The mandelamide keto-enol system in aqueous solution. Generation of the enol by hydration of phenylcarbamoylcarbene

Article abstract of DOI:10.1021/ja021017f

Flash photolysis of diazophenylacetamide in aqueous solution produced phenylcarbamoylcarbene, whose hydration generated a transient species that was identified as the enol isomer of mandelamide. This assignment is based on product identification and the shape of the rate profile for decay of the enol transient, through ketonization to its carbonyl isomer, as well as by the form of acid - base catalysis of and solvent isotope effects on the decay reaction. Rates of enolization of mandelamide were also determined, by monitoring hydrogen exchange at its benzylic position, and these, in combination with the ketonization rate measurements, gave the keto - enol equilibrium constant pK_E = 15.88, the acidity constant of the enol ionizing as an oxygen acid, pQ_a^E = 8.40, and the acidity constant of the amide ionizing as a carbon acid pQ_a^K = 24.29. (These acidity constants are concentration quotients applicable at ionic strength = 0.10 M.) These results show the enol content and carbon acid strength of mandelamide, like those of mandelic acid and methyl mandelate, to be orders of magnitude less than those of simple aldehydes and ketones; this difference can be attributed to resonance stabilization of the keto isomers of mandelic acid and its ester and amide derivatives, through electron delocalization into their carbonyl groups from the oxygen and nitrogen substituents adjacent to these groups. The enol of mandelamide, on the other hand, again like the enols of mandelic acid and methyl mandelate, is a substantially stronger acid than the enols of simple aldehydes and ketones. This difference can be attributed to the electronegative nature of the oxygen and nitrogen substituents geminal to the enol hydroxyl group in the enols of mandelic acid and its derivatives; in support of this, the acidity constants of these enols correlate well with field substituent constants of these geminal groups.

Full text of DOI:10.1021/ja021017f

Published on Web 12/06/2002
The Mandelamide Keto-Enol System in Aqueous Solution.
Generation of the Enol by Hydration of
Phenylcarbamoylcarbene
Y. Chiang,† H.-X. Guo,† A. J. Kresge,*,† J. P. Richard,‡ and K. Toth‡
Contribution from the Department of Chemistry, UniVersity of Toronto,
Toronto, Ontario, M5S 3H6, Canada, and Department of Chemistry, UniVersity at Buffalo,
SUNY, Buffalo, New York 14260-3000
Received July 25, 2002; Revised Manuscript Received October 2, 2002;
E-mail: akresge@chem.utoronto.ca
Abstract: Flash photolysis of diazophenylacetamide in aqueous solution produced phenylcarbamoylcarbene,
whose hydration generated a transient species that was identified as the enol isomer of mandelamide.
This assignment is based on product identification and the shape of the rate profile for decay of the enol
transient, through ketonization to its carbonyl isomer, as well as by the form of acid-base catalysis of and
solvent isotope effects on the decay reaction. Rates of enolization of mandelamide were also determined,
by monitoring hydrogen exchange at its benzylic position, and these, in combination with the ketonization
rate measurements, gave the keto-enol equilibrium constant pKE ) 15.88, the acidity constant of the enol
ionizing as an oxygen acid, pQEa ) 8.40, and the acidity constant of the amide ionizing as a carbon acid
pQKa ) 24.29. (These acidity constants are concentration quotients applicable at ionic strength ) 0.10 M.)
These results show the enol content and carbon acid strength of mandelamide, like those of mandelic acid
and methyl mandelate, to be orders of magnitude less than those of simple aldehydes and ketones; this
difference can be attributed to resonance stabilization of the keto isomers of mandelic acid and its ester
and amide derivatives, through electron delocalization into their carbonyl groups from the oxygen and
nitrogen substituents adjacent to these groups. The enol of mandelamide, on the other hand, again like
the enols of mandelic acid and methyl mandelate, is a substantially stronger acid than the enols of simple
aldehydes and ketones. This difference can be attributed to the electronegative nature of the oxygen and
nitrogen substituents geminal to the enol hydroxyl group in the enols of mandelic acid and its derivatives;
in support of this, the acidity constants of these enols correlate well with field substituent constants of
these geminal groups.
The enol isomers and enolate ions of simple carboxylic acids,
esters, and amides play prominent roles in many important
chemical and biological reactions. Such enols and enolate ions,
however, are also very unstable, both kinetically and thermo-
dynamically,1 and relatively little directly obtained information
about these elusive, short-lived substances has consequently
become available. This stands in striking contrast to the
remarkable expansion in the chemistry of enols and enolate ions
of simple aldehydes and ketones that has taken place over the
past two decades.2 The difference is undoubtedly due to the
much greater instability of carboxylic acid enols over the already
quite unstable enols of simple aldehydes and ketones.
we have employed such techniques to investigate the enol and
enolate ions of mandelic acid, 1,3 and methyl mandelate, 2.4
We now add to that a corresponding study of the mandelamide,
3, keto-enol/enolate ion system.
We generated the enol of mandelamide by photolysis of
diazophenylacetamide, 4. Irradiation of a diazocarbonyl com-
pound such as this generally gives a Wolff rearrangement
producing a ketene, 5, which, in aqueous solution, would
become hydrated to a carboxylic acid, in the present case to
phenylglycine, 6, eq 1. However, when the migratory aptitude
Flash photolytic fast reaction techniques have proven to be
especially useful in studying unstable short-lived species, and
† University of Toronto.
‡ University at Buffalo.
(1) Hegarty, A. F.; O’Neil, P. In The Chemistry of Enols; Rappoport, Z., Ed.;
Wiley: New York, 1990; Chapter 10. Kresge, A. J. Chem. Soc. ReV. 1996,
25, 275-280.
(2) See, for example: The Chemistry of Enols; Rappoport, Z., Ed.; Wiley: New
York, 1990.
(3) Chiang, Y.; Kresge, A. J.; Popik, V. V.; Schepp, N. P. J. Am. Chem. Soc.
1997, 119, 10203-10212.
(4) Chiang, Y.; Kresge, A. J.; Schepp, N. P.; Xie, R.-Q. J. Org. Chem. 2000,
65, 1175-1180.
9
10.1021/ja021017f CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 187-194
187

A R T I C L E S
Chiang et al.
and mandelate ion; 60 transients were usually collected. The baselines
of the NMR spectra were subjected to a first-order drift correction before
determination of integrated peak areas, and areas were measured relative
to an internal tetramethylammonium ion standard.
of the potentially migrating group is poor, another process
intervenes,5 in which loss of nitrogen gives an R-carbonylcar-
The mandelamide hydrolysis reaction consumes hydroxide ion, and,
to minimize the effect of this on hydroxide ion concentrations of the
hydrogen exchange reaction mixtures, mandelamide substrate was
always supplied at initial concentrations at least 10 times less than
hydroxide ion concentrations. For sodium hydroxide concentrations of
0.08 M and greater, this was achieved by using an initial mandelamide
concentration of 0.008 M and performing the NMR analysis directly
on these reaction mixtures. For sodium hydroxide concentrations less
than 0.08 M, reaction mixtures with initial mandelamide concentrations
of 0.0008 M were used, and samples of these mixtures were concen-
trated before NMR spectra were taken. In applying this latter method,
5 mL aliquots of reaction mixture were withdrawn at appropriate
reaction times, their pD was adjusted to ca. 6 with 2 M CH3CO2D in
D2O solution, and these samples were then concentrated under reduced
pressure to a volume of ca. 0.7 mL. These concentrates were either
subjected to NMR analysis directly or they were kept frozen until NMR
analysis was performed a few days later. Control experiments showed
that such frozen storage had no significant effect on the NMR analysis
and also that the two kinetic methods produced identical results.
Observed first-order rate constants were determined from slopes of
semilogarithmic plots of NMR peak area versus time; such plots were
accurately linear for at least two reaction halftimes.
bene, 7, whose hydration provides an enol, 8, that in this case
would be the enol of mandelamide, 3, eq 2. We have found
that mandelamide is in fact the major product formed by flash
photolysis of diazophenylacetamide; some phenylglycine is
produced as well, but this is made in very minor amounts.
Experimental Section
Materials. Diazophenylacetamide (4) was prepared by a Bamford-
Stevens reaction6 on benzoylformamide7 by converting the latter to its
tosyl hydrazone and then cleaving the hydrazone with sodium
hydroxide. The product, a bright yellow solid, mp ) 129-130 °C, was
obtained in 80% yield. 1H NMR (200 MHz, CDCl3): δ/ppm ) 7.46-
7.24 (m, 5H), 5.41 (bs, 2H). 13C NMR (50 MHz, CDCl3): δ/ppm )
167.3, 129.8, 127.8, 127.4, 126.5, 64.7. HRMS: m/e ) 161.0603
(calcd), 161.0591 (found).
Product Analysis. Product analyses were conducted by HPLC using
a Varian Vista 5500 instrument with a Novopak C18 reverse-phase
column and methanol-water (30/70 ) v/v) as the eluent. Reaction
solutions, containing diazophenylacetamide substrate at concentrations
similar to those used for the rate measurements, were subjected to a
single flash from our conventional flash photolysis system,8 and
products were identified by comparing retention times and UV spectra
with those of authentic samples.
All other materials were best available commercial grades.
Kinetics, Flash Photolysis. Flash photolytic rate measurements were
made using conventional (flash lamp)8 and laser (λ ) 248 nm)9 systems
that have already been described.8,9 Substrate concentrations in the
solutions upon which the rate measurements were made were on the
order of 10-4 M, and the temperature of these solutions during the rate
measurements was controlled at 25.0 ( 0.05 °C. Reactions were
monitored by following UV light absorbance changes, for the carbo-
nylcarbene hydration reaction at λ ) 340 nm (absorbance decay in
HClO4 solutions and in CH3CO2H and H2PO4 buffers; absorbance rise
Results
Product Identification. HPLC product analyses were con-
ducted in wholly aqueous 1 × 10-4 M HClO4 and 1 × 10-3
NaOH solutions and also in H2PO4- ([H+] ) 2 × 10-7 M) and
(CH2OH)3CNH3+ ([H+] ) 8 × 10-9 M) buffers. These analyses
showed that a single pulse from our conventional flash pho-
tolysis apparatus converted an average of 27% of the diazo-
phenylacetamide substrate into products. Of the 27%, 25% was
mandelamide (3) and 2% was phenylglycine (6); there were
also trace amounts of other unidentified substances.
+
in NH4 buffers and NaOH solutions) and for the enol ketonization
reaction at λ ) 300-310 nm (absorbance decay in all solutions
examined). Observed first-order rate constants were obtained by least-
squares fitting of a single exponential function when rates of the
carbonylcarbene hydration and enol ketonization reaction were suf-
ficiently different, and by least-squares fitting of a double exponential
function when they were not.
Kinetics, Hydrogen Exchange. Rates of deuterium incorporation
from D2O solvent into the benzylic carbon-hydrogen bond of man-
delamide were measured by monitoring the decrease in intensity of
the benzylic 1H NMR signal of mandelamide at δ ) 5.13 ppm. In the
sodium hydroxide solutions used for this purpose, hydrolysis of the
amide function of mandelamide also took place, and, to enable
determination of hydrogen exchange rate constants from observed rate
constants for loss of the mandelamide NMR signal, the 1H NMR signal
of mandelate ion produced by hydrolysis at δ ) 4.93 ppm was
monitored as well. Measurements were made at 25 °C using a Varian
Unity Inova 500 NMR spectrometer operating at 500 MHz. A relaxation
delay between pulses of 57 s was used, which is 10-fold greater than
the measured relaxation times of the benzylic protons of mandelamide
These results show that the major process by far occurring
in the present flash photolysis study is that shown in eq 2,
involving carbonylcarbene formation and hydration followed
by enol ketonization. A very minor amount of the photo-
Wolff reaction of eq 1, however, does take place as well. This
is similar to the results obtained for flash photolysis of
diazophenylacetic acid, 9, where formation of phenylcarboxy-
carbene, 10, and hydration of that to mandelic acid enol, 11,
followed by ketonization of the enol, eq 3, was found to be the
principal reaction path, with only a minor amount of the
corresponding photo-Wolff process taking place.3 Flash pho-
(5) Chiang, Y.; Jefferson, E. A.; Kresge, A. J.; Popik, V. V.; Xie, R.-Q. J.
Phys. Org. Chem. 1998, 11, 610-613.
(6) Regitz, M.; Maas, G. Diazo Compounds Properties and Synthesis; Academic
Press: New York, 1986; Chapter 9.
(7) Photis, J. M. Tetrahedron Lett. 1980, 21, 3539-3540.
(8) Chiang, Y.; Hojatti, M.; Keeffe, J. R.; Kresge, A. J.; Schepp, N. P.; Wirz,
J. J. Am. Chem. Soc. 1987, 109, 4000-4009.
(9) Andraos, J.; Chiang, Y.; Huang, C.-G.; Kresge, A. J.; Scaiano, J. C. J. Am.
Chem. Soc. 1993, 115, 10605-10610.
tolysis of the methyl ester of diazophenylacetic acid, 12, on
the other hand, gave only the carbonylcarbene route leading to
9
188 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Mandelamide Keto−Enol System in Aqueous Solution
A R T I C L E S
acidity constants and activity coefficients recommended by
Bates.11
This rate profile, just as those for the hydration of carbon-
ylcarbenes derived from diazophenylacetic acid, eq 3, and
methyl diazophenylacetate, eq 4, shows acid-catalyzed, uncata-
lyzed, and base-catalyzed portions. It may therefore be analyzed
using the simple rate law shown as eq 6. Least-squares fitting
of this expression gave kH ) (2.97 ( 0.09) × 1010 M-1 s-1
,
+
kprof ) kH+[H+] + k + k -[HO-]
(6)
o
HO
9
ko ) (2.11 ( 0.13) × 106 s-1, and kHO ) (7.74 ( 0.23) × 10
-
M-1 s-1. These results are numerically similar to those obtained
for previous carbonylcarbene hydrations,3,4 which reinforces
assignment of the present process to the phenylcarbamoylcar-
bene hydration reaction.
Figure 1. Rate profiles for the hydration of phenylcarbamoylcarbene (∆)
and ketonization of mandelamide enol (O) in aqueous solution at 25 °C.
Enol Ketonization, Rate Profile. Rates of the slower
absorbance change produced by flash photolysis of diazo-
phenylacetamide, attributable to ketonization of mandelamide
methyl mandelate (2) shown in eq 4, with no detectable amount
of photo-Wolff reaction.4
enol (eq 2), were measured in wholly aqueous HClO4 and NaOH
+
solutions, as well as in CH3CO2H, H2PO4-, (CH2OH)3CNH3
,
+
and NH4 buffers. A range of concentrations for each kind of
solution was once again employed, replicate measurements were
made, and the ionic strength was maintained constant at 0.10
M. These data are summarized in Tables S4-S6.10
Carbonylcarbene Hydration. Flash photolysis of diazo-
phenylacetamide produced biphasic light absorbance changes,
consistent with the presence of two reactive intermediates in
the reaction scheme, eq 2, required by formation of mandelamide
as the principal reaction product. The faster of these absorbance
changes, with lifetimes in the submicrosecond range, may be
assigned, in analogy to previous studies,3,4 to hydration of
phenylcarbomaylcarbene, 7, itself formed within the laser pulse.
This hydration produces mandelamide enol, 8, whose keton-
ization gives the second somewhat slower absorbance change.
Rates of the faster carbene hydration reaction were measured
in wholly aqueous HClO4 and NaOH solutions and in CH3-
CO2H, H2PO4- and NH4+ buffers. For each kind of solution, a
range of concentrations was employed, and replicate measure-
ments were made at each concentration; the ionic strength of
all reaction solutions was maintained at 0.10 M by adding
sodium perchlorate as required. These data are summarized in
Tables S1-S3.10
The measurements in buffers were done in series of solutions
of constant buffer ratio and constant ionic strength but varying
total buffer concentration; the hydrogen ion concentrations along
a given buffer series therefore remained constant. Observed first-
order rate constants within a buffer series were found to increase
linearly with increasing buffer concentration, and the data were
therefore analyzed by least-squares fitting of the buffer dilution
expression shown in eq 5. The zero-buffer-concentration
intercepts obtained in this way, kint, together with the rate
Mandelamide enol can exist in two stereoisomeric forms as
(Z)-1-amino-2-hydroxy-2-phenylethenol, 13, or (E)-1-amino-2-
hydroxy-2-phenylethenol, 14. These two isomers would pre-
sumably ketonize at somewhat different rates, leading to
deviations from simple first-order enol decay. No such com-
plications were observed, which suggests that only one enol
stereoisomer was formed in the carbonylcarbene hydration
reaction. We have no information, however, that allows us to
decide which isomer that is.
The measurements in buffers were again done in series of
solutions of constant buffer ratio but varying total buffer
concentration, and observed first-order rate constants along a
buffer series again increased linearly with increasing buffer
concentration. The data were therefore once more analyzed by
least-squares fitting of eq 5. The zero-buffer-concentration
intercepts obtained by this treatment, together with the rate
constants measured in HClO4 and NaOH solutions, are displayed
as the lower rate profile of Figure 1.
This rate profile is similar to those observed before for
ketonization of the enols of mandelic acid3 and methyl man-
delate.4 Like those rate profiles, it may be interpreted in terms
of the known reaction mechanism for enol ketonization involv-
ing rate-determining proton transfer from any available acid to
the â-carbon atom of the enol or its enolate ion.12 Because this
profile refers to reaction through solvent related species in
aqueous solution, the available acids will be water and the
kobs ) kint + kbuff[buffer]
(5)
constants measured in HClO4 and NaOH solutions, are displayed
as the upper rate profile of Figure 1. Hydrogen ion concentra-
tions of the buffer solutions needed for this purpose were
obtained by calculation using literature values of the buffer
(11) Bates, R. G. Determination of pH Theory and Practice; Wiley: New York,
1973; p 49.
(12) Keeffe, J. R.; Kresge, A. J. In The Chemistry of Enols; Rappoport, Z., Ed.;
Wiley: New York, 1990; Chapter 7.
(10) Supporting Information: see paragraph at the end of this paper regarding
availability.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 189

A R T I C L E S
Chiang et al.
hydronium ion, written here as H+. The present enol, however,
unlike those of mandelic acid and methyl mandelate, contains
an amino group which might be protonated in acid solution,
thus introducing another enol species. The resulting reaction
scheme is shown in eq 7.
should occur with a rate constant 104 times greater than that
for the enol of mandelic acid. This gives the mandelamide enol
reaction the rate constant kKH+ ) 9 × 107 M-1 s-1, which would
provide the rate profile for this reaction with an acid-catalyzed
component that should be clearly visible at acidities of [H+] )
10-5 M and higher. Figure 1 shows that no such component
was observed, and that means that enol was being shifted from
its reactive neutral form (8) into some unreactive species. A
good candidate for this unreactive species is of course the amine-
protonated form 16. This interpretation then assigns the “un-
catalyzed” reaction plateau observed in the acid region of the
rate profile to ketonization beginning with amine-protonated enol
as the initial state. This species first ionizes to give neutral enol
and H+, and that is then followed by rate-determining â-carbon
of the enol by H+. Because H+ is first produced in a
preequilibrium and is then used up in the rate-determining step,
the overall process is independent of [H+] and has the
appearance of an uncatalyzed reaction.
The presence of the amino group in the present enol also
raises the question of whether the formally neutral enol form is
the uncharged species 8 or the zwitterion 15. Insight into this
On the low acid side of this acid-region plateau, there is a
diagonal rising section in which neutral enol is the initial state.
Ketonization here occurs by equilibrium ionization of the
hydroxyl group of the enol to give the much more reactive
enolate ion,16 followed by rate-determining â-carbon protonation
of that by H2O. Because the H+ produced by the ionization is
now not used up in the rate-determining step, the overall process
is inversely proportional to [H+], giving an apparent hydroxide
ion catalysis. Finally, at sufficiently low acidities, the position
of the enol to enolate ionization equilibrium shifts to the ion
side, and the advantage of converting a less reactive to a more
reactive substrate species is lost. The result is simple protonation
of the enolate ion by H2O, giving the final horizontal profile
segment.
matter can be obtained through knowledge of the relative acid
strengths of the ammonio and hydroxyl groups in the fully
protonated enol form, 16: if the ammonio group is the stronger
acid, its ionization will occur in preference to that of the
hydroxyl group, and the neutral enol form 8 will predominate;
if, on the other hand, the hydroxyl group is the stronger acid,
its ionization will occur, and the zwitterionic form 15 will
predominate. Simple ammonium ions with pKa = 10 are
considerably more acidic than simple alcohols with pKa = 16.
This difference, however, will not carry over intact to the
ammonio and hydroxyl groups of 16 because each of these
groups in 16 has a (different) geminal substituent that will
modify its acid strength.
These acid strength modifying effects can be estimated using
a Hammett-type σ-F relationship. A critical analysis of the use
of such relationships to correlate geminal substituent effects on
the pKa’s of simple ammonium ions and alcohols concluded
that F ) 8.4 for both kinds of acid and that inductive sigma
constants, σI, work well for this purpose.13 Using σI(NH3+) )
0.60 and σI(OH) ) 0.2514 then gives ∆pKa ) 5.0 as the acid
strengthening effect of the geminal ammonio group on the
acidity of the hydroxyl group of 16 and ∆pKa ) 2.1 as the acid
strengthening effect of the geminal hydroxyl group on the acidity
of the ammonio group of 16. These modifying effects reduce
the 6 pK unit difference between the acid strength of simple
ammonium ions and alcohols noted above, but they still leave
the ammonio group of 16 a stronger acid than its hydroxyl group
by 3 pK units. This difference would seem to be large enough
to compensate for any deficiency in the method used to estimate
the modifying effects and to leave little doubt that the uncharged
species 8 and not the zwitterion 15 is the predominant form of
the neutral enol.
The rate law that applies to this reaction scheme is shown in
eq 8, whose rate and equilibrium constants are defined by eq 7.
kHK+QN[H+]2 + k′KQN QE
[H+]2 + QaN[H+] + QaN QaE
a
o
a
a
kobs
)
(8)
Least-squares fitting of this expression gave kKH ) (1.74 (
+
0.56) × 106 M-1 s-1, k′K ) (2.56 ( 0.06) × 106 s-1
,
o
QNa ) (8.49 ( 3.13) × 10-5, pQaN ) 4.07 ( 0.16,17 and QaE )
(3.95 ( 0.18) × 10-9 M, pQaE ) 8.40 ( 0.02.17 Because
kKH and QNa are somewhat correlated,18 they are not well
+
determined and have sizable associated uncertainties; k′K and
o
QEa , on the other hand, are well determined.
Enol Ketonization, Buffer Catalysis. The slopes of the
buffer dilution plots kbuff, obtained by least-squares fitting of
eq 5, were separated into their general acid, kHA, and general
base, kB, components with the aid of eq 9, in which fA is the
fraction of buffer present in the acid form.19 Least-squares fitting
kbuff ) kB + (kHA - kB)fA
(9)
of this expression gave results that showed general base catalysis
+
in the CH3CO2H, H2PO4-, and (CH2OH)3CNH3 buffers and
A σ-F relationship that successfully correlates rates of
protonation of a large number of carbon-carbon double bonds15
predicts that â-carbon protonation of mandelamide enol by H+
(15) Csizmadia, V. M.; Koshy, K. M.; Lau, K. C. M.; McClelland, R. A.;
Nowlan, V. J.; Tidwell, T. T. J. Am. Chem. Soc. 1979, 101, 974-979.
(16) Pruszynski, P.; Chiang, Y.; Kresge, A. J.; Schepp, N. P.; Wirz, J. J. Phys.
Chem. 1986, 90, 3760-3766. Chiang, Y.; Kresge, A. J.; Santaballa, J. A.;
Wirz, J. J. Am. Chem. Soc. 1988, 110, 5506-5510.
(17) This is a concentration acidity constant applicable at ionic strength ) 0.10
M.
(13) Fox, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1974, 96, 1436-1449.
(14) Hine, J. Structural Effects on Equilibria in Organic Chemistry; Wiley-
Interscience: New York, 1975; p 98.
9
190 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Mandelamide Keto−Enol System in Aqueous Solution
A R T I C L E S
Table 2. Summary of Rate and Equilibrium Constantsa
Table 1. General Acid Catalytic Coefficients for the Ketonization
of Mandelamide Enol in Aqueous Solution at 25 °Ca
HA
k′ K/M-1 s-1
HA
CH3CO2H
H2PO4
7.10 × 108
3.32 × 108
1.40 × 108
4.37 × 107
-
+
(CH2OH)3CNH3
+
NH4
a Ionic strength ) 0.10 M.
general acid catalysis in the NH4+ buffers. These forms of buffer
catalysis are consistent with the fact that the H+ concentrations
+
of the CH3CO2H, H2PO4-, and (CH2OH)3CNH3 buffers lay
in the region of the ketonization rate profile showing apparent
hydroxide ion catalysis, whereas those of the NH4+ buffers lay
in the final plateau profile region. The general base catalytic
coefficients may therefore be attributed to an analogue of the
apparent hydroxide-ion-catalyzed process, that is, to general
base-assisted ionization of the enol to enolate ion followed by
rate-determining carbon protonation of enolate by the general
base conjugate acid, eq 10. The general acid catalytic coefficient
obtained from the NH4+ buffer data, on the other hand, may be
attributed to direct carbon protonation of the enolate ion by
+
NH4
.
The rate law for the reaction route of eq 10 leads to the
relationship kH′KA ) kBQHA/QE, in which kB is the general base
a
a
catalytic coefficient obtained by applying eq 9. Values of kH′KA
obtained in this way are summarized in Table 1. It may be seen
that kH′KA diminishes progressively with diminishing general
acid strength, as is expected for a rate-determining proton
transfer reaction. Although the acids are limited in number and
of several different charge types, they nevertheless give a fairly
good Bro¨nsted relation with a low exponent, R ) 0.33 ( 0.04,
as is expected for such very fast reactions. This provides good
support for the molecular interpretation of the experimental data
given.
Enol Ketonization, Isotope Effects. Further support for this
interpretation is given by solvent isotope effects on the
ketonization reaction. These were provided by rate measure-
ments made in the acidic and basic plateau regions of the
ketonization rate profile using D2O solutions of DClO4 and
NaOD. These data are summarized in Tables S4 and S5.10
The measurements in basic solution gave the solvent isotope
effect kH O/kD O ) 5.88 ( 0.13. This is similar to isotope effects
a 25 °C; ionic strength ) 0.10 M.
direction, generated in the solvation shell of the hydroxide ion
being formed by proton donation from a water molecule.20
The measurements in acidic solutions gave the solvent isotope
effect kH O/kD O ) 6.71 ( 0.39. The large value of this isotope
2
2
2
2
effect is also consistent with the molecular mechanism assigned
to this region of the rate profile, i.e., acid ionization of the amine-
protonated enol followed by carbon protonation of the neutral
enol by a hydronium ion. That makes this isotope effect a
composite ratio consisting of an isotope effect on the acid
in this profile region determined for the ketonization of mandelic
acid enol (kH O/kD O ) 6.9)3 and methyl mandelate enol (kH O
/
2
2
2
kD O ) 7.7).4 The large values of these effects are as expected
2
for the molecular mechanism assigned to this part of the rate
profile, i.e., carbon protonation of an enolate ion by a water
molecule. The primary kinetic isotope effect on such a process
is augmented by a secondary effect in the normal (kH/kD > 1)
ionization constant QaN times that on the rate constant kHK
.
+
Solvent isotope effects on the ionization of acids with acid
strengths comparable to that of the present substance are on
the order of 3,21 and application of that to the observed ratio
(18) The numerical value of the product kKH+ QaN is fixed by that of the acid-
region rate profile plateau. Separation of this product into its constituent
parts requires the acid-region plateau to be connected to the upward sloping
diagonal which follows by another downward sloping diagonal segment,
but such an additional diagonal segment can barely be discerned.
(19) Values of kbuff, before analysis by eq 9, were adjusted to reaction through
either wholly ionized or wholly un-ionized enol, as required.
leaves kKH+/kK ) 2.2. This is a reasonable value for such an
+
D
(20) 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, Chapter 4.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 191

A R T I C L E S
Chiang et al.
effect inasmuch as isotope effects on hydron transfer from the
hydronium ion have an inverse secondary component and
consequently are small.22
Amide Enolization. Rates of enolization of mandelamide
were determined by measuring rates of hydrogen exchange at
the benzylic position of the amide. This was done by monitoring
the reduction of the 1H NMR signal of the mandelamide benzyl
proton at δ ) 5.13 ppm as this proton was replaced by deuterium
incorporated from a D2O solvent. Measurements were made in
KOD solutions over the concentration range [KOD] ) 0.01-
0.08 M at a constant ionic strength of 0.10 M and at [KOD] )
0.08 and 0.20 at a constant ionic strength of 0.20 M. The data
are summarized in Table S7.10
In the solutions used for these measurements, hydrolysis of
the amide functional group of mandelamide also occurred, eq
11, at rates comparable to those of hydrogen exchange. This
Figure 2. Relationship between deuterioxide ion concentration and rates
of incorporation of deuterium into the benzylic position of mandelamide in
D2O solutions of KOD at 25 °C.
ionization would put negative charge on the substrate and
thereby inhibit the negative-charge-forming enolization reaction.
The reaction scheme that applies to such a situation is shown
in eq 12, and its rate law is given by eq 13; Kb is the basicity
constant of this inhibiting group, and kEDO is the deuterioxide
-
produced mandelate ion, which gave a separate 1H NMR signal
at δ ) 4.93 ppm. Observed rates of loss of the mandelamide
signal at δ ) 5.13 ppm were therefore sums of rate constants
for the hydrogen exchange and amide hydrolysis reactions: kobs
) kexch + khydr . As the reaction scheme of eq 11 illustrates,
both exchanged and unexchanged mandelamide eventually
underwent complete hydrolysis, but, whereas the exchanged
substrate gave mandelate ion with deuterium at the benzylic
position, unexchanged substrate gave mandelate ion with
protium at this position. The fraction of observed rates due to
hydrolysis, fhydr , could therefore be ascertained by determining
the amount of protium remaining at the benzylic position of
the mandelate ion hydrolysis product after hydrolysis was
complete. This fraction could then be used to separate observed
rate constants into their constituent hydrolysis and exchange
kEDO
Kb
\
S- + D2O {
} SD + DO-
-8 enol
(12)
(13)
kexch ) kEDO- K [DO-]/(K + [DO-])
b
b
ion catalytic coefficient for the enolization reaction. Least-
squares fitting of this expression gave Kb ) (8.41 ( 0.38) ×
10-2 M24 and kEDO ) (1.17 ( 0.02) × 10-4 M-1 s-1. These
-
parameters were used to draw the correlation line shown in
Figure 2; it may be seen that the experimental data fit this
reaction scheme well.
This inhibition of hydroxide-ion-catalyzed enolization could
be produced by either ionization of the benzylic hydroxyl group
of mandelamide, eq 14, or ionization of its amide group, eq 15.
parts: khydr ) fhydr kobs and kexch ) (1 - fhydr)kobs
.
Values of fhydr were obtained by comparing the areas of the
benzylic, Abz, and aromatic, Aar, 1H NMR signals of mandelate
ion obtained after 10 hydrolysis halftimes: fhydr ) Abz/(Aar/5);
the factor 5 accounts for the fact that there are 5 times as many
aromatic as benzylic hydrogens. These data are also summarized
in Table S7. Six determinations gave the average value fhydr
)
0.500 ( 0.005.
Hydrogen exchange rate constants obtained in this way
increased with increasing hydroxide ion concentration, as
expected for exchange occurring through base-catalyzed eno-
lization. As Figure 2 shows, however, the relationship between
kexch and [DO-] was not linear, but rather showed impending
saturation of hydroxide ion catalysis at the higher values of
[DO-] employed.23 Such behavior is consistent with a hydroxide-
ion-assisted, rapid equilibrium ionization of a weakly acidic
group of the substrate other than its benzylic hydrogen: such
Some insight into which of these it is can be obtained through
estimates of the acidity constants of the two groups. Use of a
σ-F relationship that correlates acidity constants of secondary
(23) To better define this saturation, a set of measurements was made at [DO-]
) 0.20 M. This required an ionic strength, 0.20 M, greater than that, 0.10
M, of the more dilute KOD solutions employed, but control experiments,
done at [DO-] ) 0.082 M with ionic strength ) 0.10 and 0.20 M, showed
that this difference had no significant effect on measured rate constants.
(24) This, strictly speaking, is a concentration quotient, Qb. It is likely, however,
that the activity coefficients of the S- and DO- species of eq 12 are similar,
and, because these species appear on opposite sides of the equilibrium
equation, their activity coefficients will cancel out of the expression for
the thermodynamic basicity constant Kb; Qb is therefore likely to be a good
surrogate for Kb.
(21) Laughton, P. M.; Robertson, R. E. In Solute-SolVent Interactions; Coetzee,
J. F., Ritchie, C. D., Eds.; M. Dekker: New York, 1969; Chapter 7.
(22) Kresge, A. J.; Sagatys, D. S.; Chen, H. L. J. Am. Chem. Soc. 1977, 99,
7228-7233.
9
192 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003

Mandelamide Keto−Enol System in Aqueous Solution
A R T I C L E S
acetone,31 and kDO /kHO ) 1.39 for the enolization of mandelate
alcohols25 gives pKa ) 12.5 for the benzylic hydroxyl group.
No such correlation appears to be available for the acidity
constants of amide N-H bonds, but the pKa of acetamide is
given as 15.1,26 and, on the assumption that the difference
between this value and that for mandelamide is the same as the
pKa difference for the carboxylic acid groups of acetic acid and
mandelic acid, pKa ) 13.8 may be estimated for the N-H bond
of mandelamide.
-
-
ion.32 Application of the rounded average value kDO /kHO
)
)
-
-
1.4 to kEDO- ) (1.17 ( 0.02) × 10-4 M-1 s-1 then gives kEHO
-
(8.32 ( 0.16) × 10-5 M-1 s-1 as the hydroxide ion catalytic
coefficient for the enolization of mandelamide in H2O solution.
Keto-enol equilibration of mandelamide catalyzed by hy-
droxide ion may be formulated as shown in eq 17, and
These estimates predict that the alcoholic hydroxyl group of
mandelamide is more acidic than is its amide N-H bond, which
suggests that it is ionization of the hydroxyl group that produces
inhibition of hydroxide-ion-catalyzed enolization of mandela-
mide. This conclusion is supported by reasonably good agree-
ment between the predicted pKa of this group and the pKa that
can be estimated from the basicity constant, Kb, obtained from
the present experimental results for inhibition of the enolization
reaction.
evaluation of the expression for its equilibrium constant, KE )
(kEHO-/k′K)(Qw/QE),33 then gives KE ) (1.31 ( 0.07) × 10-16
,
pKE ) o15.88 ( a0.02.
The first part of the relationship of eq 17 includes the
ionization of mandelamide as a carbon acid, and the equilibrium
constant for this ionization can be evaluated as QKa ) (kHEO
/
This basicity constant refers to the base ionization reaction
in D2O solution, and it can be converted into an acidity constant
in the same medium by applying the relationship pKa + pKb )
pKw , with pKw ) 14.87 as the autoprotolysis constant of D2O.27
Conversion of this to an acidity constant in H2O solution requires
knowledge of the solvent isotope effect on the ionization of
acids as weak as the present one. Two such values are
available: (Ka)H O/(Ka)D O ) 4.5 for CF3CH2OH (pKa ) 12.4)28
-
k′K)Qw ) (5.16 ( 0.17) × 10-25 M, pQKa ) 24.29 ( 0.01.17
o
Discussion
The keto-enol equilibrium constant for mandelamide deter-
mined in the present study, pKE ) 15.88, is many orders of
magnitude less than those for simple aldehydes and ketones,
for example, pKE ) 6.23 for acetaldehyde,8 pKE ) 8.33 for
acetone,34 and pKE ) 7.96 for acetophenone.35 This striking
difference may be attributed to resonance stabilization of the
keto isomer in the amide system by interaction of its amino
and carbonyl moieties; this interaction is lost in the enol isomer,
and it consequently serves to increase the energy difference
between the keto and enol forms making, KE as small as it is.
The keto-enol equilibrium constant for mandelamide, 3, on
the other hand, is some 3 orders of magnitude greater than that
for its unsubstituted analogue, acetamide, 17, for which the
reliable estimate pKE ) 19.2 has recently been made.36 This is
2
2
and (Ka)H O/(Ka)D O ) 4.8 for CH2ClCH2OH (pKa ) 14.3).29
2
2
These isotope effects, however, are based on a value of pKw
for D2O now known to be in error, and adjustment of the results
using a value of pKw free of this error27 gives isotope effects
whose average is (Ka)H O/(Ka)D O ) 5.2. Application of that to
2
2
the present system then gives pKa ) 13.1 for ionization of the
hydroxyl group of mandelamide in H2O solution. This result is
gratifyingly consistent with the estimate pKa ) 12.5.
Keto-Enol Equilibrium and Carbon Acid Acidity Con-
stants. The equilibrium constant, KE, for mandelamide isomer-
izing to its enol in aqueous solution, eq 16, may be obtained
from the presently determined rate constants for the enolization
and ketonization reactions. To do this, the hydroxide ion
similar to the difference in KE for mandelic acid, 1, and acetic
acid, 18, noted before.3 It can be attributed to the combined
influence of the phenyl and hydroxyl groups in 3 and 1,
inasmuch as comparisons of the phenylacetaldehyde, 19,37 with
acetaldehyde, 20,8 systems and the isochroman-4-one, 21,38 with
enolization catalytic coefficient must first be converted from a
value for the D2O solvent in which it was measured to a value
for the H2O solvent in which the ketonization rate measurements
were made. Solvent isotope effects on hydroxide-ion-catalyzed
reactions have no primary component, but they do have a
secondary component produced by release of the water mol-
ecules solvating the hydroxide ion.20 For the enolization of
carbonyl compounds, this usually gives an inverse isotope effect
(30) Keeffe, J. R.; Kresge, A. J. Can. J. Chem. 1988, 66, 2440-2442.
(31) Pocker, Y. Chem. Ind. (London) 1959, 1383.
(32) Pocker, Y. Chem. Ind. (London) 1958, 1117-1118.
(33) Qw is the concentration quotient autoprotolysis constant of water at the
ionic strength, 0.10 M, of the present measurements. It was evaluated as
Qw ) 1.59 × 10-14 M2 using the thermodynamic value Kw ) 1.00 × 10-14
M2 and activity coefficients recommended by Bates.11
(34) Chiang, Y.; Kresge, A. J.; Schepp, N. P. J. Am. Chem. Soc. 1988, 110, 1,
3977-3980.
-
-
of about 40%. For example, kDO /kHO ) 1.37 for the enolization
of acetaldehyde,30 kDO /kHO ) 1.47 for the enolization of
-
-
(25) Perrin, D. D.; Dempsey, B.; Serjeant, E. P. pKa Predictions for Organic
Acids and Bases; Chapman and Hall: New York, 1981; p 127.
(26) Stewart, R. The Proton: Applications to Organic Chemistry; Academic
Press: New York, 1985; p 71.
(27) Covington, A. K.; Robinson, R. A.; Bates, R. G. J. Phys. Chem. 1966, 70,
3820-3824. Gold, V.; Lowe, B. M. J. Chem. Soc. A 1967, 936-943.
(28) Ballinger, P.; Long, F. A. J. Am. Chem. Soc. 1959, 81, 1050-1053.
(29) Ballinger, P.; Long, F. A. J. Am. Chem. Soc. 1959, 81, 2347-2352.
(35) Keeffe, J. R.; Kresge, A. J.; Toullec, J. Can. J. Chem. 1986, 64, 1224-
1227.
(36) Richard, J. P.; Williams, G.; O’Donoghue, A. C.; Amyes, T. L. J. Am.
Chem. Soc. 2002, 124, 2957-2968.
(37) Chiang, Y.; Kresge, A. J.; Yin, Y. J. Chem. Soc., Chem. Commun. 1989,
869-871.
(38) Chiang, Y.; Kresge, A. J.; Meng, Q.; More O’Ferrall, R. A.; Zhu, Y. J.
Am. Chem. Soc. 2001, 123, 11562-11569.
9
J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003 193

A R T I C L E S
Chiang et al.
A group of simple primary alcohols of structure RCH2OH,
whose pKa’s were deemed to be reliable,42 also gives a good
correlation using these field constants: pKa ) (15.68 ( 0.19)
- (8.92 ( 1.00)F. It is interesting that the sensitivity parameters
obtained in these two correlations are not significantly different,
which implies that substituent effects are transmitted as well
through the sp2 carbon center of enols as through the sp3 carbon
center of alcohols.
The acidity constant of mandelamide ionizing as a carbon
acid determined here, pQaK ) 24.29, is less than those for
simple aldehydes and ketones, for example, pQKa ) 16.73 for
acetaldehyde,8 pQKa ) 19.27 for acetone,34,39 and pQKa ) 18.36
for acetophenone.35,40 The direction of these differences is the
same as that for the keto-enol equilibrium constants (vide
supra), but their magnitude is smaller. This follows from the
formulation of the carbon acid ionization reaction as the closing
leg of a thermodynamic cycle, the other two legs of which are
keto-enol equilibration and ionization of the enol as an oxygen
acid, eq 18. The carbon acid ionization constant is therefore
Figure 3. Correlation of enol acidity constants for mandelic acid, methyl
mandelate, and mandelamide using the field constant F.
tetralin, 22,38 systems show that both phenyl and oxygen
substituents in the â-position raise keto-enol equilibrium
constants.
The acidity constant of mandelamide enol determined here,
pQEa ) 8.40, on the other hand, is several orders of magnitude
greater than those for the enols of simple aldehydes and ketones,
equal to the product of keto-enol equilibrium and enol acidity
constants, QK ) KE QE, and differences in KE QE will be equal
for example, pQEa ) 10.50 for acetaldehyde enol,8 pQaE
)
a
a
a
to the product of differences in KE and QK. Because KE for
10.94 for acetone enol,39 and pQEa ) 10.40 for acetophenone
enol.40 This difference may be attributed to the electronegative
nature of the amino substituent in the mandelamide system, i.e.,
to its electron-withdrawing inductive or field effect: this will
stabilize the negatively charged enolate ion, making the enol a
stronger acid.
a
mandelamide is less than KE for simple aldehydes and ketones
by very large factors, while the differences in QEa are in the
opposite direction by considerably smaller factors, the differ-
ences in QK must be in the same direction as those for KE but
a
by smaller though still quite large factors. This is, of course, is
as observed.
The hydroxyl and methoxyl substituents in the enols of
mandelic acid and methyl mandelate also have electron-
withdrawing inductive or field effects, and these enols are also
more acidic than those of simple aldehydes and ketones. As
Figure 3 shows, the acidity constants of all three of these more
acidic enols are correlated well by the Swain-Lupton field
constants F as modified by Hansch, Leo, and Taft.41 Linear least-
squares analysis gives the relationship pQEa ) (9.08 ( 0.17) -
(8.38 ( 0.66)F.
Acknowledgment. We are grateful to the Natural Sciences
and Engineering Research Council of Canada and the United
States National Institutes of Health for financial support of this
work.
Supporting Information Available: Tables S1-S7 of rate
data (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org.
(39) Chiang, Y.; Kresge, A. J.; Tang, Y. S.; Wirz, J. J. Am. Chem. Soc. 1984,
106, 460-462.
JA021017F
(40) Jefferson, E. A.; Keeffe, J. R.; Kresge, A. J. J. Chem. Soc., Perkin Trans.
2 1995, 2041-2046.
(42) Takahashi, S.; Cohen, L. A.; Miller, H. K.; Peake, E. G. J. Org. Chem.
1971, 36, 1205-1209.
(41) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165-195.
9
194 J. AM. CHEM. SOC. VOL. 125, NO. 1, 2003