Flash Photolytic Decarbonylation and Ring-Opening of 2-(N-(Pentafluorophenyl)amino)-3-phenylcyclopropenone. Isomerization of the Resulting Ynamine to a Ketenimine, Hydration of the Ketenimine, and Hydrolysis of the Enamine Produced by Ring-Opening

Article abstract of DOI:10.1021/jo970765y

Flash photolysis of 2-(N-(pentafluorophenyl)amino)-3-phenylcyclopropenone, 4, in aqueous solution was found to produce N-(pentafluorophenyl)phenylethynamine, 3, by the expected photodecarbonylation reaction and also 2-phenyl-3-(N-(pentafluorophenyl)amino)acrylic acid, 5, by an apparently unprecedented photochemical ring-opening process. The ynamine underwent rapid isomerization to N-(pentafluorophenyl)phenylketenimine, 9, by an acid-catalyzed route that involves rate- determining proton transfer to the β-carbon atom of the ynamine and also by a base-catalyzed route involving equilibrium ionization of the N-H bond of the ynamine to give an ynamide ion followed by rate-determining β-carbon protonation of this ion. Saturation of the base catalysis allowed determination of the acidity constant of the ynamine; the result, pQ_a = 10.23, makes this amine a remarkably strong nitrogen acid. Hydration of the ketenimine 9 gave N-(pentafluorophenyl)phenylacetamide, 6, as the ultimate product produced by this reaction route, and hydrolysis of the aminoacrylic acid 5 gave pentafluoroaniline, 7, and 2-phenylformylacetic acid, 10, which underwent decarboxylation to phenylacetaldehyde, 8, as the ultimate products of this route.

Full text of DOI:10.1021/jo970765y

J . Org. Chem. 1997, 62, 5363-5370
5363
F la sh P h otolytic Deca r bon yla tion a n d Rin g-Op en in g of
2-(N-(P en ta flu or op h en yl)a m in o)-3-p h en ylcyclop r op en on e.
Isom er iza tion of th e Resu ltin g Yn a m in e to a Keten im in e,
Hyd r a tion of th e Keten im in e, a n d Hyd r olysis of th e En a m in e
P r od u ced by Rin g-Op en in g
Y. Chiang, A. S. Grant, H.-X. Guo, A. J . Kresge,* and S. W. Paine
Department of Chemistry, University of Toronto, Toronto, Ontario M5S 3H6, Canada
Received April 30, 1997X
Flash photolysis of 2-(N-(pentafluorophenyl)amino)-3-phenylcyclopropenone, 4, in aqueous solution
was found to produce N-(pentafluorophenyl)phenylethynamine, 3, by the expected photodecar-
bonylation reaction and also 2-phenyl-3-(N-(pentafluorophenyl)amino)acrylic acid, 5, by an appar-
ently unprecedented photochemical ring-opening process. The ynamine underwent rapid isomer-
ization to N-(pentafluorophenyl)phenylketenimine, 9, by an acid-catalyzed route that involves rate-
determining proton transfer to the â-carbon atom of the ynamine and also by a base-catalyzed
route involving equilibrium ionization of the N-H bond of the ynamine to give an ynamide ion
followed by rate-determining â-carbon protonation of this ion. Saturation of the base catalysis
allowed determination of the acidity constant of the ynamine; the result, pQa ) 10.23, makes this
amine a remarkably strong nitrogen acid. Hydration of the ketenimine 9 gave N-(pentafluorophen-
yl)phenylacetamide, 6, as the ultimate product produced by this reaction route, and hydrolysis of
the aminoacrylic acid 5 gave pentafluoroaniline, 7, and 2-phenylformylacetic acid, 10, which
underwent decarboxylation to phenylacetaldehyde, 8, as the ultimate products of this route.
We recently discovered that phenylaminocycloprope-
nones, 1, undergo photolytic decarbonylation to acetylenic
amines, 2, commonly called ynamines, eq 1, and that the
is accompanied by yet another photochemical process
which produces the enamine 5, eq 3. A similar ring-
opening is a well known thermal reaction of cycloprope-
nones,5 but the photochemical process that we have found
appears to be unprecedented.
Part of this work has been published in preliminary
form.6
chemistry of these substances in aqueous solution, where
they are very short-lived, can be examined using flash
photolytic techniques.1 Our studies have shown that the
phenylethynyl group has a powerful weakening effect on
the basicity of ynamines. This base-weakening influence
is the counterpart of a very strong acid-strengthening
effect that the phenylethynyl group has on the ionization
of acetylenic alcohols (ynols),2 and, in keeping with that,
this group confers acidic properties on ynamines as well.
We report here that introduction of yet another acidifying
group, the pentafluorophenyl moiety, has the remarkable
effect of producing an ynamine, 3, that ionizes as an acid
in dilute aqueous solution, eq 2.3,4
Exp er im en ta l Section
Ma ter ia ls. 2-(N-(P en ta flu or op h en yl)a m in o)-3-p h en yl-
cyclop r op en on e, 4, was a sample that had been prepared
before.1b N-(P en ta flu or op h en yl)p h en yla ceta m id e, 6, was
synthesized by treating phenylacetyl chloride with pentafluo-
roaniline, and m eth yl 2-p h en yl-3-(N-(p en ta flu or op h en yl)-
a m in o)a cr yla te, 11, was obtained by allowing a solution of
2-(N-(pentafluorophenyl)amino)-3-phenylcyclopropenone in 5:1
methanol:acetone to stand at room temperature overnight.7
These substances were characterized by their mass and NMR
spectra; details are given in Table S1.8 All other materials
were best available commercial grades.
Kin etics. Flash photolytic rate measurements were made
using conventional flash-lamp and eximer-laser systems that
have already been described.9 Excitation in the conventional
system9a was provided by a pair of xenon lamps that produced
We have also found that photodecarbonylation of the
cyclopropenone precursor, 4, used to generate this ynamine
(4) A brief review of ynol and ynamine chemistry appears in Chiang,
Y.; Kresge, A. J .; Paine, S. W.; Popik, V. V. J . Phys. Org. Chem. 1996,
9, 361-370.
X Abstract published in Advance ACS Abstracts, J uly 1, 1997.
(1) (a) Chiang, Y.; Grant, A. S.; Kresge, A. J .; Pruszynski, P.; Schepp,
N. P.; Wirz, J . Angew. Chem., Int. Ed. Engl. 1991, 30, 1346-1358. (b)
Chiang, Y.; Grant, A. S.; Kresge, A. J .; Paine, S. W. J . Am. Chem. Soc.
1996, 118, 4366-4372. (c) Kresge, A. J .; Paine, S. W. Can. J . Chem.
1996, 74, 1373-1376.
(2) Chiang, Y.; Kresge, A. J .; Hochstrasser, R.; Wirz, J . J . Am. Chem.
Soc. 1989, 111, 2355-2357. Chiang, Y.; Kresge, A. J .; Popik, V. V. J .
Am. Chem. Soc. 1995, 117, 9165-9171.
(5) Potts, K. T.; Baum, J . S. Chem. Rev. 1974, 74, 189-213.
(6) Andraos, J .; Chiang, Y.; Grant, A. S.; Guo, H.-X.; Kresge, A. J .
J . Am. Chem. Soc. 1994, 116, 7411-7412.
(7) Farnum, D. G.; Chickos, J .; Thurston, P. E. J . Am. Chem. Soc.
1966, 88, 3075-3081.
(8) Supporting information; see paragraph at the end of this paper
regarding availability.
(3) For theoretical analyses of these base-weakening and acid-
strengthening effects see Smith, B. J .; Radom, L.; Kresge, A. J . J . Am.
Chem. Soc. 1989, 111, 8297-8299. Smith, B. J .; Radom, L. J . Am.
Chem. Soc. 1992, 114, 36-41.
(9) (a) Chiang, Y.; Hojatti, M.; Keeffe, J . R.; Kresge, A. J .; Schepp,
N. P.; Wirz, J . J . Am. Chem. Soc. 1987, 109, 4000-4009. (b) Andraos,
J .; Chiang, Y.; Huang, C.-G.; Kresge, A. J .; Scaiano, J . C. J . Am. Chem.
Soc. 1993, 115, 10605-10610.
S0022-3263(97)00765-2 CCC: $14.00 © 1997 American Chemical Society

5364 J . Org. Chem., Vol. 62, No. 16, 1997
Chiang et al.
light over the entire spectral region from the ultraviolet
through the visible; the laser system9b operated at λ ) 248
nm.
Some reactions were too slow to allow accurate monitoring
by flash photolysis, and these were therefore followed using a
Cary 2200 spectrometer. The reactions were first initiated by
a single flash from the conventional system, and the reacting
solutions were then quickly transferred to the Cary instru-
ment.
The temperature of all reaction solutions was controlled at
25.0 ( 0.05 °C, and observed rate constants were calculated
by least-squares fitting of exponential functions.
Acid ity Con sta n t Deter m in a tion s. Acid dissociation
constants were determined spectrophotometrically, by moni-
toring the changes in UV absorbance that the substrates
underwent upon ionization. Absorbance measurements were
made with a Cary 2200 spectrometer whose cell compartment
was thermostated at 25.0 ( 0.05 °C; solutions containing a
constant stoichiometric concentration of substrate were em-
ployed. The data so obtained were analyzed by least squares
fitting of eq 4, in which AA and AB are the absorbances of the
AA[H+] + ABQa
F igu r e 1. Biphasic absorbance decay following a faster
absorbance rise produced by flash photolysis of 2-(N-(pen-
tafluorophenyl)amino)-3-phenylcyclopropenone in aqueous 0.04
M acetic acid buffer solution, buffer ratio ) 1; the inset shows
deviations from least squares fit of a double exponential rate
law.
Aobs
)
(4)
Qa + [H+]
acidic and basic forms of the substrate, respectively, and Qa
is its acidity constant.
Resu lts
Rea ction Id en tifica tion . Flash photolysis of 2-(N-
(pentafluorophenyl)amino)-3-phenylcyclopropenone in
aqueous solution produces several different reaction
products. Using HPLC analysis with authentic-sample
spiking, we have identified three of these products as
N-(pentafluorophenyl)phenylacetamide, 6, pentafluoro-
aniline, 7, and phenylacetaldehyde, 8. The first of these
products is formed immediately, whereas the second and
The more rapid component of this biphasic decay is
complete within a few seconds, which gives it the same
time scale as formation of N-(pentafluorophenyl)phenyl-
acetamide, observed as an immediately generated reac-
tion product. That suggests that this more rapid absor-
bance decay represents the ketenimine hydration reaction
of eq 6. This assignment is supported by the fact that
the reactivity and form of acid-base catalysis shown by
this absorbance decay are consistent with expectation
based upon an earlier study of ketenimine hydration
using authentic ketenimine substrates (vide infra).10
The slower component of the biphasic absorbance decay
occurs on a time scale consistent with the generation of
pentafluoroaniline and phenylacetaldehyde, which were
observed as reaction products requiring some tens of
minutes to appear, and that suggests that the slower
decay is associated with the formation of these sub-
stances. A common reaction of cyclopropenones is nu-
cleophile-assisted ring opening,5 which in the present
case would generate the enamine 2-phenyl-3-(N-(pen-
tafluorophenyl)amino)acrylic acid, 5, as shown in eq 3.
Hydrolysis of this enamine would then produce pen-
tafluoroaniline, 7, and 2-phenylformylacetic acid, 10, eq
7, which is known to decarboxylate readily giving phen-
ylacetaldehyde, 8, eq 8.11
third take some tens of minutes to appear.
This product complexity has its counterpart in the UV
spectral changes that occur following flash photolysis.
There is at first a rapid rise in absorbance in the region
λ ) 260-265 nm, and that is then followed by a
considerably slower decay. This is similar to the changes
observed before in the flash photolysis of other secondary
aminocyclopropenones,1a,b where the rise was identified
as formation of a phenylketenimine by isomerization of
a phenylethynamine, itself formed by the still more rapid
decarbonylation reaction of eq 1, and the decay was
identified as hydration of the ketenimine to a phenylac-
etamide. An analogous assignment may be made in the
present case, with the absorbance rise corresponding to
isomerization of N-(pentafluorophenyl)phenylethynamine,
3, to N-(pentafluorophenyl)phenylketenimine, 9, as shown
in eq 5. However, as Figure 1 illustrates, the absorbance
decay in the present case is biphasic, with the data
conforming well to a double exponential rate law for two
concurrent first-order reactions. This raises the question
as to which component of this biphasic decay represents
the hydration of N-(pentafluorophenyl)phenylketenimine
to N-(pentafluorophenyl)phenylacetamide, 6, eq 6, and
which corresponds to some other reaction.
This assignment receives support from a comparison
of rates of reaction based upon this slower absorbance
(10) McCarthy, D. G.; Hegarty, A. F. J . Chem. Soc., Perkin Trans.
2 1980, 579-591.
(11) Strukov, I. T. J . Gen. Chem. USSR 1952, 22, 1077-1084.

Photolytic Decarbonylation and Ring-Opening of a Cyclopropenone
J . Org. Chem., Vol. 62, No. 16, 1997 5365
decay with rates of hydrolysis of the methyl ester of the
enamine-acid, methyl 2-phenyl-3-(N-(pentafluorophenyl)-
amino)acrylate, 11, eq 9, to whose hydrolysis this absor-
bance decay is being attributed. These rates of reaction
were measured in aqueous perchloric acid solution over
a range of acid concentrations at a constant ionic strength
(0.10 M). The data are summarized in Tables S2 and
S3.8 In both cases, observed first-order rate constants
were accurately proportional to acid concentration; least
squares analysis gave kobs ) (4.83 ( 2.15) × 10-5 + (1.25
( 0.03) × 10-2[HClO4] for the absorbance decay and kobs
F igu r e 2. Rate profiles for the isomerization of N-(pentafluo-
rophenyl)phenylethynamine to N-(pentafluorophenyl)phenyl-
ketenimine (O) and hydration of the ketenimine (4) in aqueous
solution at 25 °C, ionic strength ) 0.10 M.
) (4.69 ( 3.79) × 10-5 + (1.06 ( 0.06) × 10-2[HClO4] for
hydrolysis of the enamine-ester. The close correspon-
dence of these two rate laws suggests that similar
processes are being monitored in the two cases, i.e., that
both reactions are enamine hydrolyses.
Further evidence that this assignment is correct comes
from the solvent isotope effect on the slower component
of the biphasic absorbance decay. This isotope effect was
determined by measuring rates in D2O solutions of
perchloric acid, again over a range of acid concentrations.
Least squares analysis of the data, summarized in Table
S2,8 gave kobs ) -(0.37 ( 1.95) × 10-5 + (4.14 ( 0.30) ×
10-3[DClO4], which, together with the H2O result pro-
P h en yleth yn a m in e Isom er iza tion . The rise in ab-
sorbance at λ ) 260-265 nm observed upon flash
photolysis of 2-(N-(pentafluorophenyl)amino)-3-phenyl-
cyclopropenone, attributed to isomerization of N-(pen-
tafluorophenyl)phenylethynamine to N-(pentafluorophen-
yl)ketenimine, eq 5, conformed well to a single ex-
ponential rate law. Observed first-order rate constants
obtained by least-squares fitting of such an expression
were determined in aqueous perchloric acid and sodium
hydroxide solutions as well as in formic acid, acetic acid,
biphosphate ion, ammonium ion, and bicarbonate ion
buffers. All measurements were made at constant ionic
strength (0.10 M). The data so obtained are summarized
in Tables S4-S6.8
The rate measurements in buffers were made in series
of solutions of constant buffer ratio and constant ionic
strength (0.10 M) but varying buffer concentration; this
served to hold hydronium ion concentrations constant
along a given buffer series. Buffer catalysis was found,
with observed first-order rate constants conforming to the
expected linear rate law of eq 12; the data were conse-
quently analyzed by least squares fitting of this expres-
+
vides the isotope effect kH+/kD ) 3.02 ( 0.24. The
hydrolysis of deactivated enamines such as this is known
to occur by rate-determining hydron transfer to â-carbon
to give isotope effects in the normal direction,12 like
the present result.
Although the ring-opening reaction generating an
enamine found here provides products similar to those
expected from the well-known thermal ring-opening of
cyclopropenones,5 the present process cannot be a ther-
mal reaction. The cyclopropenone precursor is stable in
the solutions used for flash photolysis; no products are
formed until the system is irradiated, and the reaction
must consequently be a photochemical process. It seems
unlikely, moreover, that this photochemical reaction
occurs by a mechanism similar to the thermal process,
which is believed to be initiated by nucleophilic attack
at the carbonyl carbon atom of the cyclopropenone,5 for
it is not clear that photoexcitation would increase the
susceptibility of this carbon atom to nucleophilic attack.
Carbonyl group photochemistry is dominated instead by
radical reactions, and a plausible primary process in the
present case woud be R-cleavage, as shown in eq 10. The
ensuing biradical, 12, is a resonance hybrid of several
structures, including the ketene-carbene, 13, and addi-
kobs ) ko + kcat[buffer]
(12)
sion. The zero-buffer-concentration intercepts, ko, to-
gether with observed rate constants measured in perchloric
acid and sodium hydroxide solutions, are displayed as
the upper rate profile of Figure 2. Hydrogen ion concen-
trations of the buffers needed for this purpose were
obtained by calculation, using literature values of the
buffer acid ionization constants and activity coefficients
recommended by Bates.13
tion of water to that would give the postulated product,
2-phenyl-3-(N-(pentafluorophenyl)amino)acrylic acid, 5.
R-Cleavage is a well known photochemical primary
process, but its occurrence in cyclopropenone chemistry
appears to be unprecedented.
Buffer catalysis was strong in the more acidic buffers
(HCO2H, CH3CO2H, and H2PO4-), but it became progres-
sively weaker as the acid strength of the buffer acid
+
decreased and was practically nonexistant in NH4 and
-
HCO3 buffers. The form of the buffer catalysis was
(12) Keeffe, J . R.; Kresge, A. J . In The Chemistry of Enamines;
Rappoport, Z., Ed.; J ohn Wiley and Sons: New York, 1994; pp 1049-
1109.
(13) Bates, R. G. Determination of pH Theory and Practise; Wiley:
New York, 1973; p 49.

5366 J . Org. Chem., Vol. 62, No. 16, 1997
Chiang et al.
F igu r e 4. Sigmoid titration curve for the ionization of 2-(N-
(pentafluorophenyl)amino)-3-phenylcyclopropenone as a ni-
trogen acid in aqueous solution at 25 °C.
F igu r e 3. Rate data for the isomerization of 2-(N-(pentafluo-
rophenyl)phenylethynamine to N-(pentafluorophenyl)phenyl-
ketenimine in aqueous H2PO-/HPO42- buffer solutions plotted
according to eq 13.
(kcat
)
D2O
) 3.22 ( 0.20 and (ko)H O/(ko)D O ) 29.8 ( 1.6.
2
2
The latter value is a comparison at equal hydrogen ion
concentrations, i.e. [H+] ) [D+] () 5.53 × 10-8 M); the
isotope effect on the ionization of biphosphate ion needed
to make such a comparison was taken from the litera-
ture.14
determined with the aid of eq 13, in which kB and kHA
are general base and general acid catalytic coefficients,
kcat ) kB + (kHA - kB)fA
(13)
Hyd r a tion of P h en ylk eten im in e. Rates of reaction
based upon the faster component of the biphasic decay
in absorbance at λ ) 260-265 nm observed upon flash
photolysis of 2-(N-(pentafluorophenyl)amino)-3-phenyl-
cyclopropenone in aqueous solution, and attributed to
hydration of N-(pentafluorophenyl)phenylketenimine, were
determined in perchloric acid and sodium hydroxide
solutions as well as in acetic acid and biphosphate ion
buffers. All measurements were made at constant ionic
strength (0.10 M). The data so obtained are summarized
in Tables S7-S9.8
Rates of reaction in series of buffer solutions at
constant buffer ratio increased with increasing buffer
concentration, with observed first-order rate constants
conforming to the linear rate law of eq 12. Least squares
analysis gave zero-buffer-concentration intercepts, which,
together with the rate constants measured in perchloric
acid and sodium hydroxide solution, are displayed as the
lower rate profile of Figure 2. The form of buffer catalysis
was again evaluated using the relationship of eq 13; the
data for acetic acid buffers gave kHA ) -(0.10 ( 1.28) ×
10-1 M-1 s-1 and kB ) (7.10 ( 0.93) × 10-1 M-1 s-1, which
shows that the catalysis was again wholly of the general
base type.
Acid ity Con sta n ts. The acidity constants of two
substances, 2-(N-(pentafluorophenyl)amino)-3-phenylcy-
clopropenone and N-(pentafluorophenyl)benzamide, ion-
izing as nitrogen acids, were determined by monitoring
the UV absorbance changes that these substances un-
derwent, at λ ) 310 and 280 nm respectively, as the
acidity of their solutions was varied. Measurements were
made at constant ionic strength (0.10 M) in aqueous
solutions of hydrochloric acid, sodium hydroxide, and
biphosphate ion, monohydrogen t-butylphosphonate ion,
ammonium ion, and bicarbonate ion buffers. The data
so obtained are summarized in Tables S10 and S11.8
As Figure 4 illustrates, the results conformed well to
the expected sigmoid relationship of eq 4. Least squares
respectively, and fA is the fraction of buffer present as
acid. The data for H2PO4 buffers plotted according to
-
this equation are displayed in Figure 3; it may be seen
that they conform to this relationship well. Least
squares analysis gave kHA ) -(0.98 ( 1.07) × 103 M-1
s-1 and kB ) (2.42 ( 0.12) × 104 M-1 s-1 for these buffers,
and the data for acetic acid buffers provided kHA ) (1.02
( 2.57) M-1 s-1 and kB ) (3.47 ( 0.09) × 102 M-1 s-1
.
These results show that buffer catalysis is wholly of the
general base type.
Some rate measurements were also made in D2O
solutions of hydrochloric acid, sodium hydroxide, and
biphosphate ion buffers. These data are summarized in
Tables S4-S68 as well.
The measurements in D2O solutions of hydrochloric
acid were made over a range of acid concentration where,
as in the case of H2O solutions of perchloric acid, observed
rate constants increased linearly with acid concentration.
Linear least squares analysis gave the second-order rate
constant kD+ ) (2.71 ( 0.06) × 102 M-1 s-1, which, when
combined with its H2O counterpart, provides the isotope
+
effect kH+/kD ) 3.26 ( 0.10.
The rate measurements in D2O solutions of sodium
hydroxide solutions were made at a single hydroxide ion
concentration (0.10 M) in the region of the high basicity
plateau where observed rate constants were independent
of acid or base concentration (see Figure 2). Replicate
measurements gave the averge value kD O ) (7.43 ( 0.23)
2
× 104 s-1, which, when combined with the average rate
constants determined in H2O solutions of sodium hy-
droxide at the same concentration, gave the isotope effect
kH O/kD O ) 8.50 ( 0.41.
2
2
The measurements in D2O solutions of biphosphate
buffers were done at a single buffer ratio, exactly the
same as that for one of the sets of solutions of this buffer
in H2O. Once again, observed rate constants increased
linearly with increasing buffer concentration and ap-
plication of eq 12 gave a catalytic coefficient and zero-
concentration intercept, which, when combined with their
(14) Gary, R.; Bates, R. G.; Robinson, R. A. J . Phys. Chem. 1964,
68, 3806-3809.
H2O counterparts, provided the isotope effects (kcat
)
/
H2O

Photolytic Decarbonylation and Ring-Opening of a Cyclopropenone
J . Org. Chem., Vol. 62, No. 16, 1997 5367
Ta ble 1. Su m m a r y of Ra te a n d Equ ilibr iu m Con sta n tsa
analysis gave Qa ) (1.57 ( 0.02) × 10-9 M, pQa ) 8.803
( 0.005 for 2-(N-(pentafluorophenyl)amino)-3-phenylcy-
clopropenone,15 and Qa ) (7.35 ( 0.12) × 10-11 M, pQa )
10.134 ( 0.007, for N-(pentafluorophenyl)benzamide.15
hydrogen-ion catalytic coefficient for the present isomer-
ization, moreover, conforms well to a correlation of
catalytic coefficients for the rate-determining carbon
protonation of other phenylethynamines.1b It would seem
safe to conclude, therefore, that the presently observed
process is an ynamine-ketenimine isomerization reac-
tion that occurs by the mechanism of eq 14.
Discu ssion
P h en ylet h yn a m in e Isom er iza t ion . Our previous
studies1 have shown that secondary phenylethynamines
undergo isomerization to phenylketenimines by rate-
determining protonation on â-carbon followed by rapid
proton loss from nitrogen of the ensuing keteniminium
ion, 14, eq 14. As expected for such a reaction, the
present ynamine isomerization shows acid catalysis,
The rate profile of Figure 2, however, shows some
additional features that indicate another reaction mech-
anism is operating as well. There is a region of base
catalysis, indicated by the diagonal segment extending
from [H+] ) 10-4 to 10-10 M, which eventually becomes
saturated producing a downward bend in the profile.
Downward bends such as this are often caused by acid
ionization of the substrate,16 and this suggests that the
base catalysis is due to equilibrium ionization of the
ynamine to an ynamide ion, 15, followed by rate-
determining protonation of this ion on â-carbon by solvent
water, eq 15. Since a proton is provided in the prior
equilibrium but is not used up in the rate-determining
step, the overall rate of reaction will be inversely
proportional to hydrogen ion concentration, which is
equivalent to being directly proportional to hydroxide ion
evidenced by the diagonal segment at the high acidity
end of the rate profile of Figure 2, and it also gives a
+
hydrogen ion isotope effect in the normal direction, kH
/
kD+ ) 3.26, also as expected for the process of eq 14. The
(15) This is a concentration dissociation constant, applicable at the
ionic strength of the measurements (0.10 M).
(16) Loudon, G. M. J . Chem. Ed. 1991, 68, 973-984.

5368 J . Org. Chem., Vol. 62, No. 16, 1997
Chiang et al.
buffer acid decreased is also consistent with the mech-
anism of eq 15. The rate-determining step of this
mechanism can be expected to be a very fast reaction,
and in fact the very large general acid catalytic coef-
ficients, kHA ) 1.61 × 108 M-1 s-1 and 6.84 × 107 M-1 s-1
for acetic acid and biphosphate ion, respectively, can be
calculated for this step from the observed general base
catalytic coefficients and the acidity constant of the
ynamine, Qa ) 5.86 × 1011 M (vide infra). Such very fast
reactions will have reactant-like transition states and low
values of the Bronsted exponent R;20 the two catalytic
coefficients cited above do in fact give R ) 0.27. General
acid catalysis becomes difficult to detect in systems with
low Bronsted exponents because reaction through proton
transfer from the solvent is dominant, and this domi-
nance becomes stronger the weaker the general acid.21
General acid catalysis of the rate-determining step in the
present ynamine isomerization by such weak acids as
ammonium and biphosphate ions will consequently be
swamped by the water reaction, which, because of the
prior substrate ionization equilibrium, will be translated
into observed general base catalysis being overwhelmed
by a hydroxide ion reaction. In the early stages of our
investigation of this reaction, we in fact failed to detect
general base catalysis and in a preliminary publication
reported the process to be subject only to specifc hydrox-
ide ion catalysis.1a
The isotope effect on general base catalysis in biphos-
phate ion buffers also supports the proposed reaction
mechanism. The observed effect, kH/kD ) 3.22, is the
product of the isotope effect on the prior equilibrium
deprotonation of the substrate, whose equilibrium con-
stant is equal to the acidity constant of the substrate
divided by that of the biphosphate ion, and the isotope
effect on the rate-determining carbon-protonation step.
Since the isotope effect on the acidity constant of the
substrate has been determined here and that on the
ionization of biphosphate ion is available from the
literature,14 the isotope effect on the rate-determining
step can be evaluated. The result is kH/kD ) 3.15 ( 0.30.
The normal direction (kH/kD > 1) of this isotope effect is
as expected for rate-determing hydron transfer from
biphosphate ion to carbon, and its modest magnitude is
consistent with the great speed of this reaction (kH ) 6.8
× 107 M-1 s-1) and its consequent reactant-like, unsym-
metrical transition state.22
concentration, as observed. The ynamide ion can be
expected to be much more reactive to carbon protonation
than is the neutral ynaminesenolate ions, for example,
are many orders of magnitude more reactive than the
corresponding enols17sand this mechanism would be the
preferred route even when the equilibrium concentration
of ynamide ion is very low. Eventually, however, at
sufficiently low hydrogen ion concentrations, the equi-
librium will shift from a preponderance of ynamine to a
preponderance of ynamide ion, and the advantage of
going from a less reactive to a more reactive species will
be lost and base catalysis will become saturated, again
as observed.
Support for this reaction mechanism comes from
isotope effects. The solvent isotope effect on the reaction
in the low-acidity plateau region, where base-catalysis
has become saturated and the reaction consists of simple
carbon protonation by water, is kH O/kD O ) 8.5. This is
2
2
a reasonable value for hydron transfer from a water
molecule; it is large because the primary isotope effect
produced by the hydron in transit is augmented by a
secondary effect generated by solvation of the hydroxide
ion being left behind.18 The solvent isotope effect in the
region of base catalysis, on the other hand, is much
greater: kH O/kD O ) 29.8. This is because the reaction
2
2
now starts from unionized ynamine as its initial state,
and the observed rate constant is consequently equal to
the product of the rate constant for the rate-determining
step, k, and the equilibrium constant for the prior step,
Qa:kobs ) kQa; the observed isotope effect is therefore the
product of isotope effects on the two steps. Since the
isotope effect on k has been determined from rate
measurements made in the region of base-catalysis
saturation, that on Qa may be evaluated; the result is
(Qa)H O/(Qa)D O ) 3.51 ( 0.25. This is a reasonable value
2
2
for ionization of an acid of this strength (pQa ) 10.23,
vide infra); for example (Ka)H O/(Ka)D O ) 3.12 has been
reported for ammonium ion (pKa )2 9.27) and (Ka)H
/
2
2O
(Ka)D O ) 4.17 for phenol (pKa ) 10.00).19
2
The reaction mechanism of eq 15 is also supported by
the fact that isomerization is catalyzed by buffers and
that the catalysis is wholly of the general base type in
acetic acid and biphosphate ion buffers. The rate-
determining step of this mechanism, proton transfer from
an acid catalyst to the substrate, will be subject to general
acid catalysis. However, in the region where undissoci-
ated ynamine is the initial state of the reaction, as is the
case in acetic acid and biphosphate ion buffers, the prior
equilibrium will add an inverse dependence on hydrogen
ion concentration, and the net result will be general base
catalysis, as observed.
With this assignment of a mechanism to the base-
catalyzed portion of the present ynamine isomerization
reaction, a complete rate law interpreting the rate profile
shown in Figure 2 may be written. This is shown in eq
16, where kH+ is the rate constant for carbon protonation
kobs ) kH+[H+] + ko + k′oQa/(Qa + [H+]) (16)
of the unionized substrate by hydrogen ion, k′o is the rate
constant for carbon protonation of the ionized form by
water, and Qa is the acidity constant of the ynamine. The
rate constant ko represents the short, horizontal region
of the profile at [H+] ) 10-3-10-4 M. The molecular
The fact that buffer catalysis became progressively
weaker and more difficult to detect as the acidity of the
(17) Pruszynski, P.; Chiang, Y.; Kresge, A. J .; Schepp, N. P.; Walsh,
P. A. 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.
(18) Kresge, A. J .; More O’Ferrall, R. A.; Powell, M. F. In Isotopes
in Organic Chemistry; Buncel, E., Lee, C. C., Eds.: Elsevier, New York,
1987; Chap. 4.
(19) Laughton, P. M.; Robertson, R. E. In Solute-Solvent Interactions;
Coetzee, J . F.; Ritchie, C. D., Eds.; M. Dekker: New York, 1969; Chap
7.
(20) Kresge, A. J . In Proton Transfer Reactions; Caldin, E. F., Gold,
V., Eds.; Chapman and Hall: London, 1975; Chap. 7.
(21) Keeffe, J . R.; Kresge, A. J . In Techniques of Chemistry, Volume
VI. Investigations of Rates and Mechanisms of Reactions; Bernasconi,
C. F., Ed.; Wiley-Interscience: New York, 1986; Chap. XI.
(22) Kresge, A. J . In Isotope Effects on Enzyme-Catalyzed Reactions;
Cleland, W. W., O’Leary, M. H., Northrop, D. B., Eds.; University Park
Press: Baltimore, 1977; p 37-63.

Photolytic Decarbonylation and Ring-Opening of a Cyclopropenone
J . Org. Chem., Vol. 62, No. 16, 1997 5369
law required by such a process is shown in eq 17, and
least squares fitting of the present data using this
expression gave kH+ ) (1.88 ( 0.08) × 10-1 M-1 s-1, ko )
kobs ) kH+[H+] + ko + kHO-[HO-]
(17)
(2.00 ( 0.04) × 10-2 s-1, and kHO ) (2.05 ( 0.02) × 102
-
M-1 s-1
.
The previous investigation of ketenimine hydration10
established a mechanism for the acid-catalyzed reaction
consisting of rate-determining protonation of the sub-
strate on â-carbon followed by hydration of the ensuing
nitrilium ion, 16, giving an amidol intermediate, 17,
which rapidly tautomerized to an amide product, eq 18.
Ketenimines with aliphatic substituents on nitrogen were
F igu r e 5. Acidity scale for some amines ionizing as acids in
aqueous solution.
interpretation of this rate constant is ambiguous: it could
represent carbon protonation of the unionized substrate
by water, or carbon protonation of the ionized form by
+
hydrogen ion. Least squares fitting of eq 16 gave kH
)
(8.91 ( 0.19) × 102 M-1 s-1, ko ) 1.08 ( 0.11 s-1, k′o )
(6.52 ( 0.12) × 105 s-1, and Qa ) (5.86 ( 0.27) × 10-11
,
pQa ) 10.232 ( 0.020.15
P h en yleth yn a m in e Acid ity. The acid dissociation
constant determined here for N-(pentafluorophenyl)phen-
ylethynamine, pQa ) 10.23, shows this substance to have
a remarkably strong acidity for an amine ionizing as a
nitrogen acid in aqueous solution: most amines take on
a proton rather than give one up in this medium. This
ynamine, for example, is 23 pK units more acidic than
ammonia, for which pKa ) 35 in aqueous solution has
been estimated.23
found to react more rapidly than the substrates with a
phenyl group in this position, consistent with the greater
ability of aliphatic groups over phenyl to stabilize the
positive charge of the nitrilium ion by inductive electron
release. The pentafluorophenyl group will be even poorer
than phenyl at stabilizing this positive charge, and it is
significant therefore that kH+ for N-(pentafluorophenyl)-
phenylketenimine obtained here is 500 times less than
that determined before for N-(phenyl)phenylketenimine.
The previous data for the N-phenyl substrate and three
ketenimines with aliphatic nitrogen substituents do in
fact give a good Hammett correlation using σI constants,27
and extrapolation of this to N-pentafluorphenyl gives the
This striking increase in acidity may be divided into
contributions from the phenylethynyl group and the
pentafluorophenyl group by making comparisons with the
acidity constant of pentafluoroaniline ionizing as an acid.
A directly measured value of the latter is unfortunately
not available for aqueous solution, but an estimate can
be made using pKa ) 23.1 determined in DMSO24 and
the change of solvent increment ∆pKa ) -2.9. This
increment is based upon comparisons using two different
nitrogen acids, each giving the same result: for aniline,
prediction kH ) 0.62 M-1 s-1, consistent with the
+
measured value kH ) 0.19 M-1 s-1
.
+
Base catalysis of ketenimine hydration was investi-
gated in the previous study for only one substrate, the
N-phenyl derivative, and a quantitative prediction of the
rate constant expected for the present N-pentafluoro-
phenyl analog is therefore not possible. It is significant,
however, that the N-pentafluorophenyl derivative is more
reactive than the N-phenyl derivative, inasmuch as the
base-catalyzed reaction can be expected to occur either
by nucleophilic attack of hydroxide ion on the R-carbon
atom of the ketenimine group or by general-base-assisted
attack of a water molecule at this site; reaction through
either one of these mechanisms will be promoted by the
more strongly electron-withdrawing pentafluorophenyl
group.
pKa(DMSO) ) 30.6,25 pKa(H2O) ) 27.7,26 and ∆pKa
)
-2.9; and for N-(pentafluorophenyl)benzamide, pKa(DM-
SO) ) 13.0,24 pKa(H2O) ) 10.1 (this work), and ∆pKa )
-2.9. The resulting estimate, pKa ) 20.2 for pentafluo-
roaniline in aqueous solution, puts this substance 15 pK
units below ammonia and 10 pK units above N-(pen-
tafluorophenyl)phenylethynamine. Thus, 40% of the
difference between this ynamine and ammonia may be
assigned to the phenylethynyl group and 60% to the
pentafluorophenyl group. These comparisons are pre-
sented in graphical form in Figure 5.
Keten im in e Hyd r a tion . The rate profile of Figure
2 for the hydration of N-(pentafluorophenyl)phenylke-
tenemine to N-(pentafluorophenyl)phenylacetamide shows
both acid and base catalysis as well as an uncatalyzed
region. This is consistent with an earlier detailed study
of ketenimine hydration in which acid and base catalysis
and an uncatalyzed reaction were also found.10 The rate
The general effect of the pentafluorophenyl group has
been to suppress acid catalysis and promote base cataly-
sis, as expected. This dominance of base catalysis ex-
tends even to the uncatalyzed region of the rate profile,
where buffer catalysis was found to be of the general base
type for the N-pentafluorophenyl derivative examined
here, in contrast to the general acid buffer catalysis in
this region found for the substrates with aliphatic N-
substituents studied before.10 In this respect, the hydra-
tion of N-(pentafluorophenyl)phenylketenimine resembles
(23) Bell, R. P. The Proton in Chemistry, 2nd ed.; Cornell Univ.
Press: Ithaca, NY, 1973; pp 86-87.
(24) Vlasov, V. M.; Terekhova, M. I.; Petrov, E. S.; Shatenshtein,
A. I.; Yakobson, G. G. Zhur. Org. Khim. 1981, 17, 2025-2031.
(25) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463.
(26) Stewart, R. The Proton: Appiclations to Organic Chemistry;
Academic Press: New York, 1985; p 70.
(27) Charton, M. Prog. Phys. Org. Chem. 1981, 13, 119-251.

5370 J . Org. Chem., Vol. 62, No. 16, 1997
Chiang et al.
the hydration of ketenes themselves, where base catalysis
is strong and acid catalysis is either weak or nonexist-
ent.28
and Engineering Research Council of Canada for finan-
cial support of this work.
Su p p or tin g In for m a tion Ava ila ble: Tables S1-S11 (16
pages). This material is contained in libraries on microfiche,
immediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
Ack n ow led gm en t. We are grateful to Dr. R.-Q. Xie
for supplying spectral data characterizing N-(pentafluo-
rophenyl)phenylacetamide and to the Natural Sciences
(28) Tidwell, T. T. Ketenes; Wiley-Interscience: New York, 1995; pp
576-587.
J O970765Y