheme 1, but
without buffer components. Graphical analysis3,10 of the parti-
-
tioning of the tetrahedral intermediate (TI , eq 3) indicates that
the k3/k2 ratio is 1.05, while the k4/k2 ratio is 2.15 M-1. At low
[HO-] (<0.1 M), the pathway second order in [HO-] is
relatively unimportant, and TI undergoes substantial reversal,
-
anisms must employ a common reactant state of LaO-(Lb-
OLc)3, where La, Lb, and Lc have assumed φ values of 1.22,
0.7, and 1.0, respectively. We consider the two possible
mechanisms shown in eqs 1 and 2 for which the nucleophilic
and general base modes of action for hydroxide, in fact, have
been considered earlier by Gold and Grist14 in analyzing various
hydroxide destroying reactions. Cursory analysis of the overall
inverse dkie for the hydrolysis indicates that at n ) 1.0:
so khyd is approximated as k1[HO-]k3/(k2 + k3). However, at
high [OH-], the tetrahedral intermediate is driven forward via
rapid capture by the second hydroxide, so reversal becomes less
important, and attack ultimately becomes the rate-limiting step
with khyd being approximated as k1[HO-].11 Under the present
conditions at [LO-] ) 1.42 M, there is a significant inverse
kinetic isotope effect, kOH/kOD ) 0.77 ( 0.02, which can thus
(9) Nonlinear least-squares fitting was done using GraphPad Prism Version
2.01, GraphPad Software Inc.
(10) Kirsch, J. F. In Isotope Effects on Enzyme Catalyzed Reactions; Cleland,
W. W., O’Leary, M. H., Northrup, D. B., Eds.; University Park Press:
Baltimore, 1977; pp 100-121.
TS
TS
kOD
kOH
)
φ /(1.22)(0.7)3 )
φ /0.42
(6)
∏
∏
i
i
i
i
(11) This analysis is identical to that proposed in ref 3 and is substantially correct.
However, at [HO-] ) 1.42 M, the partitioning of the tetrahedral intermedi-
ate is ∼4.1 in favor of product formation relative to reversal. Of that ratio,
some 3.1 parts are attributed to the second-order pathway (k4), and 1 part
is attributed to the spontaneous pathway, k3. Strictly speaking, while the
dkie is substantially attributable to the attack step, k1, there is a small
additional component attributable to k4 which should be inverse given OD-
in D2O is a stronger base than is HO- in H2O.
Because the observed kOD/kOH ) 1.30, the product of the
fractionation factors in any appropriate TS must be ∼0.56. As
a starting point for the analysis of the TS fractionation factors
for the nucleophilic process, we assume that the initially
desolvated hydroxide (H-O-(H-OH)2) has fractionation factors
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1853

A R T I C L E S
SÄlebocka-Tilk et al.
Table 2. NLLSQ Generated Parameters Obtained from Fitting
Table 1 Data to Eqs 7-10a
of 1.22 and 0.7 for the lyoxide and solvating waters, respec-
tively. In principle, if removal of the single water of hydration
had led to stronger H-bonding to HO-, the fractionation factors
of the two remaining waters of solvation could have values
slightly lower than 0.7, while the value for H-O- could be
larger than 1.22.15 However, as discussed by Huskey and
Schowen in their analysis of the proton inventory for methoxide
attack on phenyl acetate,16 it seems likely that incipient
association of the oxyanion with the polarized >CdO can
partially offset the effect of the removal of the solvating water
in exposing the lone pair for attack so that the fractionation
factors of H-O-(H-OH)2 should not change as much as if
there were no electrophilic stabilization at all. For the intermedi-
ate amide hydrate oxyanion, HC(O-)(OL)(NL2), we assume the
L-O fractionation factor is 1.0 as is the case for a gem diol or
hemiacetal,17 the fractionation factor for the N-L is 1.0, and
those for the three waters of solvation of the oxyanion (if
present) are 0.7. Chemical intuition assists in assessing accept-
able computed values for the various TS fractionation factors
because appropriate mechanisms require that the La hydroxide
proton and the remaining Lb H-bonding protons should assume
respective φTS values somewhere between those of the reactant
and intermediate states, for example, 1.22 f 1.0 and 0.7 f
1.0, during their transitions to the intermediate HC(O-)(OL)(NL2)
and L2O, respectively. For the proton in flight in the GB
mechanism, a fractionation factor in the range of 0.5 or less5,12
is anticipated, suggesting it contributes a primary isotope effect
of kH/kD g 2. Inclusion of the 0.5 number in the numerator of
eq 6 requires that the product of all remaining φTS values be
∼1.1, which places an important limitation on this mechanism
because any remaining or additionally acquired waters of
solvation must have φTS values intermediate between 0.7 and
1.0. Clearly, values of the φTS for the proton in flight <0.5
require the product of the remaining fractionation factors to be
>1.1, creating an even larger constraint on the GB mechanism.
We consider below four mechanistic possibilities. Case a
refers to the minimal nucleophilic mechanism of eq 1, while
case b refers to the minimal GB mechanism of eq 2, the
appropriate formulas for transition states (1 and 2) being shown
in each equation.
parameter
eq
103 ko
φ1
φ2
φ3
φ4
7
8
9
3.10 ( 0.04 1.35 ( 0.10 0.64 ( 0.02
3.10 ( 0.04 1.12 ( 2.72 0.99 ( 1.4 0.50 ( 0.21
3.10 ( 0.04 1.20 ( 21.6 1.18 ( 10.8
0.69 ( 0.14
10 3.10 ( 0.04 1.14 ( 73
0.97 ( 132 0.55 ( 1.5 1.08 ( 112
a Unconstrained fittings.9
the GB mechanism with additional solvating waters. The proton
inventory analyses for these are given in eqs 9 and 10,
respectively, corresponding to transition states 4 and 5. In those
equations, the fractionation factors for the H-bonding waters
of solvation on the developing alkoxide are given as φ4.
kn ) ko(1 - n + φ1n)(1 - n + φ2n)2(1 - n + φ4n)3/
(1 - n + 1.22n)(1 - n + 0.7n)3 (9)
kn ) ko(1 - n + φ1n)(1 - n + φ2n)2(1 - n + φ3n)(1 - n +
φ4n)3/(1 - n + 1.22n)(1 - n + 0.7n)3 (10)
Unrestricted NLLSQ fitting of the data in Table 1 to each
equation generates the various parameters listed in Table 2, from
which it can be concluded that each has some deficiency. In
case a, the computed fractionation factors of φ1 ) 1.35 ( 0.10
and φ2 ) 0.64 ( 0.02 seem, respectively, too high and too low
to support the simple nucleophilic mechanism of eq 1, and in
the most generous analysis these would imply an extremely early
TS. For the GB mechanism in case b, the computed φ1 ) 1.12
( 2.72 and φ3 ) 0.50 ( 0.21 are appropriate for the lyoxide
L-O and the proton in flight, but the φ2 of 0.99 ( 1.4 implies
that the solvating waters are essentially lost in the TS.
Additionally, the large uncertainty in all of the computed values
suggests that these are heavily correlated. If we accept, for the
moment, that the solvating waters have fixed φ2 values of unity,
implying complete release in the TS, the mechanism is identical
to what Mata-Segreda assumed for the saponification of ethyl
acetate.5 While the φ1 and φ3 values are well within what one
anticipates for a good fit to the general base mechanism, we
see no good reason to expect that the attacking hydroxide would
completely shed all of its solvating waters in what must surely
be a very endothermic process (vide infra).
kn ) ko(1 - n + φ1n)(1 - n + φ2n)2/
(1 - n + 1.22n)(1 - n + 0.7n)3 (7)
kn ) ko (1 - n + φ1n)(1 - n + φ2n)2(1 - n + φ3n)/
(1 - n + 1.22n)(1 - n + 0.7n)3 (8)
In the corresponding proton inventory equations, eqs 7 and 8,
φ1 and φ2 refer to the lyoxide L-O- and its two H-bonding
waters, while φ3 in eq 8 refers to the proton in flight.
Case c refers to the nucleophilic mechanism with three
solvating waters on the developing alkoxide, while case d is
Our recent study7 provided the activation parameters ∆Hq
(17.9 ( 0.2) kcal/mol and ∆Sq (-11.1 ( 0.5) cal/mol K for
the base-promoted hydrolysis of formamide. The entropy term
implies some restriction of waters of solvation in the transition
state relative to the ground state.18 A more complete proton
inventory analysis therefore requires consideration of additional
(12) (a) Schowen, R. L. In Isotope Effects on Enzyme-Catalyzed Reactions;
Cleland, W. W., O’Leary, M. H., Northrop, D. B., Eds.; University Park
Press: Baltimore, 1977; pp 64-99. (b) Alvarez, F. J.; Schowen, R. L. In
Isotopes in Organic Chemistry; Buncel, E., Lee, C. C., Eds.; Elsevier:
Amsterdam, 1987; Vol. 7, pp 1-60. (c) Kresge, A. J.; More-O’Ferrall, R.
A.; Powell, M. F. In Isotopes in Organic Chemistry; Buncel, E., Lee, C.
C., Eds.; Elsevier: Amsterdam, 1987; Vol. 7, pp 177-274.
(13) Kresge, A. J. J. Am. Chem. Soc. 1973, 95, 3065.
(14) Gold, V.; Grist, S. J. Chem. Soc., Perkin Trans. 2 1972, 89.
(15) Kresge, A. J.; More-O’Ferrall, R. A.; Powell, M. F. Isot. Org. Chem. 1987,
7, 61.
(17) Bone, R.; Wolfenden, R. J. Am. Chem. Soc. 1985, 107, 4772. This work
specifically refers to the fractionation factor for (R2C(OH)OR) (φH ) 1.0),
which we take as a reasonable approximation for the fractionation factor
in the amide hydrate anion.
(16) Huskey, W. P.; Schowen, R. L. Gazz. Chim. Ital. 1987, 117, 409.
9
1854 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

Study of the Base-Catalyzed Hydrolysis of Formamide
A R T I C L E S
solvating waters on the developing alkoxide oxygen in the
transition states shown in 4 and 5 for the nucleophilic and GB
mechanisms. Inclusion of the additional solvating waters seems
appropriate from a chemical standpoint, but does not lead to
more satisfactory results because all of the computed φ values
in Table 2 for cases c and d are heavily correlated with large
uncertainties and are therefore meaningless in the present form.
Reduction of the Number of Variables in Fitting. While
transition states 4 and 5 contain a more complete account of
possible exchangeable hydrogens, the increased number of
parameters introduces large errors in the fits and therefore limits
our ability to make meaningful conclusions. It is generally
considered that the magnitudes of the various fractionation
factors are related in some way to progress along the reaction
coordinate.5,12,19 In previous works concerning the acid-20 and
base-catalyzed hydrolyses of amides21 and the base-catalyzed
hydrolysis of esters,22 we have considered that solvent dkie data
can be treated in terms of fractionation factors derived on the
basis of a percent of progress along the reaction coordinate.
Such an approach reduces the number of independent variables
in transition states 4 and 5 because if all exchangeable hydrogens
respond to approximately the same extent in passing from
reactants to transition state, the main variable becomes a
parameter related to the extent of progress along the reaction
coordinate. Schowen19c has presented a treatment based on a
free energy relationship where some weighting factor, x,
describes the structure of the transition state in terms of its
progress from reactants (x ) 0) to product or intermediate (x )
1). If some measure of the transition state structure in the vicinity
of the isotopic site is available, and if the fractionation factors
for the ground (φgs) and intermediate (φint) states are known,
φTS for a given H can be calculated as
reactants and intermediates. The φ3 for the proton in flight
cannot be derived in such a way and is treated as an independent
parameter. In applying eq 12 to the hydrolytic mechanisms
proceeding through TS 1, 2, 4, and 5, we assume that the
ground-state fractionation factors (φgs) are those presented by
Gold and Grist14 and that the intermediate is an amide hydrate
alkoxide with waters of solvation having φint values similar to
those of hydroxide (φ ) 0.7). Further, we assume that, with
the exception of the proton in flight, the various φTS values for
protons undergoing changes in bonding in the TS respond in a
similar way to progress along the reaction coordinate. This
assumption is clearly an oversimplification because in the rather
extended transition states shown in 4 or 5, the degrees of change
in bonding and solvation at the attacking site and remote
solvating site are probably not the same.23 Nevertheless, this
treatment, for the moment, has overall merit because any
acceptable mechanism must have positive x values, meaning
that the various contributing φTS’s will not be random values
produced to satisfy solely mathematical fitting criteria, but rather
have chemical meaning because acceptable values must be
between reactant and intermediate states. Conversely, negative
x values, even if these satisfy the mathematical fitting criteria,
cannot be chemically acceptable because the computed φTS will
not be between reactant and intermediate states. In what follows,
we will start from the simplest assumption that all exchangeable
protons respond to the progress along the reaction coordinate
according to x, and we will subsequently consider other
possibilities.
Cases a′ and b′. Minimal Nucleophilic and General Base
Mechanisms. From the relationship given in eq 12, the various
φTS values in eqs 7 and 8 are recast in eqs 13 and 14 as
kn ) ko(1 - n + 1.22(1-x)n)(1 - n + 0.7(1-x)n)2/
(1 - n + 1.22n)(1 - n + 0.7n)3 (13)
(1-x)
x
φTS ) φgs
φ
(11)
int
kn ) ko(1 - n + 1.22(1-x)n)(1 - n + 0.7(1-x)n)2(1 - n +
φ3n)/(1 - n + 1.22n)(1 - n + 0.7n)3 (14)
By combining eq 11 with eq 5, one obtains an expression:
TS
(1-x)
gs
k ) k
(1 - n + n(φ
φ
x))i(1 - n + nφ3)/
int
∏
n
o
i
NLLSQ fitting of the Table 1 data to these gives the parameters
listed in Table 3. The negative computed value for x in both
cases is unacceptable because it requires that the φTS’s for the
lyoxide protons are unrealistically high (φ1 ) 1.26; 1.29), while
the solvating ones are unrealistically low (φ2 ) 0.66; 0.69).24
Further, for the general base mechanism, the computed value
for φ3 is 1.07, an unrealistic value for a proton in flight.
Interestingly, NLLSQ fitting of the GB mechanism to eq 14
gives two minima as shown in Table 3,25 each being statistically
RS
(1 - n + nφ ) (12)
∏
j
j
in which the fractionation factors for all exchangeable protons
in the TS can be described in terms of known ones for the
(18) The ∆Sq of (-11.1 ( 0.5) cal/mol K looks to be rather low for a simple
process involving two species going to one in the transition state. However,
if the nucleophilic process is correct and one of the waters of solvation is
released in reaching the transition state, one would expect that the overall
translational entropy change should be close to zero in the absence of
additional solvation. The fact that there is a small negative observed value
is consistent with the developing solvation on the alkoxy C-O-. If the
general base mechanism applies, having all three solvating waters released
at the TS, then the ∆Sq is predicted to be substantially positive contrary to
the observed value.
(19) See, for example: (a) Hogg, J. L.; Phillips, M. K. Tetrahedron Lett. 1977,
3011. (b) Kershner, L. D.; Schowen, R. L. J. Am. Chem. Soc. 1971, 93,
2014. (c) Schowen, R. L. Prog. Phys. Org. Chem. 1972, 9, 275.
(20) (a) Bennet, A. J.; SÄlebocka-Tilk, H.; Brown, R. S.; Guthrie, J. P.; Jodhan,
A. J. Am. Chem. Soc. 1990, 112, 8497. (b) Bennet, A. J.; SÄlebocka-Tilk,
H.; Brown, R. S.; Guthrie, J. P.; Jodhan, A. J. Am. Chem. Soc. 1990, 112,
8497.
(22) (a) Kellogg, B. A.; Brown, R. S.; MacDonald, R. S. J. Org. Chem. 1994,
59, 4652. (b) Kellogg, B. E.; Tse, J. E.; Brown, R. S. J. Am. Chem. Soc.
1995, 117, 1731.
(23) Huskey and Schowen16 have demonstrated a substantial imbalance between
solvent reorganization and heavy atom reorganization for the attack of
methoxide on phenyl acetate. The methoxide:>CdO attack has progressed
to the extent of about 15%, while the two residual methanols of solvation
have reorganized to the extent of 55-60%. Their analysis does not
incorporate any additional solvation of the developing (-)-charge on the
carbonyl, which may be slight given the apparent early nature of the
methoxide- -CdO bonding.
(24) Even if one assumes for the nucleophilic mechanism that the partially
desolvated hydroxide (HO-(H2O)2) has fractionation factors slightly
different from those in the ground state, the computed φTS implies a very
early transition state which is unreasonable for hydroxide attack on
formamide.10 However, for the general base mechanism, where no partially
desolvated hydroxide is required, the situation is clearer because any
acceptable x value must be positive.
(21) (a) SÄlebocka-Tilk, H.; Bennet, A. J.; Keillor, J. W.; Brown, R. S.; Guthrie,
J. P.; Jodhan, A. J. Am. Chem. Soc. 1990, 112, 8507-8514. (b) SÄlebocka-
Tilk, H.; Bennet, A. J.; Hogg, H. J.; Brown, R. S. J. Am. Chem. Soc. 1991,
113, 1288. (c) Brown, R. S.; Bennet, A. J.; SÄlebocka-Tilk, H.; Jodhan, A.
J. Am. Chem. Soc. 1992, 114, 3092.
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1855

A R T I C L E S
SÄlebocka-Tilk et al.
Table 3. NLLSQ Generated Parameters Obtained by Fitting the
Table 1 Data to Eqs 12-15a
Table 4. NLLSQ Fit Values of the Table 1 Data to Nucleophilic
Mechanism According to Eq 17
parameter
103 ko
x
φ1a
φ2a
φ4b
r2
eq
103 ko
x
φ1b
φ2b
φ3c
φ4b
3.23 ( 0.04
3.23 ( 0.04
3.23 ( 0.04
3.23 ( 0.04
3.22 ( 0.04
0.9
0.8
0.7
0.6
0.5
1.02
1.04
1.06
1.09
1.10
0.96
0.93
0.90
0.87
0.84
0.83
0.85
0.86
0.88
0.89
0.9330
0.9315
0.9312
0.9324
0.9352
13 3.14 ( 0.02 -0.15 ( 0.01
14d 3.10 ( 0.04 -0.28 ( 0.09
1.26 0.66
1.29 0.64 1.07 ( 0.04
0.96 1.07 0.50 ( 0.03
3.20 ( 0.04
1.19 ( 0.13
0.141 ( 0.001 1.19 0.74
15 3.17 ( 0.01
0.95
16d 3.10 ( 0.04 -0.22 ( 0.09
1.27 0.65 0.82 ( 0.04 1.08
1.02 0.96 1.53 ( 0.09 .73
a Calculated as φ1 ) (1.22)(1-x), φ2 ) (0.7)(1-x)
. )
b Calculated as φ4
3.10 ( 0.04
0.90 ( 0.09
(0.7)y.
a Unrestricted fits. b Calculated from φ1 ) (1.22)(1-x), φ2 ) (0.7)(1-x)
,
Table 5. NLLSQ Fits of Table 1 Data to General Base
Mechanism According to Eq 18
and φ4 ) (0.7)x. c Computed as independent parameters in fits. d Fitting
provides two local minima.
103 ko
x
φ1a
φ2a
φ3c
φ4b
r2
3.11 ( 0.03
3.10 ( 0.03
3.10 ( 0.03
3.09 ( 0.03
3.08 ( 0.03
0.9
0.8
0.7
0.6
0.5
1.02
1.04
1.06
1.09
1.10
0.96
0.93
0.90
0.87
0.84
0.5
0.5
0.5
0.5
0.5
1.05
1.07
1.09
1.10
1.12
0.9728
0.9727
0.9723
0.9714
0.9696
equivalent, but the second computed minimum is also unsatis-
factory because the x value of 1.19 is greater than the theoretical
maximum of unity.
Cases c′ and d′. Nucleophilic and General Base Mecha-
nisms with Solvating Waters on Developing Alkoxide.
Equations 9 and 10 are recast as eqs 15 and 16, which yield,
after NLLSQ fitting, the various values given in Table 3:
a Calculated as φ1 ) (1.22)(1-x), φ2 ) (0.7)(1-x) b Calculated as φ4
. )
(0.7)y. c Set value at 0.5 for a primary dkie of 2.0 for the proton in flight;
values lower than 0.5 do not lead to acceptable fits, see text.
coordinate.16,26 We concur with Gold and Grist14 who state that
“it seems doubtful...that the kinetic solvent isotope effects for
hydroxide destroying reactions can be formulated on the basis
of a transition state which is simply characterized by partial
progress from reactants to products in a single process.” This
suggests that refinement of the mechanism might be ac-
complished by uncoupling the attack of hydroxide from the
resolvation of the developing oxyanion with a late transition
state somewhere between 50 and 90% along the reaction
coordinate as suggested by Kirsch.10 Thus, in transition states
4 and 5, above, the fractionation factors associated with the L1
and L2 protons are assumed to be correlated to the same extent
with a fixed progress along the reaction coordinate (x), while
the L4 solvating protons are treated as an independent variable
(y). The appropriate equations for the nucleophilic and general
base mechanisms are given in eqs 17 and 18. Given in Tables
4 and 5 are the best fit ko and φ4 values for x, five assumed
values corresponding to progress of the TS for the HO- attack
of 50-90%.
kn ) ko(1 - n + 1.22(1-x)n)(1 - n + 0.7(1-x)n)2(1 - n +
0.7xn)3/(1 - n + 1.22n)(1 - n + 0.7n)3 (15)
kn ) ko(1 - n + 1.22(1-x)n)(1 - n + 0.7(1-x)n)2(1 - n +
φ3n)(1 - n + 0.7xn)3/(1 - n + 1.22n)(1 - n + 0.7n)3 (16)
The two fits are shown as the dashed and solid lines in Figure
1, which fit the data acceptably and with nearly identical
statistics, but only the fit to eq 15 gives chemically acceptable
fractionation factors between the reactants and intermediate state
values. The negative computed value of x found for the first
local minimum for the fit to eq 16 is unrealistic, and the
computed fractionation factor for the proton in flight is high,
that is, ∼0.82. This fit too has a second local minimum as shown
in Table 3, and although the x-value is well within acceptable
values for a late TS, the proton in flight has a computed
unrealistic fractionation factor of 1.53.
Refining the Mechanism by Uncoupling Attack and
Resolvation. Only the nucleophilic mechanism (case c′ above)
gives an acceptable fit to the data under the assumed requirement
that all components of the nucleophilic reaction progress to the
same extent in the transition state. However, the computed value
of x for the preferred mechanism (eq 15, x ) 0.14) is
suspiciously low and if interpreted literally would imply a
remarkably early transition state, contradicting Kirsch’s results10
that indicate that hydroxide attack on formamide has a very
late transition state. Various studies have suggested that the
solvent isotope effects for hydroxide and water attack on
carbonyls are correlated with the bond order of the nucleo-
phile- - - -CdO bond19a and Brønsted parameters describing the
sensitivity of the nucleophilic addition to structural variations.12,19c
The latter parameters are commonly interpreted as indices of
progress of bond changes for the heavy atoms at the transition
state. Numerous cases are known where there are imbalances
of the progress of reacting components along the reaction
kn ) ko(1 - n + 1.22(1-x)n)(1 - n + 0.7(1-x)n)2(1 - n +
0.7yn)3/(1 - n + 1.22n)(1 - n + 0.7n)3 (17)
kn ) ko(1 - n + 1.22(1-x)n)(1 - n + 0.7(1-x)n)2(1 - n +
φ3n)(1 - n + 0.7yn)3/(1 - n + 1.22n)(1 - n + 0.7n)3 (18)
Nucleophilic Mechanism with Solvation. Inspection of the
data in Table 4 indicates that fits with very nearly the same
correlation coefficients can be obtained when x varies between
0.5 and 0.9. As expected, the latest TS requires more resolvation
of the developing C-O- charge than the earlier ones. When x
is set at 0.9, φ4 assumes a value of 0.83, implying that the three
resolvating waters have reorganized to the extent of ∼50%, thus
lagging behind the attack of HO- and loosening its two attendant
waters of solvation. We tried to fit the data to a model where
φ1 has assumed values corresponding to 0.5 e x e 0.9, while
φ2 and φ4 are treated as variables, but φ2 in that treatment always
assumes unacceptably high values of ∼1.06, while φ4 adopts
values of essentially unity (from 1.02 to 1.0, respectively). We
conclude that the nucleophilic mechanism can be fit only if the
(25) For each fitting to eqs 13-16, the NLLSQ procedure involved selecting
different initial estimates of the variables to probe for additional local
minima. Only in the GB cases were two minima found.
(26) Bernasconi, C. F. AdV. Phys. Org. Chem. 1992, 27, 119.
9
1856 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003

Study of the Base-Catalyzed Hydrolysis of Formamide
A R T I C L E S
L1 and L2 protons (TS 4) are assumed to vary to the same extent
with x with the L4 protons lagging behind but nevertheless
producing an essential resolvation of the developing oxyanion.
General Base Mechanism with Solvation. We have assumed
a value of 0.5 for φ3 of the proton in flight for all of our analyses
of the GB mechanism. This is in accordance with the generally
accepted notion that protons in flight should contribute primary
kinetic isotope effects of 2 or larger and because we find that
any smaller value for φ3 leads to fits which are markedly bowed
upward with poor statistics. The data presented in Table 5 clearly
indicate that with assumed values of 0.5 e x e 0.9 fixed for
the attacking hydroxide and its solvating waters, the value of
φ4 for the three resolvating waters varies between 1.12 and 1.07,
values which are too high to be acceptable. We have also tried
to fit the data to a GB model, where φ3 ) 0.5, and only φ1 has
assumed values corresponding to 0.5 e x e 0.9, while φ2 and
φ4 are treated as variables. Interestingly, φ2 in that treatment
assumes high values of 1.0-1.15, while φ4 adopts values of
∼1.0, and when x is 0.5-0.55, the φ2,4 values are both ∼1.0.
This circumstance is essentially identical to the results of the
unrestricted fit for the GB process in case b (eq 8), where
resolvation is ignored, and the two waters of solvation on the
attacking hydroxide are completely removed in the TS. We
conclude that the general base mechanism can only have
mathematically acceptable fits in any of our treatments under
the chemically unlikely scenario, where all waters of solvation
on the attacking hydroxide and developing C-O- have φ-values
of unity, with the isotope effect being dictated by φ1 and φ3.
Saponification of Ethyl Acetate. The recent report5 that the
proton inventory analysis of the saponification of ethyl acetate
fits a highly simplified general base mechanism proceeding
through TS 6 having only two contributing fractionation factors
and no solvating waters demands that we consider reanalysis
of that data within the framework of the models described in
eqs 15-18. In the original analysis,5 the TS was suggested to
be about 40% along the reaction coordinate, far earlier than that
for hydroxide attack on formamide,10 but consistent with what
Kirsch and co-workers reported on the basis of the formyl
hydrogen kinetic isotope effect for the saponification of methyl
formate.27
Figure 2. Plot of the second-order rate constants for saponification of ethyl
acetate as a function of mole fraction of D2O (n). Original data from ref 5.
Dashed line, NLLSQ fit to eq 15; solid line, NLLSQ fit to eq 16. Best fit
parameters are given in the text.
solid line in Figure 2. The small positive computed value for x
is for the nucleophilic process consistent with what we have
derived above for the nucleophilic attack on formamide with a
suspiciously early TS. However, this treatment, requiring that
all protons respond to the measure of progress along the reaction
coordinate to the same extent, is probably too simplistic in view
of the discussion presented above.
Fitting of the ester saponification data5 to the uncoupled
nucleophilic mechanism of eq 17 provides acceptable fits with
nearly identical correlation coefficients for 0.4 e x e 0.9, with
the φ4 value for the three solvating waters on the developing
C-O- being between 0.92 and 0.83 (y ) 0.27-0.51), indicating
that the resolvation of the oxyanion lags behind the nucleophilic
attack. If the value of x is varied for only the L1 hydroxide
proton, and all others are allowed to fit as independent variables,
no acceptable values for φ2 and φ4 can be computed as these
are all >1.0. In the case of the GB mechanism, when the data5
are fit to eq 18 with φ3 ) 0.5 for the proton in flight and 0.4 e
x e 0.9 pertaining to both the L1 and the L2 protons of TS 5,
the computed fractionation factors for the resolvating waters,
φ4, are invariably greater than 1.0. Interestingly, if only the L1
is assumed to correspond to a fixed x value between 0.4 and
0.9, and the φ3 and φ4 values are treated as variables, the only
acceptable fit comes at x ≈ 0.4-0.5, but in this case the
fractionation factors for all of the solvating waters are 1.0. In
the limit, the analysis is the same as that of Mata-Segreda,5
suggesting that all solvation is lost in the transition state
corresponding to TS 6 above.
Conclusion
From the above discussion, a nucleophilic mechanism involv-
ing resolvating waters can be fit to the available data for both
ethyl acetate and formamide hydrolysis. A simple model
satisfying the data considers that all fractionation factors,
including the ones for the resolvating waters, respond to the
same extent with progress (x) along the reaction coordinate.
However, in this case, the computed x values, if taken literally,
correspond to much earlier progress along the reaction coordi-
nate than the accepted transition state positions10,27 of about 0.4
for ester saponification or 0.7-0.9 for amide hydrolysis. A
refined nucleophilic mechanism giving good fits to the data is
suggested where the attack of hydroxide and resolvation of the
developing oxyanion are uncoupled with the resolvation lagging
behind the attack. However, these fits, in our hands, are only
mathematically and chemically reasonable if the attacking
Hydroxide ion is well known to possess a large hydration
enthalpy of 101.3 kcal/mol,28 and it is difficult to envision any
method of recouping the energy lost through desolvating the
hydroxide in TS 6 without some additional resolvation of the
developing (-)-charge. Fitting of the ethyl acetate data5 to our
eq 15 describing a nucleophilic process with a TS analogous to
4 gives ko ) 0.124 ( 0.001 and x ) 0.16 ( 0.03, the best fit
being shown as the dotted line in Figure 2. On the other hand,
analysis of the data according to eq 16 describing a general
base TS analogous to 5 gives ko ) 0.122 ( 0.0003, an
unacceptable x of -0.16 ( 0.02, and a fractionation factor for
the proton in flight of 0.84 ( 0.01. That fit is shown as the
(27) Bilkadi, Z.; de Lorimer, R.; Kirsch, J. F. J. Am. Chem. Soc. 1975, 97, 4317.
(28) Friedman, H. L.; Krishnan, C. V. In Water. A ComprehensiVe Treatise;
Franks, F., Ed.; Plenum: New York, 1973; Vol. 3, p 56.
9
J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003 1857

A R T I C L E S
SÄlebocka-Tilk et al.
Scheme 2
hydroxide and its waters of solvation are assumed to respond
to the same extent to progress along the reaction coordinate.
A general base mechanism can also be accommodated, but
any successful fits for the ester and amide require that all waters
of solvation have fractionation factors of unity, while that for
the proton in flight can be no smaller than 0.5.
Analysis of the proton inventory data for the base-catalyzed
hydrolysis of formamide suggests that, within the fitting
constraints delineated above, one cannot unambiguously rule
out either the nucleophilic mechanism or the general base
mechanism. In our opinion, the same situation also applies for
the analysis of the data for saponification of ethyl acetate.5 For
formamide, we believe this to be a consequence of the low
overall inverse dkie of kOH/kOD ) 0.77 ( 0.02 at [LO-] ) 1.42
M, as well as the rather featureless and nearly linear appearance
of the second-order rate constant versus n plot (Figure 1), which
does not lend itself to unique fitting with multiparameter
equations pertaining to different mechanisms. There is a possible
additional complication that the kinetic data are substantially,
but not entirely, devoted to the single step of hydroxide attack.
In our opinion, any successful analysis of these nearly linear
proton inventory data in terms of a suite of plausible mechanisms
cannot rely exclusively on the goodness of mathematical fit to
one model or another but rather must take cognizance of
chemical reality. The clearest indication of this is the fact that
a straight line accommodates the data with a correlation
coefficient which is as good or better than most of the more
complex fits, but there is no reasonable hydrolytic mechanism
involving bonding changes of a single proton.
forming (HO-(H2O)3) is too large to have all of the stabilizing
waters lost in any conceivable transition state for a nucleophilic
or general base process without compensating stabilization of
the negative charge in the transition state through H-bonding
interactions with additional solvent molecules.
Therefore, our currently favored mechanism for the basic
hydrolysis of formamide is a nucleophilic one, which we present
in Scheme 2 and suggest is consistent with the heavy atom
isotope effects. The first step results in the formation of an
encounter complex (EC1) where a formamide appears next to
one of the solvating waters of hydroxide. This can occur either
by diffusion of formamide or by chainlike proton transfer for
the hydroxide through the solvent, which is consistent with the
high ionic mobility of hydroxide in water.14 In either event,
formation of EC1 requires that the CdO of the formamide is
exposed to a lone pair of electrons on one of the waters of
hydration of the HO-. In the next step, proton transfer occurs
from this water of solvation to the hydroxide, generating EC2
with a partially solvated or “hot” hydroxide poised for attack
on the carbonyl. Part of the activation energy for the attack must
result from this desolvation, but, following Huskey and Scho-
wen,16 we suggest that this is offset by electrophilic stabilization
of the hot hydroxide by its association with the CdO dipole.
Because this hot hydroxide is immediately derived from water,
the heavy atom distribution must be similar to that of the bulk
water, thus accounting for Marlier’s observations.3 In essence,
this mechanism is reminiscent of the GB mechanism, but the
requisite proton transfer is completed prior to the attack of the
hot hydroxide on the formamide, making it a nucleophilic
mechanism consistent with the slightly inverse dkie observed.
The nucleophilic mechanism for hydroxide addition to esters
and amides has been widely accepted in explaining all extant
data prior to 1994. The main reason for now invoking the GB
mechanism stems from Marlier’s heavy atom isotope effects
for the base-catalyzed hydrolysis of methyl formate4 and
formamide,3 which are consistent with the attacking oxygen
being derived from H2O and not hydroxide. Those reports
prompted Mata-Segreda5 and us to conduct proton inventory
studies in the hopes of supporting one mechanism or the other.
From our current study, it seems unlikely that fractionation factor
analysis alone will allow one to distinguish between these two
mechanisms, but chemical intuition suggests that any acceptable
mechanism must incorporate solvent molecules. Solvation must
be an important component in hydroxide attack on formamide
because in the gas phase, where there is no additional solvent,
Acknowledgment. The authors gratefully acknowledge the
financial assistance of the Natural and Engineering Research
Council of Canada and Queen’s University. In addition, they
are grateful to Dr. Franc¸oise Sauriol for her assistance in
conducting the NMR experiments for the hydrolysis reactions
and to the referees for their insightful comments.
the attack is barrierless,29 but the experimental ∆Hq and ∆G25
q
values for base-promoted hydrolysis of formamide in solution
are 17.9 and 21.2 kcal/mol.7 Further, it seems reasonable that
the 101.3 kcal/mol enthalpy of hydration28 of hydroxide in
(29) Alagona, G.; Scrocco, E.; Tomasi, J. J. Am. Chem. Soc. 1975, 97, 6976.
JA021055Z
9
1858 J. AM. CHEM. SOC. VOL. 125, NO. 7, 2003