Flash photolysis of 10-diazo-9(10H)-phenanthrenone in aqueous solution. Hydration of fluorenylideneketene and the fluorene-9-carboxylic acid keto-enol system

Article abstract of DOI:10.1021/ja971381s

Flash photolysis of 10-diazo-9(10H)-phenanthrenone in aqueous solution was found to give two successively formed transient species and to produce fluorene-9-carboxylic acid as the major reaction product. These transients were identified, through solvent isotope effects and the form of acid-base catalysis, as fluorenylideneketene, formed by photo-Wolff reaction of the diazophenanthrenone, and fluorene-9-carboxylic acid enol, formed by hydration of this ketene. Analysis of the rate profile of the enol ketonization reaction produced the first and second ionization constants for the enol ionizing as an oxygen acid, pQ(a)(E) = 2.01 and pQ'(a)(E) = 9.61, respectively. The rate of enolization of fluorene-9-carboxylic acid was also determined, by bromine scavenging, and that, coupled with a literature value of the acidity constant of this acid, allowed evaluation of the two keto-enol equilibrium constants (pK(E) = 9.67 for interconverting un-ionized carboxylic acid and enol and pK'(E) = 8.24 for interconverting singly ionized acid and enol), and it also allowed evaluation of the two carbon acid acidity constants (pQ(a)(K) = 11.67 for ionization of the un-ionized carboxylic acid as a carbon acid and pQ'(a)(K) = 17.85 for ionization of its carboxylate ion as a carbon acid). (All acidity constants are concentration quotients applicable at ionic strength 0.10 M.) These keto-enol equilibrium constants and acid dissociation constants are large because of the enol and enolate ion stabilizing effects of the cyclopentadienyl ring of the fluorenyl group; this ring also makes fluorenylideneketene an unusually reactive substance.

Full text of DOI:10.1021/ja971381s

J. Am. Chem. Soc. 1997, 119, 8417-8424
8417
Flash Photolysis of 10-Diazo-9(10H)-phenanthrenone in
Aqueous Solution. Hydration of Fluorenylideneketene and the
Fluorene-9-carboxylic Acid Keto-Enol System
J. Andraos, Y. Chiang, A. J. Kresge,* and V. V. Popik
Contribution from the Department of Chemistry, UniVersity of Toronto,
Toronto, Ontario M5S 3H6, Canada
X
ReceiVed April 30, 1997
Abstract: Flash photolysis of 10-diazo-9(10H)-phenanthrenone in aqueous solution was found to give two successively
formed transient species and to produce fluorene-9-carboxylic acid as the major reaction product. These transients
were identified, through solvent isotope effects and the form of acid-base catalysis, as fluorenylideneketene, formed
by photo-Wolff reaction of the diazophenanthrenone, and fluorene-9-carboxylic acid enol, formed by hydration of
this ketene. Analysis of the rate profile of the enol ketonization reaction produced the first and second ionization
E
E
constants for the enol ionizing as an oxygen acid, pQa ) 2.01 and pQ′a ) 9.61, respectively. The rate of
enolization of fluorene-9-carboxylic acid was also determined, by bromine scavenging, and that, coupled with a
literature value of the acidity constant of this acid, allowed evaluation of the two keto-enol equilibrium constants
(
pKE ) 9.67 for interconverting un-ionized carboxylic acid and enol and pK′E ) 8.24 for interconverting singly
K
ionized acid and enol), and it also allowed evaluation of the two carbon acid acidity constants (pQ ) 11.67 for
ionization of the un-ionized carboxylic acid as a carbon acid and pQ′ ) 17.85 for ionization of its carboxylate ion
a
K
a
as a carbon acid). (All acidity constants are concentration quotients applicable at ionic strength 0.10 M.) These
keto-enol equilibrium constants and acid dissociation constants are large because of the enol and enolate ion stabilizing
effects of the cyclopentadienyl ring of the fluorenyl group; this ring also makes fluorenylideneketene an unusually
reactive substance.
During the past 15 years there has been a remarkable
development of methods for generating the enol isomers of
simple aldehydes and ketones in solution, and as a result much
1
is now known about the chemistry of these unstable substances.
Much less, however, is known about enols of carboxylic acids,
undoubtedly because these substances are very much less stable
than the already quite labile aldehyde and ketone enols. Most
We generated the enol of fluorene-9-carboxylic acid by
hydration of fluorenylideneketene, 8, obtained by photo-Wolff
reaction of the diazoquinone 10-diazo-9(10H)-phenanthrenone,
, eq 4. We were able to measure rates of this hydration reaction
2
of the relatively little information that is available has been
obtained from systems in which the carboxylic acid enol is
stabilized, either sterically3 or electronically.
,4
5-7
The cyclo-
7
pentadienylidene group has proved to be especially good at
stabilizing carboxylic acid enols, and this effect has enabled
5
studies of fulvenediol, 2, the enol of cyclopentadienecarboxylic
acid, 1, eq 1,and its benzo analog 4, the enol of indenecarboxylic
as well as rates of ketonization of the enol, and we consequently
gained information about the chemistry of the ketene in addition
to that of the enol.
(
3) O’Neill, P.; Hegarty, A. F. J. Chem. Soc., Chem. Commun. 1987,
44-745. Allen, B. M.; Hegarty, A. F.; O’Neill, P.; Nguyen, M. T. J.
Chem. Soc., Perkin Trans. 2 1992, 927-934.
7
(
4) Frey, J.; Rappoport, Z. J. Am. Chem. Soc. 1995, 117, 1161-1162;
996, 118, 5169-5181, 5182-5191.
5) (a) Urwyler, B.; Wirz, J. Angew. Chem., Int. Ed. Engl. 1990, 29,
1
(
7
1
90-792. (b) Almstead, J.-I. K.; Urwyler, B.; Wirz, J. J. Am. Chem. Soc.
994, 116, 954-960.
acid, 3,5b,6b,d,7 eq 2. We have now completed this series by
investigating the dibenzo analog 6, which is the enol of fluorene-
(6) (a) Chiang, Y.; Kresge, A. J.; Pruszynski, P.; Schepp, N. P.; Wirz, J.
Angew. Chem., Int. Ed. Engl. 1990, 29, 792-794. (b) Andraos, J.; Chiang,
9
-carboxylic acid, 5, eq 3.
Y.; Huang, C.-G.; Kresge, A. J.; Scaiano, J. C. J. Am. Chem. Soc. 1993,
1
15, 10605-10610. (c) Andraos, J.; Chiang, Y.; Kresge, A. J.; Pojarlieff,
X
Abstract published in AdVance ACS Abstracts, August 15, 1997.
I. G.; Schepp, N. S.; Wirz, J. J. Am. Chem. Soc. 1994, 116, 73-81. (d)
Andraos, J.; Kresge, A. J.; Popik, V. V. J. Am. Chem. Soc. 1994, 116,
961-967.
(1) See, for example: The Chemistry of Enols; Rappoport, Z., Ed.;
Wiley: New York, 1990.
2) For a brief review: see Kresge, A. J. Chem. Soc. ReV. 1996, 25,
75-280.
(
(7) Barra, M.; Fisher, T. A.; Cernigliaro, G. J.; Sinta, R.; Scaiano, J. C.
J. Am. Chem. Soc. 1992, 114, 2630-2634.
2
S0002-7863(97)01381-4 CCC: $14.00 © 1997 American Chemical Society

8418 J. Am. Chem. Soc., Vol. 119, No. 36, 1997
Andraos et al.
Experimental Section
by photodediazotization of the diazoquinone, eq 6,in competition
Materials. 10-Diazo-9(10H)-phenanthrenone was prepared by treat-
ing 9,10-phenanthrenequinone with (p-tolylsulfonyl)hydrazine.8 Other
materials were the best available commercial grades.
Flash Photolysis. Flash photolysis experiments were done using
two different systems: (1) a conventional apparatus with excitation
flash produced by capacitor discharge through a pair of xenon flash
9
lamps, 50 µs pulse width, and (2) a KrF eximer laser system, λexc
)
6
b
2
48 nm, 25 ns pulse width.
Substrate concentrations were ca.
-
5
1
0
M.
Enolization Rate Measurements. Rates of enolization of fluorene-
-carboxylic acid were determined by using bromine to scavenge the
9
enol as it formed. Measurements were made in sodium hydroxide
-
solutions using the absorbance of OBr at λ ) 330 nm to monitor
bromine concentration. Initial concentrations of the carboxylic acid
-
4
-4
substrate were 1.4 × 10 M and of bromine 1.7 × 10 M. The
reaction is pseudo-first-order under these conditions, and the first-order
rate law was obeyed accurately; observed rate constants were deter-
mined by least squares fitting of an exponential function.
with Wolff rearrangement of this intermediate to fluore-
Further experimental details are available in the theses upon which
part of this paper is based.
nylideneketene, 8; such addition of water to ketocarbenes is
1
0
1
2
known to occur in aqueous solution.
Further support for production of fluorene-9-carboxylic acid
via hydration of fluorenylideneketene and ketonization of the
ensuing enol as the main process occurring upon flash photolysis
of 10-diazo-9(10H)-phenanthrenone in aqueous solution is
provided by the behavior of transient species observed during
the flash photolysis experiments (Vide infra).
Results
Stability of 10-Diazo-9(10H)-phenanthrenone. The UV
spectrum of this substance in aqueous solution, which consists
of bands at λmax ) 245 and 375 nm, showed no change over
the acidity range used in the present work, pCH+ ) 1-13, and
it was also invariant for periods of time longer than those taken
to prepare solutions and subject them to flash photolysis. This
shows that this substrate was stable under the flash photolytic
conditions and that it underwent no thermal reaction before
photolysis.
-5
Flash Photolysis. Irradiation of dilute (10 M) aqueous
solutions of 10-diazo-9(10H)-phenanthrenone, with a pulse of
light from either our conventional flash system or the laser
apparatus operating at λexc ) 248 nm, produced an intermediate
bleaching of the substrate absorption bands at λmax ) 245 and
Product Study. HPLC analysis of aqueous solutions of 10-
diazo-9(10H)-phenanthrenone at concentrations used for flash
3
75 nm, followed by a growth and subsequent decay of
absorbance in the region λ ) 305-320 nm. The relative times
required for these absorbance changes at λ ) 305-320 nm
depended strongly on the acid-base properties of the medium.
When they were well separated in time, they were analyzed
separately using a single-exponential form of the first-order rate
law, to which they both conformed well, and when they occurred
at comparable times, they were analyzed using a double-
exponential expression, to which the data also conformed well.
Absorbance changes such as these were observed before in
-5
photolysis (10 M), which had been subjected to one flash from
our conventional flash photolysis apparatus, showed two reaction
products, formed in major and minor amounts, in addition to
smaller quantities of unreacted starting material. Spiking with
authentic samples identified the major product as fluorene-9-
carboxylic acid, 5, and the minor product as 9,10-phenanthrene-
quinone, 10. In 0.001 M HClO4 these products were formed
in 80% and 15% yields, and in 0.002 M NaOH, in 53% and
2
0% yields.
the flash photolysis of other diazoquinones in aqueous
Fluorene-9-carboxylic acid is the product expected from
5,6b,d
solution,
where they have been assigned to the occurrence
ketonization of the enol 6, formed by photo-Wolff reaction of
0-diazo-9(10H)-phenanthrenone, 7, and hydration of the result-
ing ketene 8 according to eq 4. It is likely that the minor product
,10-phenanthrenequinone, 10, was formed from 9,10-phenan-
threnediol, 9, which is known to oxidize readily to the quinone,
of two consecutive reactions: (1) hydration of a ketene formed
within the time of the light pulse by photo-Wolff reaction of
the diazoquinone and (2) ketonization of the carboxylic acid
enol generated by this ketene hydration. A similar assignment
is made here; this is supported by the form of acid-base
catalysis and the magnitude of solvent isotope effects shown
by these absorption changes (Vide infra).
1
9
1
1
eq 5. The diol, in turn, was probably produced by conjugate
Ketene Hydration. Rates of the absorbance change at-
tributed to hydration of fluorenylideneketene were measured
in aqueous solutions of perchloric acid, sodium hydroxide, and
ammonium ion buffers, as well as in solutions with no added
acid or base. The ionic strength of these solutions was
maintained at 0.10 M through the addition of sodium perchlorate
as required. The results are summarized in Tables S1-S4 in
1
3
addition of water to the ketocarbene intermediate 11, formed
the Supporting Information.
The measurements in ammonium ion buffers were done in
series of solutions of constant buffer ratio but varying buffer
concentration. Buffer catalytic coefficients were evaluated as
(
8) Cava, M. P.; Litle, R. L.; Napier, D. R. J. Am. Chem. Soc. 1958, 80,
2
1
257-2263. Sus, O.; Steppan, H.; Dietrich, R. Justus Liebigs Ann. Chem.
958, 617, 20-25.
(
9) Chiang, Y.; Hojatti, M.; Keeffe, J. R.; Kresge, A. J.; Schepp, N. P.;
Wirz, J. J. Am. Chem. Soc. 1987, 109, 4000-4009.
10) Andraos, J. M.Sc. Thesis, University of Toronto, 1989; Ph.D. Thesis,
University of Toronto, 1992.
11) Qesch, F.; Sparrow, A. J.; Platt, K. L. J. Labelled Compd.
Radiopharm. 1980, 17, 93-100.
(12) (a) Chiang, Y.; Kresge, A. J.; Pruszynski, P.; Schepp, N. P.; Wirz,
J. Angew. Chem., Int. Ed. Engl. 1991, 30, 1366-1368. (b) Chiang, Y.;
Jefferson, E. A.; Popik, V. V. Unpublished work.
(13) Supporting Information; see paragraph at the end of this paper
regarding availability.
(
(

Flash Photolysis of 10-Diazo-9(10H)-phenanthrenone
J. Am. Chem. Soc., Vol. 119, No. 36, 1997 8419
these rate constants is considerably greater than its counterpart
for the hydration of the closely structurally related substance
2
-1
diphenylketene, 12, for which ko ) 2.75 × 10 s and kHO-
)
4
-1
-1 17b
6
.11 × 10
M
s .
This reactivity difference is a
consequence of the cyclopentadienyl nature of the central ring
in fluorenylideneketene, which takes on aromatic character when
it stabilizes the negative charge assumed by the substrate in
the transition state of the reaction, 13,as it undergoes nucleo-
Figure 1. Rate profiles for the hydration of fluorenylideneketene, 4,
and ketonization of the enol of fluorene-9-carboxylic acid, O, in aqueous
solution at 25 °C.
philic attack by a water molecule or hydroxide ion. Another
contributing factor is steric crowding of the benzene rings in
diphenylketene, which forces these substituents out of copla-
narity with the ketene group and reduces their ability to accept
negative charge.
slopes of buffer dilution plots, but buffer catalysis was weak,
and, although it appeared that the reaction involved only the
basic component of the buffer, the results were too uncertain
to allow detailed analysis. This is the expected result, for bases
are known to react with ketenes by direct nucleophilic attack
on the ketene carbonyl carbon atom, with the direction of
attack being in the ketene molecular plane:15 such attack in the
case of fluorenylideneketene will be hindered by steric interfer-
ence from hydrogens in the peri positions of the benzene rings
flanking the ketene group, and the buffer reaction will conse-
quently be weak.
The intercepts of these ammonium ion buffer solution plots,
on the other hand, were well defined, and they provide good
estimates of rates of ketene hydration through solvent-related
species in these solutions. They are displayed as triangles in
Figure 1, and, together with rate constants measured in all of
the other solutions, they provide the rate profile shown there.
Hydrogen ion concentrations needed for this purpose were
obtained by calculation using acidity constants from the literature
and activity coefficients recommended by Bates.16
Rates of hydration of fluorenylideneketene were also mea-
sured in D2O solutions of perchloric acid and in D2O solution
with no added acid or base. These data summarized in Tables
1
4
1
3
S1 and S3.
The observed rate constants determined in D2O solution
showed no systematic variation with acidity. They were
therefore simply averaged to give the result (ko)D O ) (7.92 (
2
4
-1
0
.12) × 10 s . Combination of this with its H2O counterpart
2 2
gives the isotope effect kH O/kD O ) 1.35 ( 0.03. The small
magnitude of this isotope effect is again typical of ketene
hydration reactions.
6
b,10,17b
It is also consistent with the
nucleophilic nature of this process, which involves no breaking
of isotopically substituted bonds but does include some weaken-
ing of the nonreacting bonds of the attacking water molecule
as it takes on positive charge in the transition state of the process,
1
9
eq 8.
It may be seen that this ketene hydration consists of a wide
region of uncatalyzed reaction, some, albeit weak, hydroxide
ion catalysis, and no acid catalysis. Such behavior is typical
of ketene hydrations, whose rate profiles commonly show large
5,6b-d,17
uncatalyzed regions and only weak acid and base catalysis.
Acid catalysis, moreover, is suppressed by aromatic substituents,
such as the fluorenyl residue present here, and in such systems
it can be detected only in concentrated acid solutions.
The rate law that corresponds to this behavior is shown in
eq 7, where ko is the rate constant for the uncatalyzed reaction
and kHO- is that for the hydroxide ion catalyzed process. Least
Enol Ketonization. Rates of the absorbance change at-
tributed to ketonization of the enol of fluorene-9-carboxylic acid
were measured in aqueous solutions of perchloric acid and
sodium hydroxide. Measurements were also made in aqueous
biphosphate ion and ammonium ion buffers using series of
solutions of constant buffer ratio but varying buffer concentra-
tion. The ionic strength of all solutions was maintained at 0.10
M using sodium perchlorate as required. The results obtained
1
7b,18
kobs ) k + k
[HO-]
HO
-
(7)
o
1
3
are summarized in Tables S5-S7.
squares fitting of this expression gave ko ) (1.07 ( 0.02) ×
0 s and kHO- ) (3.47 ( 0.10) × 10 M s . Each one of
5
-1
6
-1 -1
The rate constants determined in perchloric acid and sodium
hydroxide solutions, together with the intercepts of buffer
dilution plots, are also displayed as the rate profile defined by
circles in Figure 1. Hydrogen ion concentrations of the buffer
solutions needed to construct this rate profile were again
obtained by calculation using literature acidity constants and
1
(
14) Andraos, J.; Kresge, A. J. J. Am. Chem. Soc. 1992, 114, 5643-
646.
15) Schiess, P.; Eberle, M.; Huys-Francotte, M.; Wirz, J. Tetrahedron
5
(
Lett. 1984, 25, 2201-2204. Seikaly, H. R.; Tidwell, T. T. Tetrahedron
1
986, 42, 2587-2613. Tidwell, T. T. Acc. Chem. Res. 1990, 23, 273-
2
79.
1
6
activity coefficients recommended by Bates.
(
16) Bates, R. G. Determination of pH Theory and Practice; Wiley: New
York, 1973, p 49.
17) (a) Bothe, E.; Dessouki, A. M.; Schulte-Frohlinde, D. J. Phys. Chem.
980, 84, 3270-3272. (b) Allen, A. D.; Kresge, A. J.; Schepp, N. P.;
Figure 1 shows that in neutral and basic solution the ketene
hydration and enol ketonization reactions were well separated
(
1
Tidwell, T. T. Can. J. Chem. 1987, 65, 1719-1723. (c) Allen, A. D.;
Tidwell, T. T. J. Am. Chem. Soc. 1987, 109, 2774-2780. (d) Allen, A.
D.; Stevenson, A.; Tidwell, T. T. J. Org. Chem. 1989, 54, 2843-2848. (e)
Chiang, Y.; Kresge, A. J.; Popik, V. V. J. Am. Chem. Soc. 1995, 117, 9165-
(18) (a) Allen, A. D.; Baigrie, L. M.; Gong, L.; Tidwell, T. T. Can. J.
Chem. 1991, 69, 138-145. (b) Andraos, J.; Kresge, A. J.; Schepp, N. P.
Can. J. Chem. 1995, 73, 539-543.
(19) Kresge, A. J.; More, O’Ferrall, R. A.; Powell, M. F. In Isotopes in
Organic Chemistry; Buncel, E., Lee, C. C., Eds.; Elsevier: Amsterdam,
1987; Vol. 7, pp 177-273.
9
171. Chiang, Y.; Guo, H.-X.; Kresge, A. J.; Tee, O. S. J. Am. Chem.
Soc. 1996, 118, 3386-3391.

8420 J. Am. Chem. Soc., Vol. 119, No. 36, 1997
Andraos et al.
in time. Rate constants for the two reactions could therefore
be calculated separately by analyzing the absorbance changes
using single exponential functions. Since the chemistry requires
ketene hydration to occur before enol ketonization and, if enol
ketonization is to be seen, also to be the faster process,
assignment of the absorbance changes and the rate constants
derived from them to the respective reactions in these media
was straightforward. In acidic solutions, however, rates of the
two reactions became comparable, and the data had to be
analyzed by fitting of double exponential functions. In such a
2
0
situation, assignment of the rate constants is more difficult,
because the initial phase of the biphasic absorption change no
longer necessarily represents the first reaction: this initial phase
instead always corresponds to the faster reaction, regardless of
whether that reaction occurs first or is second.
This difficulty was overcome in the present case by noting
how the maximum concentration of enol formed in the overall
process changed as the relative rates of the two reactions varied.
The question is whether, as the reaction medium becomes more
acidic, the rate of ketene hydration remains constant and that
of enol ketonization continues to increase, eventually crossing
Figure 2. Comparison of rate profiles for the ketonization of fluorene-
9
-carboxylic acid enol in H
2 2
O (O) and D O (4) solutions of perchloric
acid at 25 °C.
the enol.5,6 The corresponding reaction scheme for the present
case is shown in eq 9,where ketonization occurs by rate-
+
the ketene rate at [H ] = 0.001 M and becoming the faster
process (as shown in Figure 1), or whether this crossing is
+
avoided and the enol rate levels off at [H ] = 0.001 M and the
ketene rate begins to increase at this point. If the rate of enol
ketonization crosses that of ketene hydration and then becomes
progressively faster, the maximum concentration of enol
intermediate will progressively drop. If, on the other hand,
crossing is avoided and ketene hydration becomes progressively
faster, then the maximum concentration of enol will rise. At
the wavelengths employed for these rate measurements, the enol
absorbed more strongly than either its ketene precursor or the
carboxylic acid product, and the observed absorbance conse-
quently first rose and then fell, with the maximum absorbance
representing the maximum concentration of enol. We found
that this maximum absorbance decreased with increasing acidity
determining â-carbon protonation of the more reactive ionized,
enolate-ion form. The horizontal regions of the rate profiles
then represent reaction starting with un-ionized enol as the initial
state, and the diagonal regions that with enolate ion as the initial
state. Observed rate constants in the horizontal region are
consequently equal to the enol ionization constant times the rate
E
a
constant for the subsequent step, kobs ) Q k′
H
+
, whereas those
in the diagonal region are equal to the rate constant for the
+
of the medium as the system passed through [H ] = 0.001 M,
and we were therefore able to conclude that the two rates did
in fact cross, as indicated in Figure 1.
subsequent step alone, kobs ) k′H+. Observed isotope effects in
the horizontal region will therefore consist of isotope effects
on equilibrium constants times those on rate constants and will
thus be greater than observed isotope effects in the diagonal
region, which are on rate constants alone.
This assignment is supported by the fact that the alternative
would have the ketene hydration reaction become acid catalyzed
+
+
at [H ] = 0.001 M and then, at about [H ] ) 0.01 M, have the
acid catalysis become saturated. As was pointed out above,
acid catalysis of ketene hydration is weak at best and is
suppressed by aromatic substituents; for mesitylketene in
aqueous perchloric acid, it does not become apparent until the
acid concentration reaches 2 M.1 It is difficult, moreover, to
see why such catalysis should become saturated.
The rate law that applies to the reaction scheme of eq 9 is
shown in eq 10. Least squares fitting of the data using this
E
-3
expression produced the results Q ) (9.96 ( 3.30) × 10
a
22
8
-1 -1
M
and k′H ) (1.24 ( 0.03) × 10 M
+
s
for H2O solution,
plus D2O values that gave the isotope effects k′H /k′D ) 2.38
+
+
8b
E
a
E
a
(
0.08 and (Q ) /(Q )
) 2.65 ( 0.21. Both of these
H O
D O
2
2
E
+
The assignment made, in addition, produces sensible rate and
equilibrium constants that show reasonable isotope effects. These
isotope effects are based upon further rate measurements made
a
H+
Q k′ [H ]
k
)
(10)
obs
E
a
+
Q + [H ]
in D2O solutions of perchloric acid. The results are summarized
in Table S513 and are displayed, along with the corresponding
isotope effects are reasonable values for the processes they
represent. Isotope effects on rate-determining hydron transfer
to carbon should be in the normal direction (kH/kD > 1), and
those on hydron transfer from hydronium ion should be weaker
than those involving other acids because, in addition to their
normal primary component, they contain an inverse secondary
data for H2O solutions, in Figure 2.
It may be seen that the rates in D2O are slower than those in
H2O at all of the acid concentrations investigated, but that the
difference is greater in the plateau regions above the bends in
these profiles than in the diagonal regions below the bends.
Bends in rate profiles such as these are commonly produced
+
19
contribution from the nonreacting bonds of H3O . The large
value of the present rate constant, moreover, implies an
unsymmetrical, reactant-like transition state, which will reduce
21
by ionization of acid or basic groups in the system, and similar
bends in the rate profiles for ketonization of other carboxylic
acid enols have been shown to be caused by acid ionization of
19,23
the magnitude of this isotope effect still further.
The isotope
(
22) This is a concentration dissociation constant applicable at ionic
(
20) Bonneau, R.; Wirz, J.; Zuberb u¨ hler, A. D. Pure Appl. Chem. 1997,
9, 979-992.
21) Loudon, G. M. J. Chem. Educ. 1991, 68, 973-984.
strength 0.10 M.
6
(23) Kresge, A. J.; Sagatys, D. S.; Chen, H.-L. J. Am. Chem. Soc. 1977,
99, 7228-7233.
(

Flash Photolysis of 10-Diazo-9(10H)-phenanthrenone
J. Am. Chem. Soc., Vol. 119, No. 36, 1997 8421
E
a
effect on Q is also consistent with expectation, inasmuch as
kcat ) k + (k - kB)fA
(13)
B
HA
oxygen acids are less ionized in D2O than in H2O and ratios in
the vicinity of 3 are not uncommon.1
9,24
-
2-
present as acid. The data for H PO /HPO
4
2
4
buffers showed
In contrast to these reasonable isotope effects, the alternative
assignment of ketene hydration and enol ketonization rate
both general acid and general base catalysis, in keeping with
the fact that these buffers were used in a region where the rate
profile showed both acid and base catalysis; least squares
+
constants, that which avoids crossing of rates at [H ] = 0.001
E
a
E
a D
4
-1 -1
M, produces (Q ) /(Q ) ) 0.65. We know of no example of
analysis gave kHA ) (3.95 ( 0.10) × 10 M
s
and kB )
H
3
-1 -1
+
an inverse solvent isotope effect on the ionization of an oxygen
acid, and such a value is in fact difficult to rationalize on the
basis of solvent isotope effect theory.1
(6.15 ( 1.05) × 10 M s . The data for NH4 /NH3 buffers,
on the other hand, showed only general base catalysis, again in
keeping with the fact that these buffers were used in a more
basic region of the rate profile where only base catalysis
occurred; least squares analysis gave kHA ) (1.49 ( 2.07) ×
9
The second bend in the rate profile for ketonization of
fluorene-9-carboxylic acid enol seen in Figure 1 is similar to
that shown by enols of other carboxylic acids, where it has been
4
-1 -1
5
-1 -1 26
10 M
s
and kB ) (6.26 ( 0.27) × 10 M s .
5,6c
attributed to the second ionization of the enol.
The diagonal
These general acid catalytic coefficients refer to direct
portion of slope 1 before this bend may then be assigned to
ketonization via â-carbon protonation of the enolate dianion by
a water molecule under conditions where the enolate monoanion
is the initial state, eq 11. The overall rate of this reaction will
â-carbon protonation of the enolate monoanion by the acid
-
- 25
involved; i.e., for H2PO4 , kHA ) k′HA(H2PO4 ). The general
base catalytic coefficients, on the other hand, may be assigned
to a reaction mechanism where the base involved first removes
the remaining oxygen-bound proton from enolate monoanion
in a rapidly established equilibrium, and the enolate dianion
and general base conjugate acid thus produced then react by
rate-determining proton transfer to â-carbon, eq 14. This leads
be inversely proportional to hydrogen ion concentration, produc-
ing an apparent hydroxide ion catalysis. This catalysis, how-
ever, will become saturated when the initial state of the reac-
tion shifts to enolate dianion, and that will give rise to the
horizontal portion of the rate profile that appears beyond the
second bend.
E
a
BH
a
to the expression kB ) Q′ k′′HA/Q , from which values of
k′′HA, the rate constant for general acid, â-carbon protonation
of the enolate dianion may be calculated. The results so
The rate law that corresponds to this reaction scheme is shown
as eq 12. When this expression plus the rate law for the acidic
portion of the rate profile (eq 10) were used to analyze the data,
it was found that a better fit could be obtained if an additional
term, k′o, was added to include an uncatalyzed reaction at the
-
6
-1 -1
obtained are k′′HA(H2PO4 ) ) (4.14 ( 0.86) × 10 M
s
+
6
-1 -1 25
and k′′HA(NH4 ) ) (1.30 ( 0.16) × 10 M s .
The pattern of values that these rate constants furnish provides
support for the reaction schemes involved. The value of
-
-
+
-7
2 4 HA 2 4
k′′HA(H PO ) is considerably greater than that of k′ (H PO ),
consistent with the expected greater reactivity of the dianion to
bottom of the profile near [H ] ) 10 M.
E
a
-
Q′ k′′
electrophilic reagents, and the value of k′′(H2PO4 ) is also
o
+
kobs
)
(12)
greater than that of k′′(NH4 ), consistent with the greater acid
E
a
+
Q′ + [H ]
-
strength of H2PO4 .
Further support for these reaction schemes comes from
additional isotope effects determined in basic solutions. These
isotope effects are based upon additional rate measurements
Least squares fitting of the entire expression gave k′H+ ) (1.25
8
-1 -1
-1 -1
(
0.03) × 10 M s , k′o ) (2.98 ( 0.47) × 10 s , k′′o )
-
3
-1
E
a
-3
(
(
(
3.22 ( 0.13) × 10 s , Q ) (9.87 ( 0.53) × 10
M
made in D O solutions of sodium hydroxide, at the concentration
2
E
22
E
a
-10
pQ ) 2.01 ( 0.02), and Q′ ) (2.46 ( 0.29) × 10
M
a
[NaOD] ) 0.001 M, and also in D2O solutions of ammonium
E
22,25
pQ′ ) 9.61 ( 0.05).
a
ion buffers at a single buffer ratio. The data so obtained are
The uncatalyzed region at the bottom of the ketonization rate
13
summarized in Tables S6 and S7.
profile could represent â-carbon protonation of the enolate
monoanion by a water molecule, or â-carbon protonation of
the enolate dianion by hydrogen ion. The latter mechanism,
In these basic solutions, the deuterium analog of eq 12 applies.
The average of the rate constants determined in sodium hy-
droxide solution plus the average intercept of the ammonium
however, appears unlikely because the rate law that would then
ion buffer dilution plot then provides values of k at two dif-
obs
E
a
, and the value of k′′H+ required
+
apply to this region is Q′ k′′
H
+
ferent values of [D ], from which (k′′ )
) (4.56 ( 1.24) ×
o D2O
102 s-1 and (Q′ )
E
-11
E
a
by the data is the improbably high rate constant k′′H+ ) (1.21
) (7.37 ( 1.43) × 10
M, p(Q′ )
a
22
D O
D O
2
2
1
1
-1 -1
(
0.245) × 10
M
s .
) 10.13 ( 0.08, may be calculated. The slope of the buffer
In contrast to the weak buffer catalysis found for the hydration
dilution plot leads to the additional result k′′ (ND +) )
DA
4
of fluorenylideneketene, buffer catalysis of the ketonization of
fluorene-9-carboxylic acid enol was strong, and reliable values
of buffer catalytic coefficients could be obtained. These were
evaluated as slopes of buffer dilution plots, kcat, and values of
kcat determined at several different buffer ratios were then
separated into general acid, kHA, and general base, kB, compo-
nents through the use of eq 13,where fA is the fraction of buffer
(1.16 ( 0.23) × 105 M-1 s-1. Calculation of these results
used literature values of the solvent isotope effect on the
autoprotolysis constant of water27 and that on the acid ionization
(25) Primed symbols are used for reactions of the monoanion and doubly
primed symbols for those of the dianion.
(26) These results refer to reactions beginning with enolate monoanion
as the initial state; values of kcat were corrected to account for small quantities
of dianion present in some of the buffer solutions.
(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.
(
24) Laughton, P. M.; Robertson, R. E. In Solute-SolVent Interactions;
Coetzee, J. F., Ritchie, C. D., Eds.; Marcel Dekker: New York, 1969;
Chapter 7.

8422 J. Am. Chem. Soc., Vol. 119, No. 36, 1997
Andraos et al.
of ammonium ion.28 Combination of these results with their
Scheme 1
E
E
H2O counterparts gives the isotope effects (Q′ ) /(Q′ ) )
a D O
2
a
H O
2
+
3
.34 ( 0.76, (k′′o)H O/(k′′o)D O ) 7.07 ( 1.94, and k′′HA(NH4 )/
2
2
+
k′′HA(ND4 ) ) 11.2 ( 2.6.
These isotope effects agree well with values expected for the
processes they represent. The effect on the second ionization
of the enol is near 3, which is typical for the ionization of oxygen
2
4
acids. It is significant, moreover, that this isotope effect is
E
a
stronger than that on the first ionization of the enol, (Q )
/
H O
2
E
a
(
Q )
) 2.65, for isotope effects on the ionization of
D O
2
structurally similar acids generally increase with decreasing acid
strength.29 The two kinetic isotope effects are also as expected.
They are both in the normal direction (kH/kD > 1), as they should
be for rate-determining hydron transfer to carbon. They are
also both quite large, which implies that they refer to reactions
with central, fairly symmetrical transition states; application of
Marcus theory relationships governing variation in the magni-
tude of isotope effects with transition state structure2 confirms
this, suggesting that both of these isotope effects lie within 90%
of their maximum possible value.
3,30
Enolization of Acid. Rates of enolization of fluorene-9-
carboxylic acid were measured in aqueous sodium hydroxide
solutions over the concentration range 0.02-0.10 M. The data
are summarized in Table S8.13 Observed first-order rate con-
stants were accurately proportional to hydroxide ion concentra-
tion, and linear least squares analysis gave the hydroxide ion
ammonia buffers, again as expected from the nucleophilic nature
of this reaction and the steric hindrance to such a process offered
by the fluorenylidene group. The overall rates of these reactions
are nevertheless high in comparison to other structurally related
ketenes, again as expected because of the special ability of the
fluorenylidene nucleus to stabilize the negative charge generated
on the substrate by nucleophilic attack.
Disappearance of the second transient gives a rate profile
different from the first, but one that is typical of enol
ketonizations. In particular, it shows two bends, identified as
first and second ionizations of fluorene-9-carboxylic acid enol
by the fact that they produce sizable solvent isotope effects
whose absolute as well as relative magnitudes are typical of
oxygen acid ionization. Rate constants in various parts of this
rate profile, assigned to ketonization of fluorene-9-carboxylic
acid enol through rate-determining â-carbon protonation of the
enolate mono- or dianion by either hydrogen ion or water, also
show solvent isotope effects whose various magnitudes are
completely consistent with isotope effect theory. Disappearance
of this second transient is also strongly catalyzed by buffers,
and analysis of the catalysis shows it to be of the general acid
type; this again is typical of ketonization reactions and is wholly
expected on the basis of their rate-determining proton-transfer
nature.
E
-1
-1 -1
catalytic coefficient, k′ HO- ) (2.88 ( 0.05) × 10
M
s .
Discussion
Reaction Scheme. We have found that flash photolysis of
0-diazo-9(10H)-phenanthrenone, 7, in aqueous solution gives
1
fluorene-9-carboxylic acid, 5, as the major product via two
transient intermediate species which we have identified as
fluorenylideneketene, 8, and the enol of fluorene-9-carboxylic
acid, 6, eq 15. Identification of these intermediates is based
upon a number of pieces of evidence.
These arguments, elaborated in the preceding section where
the experimental data upon which they are based are presented,
leave little doubt that the transient species observed in these
flash photolysis experiments are fluorenylideneketene and
fluorene-9-carboxylic acid enol.
Conversion of the first transient into the second gives a rate
profile typical of aromatic ketene hydration, e.g., a long region
of uncatalyzed reaction accompanied by weak hydroxide ion
catalysis and no acid catalysis. The uncatalyzed reaction shows
only a weak solvent isotope effect in the normal direction (kH/
kD > 1), consistent with the fact that uncatalyzed ketene
hydration involves nucleophilic attack by a water molecule with
some weakening but no breaking of hydrogen-oxygen bonds.
This intermediate also reacts only weakly with ammonium ion/
Determination of Equilibrium Constants. The keto-enol
system of a simple monofunctional carboxylic acid consists of
five species, the un-ionized and ionized forms of the acid and
the un-ionized, monoionized, and diionized forms of the enol,
as illustrated in Scheme 1 for the present case. Some of the
equilibrium constants connecting these various species have
already been evaluated, from the ketonization rate profile and
by measurement of the ionization constant of fluorene-9-
(
28) Salomaa, P.; Schaleger, L. L.; Long, F. A. J. Phys. Chem. 1964,
8, 410-411.
29) Bell, R. P. The Proton in Chemistry; Cornell University Press:
Ithaca, NY, 1959; pp 187-189.
30) Kresge, A. J. In Isotope Effects on Enzyme-Catalyzed Reactions;
31
carboxylic acid. The rest can be determined in the following
6
way.
(
As eq 16 shows, fluorene-9-carboxylate ion and its enolate
dianion are interconverted by reaction of the carboxylate ion
(
Cleland, W. W.; O’Leary, M. H., Northrop, D. B., Eds.; University Park
Press: Baltimore, MD, 1977; pp 37-63. Kresge, A. J. J. Am. Chem. Soc.
(31) Kresge, A. J.; Pojarlieff, I. G.; Rubinstein, E. M. Can. J. Chem.
1993, 71, 227-229.
1
980, 102, 7797-7798.

Flash Photolysis of 10-Diazo-9(10H)-phenanthrenone
J. Am. Chem. Soc., Vol. 119, No. 36, 1997 8423
with hydroxide ion and reaction of the enolate dianion with
water. Rate constants for both of these reactions have been
Table 1. Summary of Rate and Equilibrium Constants for the
Fluorenylideneketene and Fluorene-9-carboxylic Acid Systems
a
determined here, and their ratio can therefore be used to provide
K
a value of the equilibrium constant, Q′ for the overall
a
K
a
-18
K
process: Q′ ) (1.42 ( 0.06) × 10
M, pQ′a ) 17.85 (
0
.02.22 This equilibrium constant can then be used in conjunc-
E
a
tion with Q′ , evaluated from the ketonization rate profile, to
calculate K′E, which is the third member of the triangular cycle
that makes up the lower part of Scheme 1. This gives K′E )
-
9
(
5.75 ( 0.72) × 10 , pK′E ) 8.24 ( 0.06. Knowledge of
31
E
a
K′E, together with Qa and Q (rate profile) then allows
-
10
evaluation of KE as KE ) (2.15 ( 0.31) × 10 , pKE ) 9.67
(
1
(
0.06, and once all four legs of the upper square in Scheme
K
a
K
a
are known, the diagonal Q can be calculated as Q ) (2.13
M, pQ ) 11.67 ( 0.06.
Comparison with Other Systems. The present results show
-
12
K
a
22
0.32) × 10
the enol content of fluorene-9-carboxylic acid to be remarkably
large for a carboxylic acid keto-enol system. The keto-enol
equilibrium constant for acetic acid, the prototype carboxylic
acid, has not been determined experimentally, but reliable
estimates have recently been made by several different methods
3
2
which give consistent results and put the value at pKE ) 20.
This is 10 orders of magnitude less than the presently determined
equilibrium constant, pKE ) 9.67. This large difference is of
course due to the fact that the enolization of fluorene-9-
carboxylic acid places a double bond exocyclic to a cyclopen-
tadienyl ring, which creates fulvenoid resonance. A similar
enol-content-enhancing effect has been found for simple alde-
33
hydes and ketones. In the case of 9-formylfluorene, eq 17,for
a
Aqueous solution, 25 °C, ionic strength 0.10 M.
Table 2. Comparison of Equilibrium Constants for Some
Reactions of Carboxylic Acid Enolsa
example, the enol is actually the more stable isomer, and with
pKE ) -1.22, the enol content in this system is 7 orders of
9
magnitude greater than that of acetaldehyde (pKE ) 6.23).
The enol of fluorene-9-carboxylic acid is also a remarkably
strong acid. With pQ ) 2.01, it is more acidic than most car-
E
a
boxylic acids, and it is in fact a stronger acid than the carboxylic
acid group of its keto isomer, for which pQa ) 3.43. A
comparison with acetic acid enol is again instructive. The
acidity of this enol has also not been measured experimentally,
E
but pQ ) 7 has recently been estimated for the enol of ethyl
a
a
Aqueous solution, 25 °C, ionic strength 0.10 M. b Reference 5b.
34
acetate, and the fact that the acidities of the enols of meth-
c
d
e
Reference 5a. Revised value provided by Professor Wirz. Reference
6c. Reference 6a.
E
f
yl mandelate and mandelic acid are closely similar, pQ )
a
1
2a
6a
E
a
6
.45 and 6.62, respectively, suggests that pQ ) 7 is a
good estimate for acetic acid as well. The enol of fluorene-9-
carboxylic acid is 5 orders of magnitude more acidic than this.
The difference, of course, is due to the fact that the negative
charge of the enolate ion produced upon ionization of the enol
of fluorene-9-carboxylic acid can be delocalized into the five-
membered ring, thus creating an aromatic cyclopentadienyl
anion. This ability of the cyclopentadienyl group to stabilize a
negative charge is aptly illustrated by the unusual carbon acidity
of cyclopentadiene itself: pKa ) 16.
As the data summarized in Table 2 show, this ability of the
cyclopentadienyl group to stabilize enols and enolate ions is
apparent also in the indenecarboxylic acid system and in
cyclopentadienecarboxylic acid itself. The effects, in fact, grow
progressively stronger as the benzo groups are removed. This,
35
(
32) Guthrie, J. P. Can. J. Chem. 1993, 71, 2123-2128. Guthrie, J. P.;
Liu, Z. Can. J. Chem. 1995, 73, 1395-1398.
33) Harcourt, M. P.; More, O’Ferrall, R. A. J. Chem. Soc., Perkin Trans.
1995, 1415-1425.
34) Amyes, T. L.; Richard, J. P. J. Am. Chem. Soc. 1996, 118, 3129-
141.
(
2
(
(35) Streitwieser, A.; Nebenzahl, L. L. J. Am. Chem. Soc. 1976, 98,
2188-2190.
3

8424 J. Am. Chem. Soc., Vol. 119, No. 36, 1997
Andraos et al.
of course, is to be expected, for benzo substitution will reduce
the resonance energy of the cyclopentadienyl ring, as evidenced,
acidities of carboxylic acid and ester enols ionizing as oxygen
acids might be similar as well, based upon the close similarity
3
6
E
for example, by the lower acidity of fluorene (pKa ) 22.6)
of Qa for mandelic acid and methyl mandelate mentioned
36
K
than that of indene (pKa ) 20.1).
above, suggests, by virtue of the relationship pQa ) pKE +
E
a
The data of Table 2 also show that the combined effects of
the phenyl and cyano groups in the enol of phenylcyanoacetic
acid, 17, are just about as effective in stabilizing enols and
enolate ions as are the cyclopentadienyl groups in the enols of
fluorenecarboxylic acid, indenecarboxylic acid, and cyclopen-
tadienecarboxylic acid. The stabilization provided by the phenyl
and hydroxyl groups in the enol of mandelic acid, 18, however,
is considerably less strong, which shows that a cyano group is
much better than a hydroxyl group in enhancing enol content
and increasing enol acidity.
pQ , that enol contents of carboxylic acids and their esters
might be similar also. Carboxylic acids could therefore serve
as useful surrogates for esters in studies of enol chemistry; this
would be especially advantageous in the case of investigations
conducted in aqueous solution, where hydrolysis of the ester is
34
often a troublesome competing reaction.
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
research.
It is interesting that the acidity of fluorene-9-carboxylic acid
K
ionizing as a carbon acid, eq 18, R ) H, pQa ) 11.67, is
Supporting Information Available: Tables S1-S8 of rate
data (15 pages). See any current masthead page for ordering
and Internet access instructions.
practically the same as that of its methyl ester, eq 18,R ) Me,
JA971381S
(
(
36) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463.
37) Chiang, Y.; Jones, J., Jr.; Kresge, A. J. J. Am. Chem. Soc. 1994,
pQa ) 11.52.37 This, coupled with the expectation that
K
116, 8358-8359.