A heavy-atom isotope effect study of the hydrolysis of formamide

Article abstract of DOI:10.1021/ja9901819

Isotope effects were measured for all the atoms at the reactive center of formamide during hydrolysis in dilute alkaline solution. Most of the reaction proceeds by a pathway that is first-order in hydroxide, although a small amount proceeds by a pathway that is second-order in hydroxide. For alkaline hydrolysis at 25 °C the carbonyl carbon isotope effect is ¹³k(obs) = 1.0321, the carbonyl oxygen isotope effect is ¹⁸k(obs)(C=O)) = 0.980, the formyl hydrogen isotope effect is (D)k(obs) = 0.80, and the nitrogen leaving group isotope effect is ¹⁵k(obs) = 1.0040. The ratio of the rate of hydrolysis to the rate of exchange for the alkaline hydrolysis of formamide was shown to be linearly dependent on the hydroxide concentration, ranging from an extrapolated value of k(h)/k(ex) = 2.1 at very low hydroxide concentrations to k(h)[k(ex) = 8.4 at 1.5 M hydroxide. This is consistent with a mechanism in which an increasing fraction of the tetrahedral intermediate pool is trapped as a dianion at high pH, effectively lowering the rate of exchange. These results also indicate that the transition states leading into and out of the tetrahedral intermediate are of comparable energy for the pathway which is first-order in hydroxide. The solvent nucleophile isotope effect is ¹⁸k(obs(nuc)) = 1.022 for water as the attacking nucleophile or ¹⁸k(obs(nuc)) = 0.982 for hydroxide as the attacking nucleophile. These results strongly suggest that one of the water molecules hydrating the hydroxide ion is the actual attacking nucleophile instead of hydroxide ion itself.

Full text of DOI:10.1021/ja9901819

4356
J. Am. Chem. Soc. 1999, 121, 4356-4363
A Heavy-Atom Isotope Effect Study of the Hydrolysis of Formamide
John F. Marlier,*,† Nancy Carter Dopke,‡ Karyn R. Johnstone,† and Todd J. Wirdzig†
Contribution from the Department of Chemistry and Biochemistry, California Polytechnic State UniVersity,
San Luis Obispo, California 93407, and Department of Chemistry, UniVersity of Wisconsin,
Madison, Wisconsin 53706
ReceiVed January 19, 1999. ReVised Manuscript ReceiVed March 8, 1999
Abstract: Isotope effects were measured for all the atoms at the reactive center of formamide during hydrolysis
in dilute alkaline solution. Most of the reaction proceeds by a pathway that is first-order in hydroxide, although
a small amount proceeds by a pathway that is second-order in hydroxide. For alkaline hydrolysis at 25 °C the
carbonyl carbon isotope effect is 13kobs ) 1.0321, the carbonyl oxygen isotope effect is 18kobs(CdO) ) 0.980, the
formyl hydrogen isotope effect is Dkobs ) 0.80, and the nitrogen leaving group isotope effect is 15kobs ) 1.0040.
The ratio of the rate of hydrolysis to the rate of exchange for the alkaline hydrolysis of formamide was shown
to be linearly dependent on the hydroxide concentration, ranging from an extrapolated value of kh/kex ) 2.1
at very low hydroxide concentrations to kh/kex ) 8.4 at 1.5 M hydroxide. This is consistent with a mechanism
in which an increasing fraction of the tetrahedral intermediate pool is trapped as a dianion at high pH, effectively
lowering the rate of exchange. These results also indicate that the transition states leading into and out of the
tetrahedral intermediate are of comparable energy for the pathway which is first-order in hydroxide. The solvent
nucleophile isotope effect is 18kobs(nuc) ) 1.022 for water as the attacking nucleophile or 18kobs(nuc) ) 0.982 for
hydroxide as the attacking nucleophile. These results strongly suggest that one of the water molecules hydrating
the hydroxide ion is the actual attacking nucleophile instead of hydroxide ion itself.
Introduction
withdrawing substituents, show a rate law term which is second-
order in hydroxide.4 One explanation is that electron-withdraw-
ing groups on the amine serve to lower the pKa of the -OH
group of the tetrahedral intermediate, facilitating the creation
of an oxydianion which can eliminate the amine with or without
prior protonation.4 General-base catalysis, although uncommon,
has been observed for reactive anilides such as trifluoroacet-
anilides and formanilides.3 Formamide alkaline hydrolysis has
been studied at several concentrations of hydroxide5,6 and found
to have a rate law containing both first- and second-order terms
in hydroxide, consistent with the mechanism in eq 1.
Amides play a key role in both organic chemistry and
biochemistry. Important biological reactions such as the deg-
radation of proteins, as well as other metabolic reactions,
underscore the importance of understanding the mechanisms
of amide hydrolysis. Amides are the least reactive of the
common acyl groups toward basic nucleophiles, due to the poor
ability of an unprotonated nitrogen to act as a leaving group.
The most widely accepted mechanism under basic conditions
involves tetrahedral intermediates1-3 as shown in eq 1.
Amides are known to exchange the carbonyl oxygen with
solvent during alkaline hydrolysis.4 The results of such exchange
experiments are usually reported as a ratio of the rate constant
for hydrolysis to that for exchange (kh/kex). In turn, this result
is used to determine the partitioning ratios (Kk4/k2 or k3/k2,
eq 1) for the tetrahedral intermediate, assuming the proton-
transfer steps are rapid. It was initially expected that the
tetrahedral intermediate would partition back to amide much
faster than formation of products (kh/kex < 1) because the
hydroxyl group is a better leaving group than most amines. This
is true for benzamide, where kh/kex ) 0.21 at 109 °C, but
increasing the number of alkyl groups on the nitrogen decreases
the relative amount of exchange from kh/kex ) 0.74 for
N-methylbenzamide to an undetectable level of exchange in N,N-
dimethylbenzamide.7,8 It was postulated that tertiary amides are
(1)
Kinetic studies have shown that the alkaline hydrolysis of
many amides with simple aliphatic amines usually follows a
rate law which is first-order in hydroxide.1,2,4 On the other hand,
some anilides, and other amides containing strong electron-
(4) Brown, R. S.; Bennet, A. J.; Slebocka-Tilk, H. Acc. Chem. Res. 1992,
25, 481 and references therein.
(5) 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, MD, 1977; pp 100-121.
(6) Hine, J. S.; King, R. S.; Midden, W. R.; Sinha, A. J. Org. Chem.
1981, 46, 3186.
†
California Polytechnic State University.
‡ University of Wisconsin.
(1) O′Conner, C. J. Q. ReV. Chem. Soc. 1971, 24, 553.
(2) Bender, M. L.; Thomas, R. J. J. Am. Chem. Soc. 1961, 83, 4183.
(3) (a) Schowen, R. L.; Jayarmann, H.; Kershner, L. J. Am. Chem. Soc.
1966, 88, 3373. (b) Mader, P. M. J. Am. Chem. Soc. 1965, 87, 3191. (c)
DeWolfe, R. H.; Newcomb, R. C. J. Org. Chem. 1971, 36, 3870.
(7) Bender, M. H.; Ginger, R. D.; Unik, J. P. J. Am. Chem. Soc. 1958,
80, 1044.
10.1021/ja9901819 CCC: $18.00 © 1999 American Chemical Society
Published on Web 04/27/1999

Hydrolysis of Formamide
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4357
reaction needed in the calculation of the nitrogen isotope effect, the
formyl hydrogen isotope effect, and the carbonyl oxygen exchange with
solvent. For the carbonyl carbon and carbonyl oxygen isotope effects
the fraction of reaction was determined by manometric measurements
of the amount of CO2 produced via oxidation of formate. In separate
control experiments the two methods of measuring the fraction of
reaction were shown to be in close agreement.
incapable of exchange because at least one hydrogen must be
present on the nitrogen of the tetrahedral intermediate to catalyze
proton transfer between oxygen atoms. Recent studies by Brown
and co-workers4 on a series of toluamides included several
tertiary amides that did exchange the carbonyl oxygen with the
solvent. The current interpretation is that the extent of carbonyl
oxygen exchange of an amide is mostly dependent on the
basicity of the amine.4
Isotope ratios for the heavy-atom isotope effects were determined
on a Finnigan Delta-E isotope ratio mass spectrometer; isotopic
compositions are given as δ(13C) for carbon, δ(18O) for oxygen, and
δ(15N) for nitrogen. The δ value represents a per mil (‰) difference in
the isotope ratio compared to a tank standard. In the oxygen exchange
experiments the 18O/16O ratio of the carbonyl oxygen of unreacted
formamide was measured on a Shimadzu QP or a Hewlett-Packard
1800A GC-MS, equipped with an XT-1 nonpolar column; the 18O/
16O ratio of the product, formate, was measured with a Bruker 360
NMR spectrometer, utilizing the isotope shift of the carbonyl carbon.
The ratio of 1-d-formamide to 1-h-formamide in the formyl hydrogen
isotope effects was determined on a Hewlett-Packard Series II (model
5890) gas chromatograph, equipped with a 60 m carbowax column.
Syntheses. Knowledge of the isotopic composition of the carbonyl
oxygen of formamide is needed to determine the oxygen nucleophile
isotope effect. This δ(18O) cannot be determined directly. Consequently,
formamide was synthesized from methyl formate, for which the isotopic
composition of the carbonyl oxygen can be determined by a published
decarbonylation procedure.11 To synthesize this formamide, a dry
methanolic solution which was 0.89 M in methyl formate and 1.88 M
in ammonia was allowed to react under dry nitrogen at room
temperature for 21 h. The methanol and ammonia were removed by
rotary evaporation, and the remaining material (crude yield 100%) was
distilled under reduced pressure. A sample of this purified formamide
was hydrolyzed completely and then oxidized by I2/DMSO. Finally,
the δ(13C) of the carbonyl carbon was determined by isotope ratio mass
spectrometry of the resulting CO2 (see below). The fact that the isotope
ratio of the carbonyl carbon of formamide did not change significantly
(<1 per mil) from that of the starting methyl formate infers that the
conversion was complete and the isotope ratio for the carbonyl oxygen
must also be unchanged during the synthesis and purification of
formamide.
The nature of the steps for decomposition of the tetrahedral
intermediates (k3 and k4) is not fully understood. Solvent isotope
effects, which are often complicated to interpret, tend to indicate
that the nitrogen of fairly basic amines (such as aliphatic amines)
is fully protonated prior to leaving;4 however, a concerted proton
transfer with a late transition state remains a possible alternative.
Isotope effects have provided detailed information concerning
the transition-state structure of many acyl group transfer
reactions. However, only a few such studies have been reported
for the reactions of amides. Kirsch determined the formyl
hydrogen isotope effect on the alkaline hydrolysis of form-
amide.5 This isotope effect is dependent on the concentration
of hydroxide, consistent with a term in the rate law that is both
first- and second-order in hydroxide. Kirsch fitted these data to
the mechanism of eq 1 and calculated the kinetic isotope effects
for formation of T- (Dk1 ) 0.65), breakdown of T- (Dk3 ) 1.58),
and breakdown of T2- (Dk4 ) 1.41), indicating fairly late
transition states in all cases. There are only a few reports of
heavy-atom isotope effects on amide hydrolysis. Leaving group
nitrogen isotope effects for chymotrypsin-catalyzed hydrolysis
of N-acetyl-L-tryptophanamide9 (15kobs ) 1.010 at pH 8) and
papain-catalyzed hydrolysis of N-benzoyl-L-argininamide10 (15kobs
) 1.023 at pH 8) were carried out by O′Leary. The difference
between the isotope effects for the two enzyme-catalyzed
reactions appears to be due to different partitioning ratios of
the intermediates. The leaving group nitrogen isotope effect on
the alkaline hydrolysis of benzamide (15kobs ) 1.004) was
reported as an unpublished result in ref 9.
1-d-Formamide was synthesized in an identical manner from 1-d-
methyl formate. The purified product was mixed in a 1:1 ratio with
natural 1-h-formamide and used in the formyl hydrogen isotope effect
experiments.
In this paper we present isotope effects for all of the atoms
at the reactive center of formamide during hydrolysis in dilute
alkaline solution, where the reaction is largely first-order in
hydroxide ion concentration. We are primarily interested in the
mechanism of the first-order reaction because of its similarity
to the mechanism of enzyme-catalyzed amide bond hydrolysis
and, hence, its suitability as a model. The results of these
experiments allow a detailed picture of the transition-state
structure for this reaction. In addition, the solvent oxygen
nucleophile isotope effect allows determination of the actual
nucleophile (water or hydroxide) for the reaction.
Carbonyl Carbon and Carbonyl Oxygen Isotope Effect Proce-
dures. The carbonyl carbon and carbonyl oxygen isotope effects were
measured by a slight modification of the previously published
procedure.11 In a typical experiment a low conversion sample contained
0.10-0.20 M KOH and 0.20 M formamide in a total volume of 2.0
mL. After a reaction time of 15 min (10-32% reaction) the reaction
was quenched by addition of 1.0 mL of HEPES buffer, pH 7.5. This
solution was transferred to a round-bottom flask equipped with a side
arm which contained a stopcock. Water was removed by heating this
flask to 85 °C under high vacuum for a minimum of 3 h, after which
2.0 mL of anhydrous DMSO containing 300 mg of I2 was added through
a septum attached to the side arm. The CO2 was collected into a U-tube
at liquid nitrogen temperature after first being passed through two liquid
nitrogen-pentane traps. Control experiments with formamide in the
absence of KOH failed to produce any CO2 when subjected to the drying
and oxidation procedure, making removal of unreacted formamide
unnecessary. The δ(18O) and δ(13C) of the resulting CO2 were
determined by isotope ratio mass spectrometry. High conversion
samples were prepared by allowing a solution which was 0.20 M
formamide and 1.4 M KOH to react for a minimum of 50 min. This
solution was quenched with 0.6 mL of 1.0 M HCl and 1.0 mL of 1.0
M HEPES, pH 7.5. The solution was then dried, oxidized, and analyzed
as above.
Experimental Section
Materials and Methods. Formamide and HEPES buffer were from
Sigma Chemical Co. Methyl formate (anhydrous), DMSO (anhydrous),
triphenylmethane, sodium hydride (60% oil suspension), and 2.0 M
ammonia in methanol were obtained from Aldrich Chemical Co. Iodine
(sublimed) was from Mallinckrodt Chemical Co. All were of reagent
grade or better and were used without further purification. Water
containing 99 atom % 18O was obtained from Isotec Inc. for the GC-
MS method, and that containing 96 atom % 18O was obtained from
Icon Services Inc. for the NMR method. 1-d-Methyl formate (99 atom
%) was obtained from Aldrich Chemical Co.
The UV spectra were measured on a Cary 4 UV-vis spectropho-
tometer. A UV assay at 240 nm was used to determine the fraction of
Leaving Nitrogen Isotope Effect Procedure. Low conversion
samples for the nitrogen isotope effects were prepared by allowing a
5.0 mL solution which was 0.20 M in formamide and 0.10-0.20 M in
(8) Bunton, C. A.; Nayak, B. O′Conner, C. J. Org. Chem. 1968, 33,
572.
(9) O′Leary, M. H.; Kluetz, M. D. J. Am. Chem. Soc. 1972, 94, 3585.
(10) O′Leary, M. H.; Urberg, M.; Young, A. P. Biochemistry 1974, 13,
2077.
(11) Marlier, J. F. J. Am. Chem. Soc. 1993, 115, 5953.

4358 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Marlier et al.
Table 1. Carbonyl Oxygen Exchange with Solvent during the
KOH to react until enough KOH was consumed that reaction effectively
had ceased (39-57% reaction by the UV assay). A 1.0 mL aliquot
was immediately applied to a 2 mL Dowex 50 column (H+-form) and
eluted with water. Fractions of 3.0 mL volume were collected. In
separate control experiments it was shown that the unreacted formamide
eluted in the first three fractions, whereas the ammonium ion was tightly
bound to the column. The pooled fractions containing the unreacted
formamide were treated with enough 10 M KOH to bring the final
concentration of hydroxide to 1.0 M. The hydrolysis to ammonia and
formate was allowed to proceed for 1.75 h, after which the solution
was made mildly acidic (pH paper) with 9 M H2SO4. The reaction
mixture was then placed in a round-bottom flask equipped with a side
arm which was attached to a small glass vessel containing a several-
fold molar excess of a concentrated NaOBr solution.9 After several
freeze-thaw cycles under high vacuum to remove atmospheric nitrogen,
the NaOBr was added from the side arm to the main flask. The nitrogen
produced by oxidation of the ammonia was collected under vacuum
into molecular sieves at liquid nitrogen temperature after passing
through two liquid nitrogen traps. The δ(15N) of the collected nitrogen
was then determined by isotope ratio mass spectrometry. High
conversion samples were generated from solutions that were 0.20 M
in formamide and 1.4 M in KOH. After complete hydrolysis the
ammonium ion produced was oxidized and analyzed as above. A
column was not necessary in the case of the high conversion samples
because no unreacted formamide remained.
Nucleophile Oxygen Isotope Effect Procedure. The oxygen
nucleophile isotope effect was measured by reacting a solution that
was 0.27 M in formamide of known δ(18O) at the carbonyl oxygen
(see above synthesis section) and 1.7 M in KOH for at least 1 h. The
formate produced by this hydrolysis was neutralized by addition of
1.0 mL of 1.0 M HEPES (acid form) and then dried and oxidized by
I2/DMSO as described above. The isotope ratio of the water (and
hydroxide) used in these experiments was determined by exchanging
the oxygen atoms of the water (10 mL) with a 0.1 mmol sample of
CO2 for 20 h with stirring. The exchanged CO2 was then collected
under vacuum and analyzed by isotope ratio mass spectrometry in the
usual manner.
Alkaline Hydrolysis of Formamide in Water
av [OH-],
Ma
GC-MS
method kh/kex
13C NMR
method kh/kex
b
c
0.092
0.11
0.22
1.46
2.5
2.7
2.9
8.3
a (Initial + final hydroxide concentration)/2. b At 25 °C, unbuffered.
c At room temperature: 22-24 °C, unbuffered.
resonances for formamide and formate. A T1
study established the
optimum setting for the integration of the carbonyl carbon resonances;
a control experiment containing equal amounts of formate and
formamide yielded the expected 1:1 integration. The extent of isotope
exchange was determined by separate integration of the 13C carbonyl
carbon resonances containing zero, one, and two 18O atoms.
Results
Formamide was shown to exchange its carbonyl oxygen with
the solvent during alkaline hydrolysis in water. Two methods
were employed to measure the ratio of the rate constant for
hydrolysis to that for exchange (kh/kex); the results are given in
Table 1. In the first method (GC-MS) the fraction of reaction
was determined either by the UV assay or by integration of the
formyl hydrogen NMR resonances for formamide and formate;
the ratio of 18O/16O of the carbonyl oxygen in unreacted formate
was determined by GC-MS. The ratio, kh/kex, was calculated
using equations published elsewhere.7
In the second method the oxygen isotope shift on the 13C
resonance for the isotopically enriched carbonyl carbon was used
to determine the 18O/16O ratio in the carbonyl oxygen of the
product, formate. In the NMR method, formate containing two
18O atoms was produced exclusively via exchange with solvent;
formate with one 18O and one 16O arose mostly from hydrolysis
(a small amount appears from exchange with the small amount
of 16O water); the small amount of formate with two 16O atoms
came exclusively from the hydrolysis reaction with the small
amount of 16O water in the reaction mixture. Simple integration
of the carbonyl carbon resonances of formamide and formate
served to determine the fraction of reaction. From the limited
data collected, the 13C NMR method appears to have an error
limit of approximately (5%, which is only slightly higher than
the error limit for determining kh/kex by the GC-MS method
((2-3%). Combining the results of both methods, kh/kex was
found to be linearly dependent on the average concentration of
hydroxide (see Table 1 and Discussion).
Formyl Hydrogen Isotope Effect Procedure. The formyl hydrogen
isotope effect was measured using a 1:1 mixture of 1-d-formamide/1-
h-formamide (see synthesis section). In these experiments a solution
that was 0.20 M in total formamide and 0.20-0.40 M in KOH was
followed by the UV assay at 240 nm. After the reaction nearly stopped
due to low levels of KOH (31-68% reaction), the solution was
immediately extracted three times with 2 mL of ethyl acetate. The
pooled ethyl acetate was dried with anhydrous MgSO4 and analyzed
by capillary GC. The two isotopic formamides separated on the 60 m
carbowax column, with the deuterium-containing isotopomer eluting
5 s later than natural abundance material. The area under each peak
was determined by a Grams 95 program. The isotope effect was then
calculated in the usual way from the fraction of reaction and the isotope
ratios.
Carbonyl Oxygen Isotope Exchange Procedures. The extent of
carbonyl oxygen isotope exchange was measured by two methods. In
the first method a solution containing 0.40 mL of 1.0 M formamide,
0.20 mL of H2O (99 atom % 18O), and 0.100-0.300 mL of 1.0 M
KOH in a total volume of 1.000 mL was allowed to react and then
extracted with ethyl acetate and dried just as in the formyl hydrogen
isotope effect experiment (see preceding paragraph). In this case the
18O/16O ratio of the unreacted formamide was determined on the
Shimadzu QP GC-MS. The isotope exchange at 2.0 M hydroxide was
determined in a similar manner, except the Hewlett-Packard GC-MS
was used for isotopic analysis. The extent of reaction was determined
by either the UV assay or integration of the formyl hydrogen resonances
in the proton NMR.
The extent of carbonyl oxygen exchange was also measured by a
novel method utilizing the 18O isotope shift on the carbonyl carbon of
[13C]formate. For this exchange experiment 10 µL of 9.4 M KOH, 10
µL of [13C]formamide, 480 µL of [18O]water (96 atom %), and a few
crystals of EDTA were sealed in an NMR tube and allowed to sit for
30 min, after which the 13C NMR spectrum was taken. The extent of
reaction was determined by integration of the carbonyl carbon
The carbonyl carbon and carbonyl oxygen isotope effects on
the alkaline hydrolysis were measured in aqueous solution at
25 °C by a slight modification of our published procedure.12
This procedure was subjected to controls that ensured unreacted
formamide did not interfere with collection and oxidation of
formate produced during hydrolysis. The carbonyl carbon
isotope effect was determined directly from the measured
δ(13C) for low and complete conversion samples (Table 2).
Calculation of the carbonyl oxygen isotope effect from the
measured δ(18O) was not as straightforward. The solvent (water),
representing a vast molar excess of oxygen atoms compared to
the carbonyl oxygen of formamide, maintains a constant δ(18O)
during the course of hydrolysis. One of the oxygen atoms of
formate (which is subsequently converted to CO2) was derived
from the solvent and the other from either the original carbonyl
oxygen of formamide or the solvent via the exchange reaction.
(12) Marlier, J. F.; Haptonstall, B. A.; Johnson, A. J.; Sacksteader, K.
A. J. Am. Chem. Soc. 1997, 119, 8838.

Hydrolysis of Formamide
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4359
Table 2. Isotope Effects on the Alkaline Hydrolysis of Formamide
in Water at 25 °
Table 3. Isotopic Composition of Oxygen Atoms Involved in
Calculation of the Nucleophile Isotope Effect on the Alkaline
Hydrolysis of Formamide in Water at 25 °C
atom
k(light)/k(heavy)a,b
oxygen source
δ(18O)a
carbonyl carbon
carbonyl oxygenc
leaving nitrogen
formyl hydrogen
1.0321 ( 0.0008 (6)
0.980 ( 0.001 (5)
1.0040 ( 0.0005 (6)
0.80 ( 0.02 (4)
CO2 from complete hydrolysis of formamide
CO2 from carbonyl oxygen of formamide
CO2 exchanged with water
water
-37.2 ( 1.0 (6)
-6.3 ( 0.3 (6)
+1.4, range 0.4 (2)
-39.6 (calc)
a Corrected for fraction of reaction. b Number of determinations in
parentheses. c Corrected for carbonyl oxygen exchange (see the text).
hydroxide
-79.6 (calc)
a Compared to a PDB standard.
Therefore, calculation of the carbonyl oxygen isotope effect
required a correction for δ(18O) of those oxygens which entered
formate either via exchange with solvent (i.e., kh/kex) or by direct
attachment of the nucleophile. Differential equations to ac-
complish this correction13 are given in detail in the Supporting
Information; the oxygen isotope effects in Table 2 reflect this
correction. The rather high standard error on the carbonyl
oxygen isotope effect reflects not only the normal uncertainty
in measuring a heavy-atom isotope effect, but also the additional
uncertainty on measuring the extent of oxygen exchange with
solvent. The observed kinetic isotope effects did not vary with
pH within the narrow range of hydroxide ion concentrations
used in these experiments.
The nitrogen isotope effect was measured by first separating
unreacted formamide from ammonia on a Dowex 50 (H+-form)
column and then hydrolyzing this formamide quantitatively to
formate plus ammonia. The ammonia was subsequently oxidized
to nitrogen with NaOBr by a published procedure.9 Control
experiments demonstrated that formamide was not fractionated
on this short column and that unreacted formamide was
quantitatively separated from the ammonia formed during
hydrolysis. The isotope effects are also given in Table 2. The
observed kinetic isotope effects did not vary with pH within
the narrow range of hydroxide ion concentrations used in these
experiments.
CO2 was found to have a δ(18O) of -37.2 (Table 3). The origin
of the two oxygen atoms of the CO2 (O1dCdO2) produced in
this manner was as follows: (1) one of the oxygen atoms (O1)
was completely derived from the solvent nucleophile (water or
hydroxide). The second oxygen atom (O2) was derived from
either (2) the solvent nucleophile via exchange or (3) the original
unexchanged carbonyl oxygen of formamide (still having a
δ(18O) of -6.3 because hydrolysis was quantitative).
In the present experiment it is not necessary to measure the
isotopic composition of the oxygen atoms of CO2 at both high
and low conversions, as is typical in a competitive experiment;
the above-mentioned quantitative hydrolysis is sufficient. This
is due to the fact that the ultimate source of O1 is the solvent
oxygen nucleophile (water or hydroxide), which is incorporated
into formamide during hydrolysis. Since the solvent is present
in a very large excess over the substrate (formamide), it can be
treated as essentially an infinite reservoir of oxygen atoms,
which is of constant concentration and isotopic composition.
All that is required to determine the nucleophile isotope effect
is a way to determine the δ(18O) for O1 and to compare it to
the invariant δ(18O) of the possible nucleophiles (water or
hydroxide).
This calculation can be accomplished using eq 2, where the
per mil isotopic abundances δ(obs), δ1, δ2(ex), and δ2
are,
(un)
respectively, those for the entire CO2 molecule, for O1 (which
The formyl hydrogen isotope effect (Table 2) was measured
at an average 0.2 M hydroxide ion concentration by a capillary
GC method. The 1-d-formamide was shown to elute 5 s after
the natural abundance formamide on a 60 m carbowax capillary
GC column. Therefore, the area under the peaks was used to
calculate the change in the ratio of these two isotopically
substituted formamides at various fractions of reaction. The
fraction of reaction was determined by the UV assay.
Measurement of the oxygen isotope effect on the nucleophile
required synthesis of formamide with a known δ(18O) for the
carbonyl oxygen. This was accomplished by measuring the
δ(18O) for the carbonyl oxygen of methyl formate11 and then
quantitatively converting this methyl formate to formamide in
the presence of dry methanolic ammonia. This synthesis
preserved the known isotopic composition of the carbonyl
oxygen during conversion of methyl formate to formamide
because there was no source of exchangeable oxygen atoms
and because the conversion was quantitative. The carbonyl
oxygen of this methyl formate (and hence formamide) was found
to have a δ(18O) of -6.3 (Table 3).
δ(obs) ) 0.5δ1 + 0.5δ2(ex)f(ex) + 0.5δ2(un)f(un)
(2)
is directly from the nucleophile), for O2 (which is from the
carbonyl oxygen of formamide after exchange with solvent),
and for O2 (which is from the carbonyl oxygen of formamide
without exchange with the solvent). Additionally, f(ex) is the
fraction of formamide molecules which has undergone oxygen
exchange, whereas f(un) is the fraction that has not exchanged.
The one assumption in this calculation is that the isotope ratio
of the oxygen atoms entering formate via the exchange pathway
is the same as that for the nucleophile (i.e., δ1 ) δ2(ex)). This is
a reasonable assumption since the exchange reaction begins with
the same nucleophilic attack step (k1, eq 1) as the hydrolysis
pathway and because the amount of exchange is so small (kh/
kex ) 8.7) at the high level of hydroxide employed for these
particular experiments. Using this assumption and solving for
δ1 gives eq 3.
δ1 ) [δ(obs) - 0.5δ2(un)f(un)]/[0.5 + 0.5f(ex)
]
(3)
A sample of this formamide was completely hydrolyzed in
aqueous 1.7 M KOH, dried, and oxidized to CO2 by I2/DMSO.
This higher hydroxide concentration was chosen because at the
lower hydroxide concentrations quantitative conversions are
more difficult (slow reaction rates) and because a large amount
of carbonyl oxygen exchange makes determination of the
δ(18O) inherently less precise. The isotopic composition of this
The magnitude of δ(obs) is -37.2 per mil (Table 3); that for
δ2
is -6.3 per mil. The values of f(un) and f(ex) can be
(un)
calculated from the differential equations cited earlier13 (and
detailed in the Supporting Information), plus the known value
of kh/kex (Table 1). The above information allows the δ(18O) of
the attacking nucleophile (or δ1) to be calculated, which is -62.3
per mil. The δ(18O) values of -39.6 per mil for the water and
-79.6 per mil for the hydroxide used in these experiments were
(13) Marlier, J. F. Ph.D. Dissertation, University of WisconsinsMadison,
Madison, WI, 1978.

4360 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Marlier et al.
kh/kex ) (2k3 + 2k4K[OH-])/k2
(4)
fitted to eq 4; linear regression analysis gave an excellent fit
(r2 ) 0.999). The y-intercept of this plot (2k3/k2) equals 2.10,
making the partition ratio for T- (k3/k2) equal to 1.05. The small
magnitude of k3/k2 normally indicates the transition states leading
to and from T- are of comparable energy and, therefore,
resemble T- in structure. However, the proton transfer required
to make the nitrogen a better leaving group appears to cause
the transition state for decomposition of T- to become less sp3-
like than expected (see below). In contrast, esters such as methyl
formate show much less oxygen exchange (kh/kex ) 18.3),
arguing for a largely rate-determining formation of T-. The
slope of the above plot (2k4K/k2) equals 4.29. From these results
it follows that, under conditions of 0.086-0.15 M average
hydroxide concentrations used in the present work, only 14-
23% of the formate is produced via the pathway which is
second-order in hydroxide. Therefore, the subsequent simplifica-
tion of the discussion of the remaining isotope effects solely in
terms of the first-order pathway is justified.
Figure 1. Summary of the isotope effects on alkaline hydrolysis of
formamide and methyl formate.11,17,18
determined by completely exchanging the oxygen atoms of
water with CO2 and then correcting for the known fractionation
factors between CO2 and water (-41.0 per mil)14 and between
water and hydroxide (-40.0 per mil).15 As a result, either water
18
is the attacking nucleophile with an isotope effect of
k
obs(nuc)
) 1.022 ( 0.001 or hydroxide is the nucleophile with an isotope
18
effect of
k
) 0.982 ( 0.001.
obs(nuc)
Discussion
The alkaline hydrolysis of amides is widely believed to occur
by the mechanism shown in eq 1. The product is formed by
pathways which are either first-order or second-order in
hydroxide. At the low initial hydroxide concentrations used in
the present isotope effect experiments (0.1-0.2 M) the hy-
drolysis proceeds mainly through the pathway which is first-
order in hydroxide (involving only T-). By contrast only a very
few esters16a were thought to undergo alkaline hydrolysis by a
second-order pathway in a mechanism analogous to that given
in eq 1. Recent studies have cast doubt on these earlier
findings;16b the current evidence supports only the first-order
pathway for ester alkaline hydrolysis. To more fully characterize
the bonding in the transition state for the alkaline hydrolysis of
amides, we have measured isotope effects for all the reacting
atoms of formamide under conditions where the first-order
pathway greatly predominates. The solvent oxygen nucleophile
isotope effect was measured under highly alkaline conditions,
where the attack of the nucleophile (step 1, eq 1) was largely
rate-determining. This nucleophile isotope effect was important
in correcting the carbonyl oxygen isotope effect for the presence
of exchanged oxygen atoms from the solvent and for determi-
nation of the actual nucleophile (water or hydroxide) in the
hydrolysis reaction. In addition, the extent of carbonyl oxygen
exchange with solvent was determined at various hydroxide
concentrations. The results of the isotope effect study are
summarized in Figure 1, and compared to those previously
published for the alkaline hydrolysis of methyl formate.11,17,18
Carbonyl Oxygen Isotope Exchange. Formamide undergoes
carbonyl oxygen exchange with solvent; this exchange is
dependent on the concentration of hydroxide. These results are
consistent with the mechanism in eq 1, where higher concentra-
tions of hydroxide effectively trap more of the intermediate as
Formyl Hydrogen Isotope Effect. The secondary formyl
hydrogen isotope effect is sensitive to changes in the hybridiza-
tion of the adjacent carbonyl carbon atom. Kirsch reported an
inverse formyl hydrogen isotope effect on the alkaline hydrolysis
of formamide5 which was dependent on the initial hydroxide
concentration (in the range 0.05-1.0 M). This result was
interpreted in terms of an increasing proportion of the reaction
proceeding through T2- instead of T-, in agreement with the
observed hydroxide dependence of kh/kex (see above). From the
observed isotope effects Kirsch calculated the isotope effects
on the individual steps of the mechanism. The isotope effect
D
on formation of T- was found to be k1 ) 0.65, arguing for a
very late sp3-like transition state (structure I). The isotope effect
on breakdown of T- (structure III) was calculated to be Dk3 )
D
1.58; that for breakdown of T2- (structure II) was k4) 1.41,
indicating late, sp2-like transition states in both cases. The
transition state for the breakdown of T- (structure III) is shown
as involving a simultaneous proton transfer (through water) from
the -OH of T- to the leaving nitrogen. On the basis of solvent
deuterium isotope effects, Brown4 has proposed an alternative
mechanism involving stepwise protonation of the nitrogen prior
to leaving. These solvent isotope effects are also consistent with
a late transition state III. The transition states for the breakdown
of T- in both mechanisms avoid the unlikely formation of the
T
2-, increasing hydrolysis relative to exchange (assuming k4 >
k3). The steady-state equations for the hydrolysis and exchange
reactions of eq 1 have been published.4 From these equations
an expression for kh/kex as a function of hydroxide concentration
can be developed and is given in eq 4, assuming rapid proton
transfers. The oxygen isotope exchange results (Table 1) were
-
highly basic NH2 ion. Neither mechanism can be strictly
eliminated on the basis of the isotope effects presented in this
paper.
Our determination of Dk(obs) ) 0.80 (Table 2) was performed
only at an average hydroxide concentration of 0.20 M and is in
general agreement with those reported by Kirsch at that
hydroxide concentration. The relationship between the isotope
effects on the individual steps of the mechanism and the
observed isotope effect is given by eq 5. In eq 5, the asterisk
(14) Friedman, I.; O′Neill, J. R. U.S. Geological Survey Professional
Paper No. 440-KK, 1977.
(15) Green, M.; Taube, H. J. Phys. Chem. 1963, 67, 1565.
(16) (a) Khan, M. N.; Olagbemiro, T. O. J. Org. Chem. 1982, 47, 3695.
(b) Kellogg, B. A.; Brown, R. S.; McDonald, R. S. J. Org. Chem. 1994,
59, 4652.
(17) Sawyer, C. B.; Kirsch, J. F. J. Am. Chem. Soc. 1973, 95, 7375.
(18) Bilkadi, Z.; deLorimier, R.; Kirsch, J. F. J. Am. Chem. Soc. 1975,
97, 4317.
*kobs ) [*Keq*k3 + *k1(k3/k2)]/(1 + k3/k2)
(5)

Hydrolysis of Formamide
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4361
a calculated magnitude of 13k3 ) 1.05 (Table 4). This calculated
isotope effect is a bit higher than expected. Given the difficulty
of accurately estimating 13Keq and 13k1, the calculated magnitude
of 13k3 is at least in the ballpark. In general, large normal kinetic
isotope effects such as these on k1 and k3 are indicative of steps
involving a high degree of reaction coordinate motion on the
part of the carbonyl carbon atom as expected in both steps of
the mechanism. However, this large normal observed isotope
effect effectively eliminates any mechanism in which T- (or
an N-protonated form of T-) is at equilibrium with the ground
state (formamide). If this were the case, the observed carbonyl
Table 4. Calculated Kinetic Isotope Effects on the Various
Steps in the Mechanism of Alkaline Hydrolysis of Formamide
(See the Text)
isotope effect
assumed
calculated
formyl H
carbonyl C
nucleophilic O
Dk1 ) DKeq ) 0.69
Dk3 ) 1.34
13
18
18
18
13
18
18
18
k
k
k
k
) 1.032, 13Keq(CdO) ) 0.983
k
k
k
k
) 1.050
1(CdO)
3(CdO)
) 0.995, 18Keq(nuc) ) 1.033
) 1.000, 18Keq(nuc) ) 1.033
) 1.005, 18Keq(nuc) ) 1.033
) 1.017
) 1.012
) 1.007
3(nuc)
3(nuc)
3(nuc)
1(nuc)
1(nuc)
1(nuc)
15
15
15
leaving N
15k1 )
15k1 )
15k1 )
K
K
K
) 1.000
) 1.005
) 1.010
15k3 ) 1.008
15k3 ) 0.998
15k3 ) 0.988
eq
eq
eq
carbon isotope effect would need to be inverse, approximately
13
in the
K
) 0.983 range.12
eq
refers to an isotope effect which is always given as a ratio of
the rate (or equilibrium) constants for the light/heavy isotopes
(i.e., in the case of hydrogen isotope effects *k symbolizes Dk,
which in turn refers to Hk/Dk). In particular, *kobs is the observed
isotope effect, *Keq is the equilibrium isotope effect on step 1,
and *k1 and *k3 are the kinetic isotope effects on formation
and decomposition of T-, respectively. The ratio k3/k2 ) 1.05
results from the isotope exchange experiments extrapolated to
very low hydroxide concentrations (see above).
Carbonyl Oxygen Isotope Effect. The carbonyl oxygen
isotope effect is a secondary isotope effect because the con-
nection between atoms is not severed during the reaction. A
mathematical calculation of the oxygen isotope effects on the
individual steps from eq 5 is not possible at this time because
18
there are no reasonable models to help estimate
K
,
eq(CdO)
18
18
k
, or
1(CdO)
k
. However, a qualitative description is
3(CdO)
possible. The bond order to the carbonyl oxygen is reduced on
going to either of the possible transition states (structure I or
III), leading to the expectation of a normal isotope effect. In
fact, most of the carbonyl oxygen isotope effects reported for
the reaction of esters with basic nucleophiles are small, but
D
One can obtain a reasonable estimate of Keq from the
hydrogen fractionation factor between the formyl hydrogen of
formic acid and the hydrogen of a secondary alcohol, which is
DKeq ) 0.69.19 The equilibrium isotope effect on formation of
T- should be of a similar magnitude. Because the formyl
hydrogen isotope effect is a secondary one and because the
transition state for formation of T- appears to be a late one, it
is reasonable to expect that DKeq should be similar in magnitude
to Dk1. Using these estimated values, Dk3 is calculated to be
1.34 (Table 4), in approximate agreement with that calculated
by Kirsch (Dk3 ) 1.58). Finding a late, sp2-like transition state
for breakdown of T- is somewhat of a surprise because it is
expected that the transition state would resemble the reactive
tetrahedral intermediate. However, the results of these formyl
hydrogen isotope effect studies strongly suggest any viable
transition state for breakdown of T- must involve considerable
C-N bond cleavage.
normal; the alkaline hydrolysis of methyl formate11 is an
18
exception, where
k
) 0.999. Therefore, the moderate
obs(CdO)
inverse isotope effect observed for formamide is somewhat of
a surprise, and the explanation must involve more than simple
π bond breaking.
A more detailed explanation requires a qualitative analysis
of the vibrational modes in the ground state and in transition
states I and III (recall that the transition states for formation
and destruction of T- are of nearly equal energy). In either
transition state I or III there is a gain of new O-C-O bending
and O-C-O-H torsional modes which stiffens the bonding
to the carbonyl oxygen. This stiffening gives an inverse isotope
effect which counters the normal isotope effect on carbonyl π
bond breaking. Why is the carbonyl oxygen isotope effect on
the alkaline hydrolysis of formamide (18kobs(CdO) ) 0.980)
considerably more inverse than that of the methyl formate case
(18kobs(CdO) ) 0.999)? The answer must be due to significantly
different changes in bonding on going to the transition states.
The transition state for step 1 (k1) in formamide hydrolysis has
more sp3 character than that for methyl formate (see Formyl
Hydrogen Isotope Effects, Figure 1), and consequently, the bond
to the nucleophile will be more fully developed in this transition
state. Therefore, any bond stiffening in the transition state
resulting from introduction of new vibrational modes (an inverse
effect) will be proportionally greater in the formamide case.
On the other hand, one might expect the corresponding reduction
in π bond order (a normal effect) to also be greater in the later
transition state of formamide hydrolysis, giving more compa-
rable observed isotope effects. However, the nitrogen lone pair
in amides is more delocalized by resonance than the corre-
sponding lone pair on the alkoxy oxygen of esters, which gives
the amide ground state a lower carbonyl π bond order and leads
to a lower than expected normal isotope effect. Thus, in
formamide hydrolysis the bond stiffening created by the new
vibrational modes in the transition state contributes more to the
overall observed isotope effect than in the case of methyl
formate hydrolysis. Similar effects are seen in secondary
Carbonyl Carbon Isotope Effect. Carbonyl carbon isotope
effect data on the hydrolysis of amides are lacking in the
literature. However, several carbonyl carbon isotope effects have
been measured for the alkaline hydrolysis of esters. The carbonyl
carbon isotope effect for the alkaline hydrolysis of methyl
formate11 (13kobs ) 1.0324) compares closely with that reported
here for formamide (13kobs ) 1.0321). The isotope effects on
the individual rate constants can be approximated from eq 5.
An estimate of 13Keq ) 0.979 has been reported for the addition
of hydrazine to methyl formate.12 After the small secondary
contribution of the outer nitrogen is removed (making 13Keq
)
0.983), this will serve as a crude model for the present
equilibrium carbon isotope effect. The best model for the
carbonyl carbon isotope effect on k1 is the alkaline hydrolysis
of methyl formate, where 13kobs ) 1.0324. In the methyl formate
case the first step (attack of hydroxide) is largely rate-
13
determining, making
k
nearly equal to the kinetic isotope
obs
effect on step 1 (13k1). Unfortunately, the transition state is
considerably earlier for the alkaline hydrolysis of methyl formate
when compared to formamide. However, the carbonyl carbon
isotope effect for the alkaline hydrolysis of a large number of
esters appears to be rather insensitive to large changes in the
structure of the ester (and hopefully the transition state);11,12,20
13
all such carbonyl carbon isotope effects are in the range
k
obs
(20) (a) Marlier, J. F.; O′Leary, M. H. J. Am. Chem. Soc. 1990, 112,
5996. (b) O′Leary, M. H.; Marlier, J. F. J. Am. Chem. Soc. 1979, 101,
3300.
) 1.03-1.04. Using these estimated isotope effects in eq 5 gives
(19) Cleland, W. W. Methods Enzymol. 1980, 64, 104.

4362 J. Am. Chem. Soc., Vol. 121, No. 18, 1999
Marlier et al.
hydrogen isotope effects for acyl group transfers, where the
carbonyl carbon goes from an sp2 ground state to an sp3-like
transition state.5,11,12,18
Oxygen Nucleophile Isotope Effect. The oxygen of formate
which is derived from the solvent during the alkaline hydrolysis
is 22 per mil lighter than that for water and 18 per mil heavier
the nucleophile isotope effect measured at high hydroxide
concentration is similar to that at low hydroxide concentration.
The known oxygen fractionation factor between water and a
secondary alcohol (18Keq(nuc) ) 1.033) is a reasonable model
for the equilibrium isotope effect on formation of T-. Estimation
18
18
of the magnitude of
estimation of 18k3(nuc) is probably easiest. The isotope effect on
k3(nuc) is composed of (1) a small normal primary isotope effect
k
or
k
is more difficult, but
1(nuc)
3(nuc)
than that for hydroxide (see Table 3). Therefore, water as the
18
nucleophile gives an isotope effect of
k
) 1.022;
obs(nuc)
hydroxide as the nucleophile gives an isotope effect of 18kobs(nuc)
) 0.982. Which of these is the actual nucleophile? In either
case a new bond is formed to the nucleophilic oxygen, making
the isotope effect on the nucleophile a primary one. Primary
isotope effects are composed of a temperature-independent factor
(TIF) which is due to reaction coordinate motion and a
temperature-dependent factor (TDF) which is due to creation
of new vibrational modes in the transition state.21 In almost all
known cases, isotope effects on attacking nucleophiles are
normal because they are dominated by reaction coordinate
motion (the TIF). One known exception is the reaction of
unhydrated chloride ion as a nucleophile.22 The observed
nitrogen nucleophile isotope effect on the hydrazinolysis of
methyl formate is slightly inverse and appears to be another
exception.12 However, this observed isotope effect is inverse
because of a secondary isotope effect on the nonnucleophilic
(or outer) nitrogen of hydrazine, which is analyzed along with
the nucleophilic one. There is no way to quantitatively correct
for the secondary isotope effect of this outer nitrogen and for
the concerted proton transfers in the transition state, but such
effects would tend to make the observed isotope effect appear
on the breaking of the O-H bond, (2) a small inverse secondary
isotope effect on formation of the partial π bond of formate,
and (3) a small normal isotope effect from loss of a vibrational
mode (O-C-N) as the leaving group departs. Since all of these
effects are expected to be small, it is likely that the overall
isotope effect will be a small inverse or small normal one,
18
perhaps in the range
k
) 0.995-1.005. Application of
3(nuc)
18
eq 5 gives a magnitude of
(Table 4).
k
in the range 1.007-1.017
1(nuc)
18
Are these calculated magnitudes for
k
reasonable?
1(nuc)
During formation of T- the nucleophilic oxygen loses an O-H
bond to the general base (hydroxide) and simultaneously gains
a new C-O bond, giving a substantial normal isotope effect.
How large can this normal isotope effect on k1 be? The
nucleophile oxygen isotope effect on the alkaline hydrolysis of
methyl formate (18kobs(nuc) ) 1.023) is probably near the upper
limit.11 On the other hand, formation of new vibrational modes
in the transition state for formation of T-, particularly the new
O-C-O and O-C-N bending modes, will give an inverse
contribution to the overall isotope effect. In the methyl formate
case discussed above, the transition state for formation of T-
is early and the inverse contribution of forming new bending
modes is relatively unimportant. However, the transition state
for formation of T- in formamide hydrolysis is later and the
inverse contribution from creation of these new vibrational
more inverse than it actually is. In light of these results a large
18
inverse isotope effect of
k
) 0.982 is clearly not
obs(nuc)
expected, making water (probably one of the water molecules
hydrogen bonded to hydroxide ion) the actual nucleophile for
alkaline hydrolysis.
modes is greater, leading to a smaller observed isotope effect.
Water was also proposed to be the nucleophile for the alkaline
hydrolysis of esters, on the basis of a nearly identical nucleophile
isotope effect of 18kobs(nuc) ) 1.023 for methyl formate hydroly-
sis11 and the observation of general-base catalysis of the
hydrolysis of a series of formate esters.23 In the methyl formate
case the carbonyl oxygen exchange data (kh/kex ) 18.3) and
the lack of a sizable leaving group oxygen isotope effect (18kobs(lg)
) 1.009) clearly indicated that the first step (attack of the
nucleophile) was rate-determining. Furthermore, the small
inverse formyl hydrogen isotope effect indicated the transition
state for attack of the nucleophile was quite early. Theoretical
calculations have indicated that early transition states for
nucleophilic attack will give large, normal isotope effects,
whereas later transition states tend to give smaller normal or,
in extreme cases, inverse isotope effects.24,25
18
Seen in this light, the calculated range for
1.017 seems quite reasonable.
k
) 1.007-
1(nuc)
One final point underscores the earlier choice of water as
18
the nucleophile for alkaline hydrolysis with
k
) 1.022
obs(nuc)
18
over hydroxide with
k
) 0.982. If one uses the latter
observed isotope effect in eq 5, the resulting calculated
obs(nuc)
18
magnitude of
k
is in the range 0.931-0.940. Inverse
1(nuc)
oxygen isotope effects of 6-7% would be among the largest
such effects observed (near the theoretical maximum) and are
clearly not possible in this case.
Nitrogen Isotope Effect. The observed leaving group
nitrogen isotope effect is surprisingly small (15kobs ) 1.004) for
a primary isotope effect involving C-N bond breaking. The
isotope effects on the individual steps of the mechanism are
again given in eq 5. Good models for estimation of the
individual isotope effects are once again lacking. Nevertheless,
a rough estimation is possible. Nitrogen isotope effects for
reactions where breaking a C-N bond is rate-determining have
been estimated to be in the range of 15kobs ) 1.01-1.03.10 What
makes interpretation of the present nitrogen isotope effect
difficult is the possibility that proton transfer to the nitrogen
may occur either simultaneous with (structure III) or prior to
C-N bond breaking; either case will lower the observed isotope
effect below the above stated range. As a result it is probably
better to estimate 15k1 from the oxygen leaving group isotope
effects on the analogous hydrolysis reactions of methyl
formate, where formation of T- is known to be largely rate-
The alkaline hydrolysis of formamide at low hydroxide
concentrations is more complex than that of methyl formate
because the transition states for formation and destruction of
T- are of comparable energy. In addition, transition states for
both formation and destruction of T- are relatively late (see
the formyl hydrogen isotope effect discussion). At this point it
is only possible to do a crude analysis of the three individual
18
18
isotope effects (18Keq(nuc)
,
k
, and
1(nuc)
k
) using eq 5 and
3(nuc)
appropriate models. One must start with the assumption that
(21) Melander, L. Isotope Effects on Reaction Rates; Ronald Press: New
York, 1960.
(22) Cromartie, T. H.; Swain, G. G. J. Am. Chem. Soc. 1976, 98, 2962.
(23) Stefandis, D.; Jencks, W. P. J. Am. Chem. Soc. 1993, 115, 6045.
(24) Paneth, P.; O′Leary, M. H. J. Am. Chem. Soc. 1991, 113, 1691.
(25) Hogg, J. L.; Rodgers, J.; Kovach, I.; Schowen, R. L. J. Am. Chem.
Soc. 1980, 102, 79.
determining. The observed methoxyl oxygen isotope effects for
18
methyl formate range from
k
) 1.009 for alkaline
obs(lg)

Hydrolysis of Formamide
J. Am. Chem. Soc., Vol. 121, No. 18, 1999 4363
hydrolysis to 18kobs(lg) ) 1.001 for acid-catalyzed hydrolysis.17,18
These models may actually overestimate the kinetic isotope
effect on formation of T- because breakdown of T- (which
can show a large normal isotope effect near 1.06)17 might still
contribute slightly to the overall rate and the observed isotope
effect. During formation of T- the leaving oxygen experiences
a loss of π bond character on going to tetrahedral geometry (a
normal effect) and gains a new O-C-O bending mode (an
inverse effect). Both effects are small and tend to cancel one
another. For formamide the isotope effect on the loss of π bond
character to the nitrogen should be larger than in the methyl
formate case because amides are more highly conjugated than
esters and because the transition state is later in the formamide
case. On the other hand, this later transition state will also
increase the effect of creating new bending modes in the
transition state, thereby increasing the inverse contribution to
the overall observed nitrogen isotope effect. In addition, a
difference of 1 amu for the nitrogen isotopes (vs 2 amu for the
oxygen isotopes) will make maximum observed nitrogen isotope
effects smaller than those for oxygen. Therefore, it is not
unreasonable to assume that 15k1 will be in the range 1.00-
1.01, similar in magnitude to those listed above for hydrolysis
of methyl formate.
1.03 range. In this light the calculated isotope effects in Table
4 seem reasonable.
Summary. The results of the extensive isotope effect study
in this paper allow a detailed picture of the transition-state
structure for the alkaline hydrolysis of formamide under
conditions where the reaction is first-order in hydroxide. Isotope
exchange experiments demonstrate that the tetrahedral inter-
mediate partitions almost equally between formamide and
formate, indicating that the transition states for both steps of
the mechanism are of nearly equal energy. The transition state
for formation of the tetrahedral intermediate is rather late and
sp3-like. The nucleophile is water, presumably one of the water
molecules in the hydration sphere of hydroxide. This is in
agreement with earlier findings for alkaline hydrolysis of esters.
The most likely transition state for decomposition of the
tetrahedral intermediate is also somewhat late, having consider-
able sp2 character and a high degree of C-N bond cleavage. It
is likely that C-N bond cleavage is accompanied by a concerted
proton transfer to the nitrogen (via water) and that this proton
transfer is nearly complete in the transition state. A stepwise
protonation of the nitrogen prior to leaving is also possible.
Acknowledgment. I am grateful to Professor W. W. Cleland
of the University of WisconsinsMadison for helpful discussions
and for providing summer laboratory space and supplies (Grant
GM 18938). I thank Isabel Treichel (University of Wisconsin)
for help with mass spectral analyses and Professors Max Wills,
Al Censullo, and Jan Simek (California Polytechnic State
University) for help with capillary GC, GC-MS, and NMR
work, respectively. Partial support for this work was obtained
from the California Polytechnic State University through a State
Faculty Support Grant summer fellowship.
Since the transition state for formation of T- is somewhat
late and the nitrogen kinetic isotope effect on step 1 is a
secondary one, it is again possible to make the crude assumption
that 15Keq and 15k1 are of comparable magnitude, which allows
estimation of 15k3 from eq 5. The calculated isotope effect on
step 2 becomes 15k3 ) 0.988-1.008 (see Table 4). Are the
calculated isotope effects for breakdown of T- reasonable? As
mentioned above breaking a C-N bond will give an estimated
nitrogen isotope effect of 1.01-1.03, depending on the degree
of bond breaking in the transition state. The simultaneous
formation of the N-H bond in the transition state is expected
to give an inverse contribution to the overall observed isotope
effect because almost all of the reaction coordinate motion will
be on part of the proton. This will lower the observed isotope
effect considerably, again depending on the degree of N-H
bond formation in the transition state. Therefore, it is safe to
propose that 15k3 will be considerably lower than the 1.01-
Supporting Information Available: Differential equations
for determining the fraction of formamide molecules which have
undergone carbonyl oxygen exchange with solvent during
alkaline hydrolysis. (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
JA9901819